Australian Journal Of Physiotherapy

Journal of Physiotherapy 67 (2021) 263–270 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research Advice and education provide small short-term improvements in pain and disability in people with non-specific spinal pain: a systematic review Caitlin MP Jones, Christina Abdel Shaheed, Giovanni E Ferreira, Priti Kharel, Chung-Wei Christine Lin, Chris G Maher Institute for Musculoskeletal Health, The University of Sydney and Sydney Local Health District, Sydney, Australia KEY WORDS ABSTRACT Back pain Questions: What is the effect of advice/education compared with placebo or no advice/education on pain Neck pain and disability in people with non-specific spinal pain? To what extent do characteristics of the patients, trial Advice or intervention modify the estimate of the treatment effects? Design: A systematic review with meta- Education analyses of randomised controlled trials. Participants: Adults with non-specific back and/or neck pain Meta-analysis with or without radiating leg/arm pain of any duration were included. Trials recruiting pregnant women or surgical patients in the immediate postoperative phase were ineligible. Intervention: Advice or educa- tion. Outcome measures: The primary outcomes were self-reported pain and disability, and the secondary outcome was adverse events. The following potential effect modifiers were examined: risk of bias, duration of pain, location of pain, intensity of intervention and mode of intervention. Results: Twenty-seven trials involving 7,006 participants were included. Eighteen of the included trials were assessed as being at low risk of bias ( 6 on the PEDro scale). There was low-quality evidence that advice had a small effect on pain (MD 28.2, 95% CI 212.5 to 23.9, n = 2,241) and moderate-quality evidence that advice had a small effect on disability (MD 24.5, 95% CI 27.9 to 21.0, n = 2,579) compared with no advice or placebo advice in the short- term. None of the items that were assessed modified the treatment effects. Conclusion: Advice provides short-term improvements in pain and disability in non-specific spinal pain, but the effects are small and may be insufficient as the sole treatment for patients with spinal pain. Registration: PROSPERO CRD42020162008. [Jones CMP, Shaheed CA, Ferreira GE, Kharel P, Lin C-WC, Maher CG (2021) Advice and education provide small short-term improvements in pain and disability in people with non-specific spinal pain: a systematic review. Journal of Physiotherapy 67:263–270] © 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction Previous systematic reviews have assessed whether advice is effective for non-specific spinal pain. Their results ranged from no Spinal pain is a leading cause of disease burden and years lived difference to a small difference in pain and disability outcomes in with disability worldwide;1–4 it is also a major contributor to the favour of advice.12–14 All reviews are outdated: they were published  economic burden of healthcare on society5 and the individual.6 10 years ago and therefore have missed many trials that have been Disability and costs associated with chronic pain (of which spinal conducted in the interim. The two earlier reviews did not include a pain is the largest contributor) are predicted to increase from $139 rating of the quality of evidence, as this was not previously a common billion to $215 billion by 2050 in Australia.7,8 practice.13,14 They also only looked at advice to stay active and did not analyse other types of advice, such as ergonomic advice, or pain International guidelines for the management of spinal pain neuroscience education. Therefore, uncertainty remains about the endorse advice as part of first-line treatment.9,10 Most guidelines (11 effectiveness of advice and education as interventions for non- out of 12) recommend providing the patient with some form of specific spinal pain. advice but the advice varies somewhat from guideline to guideline.9 Recommendations from most contemporary guidelines include the This review aimed to evaluate the effectiveness of advice/ provision of advice to remain active and avoid bed rest, advice on education in reducing pain and disability for patients with self-management and reassurance of the good prognosis.11 Less non-specific spinal pain. It also aimed to investigate potential commonly, guidelines may encourage pain neuroscience education to effect modifiers such as patient characteristics (chronicity), help the patient better understand their pain or ergonomic advice to study characteristics (risk of bias) and intervention characteris- help the patient avoid reinjuring their spine.9 At present the optimal tics (type of advice, intensity of treatment by time and mode of content for patient advice is unclear. delivery). https://doi.org/10.1016/j.jphys.2021.08.014 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

264 Jones et al: Advice and education for spinal pain Therefore, the specific research questions for the systematic re- Box 1. Inclusion criteria. view were: 1. What is the effect of advice/education compared with placebo or Design no advice/education on pain and disability in people with non-  Randomised controlled trial specific spinal pain? Participants  People with non-specific spinal pain (back or neck) of any 2. To what extent do characteristics of the patients, trial or inter- vention modify the estimate of the treatment effects? duration  With or without radiating leg/arm pain Method Intervention  Advice, defined as any advice, education or information given This systematic review was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) by a healthcare professional to improve a patient’s statement.15 understanding of pain or appropriate management Outcomes Identification and selection of trials  Primary: pain, disability  Secondary: adverse events Searches were conducted on MEDLINE, EMBASE, Cochrane Central Comparisons Register of Controlled Trials, CINAHL, PEDro, and the World Health  Advice versus no advice or placebo advice Organization International Clinical Trials Registry Platform (ICTRP) for eligible studies from their inception to 25 September 2020. Searches a serious flaw was present using the following criteria. The Study were based on the keywords recommended by the Cochrane Back Limitation domain was downgraded by one level if . 25% of partic- and Neck Group.16 The detailed search strategy is presented in ipants were from studies at high risk of bias (ie,  5 on the PEDro Appendix 1 on the eAddenda. Two reviewers screened titles and scale).22 The Inconsistency domain was downgraded by one level if abstracts independently and read the full-text version of potentially heterogeneity was large (I2 statistic value . 50%, representing eligible articles to determine eligibility. Disagreements were resolved potentially substantial heterogeneity).23 The Indirectness domain was by discussion, with arbitration by a third researcher where necessary. not assessed because the inclusions (patients, treatments and com- Citation tracking from previous systematic reviews12–14 was also parators) closely matched the research question. The Imprecision performed and these studies were included. domain was downgraded by one level if the 95% confidence interval (95% CI) width was . 20 points (ie, twice the pre-defined minimum The eligibility criteria are shown in Box 1. Trials recruiting preg- clinically important difference of 10 points on a 100-point scale).24 nant women or surgical patients in the immediate postoperative Dichotomous outcomes were downgraded by one level if the lower phase were ineligible. Advice was defined as any advice, education or or upper limits of the 95% CI included appreciable benefit or harm (ie, information given by a healthcare professional to improve a patient’s 95% CI , 0.75 or . 1.25). The Publication Bias domain was down- understanding of pain or appropriate management. This included graded by one level if a funnel plot suggested the presence of pub- advice about being physically active, advice about how to self- lication bias or Egger’s regression score was , 0.1.25 manage spinal pain, reassurance about the good prognosis and self- limiting nature of spinal pain, and pain management education. Participants This definition is concordant with previous reviews.9,11 The advice To characterise and assess the similarity of the participants among could be delivered: in any mode (verbal, written, technology-based or a combination of these); over single or multiple sessions; and by any the included trials, the following data were extracted from each trial: health professional. There were no restrictions regarding care setting sample size, age range, location/chronicity of spinal pain, and country (eg, primary healthcare or workplace). Eligible controls included and clinic of recruitment. placebo/sham advice, no advice (eg, waitlist control) or an active treatment that was common to the advice arm (eg, advice and ex- Intervention ercise versus exercise alone). Placebo/sham advice allowed contact To characterise the experimental interventions, details of the with a health professional but not provision of information on back pain (eg, using a reflective and non-directive approach).17 Sham in- advice intervention were extracted using items from the TIDieR terventions were also included (eg, detuned ultrasound) when used checklist.26 Details of the format included session duration, session as a control for advice. frequency, program duration, group/individual format and the pro- fession of the person who delivered the intervention. Details of the content included subject matter and verbal and/or written delivery. Assessment of characteristics of trials Outcome measures Primary: One primary outcome was pain intensity, measured by self- Risk of bias and certainty of evidence report using a tool such as a visual analogue scale, numerical rating Risk of bias for each trial was assessed with the PEDro Scale.18 This scale (0 to 10) or brief pain inventory. The other primary outcome was disability measured by self-report using a tool such as the Roland was a deviation from the original protocol, which specified the Morris Disability Questionnaire or the Oswestry Disability Index. Cochrane Risk of Bias Tool 2. There were concerns about the inter- Secondary: The secondary outcome was adverse events, with data rater reliability of the Cochrane Risk of Bias Tool 2,19 whereas the extracted as the number (%) of participants reporting an adverse PEDro scale did not have this issue.20 Where available, scores were event and the nature of the adverse event. extracted from the PEDro database. The PEDro scoring process is Timing and effect measures: Outcomes were grouped by the timing of completed by two trained independent reviewers and checked by follow-up: immediate term ( 2 weeks), short term (. 2 weeks but  PEDro staff before scores are posted on the PEDro website. In the case 3 months), intermediate term (. 3 months but  12 months) and where papers were unavailable on the PEDro database, two re- long term (. 12 months). The pre-specified primary time point was searchers independently scored each study. Differences were the short term. Where there were multiple follow-up points within resolved with discussion. The quality of the evidence for each the specified time periods, we used the ones that were closest to 2 outcome was rated using the GRADE framework.21 The domains weeks (immediate term), closest to 3 months (short term), closest to evaluated by the GRADE framework include study design, risk of bias, 6 months (intermediate term) and closest to 12 months (long term). inconsistency, indirectness, imprecision and publication bias. For each domain, the quality of evidence was downgraded by one level if

Research 265 Box 2. Coding of experimental interventions. Advice  advice to keep active  advice to avoid bed rest  advice to self-manage  reassurance of a positive prognosis Ergonomic advice  biomechanical or postural advice  biomechanical or postural education Pain neuroscience education  ‘Explain pain’ interventions Box 3. Items examined for association with treatment outcomes. Risk of bias  high risk (PEDro Scale scores  5)  PEDro Scale score as a continuous variable Duration of pain  acute ( 12 weeks) versus chronic (. 12 weeks) a Location of pain  neck pain versus back pain b Intensity of intervention  high ( 1 hour or . one session) versus low (, 1 hour and one session only)  total dosage (minutes) as a continuous variable Mode of intervention  verbal, written or mixed a Studies with mixed populations were categorised based on whether the majority of participants had acute or chronic pain. b Studies with mixed populations were categorised based on whether the majority of participants had back or neck pain. Data extraction Figure 1. Flow of trials through the review. a Five trials ongoing and one author contacted with no response. Two independent researchers used a standardised form to extract b Deviation from protocol 24 July 2020, excluding studies with discordant advice as a the following data: bibliometric data, study characteristics, charac- control. teristics of included participants, outcome measures and outcomes. Any disagreement between the two researchers was resolved by within-group change scores were used rather than follow-up scores. consensus first, then by arbitration with a third independent A full list of necessary data conversions made for the purpose of researcher where necessary. The pooled difference was calculated as meta-analysis is provided in Appendix 2. a mean and 95% CI for each subgroup in the simple stratified analysis. Comprehensive Meta-Analysisa software was used.27 A significance Exploration of heterogeneity level of p , 0.01 (and 99% CI) was used for exploratory meta- The overall effect of advice was examined and studies were regression. grouped according to the ‘type’ of advice provided. This catego- Data analysis risation was added post-hoc in response to unforeseen levels of heterogeneity between the types of advice being provided. Advice Data synthesis was coded as shown in Box 2. Associations with the items shown in To improve the comparability of findings between trials, pain Box 3 on treatment outcomes were further explored. severity and disability outcomes were converted to a common 0-to- Results 100 scale and presented as mean differences (MD) rather than pro- portions of a standard deviation (SD).28,29 The pooled MD was Compliance with the review protocol calculated along with 95% CIs using a random-effects model. An effect size was considered small and not clinically important if it was All post-registration deviations to the protocol and their accom- smaller than the pre-specified threshold for clinical importance ( 10 panying justifications are listed in Appendix 3. points on a 100-point scale).28 Post-treatment means and SDs were extracted where possible. Some studies presented their results in Flow of trials through the review other formats (eg, median and inter-quartile range (IQR) or in figures only). We converted IQR to SD using the formula of IQR/1.35.30 A total of 8,396 articles were identified via searches, of which 27 Confidence intervals were converted to SD using the Cochrane articles (26 unique studies with a total of 7,006 participants) were Group’s RevMan calculator.31 Other studies had missing data (eg, no found to be eligible for inclusion (Figure 1).17,32–57 See Appendices 4 reported measure of dispersion); in these cases, SDs were borrowed and 5 for a list of the excluded studies and reasons for exclusion. from the most similar studies. Some studies compared groups that were not equal at baseline for pain and disability outcomes. Where between-group differences were  5 points on a 100-point scale,

Table 2 PEDro criteria and scores for the included trials (n = 5). Study Eligibility and Random Concealed Groups similar Participant Therapi sourcea allocation allocation at baseline blinding blindin Akca 201732 Albaladejo 201033 Y Y Y Y N N Ayanniyi 201534 Y Y Y Y N N Bodes 201835 Y Y N Y N N Bucker 201036 Y Y Y Y N N Cherkin 199637 Y Y Y Y N N Damush (long) 200338 Y Y N Y N N Damush (short) 200339 Y Y N Y N N Darlow 201940 Y Y Y Y N N Derebery 200941 Y Y Y Y N N Ibrahim 201842 Y Y N Y N N Jellema 200543 Y Y Y Y N N José 201444 Y Y N Y N N Kovacs 200745 Y Y Y N N N Lamb 201346 Y Y Y Y N N Little 200147 Y Y Y Y N N Matias 201948 Y Y Y Y N N Pengel 200749 Y Y Y Y N N Pires 201550 Y Y Y Y Y N Rantonen 201851 Y Y Y Y N N Roberts 200252 Y Y Y Y N N Saracoglu 202053 Y Y Y N N N Slater 201354 Y Y Y Y N N Téllez-García 201555 Y Y Y Y N N Traegar 201959 Y Y Y Y N N Werner 201656 Y Y Y Y N N Y Y N Y N N Y = yes, N = no. a This item relates to external validity and therefore does not contribute to the total score.

ist Assessor , 15% Intention-to-treat Between-group Point estimate and Total (0 to 10) 266 Jones et al: Advice and education for spinal pain ng blinding dropouts analysis difference reported variability reported 6 N Y N Y Y 6 N Y N Y Y 4 N N N Y Y 6 N Y N Y Y 6 N N Y Y Y 5 N Y N Y Y 4 N N N Y Y 6 N N Y Y Y 7 N Y Y Y Y 3 N N N Y N 7 N Y Y Y Y 6 N Y Y Y Y 3 N N N Y N 6 N Y N Y Y 7 N Y Y Y Y 5 N N N Y Y 6 N Y N Y Y 9 Y Y Y Y Y 7 N Y Y Y Y 6 N N Y Y Y 5 N Y N Y Y 6 N Y N Y Y 5 N N N Y Y 6 N Y N Y Y 8 Y Y Y Y Y 5 N N Y Y Y

Research 267 Study WMD (95% CI) Overall Bodes Pardo 2018 Random high Darlow 2019 low Ibrahim 2018 low Jellema 2005 mod Pires 2015 low Zhang 2014 high Pengel 2007 high Clemente da Silva 2014 high Matias 2019 low Akca 2017 high Ayanniyi 2015 low Slater 2013 (written) low Slater 2013 (mixed) mod Werner 2016 high Albaladejo 2010 low Traeger 2019 mod Little 2001 Saracoglu 2020 Publication bias (Egger)a no no Pooled no no no (0.5) no no no no (0.2) no no (0.8) no (0.9) no no no no (0.3) Quality of evidence assessment (GRADE) Imprecision no no yes no –40 –20 0 20 40 Inconsistency yes no yes no yes no no no no no no no yes no no no yes no yes no no yes no no no yes yes no Favours advice Favours control Figure 2. Weighed mean difference (95% CI) in the effect of advice versus control on Immediate-term ( 2 weeks); short-term (. 2 weeks but  3 months); intermediate-term (. 3 months but  12 months); long-term (. 12 months). pain in the short term (. 2 weeks and  3 months). Shaded row = primary analysis (pain overall at the short-term). Characteristics of included trials Study limitation a Egger two-tailed p-values; values were only calculated if  10 trials were meta-analysed. no b At primary time point (short-term). Detailed characteristics of the included studies are shown in yes c Combined types of advice. Table 1 on the eAddenda. yes no Quality yes The majority of the included studies (18 of 26) were assessed as no no being at low risk of bias, defined as  6 on the PEDro scale (Table 2). no All included trials were of parallel, randomised design. yes no Participants yes Most trials defined either acute or chronic samples; however, yes no some included participants with symptoms persisting . 12 weeks. In no these cases, they were categorised as acute or chronic, based on the yes chronicity of the majority of participants. See Table 1 on the eAd- no denda for further details. Mean difference (95% CI) Intervention Exp minus Con The trials used advice or education of various types, intensities 23.3 (28.7 to 4.5) and delivery modes. Treatment intensity ranged from 10 to 480 mi- 215.6 (224.2 to 27.1) nutes. The number of sessions ranged from one to twelve. Fourteen of 211.4 (220.6 to 22.2) the 26 included studies were classified as delivering a ‘high-intensity’ intervention. The included studies used either a verbal, written or 22.2 (28.4 to 4.1) mixed mode of delivery. The content varied across trials (see Table 1 2 8.2 (212.5 to 23.9) on the eAddenda) and included topics such as neurophysiology of pain, self-management advice, ergonomic advice and advice to 21.0 (23.7 to 1.6) remain active. No included studies used web or app-based delivery 22.2 (26.0 to 1.6) modes. 23.3 (25.9 to 20.8) 211.2 (217.0 to 25.3) Outcome measures 22.1 (25.2 to 0.9) Follow-up timepoints varied from 0 weeks (assessed immediately 211.9 (217.2 to 26.5) 28.5 (212.9 to 24.1) after the intervention was delivered) to 104 weeks after the 1.0 (215.5 to 17.5) intervention. The majority of the studies (22 of 26) included at least 28.5 (214.7 to 22.2) one outcome measurement at short-term follow-up (the primary 4.0 (26.2 to 14.2) time point). 29.1 (215.2 to 22.9) Effects of intervention Results Pain n Low-quality evidence from 18 trials with 2,241 participants indi- 1,307 385 cated that advice had a small effect on pain in the short term (MD 581 824 2,241 1,140 911 1,261 980 1,053 1,188 2,201 40 948 83 1,210 Trials 8 4 6 8 18 5 5 6 12 6 11 17 1 7 1 10 Table 3 Outcome contemporary Effect of advice on pain – subgroup analysis. ergonomic explain pain immediate short-term intermediate long-term acute chronic low high back neck verbal written mixed Subgroup Type of adviceb Time pointc Chronicityb, c Treatment intensityb, c Pain locationb, c Mode of deliveryb, c

268 Jones et al: Advice and education for spinal pain Study WMD (95% CI) Albaladejo 2010 Random Ayanniyi 2015 Overall Bodes Pardo 2018 high Bucker 2010 mod Darlow 2019 Ibrahim 2018 very low Jellema 2005 high Kovacs 2007 mod Matias 2019 mod Pires 2015 high Roberts 2002 high Werner 2016 low Zhang 2014 high Slater 2013 (written) low Slater 2013 (mixed) mod Derebery 2009 mod Traeger 2019 high Pengel 2007 mod Saracoglu 2020 mod Pooled Publication bias (Egger)a no (0.9) no no no (0.9) no (0.9) no no no no (0.2) no no (0.7) no (0.9) no no no no Quality of evidence assessment (GRADE) Imprecision no no no no Inconsistency yes yes no no yes no yes no no no no no yes no no no yes no yes no no no no no no no yes no –40 –20 0 20 40 Favours advice Favours control Figure 4. Weighed mean difference (95% CI) in the effect of advice versus control on Study limitation Immediate term ( 2 weeks); short term (. 2 weeks but  3 months); intermediate term (. 3 months but  12 months); long term (. 12 months). disability in the short term (. 2 weeks and  3 months). no Shaded row = primary analysis (disability overall at the short term). yes –8.2, 95% CI –12.5 to –3.9). A forest plot is presented in Figure 2 and yes a Egger two-tailed p-values; values were only calculated if  10 trials were meta-analysed. the GRADE rating is presented in Table 3. See Figure 3 on the no b At primary time point (short term). eAddenda for a more detailed forest plot. We found evidence of no c Combined types of advice. negligible effects of advice on pain at all secondary time points. A no summary of the main and subgroup analyses and the GRADE rat- no ings of each analysis is shown in Table 3. Meta-regression revealed no that there were no associations between type of advice, duration of yes pain, location of pain, intensity of treatment or mode of delivery no with treatment effects (p . 0.01). yes no Disability yes Moderate-quality evidence from 19 trials with 2,579 partici- no yes pants indicated that advice had a small effect on disability in the no short term (MD –4.5, 95% CI –7.9 to –1.0). A forest plot is presented in Figure 4 and the GRADE rating is presented in Table 4. See Mean difference (95% CI) Figure 5 on the eAddenda for a more detailed forest plot. We found Exp minus Con evidence of negligible effects of advice on disability at all secondary time points. A summary of the main and subgroup analyses and the 21.9 (24.3 to 0.9) GRADE ratings of each analysis is shown in Table 4. Consistent with 25.5 (29.4 to 21.6) the results for pain, meta-regression revealed that none of the 28.8 (218.0 to 0.5) putative effect modifiers were associated with treatment effects on disability (p . 0.01). 0.2 (23.0 to 2.6) 24.5 (27.9 to 21.0) Adverse events Data on adverse events were reported in two of the included 0.1 (22.9 to 3.1) 20.7 (22.7 to 1.3) trials, which had 462 participants. A total of 22 adverse events were 21.3 (23.3 to 0.6) reported across the 27 articles (n = 7,932). The risk ratio for an 27.9 (213.1 to 21.5) adverse event was 1.29 (95% CI 0.34 to 4.94), as shown in Figure 6. 20.9 (23.7 to 1.9) See Figure 7 on the eAddenda for a more detailed forest plot. This 27.7 (212.9 to 22.4) evidence was given a GRADE rating of moderate certainty (down- 25.2 (28.8 to 21.5) graded by one level for imprecision). The adverse events were all 2.5 (22.8 to 7.9) reported as non-serious. In one trial,40 one participant from the 24.5 (28.1 to 20.9) control group reported a gluteal tear related to activity. In another 0.0 (26.4 to 6.4) trial,49 all reported adverse events were considered mild and did 25.3 (211.4 to 0.8) not result in withdrawal from the study. Adverse events included muscle soreness, increased pain, tiredness, nausea, weight gain, Results itchy scalp and numbness in the legs. n 1,772 249 558 1,217 2,579 4,639 3,798 1,703 876 1,518 1,061 2,452 127 937 285 1,357 Trials 11 2 6 11 19 9 8 9 10 9 10 17 2 7 3 9 Table 4 Outcome contemporary Effect of advice on disability: subgroup analysis. ergonomic explain pain immediate short-term intermediate long-term acute chronic low high back neck verbal written mixed Subgroup Type of adviceb Time pointc Chronicityb, c Treatment intensityb, c Pain locationb, c Mode of deliveryb, c

Research 269 Study OR (95% CI) determine patients’ concordance with advice, such as by asking pa- Darlow 2019 Random tients to keep an activity diary. Pengel 2007 Another question that future research could answer is whether Pooled advice/education is more effective for people who have poor knowledge and beliefs about spinal pain. Future trials could assess 0.01 0.1 1 10 100 participant’s knowledge and beliefs prior to providing advice and education, and compare effectiveness in participants with poor Favours advice Favours control knowledge and beliefs versus those with appropriate knowledge and helpful beliefs. Figure 6. Pooled relative risk (95% CI) of adverse events with advice versus control. In conclusion, these results provided evidence that advice pro- vides a short-term benefit for non-specific spinal pain, but the effect is likely to be small and insufficient as the sole treatment in situations where more pain relief is required. Discussion This review found low-to-moderate quality evidence that advice What was already known on this topic: International may have a small effect on pain and disability in the short term guidelines for the management of spinal pain endorse advice as compared with sham/placebo advice or no advice. Advice provided a part of first-line treatment. The guidelines are inconsistent negligible effect at other time points, based on evidence of moderate- regarding what advice should be provided. to-high quality (Tables 3 and 4). Meta-regression did not provide What this study adds: Advice had beneficial short-term ef- clear evidence of effect modification of any variable on treatment fects on pain and disability, but the effects were small and may effects. Advice may not increase the risk of adverse events. be insufficient as the sole treatment for patients with spinal pain. The amount of benefit from advice did not vary substantially This review updates the evidence on the effectiveness of advice as between patients with various durations/locations of pain or a treatment for non-specific spinal pain. Previous reviews have not when the advice was delivered in a variety of ways. isolated advice as a stand-alone treatment. For example, a 2002 Cochrane review examined advice as a treatment for acute back Footnotes: a Comprehensive Meta-Analysis software Version pain12 but they used bed rest (which is known to be harmful) as a 3.3.070, Biostat, Englewood, NJ, USA. control. This design cannot reveal how effective the advice inter- vention was, only how it compared to bed rest. Other reviews such as eAddenda: Table 1, Figures 3, 5 and 7, and Appendices 1, 2, 3, 4 and Liddle et al included trials using advice as an adjunct or only one 5 can be found online at https://doi.org/10.1016/j.jphys.2021.08.014 component of a broader treatment plan.14 Therefore, those reviews could not answer the question of whether advice specifically was the Ethics approval: Not applicable. effective intervention. Given that guidelines recommend advice as a Competing interests: Flexeze provided heat wraps at no cost for stand-alone treatment, it is important that its effectiveness is the SHaPED trial, which has CM as an investigator. CM’s expenses examined accordingly. have been covered, or partially covered, by professional associations hosting conferences where he has been an invited speaker. Flexeze The current findings raise questions about many current guide- provided heat wraps at no cost for a pilot trial in which CAS is an lines that enthusiastically recommend advice as a sole first-line investigator. No other authors have competing interests to declare. treatment for non-specific spinal pain;58 advice alone is likely to Source(s) of support: No specific funding was received for this have at best only a small effect on pain and disability. The effect es- study. CM and CL are funded by research fellowships from the Na- timates for both pain and disability did not meet the pre-determined tional Health and Medical Research Foundation, Australia. minimum clinically important difference of 10 points. The effect es- Acknowledgements: Nil. timate for disability did not contain values above this 10-point Data sharing: The authors will make all relevant data available threshold, which means that our analysis could rule out any clini- upon reasonable request. cally important benefits of advice for that outcome. This suggests that Provenance: Not invited. Peer reviewed. other strategies may need to be considered, such as what is currently Correspondence: Caitlin MP Jones, Institute for Musculoskeletal recommended as second-line treatment (physical and psychological Health, The University of Sydney and Sydney Local Health District, therapies and pain medicines).58 Sydney, Australia. Email: [email protected] It was surprising to see that ‘contemporary guideline advice’ had References the lowest estimate of effect compared with ergonomic advice and ‘explain pain’ education across both outcomes. However, no firm 1. Hurwitz EL, Randhawa K, Yu H, Côté P, Haldeman S. The Global Spine Care conclusions could be drawn due to overlapping confidence intervals. Initiative: a summary of the global burden of low back and neck pain studies. Eur While there is reasonable confidence in the estimates contemporary Spine J. 2018;27:796–801. guideline advice has low efficacy, the efficacy of the other two types of advice is less certain due to ‘low’ or ‘very low’ GRADE ratings. 2. Hoy D, March L, Brooks P, Woolf A, Blyth F, Vos T, et al. Measuring the global burden of low back pain. Best Pract Res Clin Rheumatol. 2010;24:155–165. Although advice overall had a small or no effect on relieving pain and disability for non-specific spinal pain, we suggest that advice 3. Vos T, Allen C, Arora M, Barber RM, Bhutta ZA, Brown A, et al. Global, regional, and about prognosis and management options should remain part of the national incidence, prevalence, and years lived with disability for 310 diseases and clinician/patient interaction. Our rationale is that clinicians do have a injuries, 1990–2015: a systematic analysis for the Global Burden of Disease Study responsibility to ensure that their patients are adequately informed 2015. Lancet. 2016;388:1545–1602. about their condition, and patients typically want to know about their spinal pain. It is also likely to have other benefits such as providing 4. Geurts JW, Willems PC, Kallewaard J-W, van Kleef M, Dirksen C. The impact of reassurance, which has been shown to reduce anxiety and healthcare chronic discogenic low back pain: costs and patients’ burden. Pain Res Manag. use.59 However, clinicians should be aware that this advice may only 2018;2018:4696180. provide small reductions in pain and disability. 5. Walker BF, Muller R, Grant WD. Low back pain in Australian adults: The economic Most studies included in this review did not assess participant burden. Asia Pac J Public Health. 2003;15:79–87. fidelity with the advice provided (eg, whether the participant heeded the specific advice they received). Therefore, it could not be deter- 6. Schofield DJ, Callander EJ, Shrestha RN, Passey ME, Kelly SJ, Percival R. Back mined whether poor fidelity to the advice was a contributor to the problems, comorbidities, and their association with wealth. Spine J. 2015;15:34–41. small effects seen. Future trials will require a mechanism to 7. Hartvigsen J, Hancock MJ, Kongsted A, Louw Q, Ferreira ML, Genevay S, et al. What low back pain is and why we need to pay attention’. Lancet. 2018;391:2356–2367. 8. Pain Australia. The cost of pain in Australia. Canberra: Deloitte Access Economics. Available at: https://www.painaustralia.org.au/static/uploads/files/the-cost-of-pain- in-australia-final-report-12mar-wfxbrfyboams.pdf. Accessed 14 September, 2020. 9. Oliveira CB, Maher CG, Pinto RZ, Traeger AC, Lin C-WC, Chenot J-F, et al. Clinical practice guidelines for the management of non-specific low back pain in primary care: an updated overview. Eur Spine J. 2018;27:2791–2803.

270 Jones et al: Advice and education for spinal pain 10. Foster N, Anema J, Cherkin D, Chou R, Cohen S, Gross D, et al. Prevention and 37. Cherkin DC, Deyo RA, Street JH, Hunt M, Barlow W. Pitfalls of patient education: treatment of low back pain: evidence, challenges, and promising directions. Lancet. Limited success of a program for back pain in primary care. Spine. 1996;21:345– 2018;391:2368–2383. 355. 11. Stevens ML, Lin CC, de Carvalho FA, Phan K, Koes B, Maher CG. Advice for acute low 38. Damush T, Weinberger M, Perkins S, Rao J, Tierney W, Qi R, et al. The long-term back pain: a comparison of what research supports and what guidelines recom- effects of a self-management program for inner-city primary care patients with mend. Spine J. 2017;17:1537–1546. acute low back pain. Arch Intern Med. 2003;163:2632–2638. 12. Dahm KT, Brurberg KG, Jamtvedt G, Hagen KB. Advice to rest in bed versus advice 39. Damush T, Weinberger M, Perkins S, Rao J, Tierney W, Qi R, et al. Randomized trial to stay active for acute low-back pain and sciatica. Cochrane Database Syst Rev. of a self-management program for primary care patients with acute low back pain: 2010;6:20556780. Short-term effects. Arthrit Care Res. 2003;49:179–186. 13. Hagen KB, Hilde G, Jamtvedt G, Winnem MF. The Cochrane review of advice to stay 40. Darlow B, Stanley J, Dean S, Abbott J, Garrett S, Wilson R, et al. The Fear Reduction active as a single treatment for low back pain and sciatica. Spine. 2002;27:1736– Exercised Early (FREE) approach to management of low back pain in general 1741. practice: A pragmatic cluster-randomised controlled trial. PLoS Med. 2019;16:e1002897. 14. Liddle D, Gracey J, Baxter D. Advice for the management of low back pain: A systematic review of randomised controlled trials. Man Ther. 2007;12:310–327. 41. Derebery M, Giang J, Gatchel W, Erickson W, Fogarty W. Efficacy of a patient- educational booklet for neck-pain patients with workersʼ compensation: a ran- 15. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JPA, et al. The domized controlled trial. Spine. 2009;34:206–213. PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 42. Ibrahim AA, Akindele MO, Ganiyu SO. Motor control exercise and patient education 2009;339:b2700. program for low resource rural community dwelling adults with chronic low back pain: a pilot randomized clinical trial. J Exerc Rehabil. 2018;14:851–863. 16. Cochrane Back and Neck Group. Cochrane Back and Neck Group in their updated Search Strategies from January 2013. https://back.cochrane.org/sites/back. 43. Jellema P, van Der Windt D, van Der Horst H, Twisk J, Stalman W, Bouter L. Should cochrane.org/files/uploads/PDF/CBRG%20Search%20Strategies%20Jan%202013.pdf. treatment of (sub)acute low back pain be aimed at psychosocial prognostic fac- Accessed 4 November, 2020. tors? Cluster randomised clinical trial in general practice. BMJ. 2005;331:84–87. 17. Traeger AC, Lee H, Hübscher M, Skinner IW, Moseley GL, Nicholas MK, et al. Effect 44. Da Silva TMJC, Da Silva NN, de Souza Rocha SH, Marques de Oliveira D, Monte- of intensive patient education vs placebo patient education on outcomes in pa- Silva KK, Da Silva Tenório A, et al. Back school program for back pain: education or tients with acute low back pain: a randomized clinical trial. JAMA Neurol. physical exercise? Conscientiae saúde. 2014;13:506–515. 2019;76:161–169. 45. Kovacs F, Abraira V, Santos S, Díaz E, Gestoso M, Muriel A, et al. A comparison of 18. Blobaum P. Physiotherapy Evidence Database (PEDro). J Med Libr Assoc. two short education programs for improving low back pain-related disability in the 2006;94:477–478. elderly: a cluster randomized controlled trial. Spine. 2007;32:1053–1059. 19. Minozzi S, Cinquini M, Gianola S, Gonzalez-Lorenzo M, Banzi R. The revised 46. Lamb SE, Gates S, Williams MA, Williamson EM, Mt-Isa S, Withers EJ, et al. Cochrane risk of bias tool for randomized trials (RoB 2) showed low interrater Emergency department treatments and physiotherapy for acute whiplash: A reliability and challenges in its application. J Clin Epidemiol. 2020;126:37–44. pragmatic, two-step, randomised controlled trial. Lancet. 2013;381:546–556. 20. Cashin AG, McAuley JH. Clinimetrics: Physiotherapy Evidence Database (PEDro) 47. Little P, Roberts L, Blowers H, Garwood J, Cantrell T, Langridge J, Chapman J. Should Scale. J Physiother. 2020;66:59. we give detailed advice and information booklets to patients with back pain?: a randomized controlled factorial trial of a self-management booklet and doctor 21. Schünemann H, Broz_ ek J, Guyatt G, Oxman A, eds. GRADE handbook for grading advice to take exercise for back pain. Spine. 2001;26:2065–2072. quality of evidence and strength of recommendations. The GRADE Working Group; 2013. Available from: guidelinedevelopment.org/handbook. 48. Matias BA, Vieira I, Pereira A, Duarte M, Silva AG. Pain neuroscience education plus exercise compared with exercise in university students with chronic idiopathic 22. Guyatt GH, Oxman AD, Vist G, Kunz R, Brozek J, Alonso-Coello P, et al. GRADE neck pain. Int J Ther Rehabil. 2019;26:1–14. guidelines: 4. Rating the quality of evidence–study limitations (risk of bias). J Clin Epidemiol. 2011;64:407–415. 49. Pengel LH, Refshauge KM, Maher CG, Nicholas MK, Herbert RD, McNair P. Phys- iotherapist-directed exercise, advice, or both for subacute low back pain: a ran- 23. Guyatt GH, Oxman AD, Kunz R, Woodcock J, Brozek J, Helfand M, et al. GRADE domized trial. Ann Intern Med. 2007;146:787–796. guidelines: 7. Rating the quality of evidence–inconsistency. J Clin Epidemiol. 2011;64:1294–1302. 50. Pires D, Cruz EB, Caeiro C. Aquatic exercise and pain neurophysiology education versus aquatic exercise alone for patients with chronic low back pain: a random- 24. Kulig M, Perleth M, Langer G, Meerpohl JJ, Gartlehner G, Kaminski-Hartenthaler A, ized controlled trial. Clin Rehabil. 2015;29:538–547. et al. GRADE guidelines: 6. Rating the quality of evidence: imprecision. Z Evid Fortbild Qual Gesundhwes. 2012;106:677–688. 51. Rantonen J, Karppinen J, Vehtari A, Luoto S, Viikari-Juntura E, Hupli M, et al. Effectiveness of three interventions for secondary prevention of low back pain in 25. Guyatt GH, Oxman AD, Montori V, Vist G, Kunz R, Brozek J, et al. GRADE guidelines: the occupational health setting - a randomised controlled trial with a natural 5. Rating the quality of evidence–publication bias. J Clin Epidemiol. 2011;64:1277– course control. BMC Public Health. 2018;18:598. 1282. 52. Roberts L, Little P, Chapman J, Cantrell T, Pickering R, Langridge J. The back home 26. Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et al. Better trial: general practitioner-supported leaflets may change back pain behavior. Spine. reporting of interventions: template for intervention description and replication 2002;27:1821–1928. (TIDieR) checklist and guide. BMJ. 2014;348:g1687. 53. Saracoglu I, Arik MI, Afsar E, Gokpinar HH. The effectiveness of pain neuroscience 27. Borenstein M, Hedges L, Higgins J, Rothstein H. Comprehensive Meta-Analysis education combined with manual therapy and home exercise for chronic low back Version 3.3.070. Englewood, NJ: Biostat; 2013. pain: A single-blind randomized controlled trial. Physiother Theory Pract. 2020. https://doi.org/10.1080/09593985.2020.1809046 [ahead of print]. 28. Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, et al. Inter- preting the clinical importance of treatment outcomes in chronic pain clinical 54. Slater H, Briggs AM, Watkins K, Chua J, Smith AJ. Translating evidence for low back trials: IMMPACT recommendations. J Pain. 2008;9:105–121. pain management into a consumer-focussed resource for use in community pharmacies: a cluster-randomised controlled trial. PLoS One. 2013;8:8. 29. Ostelo RW, Deyo RA, Stratford P, Waddell G, Croft P, Von Korff M, et al. Interpreting change scores for pain and functional status in low back pain: towards interna- 55. Tellez-Garcia M, de-la-Llave-Rincon AI, Salom-Moreno J, Palacios-Cena M, Ortega- tional consensus regarding minimal important change. Spine. 2008;33:90–94. Santiago R, Fernandez-de-Las-Penas C. Neuroscience education in addition to trigger point dry needling for the management of patients with mechanical chronic 30. Luo D, Wan X, Liu J, Tong T. Optimally estimating the sample mean from the sample low back pain: a preliminary clinical trial. J Bodyw Mov Ther. 2015;19:464–472. size, median, mid-range, and/or mid-quartile range. Stat Methods Med Res. 2018;27:1785–1805. 56. Werner EL, Storheim K, Lochting I, Wisloff T, Grotle M. Cognitive patient education for low back pain in primary care: a cluster randomized controlled trial and cost- 31. Review Manager (RevMan) [Computer program]. Version 5.3. The Cochrane effectiveness analysis. Spine. 2016;41:455–462. Collaboration; 2019. Available at : revman.cochrane.org. 57. Zhang Y, Wan L, Wang X. The effect of health education in patients with chronic 32. Akca NK, Aydin G, Gumus K. Effect of body mechanics brief education in the clinical low back pain. J Int Med Res. 2014;42:815–820. setting on pain patients with lumbar disc hernia: a randomized controlled trial. Int J Caring Sci. 2017;10:1498–1506. 58. Almeida M, Saragiotto B, Richards B, Maher CG. ‘Primary care management of non- specific low back pain: key messages from recent clinical guidelines. Med J Aust. 33. Albaladejo C, Kovacs FM, Royuela A, Del Pino R, Zamora J. The efficacy of a short 2018;208:272–275. education program and a short physiotherapy program for treating low back pain in primary care: A cluster randomized trial. Spine. 2010;35:483–496. 59. Traeger AC, Hübscher M, Henschke N, Moseley GL, Lee H, McAuley JH. Effect of primary care-based education on reassurance in patients with acute low back pain: 34. Ayanniyi O, Ige OG. Back care education on peasant farmers suffering from chronic systematic review and meta-analysis. JAMA Intern Med. 2015;175:733–743. mechanical low back pain. J Experim Integrative Med. 2015;5:215–221. Websites 35. Bodes Pardo G, Lluch Girbes E, Roussel NA, Gallego Izquierdo T, Jimenez Penick V, Pecos Martin D. Pain neurophysiology education and therapeutic exercise for pa- PEDro www.pedro.org.au tients with chronic low back pain: a single-blind randomized controlled trial. Arch PRISMA www.prisma-statement.org Phys Med Rehabil. 2018;99:338–347. 36. Bucker B, Butzlaff M, Isfort J, Koneczny N, Vollmar HC, Lange S, et al. Effect of written patient information on knowledge and function of patients with acute uncomplicated back pain (PIK Study). Gesundheitswesen. 2010;72:e78–e88.

Journal of Physiotherapy 67 (2021) 291–297 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research An international core capability framework for physiotherapists to deliver quality care via videoconferencing: a Delphi study Luke Davies a, Rana S Hinman a, Trevor Russell b, Belinda Lawford a, Kim Bennell a, International Videoconferencing Steering Group* a Centre for Health, Exercise & Sports Medicine, The University of Melbourne, Melbourne, Australia; b RECOVER Injury Research Centre, The University of Queensland, Brisbane, Australia KEY WORDS ABSTRACT Telehealth Question: What are the core capabilities that physiotherapists need in order to deliver quality care via Physical therapy videoconferencing? Design: A three-round modified e-Delphi survey. Participants: An international Delphi Videoconferencing panel comprising a Steering Group and experts in the field, including physiotherapy researchers, physio- Telemedicine therapy clinicians, representatives of physiotherapy organisations, and consumers. Methods: The draft Rehabilitation framework was developed by the research team and Steering Group, based on relevant documents identified within the literature. The panel considered a draft framework of 73 specific capabilities mapped across eight domains. Over three rounds, panellists rated their agreement (Likert or numerical rating scales) on whether each capability was essential (core) for physiotherapists to deliver quality care via videoconferencing. Those capabilities achieving consensus, defined as 75% of the panel ratings being  7 out of 10 in Round 3, were retained. Results: A total of 130 panellists from 32 countries participated in Round 1, with retention rates of 65% and 60% in Rounds 2 and 3, respectively. The final framework comprised 60 capabilities across seven domains: compliance (n = 7 capabilities); patient privacy and confidentiality (n = 4); patient safety (n = 7); technology skills (n = 7); telehealth delivery (n = 16); assessment and diagnosis (n = 7); and care planning and management (n = 12). Conclusion: This framework outlines the specific core capabilities required of physiotherapists to provide quality care via videoconferencing. The core capability framework provides guidance for physiotherapists to deliver care via videoconferencing and will help inform future development of physiotherapy curricula and professional development initiatives in the delivery of telehealth. [Davies L, Hinman RS, Russell T, Lawford B, Bennell K, International Videoconferencing Steering Group (2021) An international core capability framework for physiotherapists to deliver quality care via videoconfer- encing: a Delphi study. Journal of Physiotherapy 67:291–297] © 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction adherence in congestive heart failure outpatient programs, equivalent Provision of physiotherapy care via telehealth is becoming clinical outcomes in surgical populations and better outcomes in increasingly common and, in some circumstances, a necessity as some musculoskeletal populations such as total hip arthroplasty.5–9 observed during the COVID-19 pandemic.1,2 Telehealth enables physiotherapists to provide care remotely to patients within their As such, telehealth is emerging as an effective and acceptable mode own home/environment or in other facilities and clinics through the of healthcare delivery.10 use of information and communication technologies.3 Videoconfer- encing technology is often used for telehealth because of its acces- Despite evidence supporting the effectiveness and acceptability of sibility, versatility and convenience, as well as the fact that it allows telehealth,11 uptake of telehealth in physiotherapy has been slow for a for visual contact in real time.4 Previous research has demonstrated that telehealth is associated with high rates of patient satisfaction variety of reasons, including lack of telehealth funding, resistance to and, compared with traditional in-person care, reduced re- hospitalisations in people with coronary artery disease, greater changing practice, and lack of physiotherapist confidence, knowledge or skills in telehealth.12 The introduction of social distancing mea- * International Videoconferencing Steering Group: Michael Billings, Carmen Cooper- Oguz, Karen Finnan, Sarah Gallagher, Daniel Kenneth Gilbertson, Lesley sures during the COVID-19 pandemic led to a dramatic uptake in Holdsworth, Anne Holland, Jeremey McAlister, Dan Miles, Robin Roots. telehealth services such as videoconferencing across Australia and worldwide.2,13 For many physiotherapists, this came with limited preparation or training in implementation of videoconferencing ser- vices. In order to effectively and safely deliver care via telehealth, physiotherapists require different skills to those used for in-person consultations and must be able to adapt their usual practice to the https://doi.org/10.1016/j.jphys.2021.09.001 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

292 Davies et al: Physiotherapist capabilities for care delivery via videoconferencing digital environment.12,14 As such, there is a need to provide physio- This draft framework was sent to the Steering Group for review. For therapists with the knowledge, confidence and skills in the delivery each capability, Steering Group members were asked to answer ‘yes’ or of telehealth. World Physiotherapy has highlighted this as a primary ‘no’ as to whether they believed it was relevant to include in the draft regulatory issue that needs to be addressed and they have called for framework. Similarly, for each domain name, members were asked to the upskilling of physiotherapists in the digital environment.14 answer ‘yes’ or ‘no’ as to whether it was suitable and relevant to include in the framework. Members were provided with the oppor- In order to facilitate the training and development of a physio- tunity to suggest edits to capabilities and domains and to also suggest therapy workforce that is skilled in the delivery of care via videocon- additional capabilities they felt had not been captured. Feedback ferencing, it is important to first identify what individual capabilities received from the Steering Group was discussed by the research team physiotherapists need to possess. A capability can be defined as an and resulted in the addition of 20 new capabilities, five amalgamated ‘integration of knowledge, skills, personal qualities and understanding capabilities, one capability separated into two and 28 other capabil- used appropriately and effectively’.15 Development of a framework of ities undergoing edits to wording. Thus, the final draft framework for the core capabilities required by physiotherapists to deliver care via the Delphi panel comprised 73 capabilities across eight domains. videoconferencing will provide physiotherapists and service providers with the knowledge of what is required to deliver quality care using Delphi panel this mode, whilst lending itself to be used as a blueprint to inform curriculum and professional development initiatives, including An international Delphi panel of experts in the physiotherapy field learning outcomes, assessment strategies and graduate attributes. was established to reach consensus on the core capability framework and comprised Steering Group members, physiotherapy researchers Competency frameworks for telehealth delivery have been involved in telehealth research, physiotherapy clinicians with expe- developed for nursing and medical professions.16,17 However, it is rience providing care for patients via videoconference, representa- suggested that frameworks based on capability need to be discipline- tives of a physiotherapy professional organisation, and consumers specific in order to capture the nuances of the specialty area.18 It is with experience of receiving care from physiotherapists via believed that there are currently no frameworks describing the core videoconferencing. capabilities for physiotherapists specifically. Thus, this study aimed to develop a core capability framework for physiotherapists to deliver All panellists were required to understand English. Further eligi- quality care via videoconferencing, using an international consensus bility criteria for researchers were being a qualified and registered process involving consumers, physiotherapy clinicians and re- physiotherapist who had at least one of the following achievements: searchers, and representatives of private health insurers and first or last author on at least three papers relating to telehealth and physiotherapy professional organisations. physiotherapy care; or invited to give a presentation on telehealth at a national or international conference within the last 5 years. Inclu- Therefore, the research question for this Delphi survey was: sion criteria for clinicians were: being registered to practise as a physiotherapist in their country, and having provided care to more What are the core capabilities that physiotherapists need in order than 50 patients via videoconferencing in the last 3 years. Inclusion to deliver quality care via videoconferencing? criteria for representatives of a physiotherapy professional organisa- tion were: currently working in a national or international Methods physiotherapy association/body, and knowledge of physiotherapy professional practice and governance issues. The inclusion criterion Overview for consumers was that they have had at least four consultations with a physiotherapist via videoconferencing in the past 12 months. This framework was developed through the lens of individual physiotherapists working potentially without the support of admin- A list of potential panel participants was compiled by the research istrative staff in their clinical practice setting. A Delphi panel was team and Steering Group, drawing on their academic, research and established and a three-round modified e-Delphi survey was con- clinical networks, as well as an internet search of physiotherapy cli- ducted between August and November 2020 to achieve expert nicians delivering telehealth, physiotherapy organisations interna- consensus on the core capability framework. tionally and physiotherapy researchers in the telehealth field. Invitations were sent to 2,315 potential panel participants via email. A Previously established methodological criteria for reporting Del- snowball method was used by inviting potential panellists to forward phi studies were used to ensure quality.19,20 The study was overseen the invitation on to their colleagues or anyone else in their network by an international Steering Group assembled by the research team. whom they believed could be eligible. We also advertised for addi- The group comprised 10 members with an interest in telehealth from tional panellists on social media (Facebook, Instagram and Twitter). Australia (n = 4), Canada (n = 4), United States of America (n = 1) and Potential panellists completed a series of screening questions at the Scotland (n = 1), and involved physiotherapy clinicians (n = 5) and beginning of the Round 1 e-Delphi survey to ensure eligibility criteria academics (n = 3), a policy representative from the Australian were met. Ineligible people were excluded from participating. For Physiotherapy Association and a representative from an Australian subsequent Delphi rounds, only those participants who had private health insurer. Ethics approval was obtained from the completed the previous Delphi round were emailed the survey. University of Melbourne. Figure 1 outlines the study phases. e-Delphi survey Survey development The Delphi panel was asked to rate each of the 73 capabilities An initial list of 55 potential core capabilities was generated by the across the eight domains. An electronic survey was developed using research team, using a number of relevant documents identified from an online software toola and administered iteratively over three a search of the literature and telehealth guidelines. These documents rounds. Round 1 was open for 4 weeks and subsequent rounds were included telehealth frameworks from the Academy of Medical Royal open for 2 weeks. Three reminder emails were sent over that time to Colleges21 and National Initiative for Telehealth Framework of non-responders, to encourage completion. Participants took approx- Guidelines,22 a Position Statement on tele-practice in speech from imately 20 to 30 minutes to complete each round of the survey. Speech Pathology Australia23 and five research articles about com- petencies required for delivering digital healthcare.16,24–27 These ca- Round 1 pabilities were then organised into eight domains: compliance; In Round 1, panellists were asked to rate each of the specific patient privacy and confidentiality; patient safety; technology skills; telehealth delivery; assessment and diagnosis; care planning and capabilities as ‘unimportant’, ‘important’ or ‘essential’ for management; and quality assurance.

Research 293 physiotherapists in the delivery of quality care via videoconferencing. did not reach consensus for retention in the final core capability The panel was invited to suggest additional capabilities that were not framework and were removed. already included in the draft framework via a free-text box option of the survey. Individual capabilities that reached panel consensus were Final phase retained for further consideration in Round 2. Consensus was defined as 75% (median threshold for consensus19) of the panel agreeing that When the research team reviewed the final framework after the capability was either ‘important’ or ‘essential’. Round 3, 13 capabilities were identified across the domains as having similar meanings and as a result were amalgamated into five capa- Round 2 bilities (Appendix 5). When the amalgamated capabilities were sent In Round 2, panellists were asked to reconsider and re-rate the to the Steering Group, the ‘quality assurance’ domain was made redundant, with the final capability being merged into the ‘telehealth capabilities from Round 1, as well as rate any new capabilities sug- delivery’ domain, and minor word edits were made to five capabil- gested by the panel in Round 1. To assist in this process, summary ities (Appendix 6). The final core capability framework comprised 60 panel data from Round 1 were displayed alongside each capability specific capabilities across seven domains (Table 2). (presented as n (%) across the three response categories) showing the proportion of agreement across panellists. For this round, the pan- Discussion ellists were asked to rate how strongly they agreed or disagreed that each capability should be included as a core capability for physio- This study aimed to develop an internationally-relevant, disci- therapists to deliver quality care via videoconferencing. Panellists pline-specific core capability framework for physiotherapists to rated their level of agreement on an 11-point numerical scale (with deliver quality care via videoconferencing. The final framework terminal anchors of 0 = strongly disagree and 10 = strongly agree). comprised 60 specific capabilities mapped across the seven domains Individual capabilities that achieved consensus (at least 75% of the shown in Table 2. This framework provides a model that communi- panel rating  6) were retained for Round 3.28 cates the key capabilities required of physiotherapists to ensure quality care via videoconferencing. The intention of this framework is Round 3 not to dictate how or when these capabilities should be implemented In Round 3, panellists were asked to reconsider and re-rate the in clinical practice. Instead, this framework provides physiotherapists with best practice recommendations about the skills and knowledge capabilities from Round 2 using the same rating scale as Round 2. required for delivery of care via videoconferencing. Summary panel data from Round 2 were presented for consideration alongside each capability in the format: number (%) of participants It is believed that no previous studies have examined the capa- rating in each of the ranges 0 to 4, 5 to 6, 7 to 8, or 9 to 10. Only those bilities required by physiotherapists to deliver quality care via capabilities that achieved consensus (at least 75% of the panel rating videoconferencing. However, our findings can be compared to other  7) were retained for inclusion in the final core capability previously developed telehealth capability frameworks for other framework.28 healthcare professions. A recent study of expert consensus from a panel of 11 leaders in telehealth investigated the skills required by Final phase medical healthcare professionals to deliver care via telehealth.17 The In the final phase, the research team (LD, RH, BL, TR, KB) reviewed skills and domains identified within this study share similarities with ours, including considerations around patient safety and privacy, the final framework for redundancies and refined wording across the performing remote clinical assessments and the use of information domains. Capabilities that were similar in nature were merged technology in the delivery of care.17 Similar findings have been re- together. Amalgamated capabilities were sent to the Steering Group ported in other studies on the development of core competency for review and agreement. The final framework was reviewed by the frameworks for physicians in the delivery of virtual health,25 nurses Steering Group to approve the edits and merged capabilities. delivering telehealth16 and generic guidelines around tele- rehabilitation delivery,26 which all identified domains relating to the Results implementation and delivery of telehealth, technology skills for both the patient and clinician and patient privacy. Characteristics of the panellists Points of difference between our framework and those discussed Table 1 describes the characteristics of the Delphi panel. In Round above16,17,25 are the additional domains with specific capabilities for 1, 130 participants from 32 countries participated, with more than assessment, diagnosis, delivery, care planning and management half of the panel (n = 79, 61%) being physiotherapy clinicians and included in our framework. One explanation for this may be because of seven (5%) being consumers. We retained 85 panellists for Round 2 the subtle differences that exist across healthcare professionals’ scope and 78 panellists for Round 3, representing 64% and 60% retention of of practice and the type of care they deliver to patients. Traditionally, Round 1 panellists, respectively. to perform an assessment of a patient, physiotherapists will often use a ‘hands-on’ approach.29 Physiotherapists may also instinctively Delphi Rounds 1 to 3 communicate through therapeutic touch in order to convey empathy and understanding to patients who are showing signs of pain.30 This Figure 1 provides a summary of each Delphi Round. In Round 1, physical aspect is removed when consulting via videoconferencing only one (1%) capability did not reach consensus as being ‘essential’ and requires physiotherapists to alter the way in which they would or ‘important’ for inclusion in the core capability framework for normally work, which may explain why additional domains are found physiotherapists to deliver quality care via videoconferencing and in our framework relative to some other frameworks. was excluded from Round 2 (Appendix 1). Additionally, 42 individual panellists provided 86 separate items of feedback/additional capa- Our framework highlights the most important factors that need to bilities for consideration. Of these, seven items were regarding be considered by physiotherapists and service providers when of- generic physiotherapy capabilities not specific to telehealth, 20 were fering a telehealth service. The domain of ‘telehealth delivery’ spans alternative suggestions for wording of capabilities already included in 16 capabilities, the largest of the seven domains, followed by 12 ca- the framework and 43 were general comments/observations about pabilities for ‘care planning and management’ and seven capabilities telehealth and the survey rather than capabilities per se. From the for ‘assessment and diagnosis’. This suggests that the delivery of remainder, an additional 11 unique capabilities were generated for telehealth (such as instructing patients to set up their camera angle inclusion in Round 2 and wording amendments were made to five placement to optimise assessment and treatment) needs to be taken capabilities in the framework (Appendix 2). In Round 2, 11 (13%) ca- into consideration when implementing physiotherapy services via pabilities (Appendix 3) did not reach consensus and were removed videoconferencing. For example, in-person care potentially allows for from the final round. In Round 3, four (5%) capabilities (Appendix 4)

294 Davies et al: Physiotherapist capabilities for care delivery via videoconferencing Table 1 Research team (LD, RH, BL, TR, KB) Characteristics of the participants in the Delphi panel. identified and analysed relevant documents • 2 frameworks Participants Round 1 Round 2 Round 3 • 1 position statement (n = 130) (n = 85) (n = 78) • 5 research articles Generation of initial draft framework Panellist classification, n (%) 22 (17) 18 (21) 17 (22) consisting of 55 capabilities researcher 79 (61) 48 (56) 43 (55) clinical physiotherapist 21 (17) 12 (15) 11 (15) Initial draft framework sent to international physiotherapy association representative Steering Group Australian private health insurer 1 (1) 1 (1) 1 (1) consumer 7 (5) 6 (7) 6 (7) Steering Group provided feedback Feedback analysed by research team Sex, n (%) 50 (38) 31 (36) 28 (36) male 80 (62) 54 (64) 50 (64) Modifications female • 28 capabilities edited 43 (33) 35 (41) 34 (44) • 1 capability separated into 2 Country of residence, n (%) 1 (1) 1 (1) 1 (1) • 5 capabilities amalgamated into 2 Australia 1 (1) 1 (1) 1 (1) • 20 new capabilities added Belgium 1 (1) 0 (0) 0 (0) Final draft framework consisting of 73 Bhutan 11 (8) 11 (13) capabilities across 8 domains Brazil 1 (1) 1 (1) 10 (13) Canada 1 (1) 1 (1) 1 (1) Delphi Survey Round 1 Chile 1 (1) 1 (1) 1 (1) • 72 capabilities reached > 75% consensus China 1 (1) 0 (0) 0 (0) • 1 capability removed Denmark 1 (1) 0 (0) 0 (0) • 11 capabilities added Egypt 2 (2) 1 (1) 0 (0) Finland 7 (5) 0 (0) 1 (1) Delphi Survey Round 2 Greece 2 (2) 1 (1) 0 (0) • 72 capabilities reached > 75% consensus India 2 (2) 0 (0) 1 (1) • 11 capabilities removed Italy 2 (2) 1 (1) 0 (0) Kenya 1 (1) 0 (0) 1 (1) Delphi Survey Round 3 Malta 2 (2) 1 (1) 0 (0) • 68 capabilities reached > 75% consensus Montenegro 1 (1) 1 (1) 1 (1) • 4 capabilities removed Nepal 1 (1) 1 (1) 1 (1) Netherlands 1 (1) 1 (1) 1 (1) New Zealand 1 (1) 1 (1) 1 (1) Norway 2 (2) 1 (1) 0 (0) Paraguay 1 (1) 1 (1) 1 (1) Philippines 1 (1) 0 (0) 1 (1) Poland 1 (1) 0 (0) 0 (0) Romania 1 (1) 1 (1) 0 (0) Saudi Arabia 1 (1) 1 (1) 1 (1) Suriname 1 (1) 1 (1) 1 (1) Sweden 8 (6) 5 (6) 1 (1) Taiwan 5 (6) United Kingdom 27 (21) 14 (16) 11 (14) United States 2 (2) 1 (1) 1 (1) Zambia 1 (1) 1 (1) 1 (1) Zimbabwe better conversational flow and the ability to provide hands-on ther- Final phase apy with a patient if required as part of the treatment plan. However, • 13 similar capabilities amalgamated into 5 delivering care via videoconferencing requires a different approach31 • 5 capabilities edited slightly such as modulating the communication style to accommodate for any • 1 domain made redundant latency with the technology to optimise conversational flow, assisting patients to set up their camera for optimal viewing of the body part Final core capability framework comprising 60 required for assessment/treatment, and providing clear instructions capabilities across 7 domains with visual aids when delivering exercises. There is currently no research outlining which capabilities physiotherapists may need the Figure 1. Development of the core capability framework. most training or support with. Further research investigating physiotherapists’ confidence in each capability may provide valuable COVID-19 revealed that only 21% of the 688 physiotherapists involved information to inform future training programs and to inform entry- agreed that they had been trained to deliver telehealth to those with to-practice curricula regarding telehealth. musculoskeletal conditions,32 further highlighting the need for edu- cation and training in the digital environment. Our framework has relevance to and can be used by a range of stakeholders, including individual physiotherapists, education/ Currently, curricular guidelines and standards are lacking for the training providers, health insurers and the wider community. The provision of care in the digital environment within entry-level and implementation of this framework provides individual physiotherapy postgraduate-level education across healthcare professions.14,33 Thus, clinicians who have little or no experience in the digital environment, our framework provides an opportunity to inform future develop- with a blueprint to what knowledge and skills are needed to deliver ment of physiotherapy curricula in the delivery of telehealth, appropriate care via videoconferencing. Internationally, the rapid providing assessment strategies and learning outcomes to better implementation of telehealth with the onset of the COVID-19 prepare graduates in tertiary education settings and to upskill prac- pandemic and the inexperience of many clinicians demonstrated a tising physiotherapists through continuing education courses. need for further education and upskilling of physiotherapists to Although established for physiotherapists, the current framework produce clinicians with appropriate understanding and skills in the could be used as a foundation for other allied health professions and delivery of care in the digital environment.14 A recent study investi- adapted to suit their own discipline. This framework may also be used gating the views of telehealth among allied health clinicians during

Research 295 Table 2 The core capability framework for physiotherapists to deliver quality care via videoconferencing. Domain 1: Compliance Physiotherapist demonstrates the ability to. a) identify any limitation to their individual scope of telehealth (videoconferencing) practice as dictated by relevant laws, registration requirements, organisational regulation, and/or the funding/reimbursement model relevant to the patient b) comply with the regulatory requirements associated with practising as a physiotherapist in the practitioners’ geographical location, the geographical restrictions associated with their professional registration and the geographical location(s) of the patient c) have professional indemnity insurance that covers the intended scope of telehealth (videoconferencing) practice d) determine a patient’s eligibility for receiving care via telehealth in accordance with federal and state regulations and/or the funding/reimbursement model relevant to the individual patient e) obtain and document informed consent from the patient and/or helper that is appropriate for the intended telehealth (videoconferencing) interactions f) align practice with relevant organisational telehealth procedures and protocols g) record and manage clinical documentation about telehealth (videoconferencing) interactions in accordance with professional association standards, state/federal regulations and medico-legal requirements Domain 2: Patient privacy and confidentiality Physiotherapist demonstrates the ability to. a) set up their and the patient’s physical environment in order to maintain patient privacy b) obtain informed consent from the patient if videos or photos are taken during the telehealth (videoconferencing) interaction (such as for assessment purposes), and explain how these will be used and stored c) inform the patient/caregiver that physiotherapist consent is required for them to take photos/videos of the consultation d) comply with the data security requirements of telehealth (videoconferencing) practice, platforms, storage and transmission (including sharing information with other health professionals) as dictated by bodies such as federal/state/professional and/or employer organisation (eg, for USA physiotherapists, The Health Insurance Portability and Accountability Act requires technical, physical and administrative safeguards; for Australia, the Australian Privacy Principles; and for Europe, the General Data Protection Regulation) Domain 3: Patient safety Physiotherapist demonstrates the ability to. a) determine whether a patient is safe to receive care via telehealth (videoconferencing), taking into consideration a patient’s health and physical environment b) inform the patient of potential risks, benefits and limitations associated with the delivery of telehealth (videoconferencing) c) describe a documented procedure in the case of a patient incident during the telehealth (videoconferencing) consultation, including being able to provide a patient’s address to emergency services if required and/or notify the patient’s emergency contact d) confirm the geographical address of the patient at the beginning of each consultation in case emergency services need to be called e) identify safety hazards related to remote care where the therapist is not in the same room as the patient f) enlist the assistance of a patient caregiver to assist with physical assessment and management tasks in order to ensure patient safety when required g) instruct the patient to set up the physical environment in a manner that is safe for performing the intended assessment and management tasks Domain 4: Technology skills Physiotherapist demonstrates the ability to. a) assess the digital literacy of the patient and suitability for a telehealth (videoconferencing) interaction b) determine if the patient has appropriate IT hardware (eg, laptop, tablet device, smart phone) to enable the delivery of the telehealth (videoconferencing) consultation c) select appropriate fit-for-purpose telehealth (videoconferencing) technology that is compliant with data security requirements d) select (if possible) a telehealth (videoconferencing) platform that is suitable for the intended assessment and management task e) competently use relevant functions of the telehealth (videoconferencing) platform to optimise delivery of care f) instruct the patient on how to use the key features of the telehealth (videoconferencing) platform g) understand and identify the potential problems and/or technical issues likely to be encountered by the patient when using the telehealth (videoconferencing) platform, and be able to assist the patient to deal with such problems Domain 5: Telehealth delivery Physiotherapist demonstrates the ability to. a) instruct the patient on how to connect to the telehealth (videoconferencing) consultation, including information about time of appointment, contact details of the physiotherapist, and provision of information for setting up technology prior to first consultation b) enact an appropriate procedure for alternative mode of contact with the patient in the event of technical/communication disruption c) set up their own physical environment ensuring optimal lighting, so the therapist is clearly visible to the patient d) set up their own physical environment ensuring optimal acoustics, including silencing and electronic notifications on the device used e) set up the camera angle so that the therapist is in centre frame with the head and shoulders visible, allowing for eye contact with the patient f) adjust their own camera angle to include the whole body or all equipment and/or props for demonstration purposes g) instruct the patient how to set up the physical environment, ensuring optimal lighting and an uncluttered neutral background (if possible), so he/she is clearly visible to the therapist h) instruct the patient how to set up the physical environment to optimise acoustics i) instruct the patient to set up the camera angle so that he/she is in centre frame with the head and shoulders visible, allowing for eye contact j) instruct the patient to set up the camera angle to visualise other patient assessment and treatment tasks appropriately (eg, walking, exercise performance) as required k) demonstrate telehealth (videoconferencing) etiquette when speaking, such as turn taking to optimise conversational flow l) modulate communication style, including clear enunciation, slower pace and lengthened pauses to reduce overlap m) utilise other means of instructions outside of verbal, such as use of hands to demonstrate angles of movements, other props to help convey the instructions n) provide written or digital information to the patient, as required, to support delivery of care o) encourage patient positive beliefs about telehealth to maximise adherence to treatment p) use the findings of evaluation to continuously improve the telehealth (videoconferencing) service Domain 6: Assessment and diagnosis Physiotherapist demonstrates the ability to. a) follow a structured process to ensure patient appropriateness for telehealth (videoconferencing) for the individual patient b) follow a structured process to identify risk of falls or other safety considerations prior to consultation c) recognise the limitations of telehealth (videoconferencing) in assessment and diagnosis d) adapt assessment processes (if required) to appropriately assess the patient via telehealth (videoconferencing) e) instruct and/or demonstrate the patient and/or helper (using videos and/or images where appropriate) on how to perform modified special tests for assessment and diagnosis if required f) determine the elements of care suitable for delivery via telehealth (videoconferencing) for the individual patient g) recognise when an in-person consultation and/or other investigations are required to supplement the telehealth assessment and/or diagnosis

296 Davies et al: Physiotherapist capabilities for care delivery via videoconferencing Table 2. Continued Domain 7: Care planning and management Physiotherapist demonstrates the ability to. a) identify and interpret the evidence for physiotherapy via telehealth (videoconferencing) b) facilitate patient choice in choosing telehealth (videoconferencing or telephone) or in-person consultation c) use the existing evidence base to deliver treatments that have been shown to have equivalence to in-person treatment d) critically apply relevant clinical practice guidelines and other best available evidence on telehealth (videoconferencing) care and service delivery, identifying where local modifications may be required e) effectively and safely adapt (if required) and deliver treatment approaches using telehealth (videoconferencing) f) develop a patient-centred management plan, which considers the digital literacy of the individual and whether a blended approach combining telehealth (videoconferencing or telephone) and in-person delivery of care is needed g) consider and use, as appropriate, written and digital resources to enhance information, sharing with the patient to increase knowledge about the condition, management options and prognosis h) proficiently use the relevant features of the chosen telehealth (videoconferencing) platform and other supporting digital tools, as appropriate, to provide effective telehealth treatment for the patient i) adapt (to the patient’s environment) and implement relevant outcome measures to monitor treatment progress to guide ongoing telehealth (videoconferencing) management j) identify opportunities for and engage in interprofessional care and collaboration via technology, where possible k) adhere to privacy, security legislative requirements when using digital mechanisms to communicate with other healthcare professionals about a patient l) provide an alternative treatment if the person is not appropriate for care delivered via telehealth (videoconferencing) by the wider public to understand what standards they should expect Footnotes: a SurveyGizmo, Alchemer LLC, Louisville, USA. from physiotherapists delivering care via videoconferencing. eAddenda: Appendices 1 to 6 can be found online at https://doi. org/10.1016/j.jphys.2021.09.001. Strengths of our study include a large Delphi panel with accept- Ethics approval: This research project was approved by the Hu- able retention rates. It assembled 130 experts and retained more than man Research Ethics Committee of The University of Melbourne half of the panel throughout the three survey rounds. While there is (2057318.1). no ideal or recommended number when it comes to the size of a Competing interests: Nil. Delphi panel, it is suggested that typical panels are between 10 to 100 Source(s) of support: This research was funded by the National panellists.20 Given the specific scope of our investigation, the Delphi Health and Medical Research Council (NHMRC) Centre of Research panel number may therefore be considered large. Another strength Excellence #1079078. RSH is supported by a National Health and was its international reach and discipline-specific breadth, Medical Research Council Senior Research Fellowship (#1154217). comprising experts from 22 countries, across seven continents and Collaborators: International Videoconferencing Steering Group: including a range of physiotherapy clinicians, researchers, represen- Michael Billings, Carmen Cooper-Oguz, Karen Finnan, Sarah Gal- tatives of physiotherapy associations, and consumers. However, the lagher, Daniel Kenneth Gilbertson, Lesley Holdsworth, Anne Holland, panel did not include experts from linguistically diverse backgrounds, Jeremey McAlister, Dan Miles, Robin Roots. while only 10% were from low-middle income economies. This may Provenance: Not Invited. Peer Reviewed. have resulted in not capturing all relevant capabilities, and limits the Correspondence: Kim Bennell, Centre for Health, Exercise and generalisability of our framework to those populations where the Sports Medicine, Department of Physiotherapy, The University of physiotherapy workforce and the healthcare contexts may differ. The Melbourne, Melbourne, Australia. Email: [email protected] under-representation of consumers in this study may also be considered a limitation. However, the breadth of our panel does References ensure that the current framework is relevant to physiotherapy practice internationally. 1. Lee AC. COVID-19 and the advancement of digital physical therapist practice and telehealth. Phys Ther. 2020;100:1054–1057. In conclusion, this framework outlines the specific core capabil- ities required of physiotherapists to provide high-quality care via 2. Dantas LO, Barreto RPG, Ferreira CHJ. Digital physical therapy in the COVID-19 videoconferencing, over and above generic physiotherapy capabil- pandemic. Braz J Phys Ther. 2020;24:381. ities. The core capability framework provides guidance for the knowledge and skills required by physiotherapists to deliver care via 3. Guise V, Wiig S. Perceptions of telecare training needs in home healthcare services: videoconferencing. This framework can help in the upskilling of a focus group study. BMC Health Serv Res. 2017;17:164. physiotherapists in the digital environment by informing future development of physiotherapy curricula and professional develop- 4. Hinman RS, Lawford BJ, Bennell KL. Harnessing technology to deliver care by ment initiatives in the delivery of telehealth. physical therapists for people with persistent joint pain: telephone and video- conferencing service models. J Appl Biobehav Res. 2019;24:e12150. What was already known on this topic: Provision of care via telehealth is emerging as an effective and acceptable mode 5. Galea MD. Telemedicine in rehabilitation. Phys Med Rehabil Clin N Am. of delivering physiotherapy. The provision of physiotherapy via 2019;30:473–483. telehealth is associated with high rates of patient satisfaction, with clinical effects sometimes surpassing those of in-person 6. van Egmond MA, van der Schaaf M, Vredeveld T, Vollenbroek-Hutten MM, van care. Although the COVID-19 pandemic led to a dramatic up- Berge Henegouwen MI, Klinkenbijl JH, et al. Effectiveness of physiotherapy with take in telehealth delivery of physiotherapy, this sometimes telerehabilitation in surgical patients: a systematic review and meta-analysis. occurred with limited training and preparation. Physiotherapy. 2018;104:277–298. What this study adds: This framework outlines the specific core capabilities that an international panel of experts recom- 7. Lawford BJ, Bennell KL, Hinman RS. Consumer perceptions of and willingness to mend for physiotherapists to provide quality care via videocon- use remotely delivered service models for exercise management of knee and hip ferencing. The capabilities cover the domains of compliance, osteoarthritis: a cross-sectional survey. Arthritis Care Res. 2017;69:667–676. patient privacy/confidentiality, patient safety, technology skills, telehealth delivery, assessment/diagnosis, and care planning/ 8. Lawford BJ, Bennell KL, Kasza J, Hinman RS. Physical therapists’ perceptions of management. telephone-and internet video-mediated service models for exercise management of people with osteoarthritis. Arthritis Care Res. 2018;70:398–408. 9. Cottrell MA, Galea OA, O’Leary SP, Hill AJ, Russell TG. Real-time telerehabilitation for the treatment of musculoskeletal conditions is effective and comparable to standard practice: a systematic review and meta-analysis. Clin Rehabil. 2017;31:625–638. 10. Nesbitt TS, Katz-Bell J. History of Telehealth. In: Rheuban KS, Krupinski EA, eds. Understanding Telehealth. New York: McGraw-Hill Education; 2018. 11. Iacono T, Stagg K, Pearce N, Hulme Chambers A. A scoping review of Australian allied health research in ehealth. BMC Health Serv Res. 2016;16:543. 12. Cottrell MA, Russell TG. Telehealth for musculoskeletal physiotherapy. Muscu- loskelet Sci Pract. 2020;48:102193. 13. Ballard E, Gallegos DF, Kluchurosky L, Scifers JR. Telehealth. Athl Train Sports Health Care. 2020;12:150–153.

Research 297 14. World Confederation for Physiotherapy, International Network of Physiotherapy 25. Sharma R, Nachum S, Davidson KW, Nochomovitz M. It’s not just Face- Regulatory Authorities. Report of the WCPT/INPTRA digital physical therapy practice Time: core competencies for the medical virtualist. Int J Emerg Med. task force. 2019. 2019;12:1–5. 15. Stephenson J, Yorke M. Capability and Quality in Higher Education. London: Rout- 26. Brennan D, Tindall L, Theodoros D, Brown J, Campbell M, Christiana D, ledge; 1998. et al. A blueprint for telerehabilitation guidelines. Int J Telerehabil. 2010;2:31–34. 16. van Houwelingen CTM, Moerman AH, Ettema RGA, Kort HSM, ten Cate O. Com- petencies required for nursing telehealth activities: A Delphi-study. Nurse Educ 27. Richmond T, Peterson C, Cason J, Billings M, Terrell EA, Lee AC, et al. American Today. 2016;39:50–62. Telemedicine Association’s principles for delivering telerehabilitation services. Int J Telerehabil. 2017;9:63–68. 17. Galpin K, Sikka N, King SL, Horvath KA, Shipman SA. Expert consensus: telehealth skills for health care professionals. Telemed J E Health. 2020;27:820–824. 28. Hinman RS, Allen KD, Bennell KL, Berenbaum F, Betteridge N, Briggs AM, et al. Development of a core capability framework for qualified health professionals to 18. Bromley P. A paradigm shift from competence to capability in neonatal nursing. optimise care for people with osteoarthritis: an OARSI initiative. Osteoarthritis J Neonatal Nurs. 2019;25:268–271. Cartilage. 2020;28:154–166. 19. Diamond IR, Grant RC, Feldman BM, Pencharz PB, Ling SC, Moore AM, et al. 29. Roger J, Darfour D, Dham A, Hickman O, Shaubach L, Shepard K. Physiotherapists’ Defining consensus: A systematic review recommends methodologic criteria for use of touch in inpatient settings. Physiother Res Int. 2002;7:170–186. reporting of Delphi studies. J Clin Epidemiol. 2014;67:401–409. 30. Hiller A, Guillemin M, Delany C. Exploring healthcare communication 20. Avella JR. Delphi panels: Research design, procedures, advantages, and challenges. models in private physiotherapy practice. Patient Educ Couns. 2015;98:1222– Int J Dr Stud. 2016;11:305–321. 1228. 21. Academy of Medical Royal Colleges. E-Health Competency Framework - Defining 31. Hinman R, Nelligan R, Bennell K, Delany C. “Sounds a bit crazy, but it was almost the role of the Expert Clinician; 2011. https://www.aomrc.org.uk/reports-guidance/ more personal”: a qualitative study of patient and clinician experiences of physical ehealth-competency-framework-0611/. Accessed 28 March, 2020. therapist-prescribed exercise for knee osteoarthritis via Skype. Arthritis Care Res. 2017;69:1834–1844. 22. MacDonald-Rencz S, Cradduck T, Parker-Taillon D. The national initiative for tele- health guidelines. Telemed J E Health. 2004;10:113–114. 32. Malliaras P, Merolli M, Williams C, Caneiro J, Haines T, Barton C. ‘It’s not hands-on therapy, so it’s very limited’: Telehealth use and views among allied health clini- 23. Speech Pathology Australia. Telepractice in Speech Pathology [Position statement]; cians during the coronavirus pandemic. Musculoskelet Sci and Pract. 2020. https://www.speechpathologyaustralia.org.au/SPAweb/Members/Position_ 2021;52:102340. Statements/spaweb/Members/Position_Statements/Position_Statements.aspx?hkey= dedc1a49-75de-474a-8bcb-bfbd2ac078b7. Accessed 27 March, 2020. 33. Jonas CE, Durning SJ, Zebrowski C, Cimino F. An interdisciplinary, multi- institution telehealth course for third-year medical students. Acad Med. 2019;94: 24. Hilty DM, Maheu MM, Drude KP, Hertlein KM. Telebehavioral health, telemental 833–837. health, e-Therapy and e-Health competencies: the need for an interprofessional framework. J Technol Behav Sci. 2017;2:171–189.

Journal of Physiotherapy 67 (2021) 314 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Appraisal of Clinical Practice Guideline: Early intervention for children aged 0 to 2 years with or at high risk of cerebral palsy: International clinical practice guideline based on systematic reviews Date of latest update: 17 May 2021. Date of next update: Not indicated. Patient eating and drinking (9 to 10); Complications prevention: vision (11 to 13), sleep group: Children aged 0 to 2 years with or at high risk of cerebral palsy. Intended (14 to 20), muscle tone (21), musculoskeletal impairments (22 and 23); and audience: Health professionals caring for neonates, infants and toddlers. Addi- Parental support: mental health and parenting (24 to 28). Three best-practice tional versions: Nil. Expert working group: Nine domain-specific working principles were supported: Early intervention: children diagnosed with or at groups of 42 health professionals, researchers and consumer representatives. high risk of cerebral palsy should be referred immediately for cerebral palsy- Funded by: Fulbright Senior Scholar Award and the Mariani Foundation of specific and age-specific interventions to optimise the critical period of plas- Milan. Consultation with: Co-chairs were representatives of the Cerebral Palsy ticity of systems; Goal-directed intervention: goals should be set that are task- Alliance Research Institute, Australia. Approved by: Lead authors C.M., L.F., and specific and context-specific, at the appropriate level of challenge and regu- I.N. Location: Guideline and supplementary material: https://doi.org/10.1001/ larly updated; and Caregiver engagement: clinicians should support parents and jamapediatrics.2021.0878. Description: This guideline was conducted to deter- caregivers to build parental capacity and expertise, prioritising a positive parent- mine evidence-based recommendations for early intervention for children aged child relationship. 0 to 2 years who have or are at high risk of cerebral palsy. Data were sought across nine domains: motor, cognitive, communication, eating/drinking, vision, Provenance: Invited. Not peer reviewed. sleep, musculoskeletal, mental health and parenting. Systematic reviews and randomised clinical trials were appraised using A Measurement Tool to Assess Emre Ilhana and Leanne M Johnstonb Systematic Reviews (AMSTAR) or Cochrane Risk of Bias tools. Recommendations aDepartment of Health Sciences, Faculty of Medicine, were formed using the Grading of Recommendations Assessment, Development, Health and Human Sciences, Macquarie University, Sydney and Evaluation (GRADE) framework and reported according to the Appraisal of Guidelines, Research and Evaluation (AGREE) II instrument. Recommendations bSchool of Health & Rehabilitation Sciences, were based on desirable/undesirable consequences of interventions, evidence The University of Queensland, Brisbane, Australia quality, family preferences and costs. Twenty-eight recommendations (24 for and four against) were presented in three categories: Skills development: motor https://doi.org/10.1016/j.jphys.2021.08.011 skills (recommendations 1 to 4), cognition (5 to 6), communication (7 to 8), 1836-9553/© 2021 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). Appraisal of Clinical Practice Guideline: Physical Therapy Evaluation and Treatment After Concussion/Mild Traumatic Brain Injury Date of latest update: April 2020. Date of next update: Within 5 years. and motor function. Key recommendations are provided in three parts: Patient group: Individuals aged  8 years with concussion/mild traumatic screening and diagnosis (screening for all individuals with a potential history brain injury, who have at most ‘mild’ cognitive impairment before or after the of concussive events, screening for indicators of emergency conditions and concussive event, can self-report symptoms and can respond to provocative differential diagnosis); examination (systems to be examined, sequencing of tests. Intended audience: Physiotherapists. Expert working group: Repre- examination based on the levels of irritability, classification of examination sentatives from the American Physical Therapy Association, Academy of Or- findings into impairment profiles, psychological and sociological factors and thopaedic Physical Therapy, American Academy of Sports Physical Therapy, outcome measure selection); and interventions (communication and educa- Academy of Neurologic Physical Therapy and Academy of Pediatric Physical tion, movement-related impairments, cervical musculoskeletal, vestibulo- Therapy. Funded by: American Physical Therapy Association grant, and oculomotor, exertional tolerance and aerobic exercise, motor function and sponsorship from the Academy of Orthopaedic Physical Therapy, American monitoring and progressing patients). Referral for further consultation is Academy of Sports Physical Therapy and Academy of Neurologic Physical recommended for patients with persistent migraine-type or other chronic Therapy for training, travel, software and librarian assistance. Consultation headaches; impairments in vision, hearing, sleep, mental health or cognition; with: Guidance from methodologists in the field. Approved by: Not or to exclude another potential medical diagnosis that may present with mentioned. Location: J Orthop Sports Phys Ther. 2020;50(4):CPG1–CPG73. concussion-like symptoms. https://doi.org/10.2519/jospt.2020.0301. Description and key recommenda- tions: This guideline aims to guide physiotherapists’ clinical decision making Provenance: Invited. Not peer reviewed. for individuals who have experienced a concussive event resulting in movement-related symptoms, impairments and functional limitations. Armaghan Dabbagh and Joy C MacDermid Strength and quality of the evidence were reviewed. The guideline covers School of Physical Therapy, Faculty of Health Sciences, Elborn College, concussion incidence, risk factors for prolonged recovery, physiotherapy ex- amination and intervention strategies. Decision trees are provided to facilitate Western University, London, Canada the triage process to determine priorities and sequencing of activities. Four overarching impairments are discussed within the physiotherapy scope of https://doi.org/10.1016/j.jphys.2021.08.010 practice: cervical musculoskeletal, vestibulo-oculomotor, exertional tolerance 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/ 4.0/).

Journal of Physiotherapy 67 (2021) 314 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Appraisal of Clinical Practice Guideline: Early intervention for children aged 0 to 2 years with or at high risk of cerebral palsy: International clinical practice guideline based on systematic reviews Date of latest update: 17 May 2021. Date of next update: Not indicated. Patient eating and drinking (9 to 10); Complications prevention: vision (11 to 13), sleep group: Children aged 0 to 2 years with or at high risk of cerebral palsy. Intended (14 to 20), muscle tone (21), musculoskeletal impairments (22 and 23); and audience: Health professionals caring for neonates, infants and toddlers. Addi- Parental support: mental health and parenting (24 to 28). Three best-practice tional versions: Nil. Expert working group: Nine domain-specific working principles were supported: Early intervention: children diagnosed with or at groups of 42 health professionals, researchers and consumer representatives. high risk of cerebral palsy should be referred immediately for cerebral palsy- Funded by: Fulbright Senior Scholar Award and the Mariani Foundation of specific and age-specific interventions to optimise the critical period of plas- Milan. Consultation with: Co-chairs were representatives of the Cerebral Palsy ticity of systems; Goal-directed intervention: goals should be set that are task- Alliance Research Institute, Australia. Approved by: Lead authors C.M., L.F., and specific and context-specific, at the appropriate level of challenge and regu- I.N. Location: Guideline and supplementary material: https://doi.org/10.1001/ larly updated; and Caregiver engagement: clinicians should support parents and jamapediatrics.2021.0878. Description: This guideline was conducted to deter- caregivers to build parental capacity and expertise, prioritising a positive parent- mine evidence-based recommendations for early intervention for children aged child relationship. 0 to 2 years who have or are at high risk of cerebral palsy. Data were sought across nine domains: motor, cognitive, communication, eating/drinking, vision, Provenance: Invited. Not peer reviewed. sleep, musculoskeletal, mental health and parenting. Systematic reviews and randomised clinical trials were appraised using A Measurement Tool to Assess Emre Ilhana and Leanne M Johnstonb Systematic Reviews (AMSTAR) or Cochrane Risk of Bias tools. Recommendations aDepartment of Health Sciences, Faculty of Medicine, were formed using the Grading of Recommendations Assessment, Development, Health and Human Sciences, Macquarie University, Sydney and Evaluation (GRADE) framework and reported according to the Appraisal of Guidelines, Research and Evaluation (AGREE) II instrument. Recommendations bSchool of Health & Rehabilitation Sciences, were based on desirable/undesirable consequences of interventions, evidence The University of Queensland, Brisbane, Australia quality, family preferences and costs. Twenty-eight recommendations (24 for and four against) were presented in three categories: Skills development: motor https://doi.org/10.1016/j.jphys.2021.08.011 skills (recommendations 1 to 4), cognition (5 to 6), communication (7 to 8), 1836-9553/© 2021 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). Appraisal of Clinical Practice Guideline: Physical Therapy Evaluation and Treatment After Concussion/Mild Traumatic Brain Injury Date of latest update: April 2020. Date of next update: Within 5 years. and motor function. Key recommendations are provided in three parts: Patient group: Individuals aged  8 years with concussion/mild traumatic screening and diagnosis (screening for all individuals with a potential history brain injury, who have at most ‘mild’ cognitive impairment before or after the of concussive events, screening for indicators of emergency conditions and concussive event, can self-report symptoms and can respond to provocative differential diagnosis); examination (systems to be examined, sequencing of tests. Intended audience: Physiotherapists. Expert working group: Repre- examination based on the levels of irritability, classification of examination sentatives from the American Physical Therapy Association, Academy of Or- findings into impairment profiles, psychological and sociological factors and thopaedic Physical Therapy, American Academy of Sports Physical Therapy, outcome measure selection); and interventions (communication and educa- Academy of Neurologic Physical Therapy and Academy of Pediatric Physical tion, movement-related impairments, cervical musculoskeletal, vestibulo- Therapy. Funded by: American Physical Therapy Association grant, and oculomotor, exertional tolerance and aerobic exercise, motor function and sponsorship from the Academy of Orthopaedic Physical Therapy, American monitoring and progressing patients). Referral for further consultation is Academy of Sports Physical Therapy and Academy of Neurologic Physical recommended for patients with persistent migraine-type or other chronic Therapy for training, travel, software and librarian assistance. Consultation headaches; impairments in vision, hearing, sleep, mental health or cognition; with: Guidance from methodologists in the field. Approved by: Not or to exclude another potential medical diagnosis that may present with mentioned. Location: J Orthop Sports Phys Ther. 2020;50(4):CPG1–CPG73. concussion-like symptoms. https://doi.org/10.2519/jospt.2020.0301. Description and key recommenda- tions: This guideline aims to guide physiotherapists’ clinical decision making Provenance: Invited. Not peer reviewed. for individuals who have experienced a concussive event resulting in movement-related symptoms, impairments and functional limitations. Armaghan Dabbagh and Joy C MacDermid Strength and quality of the evidence were reviewed. The guideline covers School of Physical Therapy, Faculty of Health Sciences, Elborn College, concussion incidence, risk factors for prolonged recovery, physiotherapy ex- amination and intervention strategies. Decision trees are provided to facilitate Western University, London, Canada the triage process to determine priorities and sequencing of activities. Four overarching impairments are discussed within the physiotherapy scope of https://doi.org/10.1016/j.jphys.2021.08.010 practice: cervical musculoskeletal, vestibulo-oculomotor, exertional tolerance 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/ 4.0/).

Journal of Physiotherapy 67 (2021) 312 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Clinimetrics: Neck Pain and Disability Scale (NPDS) Summary Description: The Neck Pain and Disability Scale (NPDS or NPAD) is Dutch, Finnish, French, German, Italian, Hindi, Iranian, Korean, a questionnaire aiming to quantify neck pain and disability.1 It is a Turkish, Japanese and Thai. patient-reported outcome measure for patients with any type of neck pain, of any duration, with or without injury.1,2 It consists of 20 items: Reliability and validity: Two systematic reviews have evaluated the clinimetric properties of 11 of the translated versions.5,8 The three related to pain intensity, four related to emotion and cognition, Finnish, German and Italian translations were particularly recom- four related to mobility of the neck, eight related to activity limita- tions and participation restrictions and one on medication.1,3 Patients mended for use in clinical practice. Face validity was established and respond to each item on a 0 to 5 visual analogue scale of 10 cm. There content validity was confirmed by an adequate reflection of all aspects is also a nine-item short version.4 of neck pain and disability.1,8 Regarding structural validity, the NPDS is Feasibility: The NPDS is published and available online (https:// a multidimensional scale, with moderate evidence that the NPDS has a mountainphysiotherapy.com.au/wp-content/uploads/2016/08/Neck- three-factor structure (with explained variance ranging from 63 to Pain-and-Disability-Scale.pdf).1 The NPDS is an easy to use ques- tionnaire that can be completed within 5 to 8 minutes.1,5 There is no 78%): neck dysfunction related to general activities; neck pain and training needed to administer the instrument but its validity is neck-specific function; and cognitive-emotional-behavioural func- compromised if the questionnaire must be read to the patient.2 tioning.4,5,9 A recent overview of four systematic reviews found Higher scores indicate higher severity (0 for normal functioning to moderate-quality evidence of high internal consistency (Cronbach’s alphas ranging from 0.86 to 0.93 for the various factors).10 Excellent 5 for the worst possible situation ‘your’ pain problem has caused you).2 The total score is the sum of scores on the 20 items (0 to 100).1 test-retest reliability was found (ICC of 0.97); however, the studies The maximum acceptable number of missing answers is three (15%).4 were considered to be of low quality.3,10 Construct validity (hypothe- Two studies found a minimum important change of 10 points ses-testing) seems adequate when the NPDS is compared with the (sensitivity 0.93; specificity 0.83) and 11.5 points (sensibility 0.74; Neck Disability Index and the Global Assessment of Change with specificity 0.70), respectively.6,7 The NPDS is available in English, moderate to strong correlations (r = 0.52 to 0.86), based on limited moderate-quality studies.3,11,12 One systematic review reported good responsiveness to change in patients (r = 0.59).12 Commentary Emmylou Beekmana and Sandra Lüttmannb aResearch Centre for Autonomy and Participation for Persons with a The advantage of the NPDS over other neck pain scales is that it Chronic Illness, Zuyd University of Applied Sciences, The Netherlands comprehensively assesses the multidimensional nature of neck pain and dysfunction in a quick and easy way.9 The NPDS has demon- bAcademy of Physiotherapy, Zuyd University of Applied Sciences strated a well-balanced distribution of items across the International The Netherlands Classification of Functioning, Disability and Health’s body function, activity and participation components. In addition, the NPDS is the References only neck pain and disability questionnaire that assesses contextual factors.3 With this the NPDS is not solely based on the illness itself, 1. Wheeler AH, et al. Spine. 1999;24:1290–1294. but emphasises functioning as a health component including envi- 2. Pietrobon R, et al. Spine. 2002;27:515–522. ronmental influences on the accomplishment of activities and tasks. 3. Ferreira ML, et al. Disabil Rehabil. 2010;32:1539–1546. In addition, the NPDS is suitable for patients taking pain medication. 4. Blozik E, et al. Eur J Pain. 2010;14:864.e1–864.e7. Calculation of the total scale is recommended in the literature, 5. Schellingerhout JM, et al. BMC Med Res Methodol. 2011;11:87. despite the proven multidimensional scale including three factors. 6. Monticone M, et al. Eur Spine J. 2015;24:2821–2827. The frequently used Neck Disability Index is favoured over the NPDS, 7. Jorritsma W, et al. Eur Spine J. 2012;21:2550–2557. with high-quality evidence for good to excellent validity and reli- 8. Pellicciari L, et al. Arch Physiother. 2016;6:9. ability. Nevertheless, the NPDS is among the four most clinimetrically 9. Pickering PM, et al. Spine (Phila Pa 1976). 2011;36:581–588. sound frequently used self-rated pain scales for neck pain (along with 10. Bobos P, et al. J Orthop Sports Phys Ther. 2018;48:775–788. the Northwick Park Neck Pain Questionnaire and the Copenhagen 11. Misailidou V, et al. J Chiroprac Med. 2010;9:49–59. Neck Functional Disability Scale).10 12. Schellingerhout JM, et al. Quality Life Res. 2012;21:659–670. https://doi.org/10.1016/j.jphys.2021.02.009 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 313 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Clinimetrics: Quadriceps muscle ultrasound Summary intraobserver reproducibility of 0.74 to 0.981,8 and for interobserver reproducibility of 0.76 to 0.96.1,8,9 Skeletal muscle atrophy is a prevalent problem associated with critically ill patients within the intensive care (ICU) setting, and can Lower limb muscle ultrasound should be performed systemati- have a significant and detrimental effect on long-term patient out- comes.1–3 Lower limb quadriceps muscle ultrasound is a valid, quick cally. However, there is currently no standardised protocol to guide and accessible tool for the assessment of muscle thickness, quality or clinicians.4 Previous studies have recommended using a linear cross-sectional area in critically ill patients.1,2,4 Quantifying muscle strength in this population is problematic, as volitional muscle testing transducer in B-mode, with the patient’s knees placed in extension can be limited by altered cognition and severe muscle wasting.1,4 The and neutral rotation. A point should be marked along the thigh, either direct relationship between muscle strength and thickness or cross- sectional area5,6 means that lower limb muscle ultrasound is an two-thirds of the distance from the anterior superior iliac spine of the ideal tool to monitor muscle wasting throughout the patient’s ICU stay. To minimise muscle wasting during periods of critical illness in pelvis to the superior pole of the patella, or at the midpoint between ICU and provide measures for targeted rehabilitation programs, these two bony landmarks.1,2,7,8 Lateral images can also be taken, with which are essential for improving patient functional outcomes, this may provide helpful clinical information for clinicians.4,7 the transducer relocated 5 cm laterally from the ‘two-thirds’ site. Placing marks on the subject’s legs can be helpful in assisting repli- Excellent intraobserver and interobserver reliability for ultra- cation of imaging in subsequent assessments.7 Basic measurements sound measurements of quadriceps muscle layer thickness have been reported in both healthy subjects and critically ill patients.1,8,9 Studies can be used on the ultrasound system using the calliper setting or have reported the intraclass correlation coefficient, with a range for images can be transferred to a computer for further measurement and analysis. Cross-sectional area, thickness and pennation angle of skeletal muscle can all be calculated. The echogenicity of muscle can also be analysed, with increased echogenicity suggestive of muscle necrosis and destruction of muscle architecture on a cellular level.7,10 Commentary Atrophy of skeletal muscle should be a major consideration for applications for physiotherapists is warranted. Whilst muscle ultra- physiotherapists in ICU when selecting assessment and treatment. Not sound provides physiotherapists with an option for early assessment only can muscle atrophy contribute to increased risk of morbidity and of skeletal muscle changes associated with critical illness, there is a mortality, it has been associated with longer time on mechanical need for better prediction of those patients who are most at risk of ventilation, longer length of stay in hospital and may have negative significant muscle wasting. Muscle ultrasound could then be a useful impacts on long-term quality of life. To be able to assess skeletal outcome measure to assess the effectiveness of physiotherapy in- muscle atrophy, physiotherapists need an assessment tool that is valid, terventions that focus on prevention or treatment of muscle wasting. reliable, noninvasive and readily available at the bedside. Skeletal muscle ultrasound, in the assessment of cross-sectional area and Provenance: Invited. Not peer reviewed. muscle thickness, is safe, quick and easy to perform and demonstrates excellent reliability and repeatability. Physiotherapists should aim to Louise Hansella,b and George Ntoumenopoulosc monitor skeletal muscle changes during a patient’s ICU stay, starting aSydney School of Health Sciences, Faculty of Medicine and Health, close to admission, which enables identification of those patients at risk of or demonstrating significant muscle atrophy. Functional reha- The University of Sydney bilitation can then be commenced as early as is feasible to counteract bPhysiotherapy Department, Royal North Shore Hospital, St Leonards the known negative effects of prolonged bed rest. cPhysiotherapy Department, St Vincent’s Hospital, Sydney, Australia Ultrasound is not without limitations, and it is paramount that clinicians have adequate training and appropriate skill to perform References imaging and interpret images. There are few courses that train physiotherapists in ultrasound relating to critical care, and they pri- 1. Tillquist M, et al. J Parenter Enteral Nutr. 2014;38:886–890. marily focus on lung and diaphragm imaging; these skills may be 2. Gruther W, et al. J Rehabil Med. 2008;40:185–189. translatable when imaging skeletal muscle. Additionally, there is a 3. Palakshappa JA, et al. Ann Am Thorac Soc. 2019;16:1107–1111. need for a standardised protocol to be developed to guide lower limb 4. Mourtzakis M, et al. Ann Am Thorac Soc. 2017;14:1495–1503. ultrasound. The extent and delineation of pre-existing muscle atro- 5. Freilich RJ, et al. Neuromuscul Disord. 1995;5:415–422. phy at admission to ICU is unclear and early muscle ultrasound on 6. Palakshappa JA, et al. J Crit Care. 2018;47:324–330. admission may assist in clarifying this. Whilst research to date has 7. Parry SM, et al. J Crit Care. 2015;30:1151.e1159–1114. demonstrated the validity, reliability and repeatability of skeletal 8. Pardo E, et al. BMC Anesthesiol. 2018;18:205. muscle ultrasound, further research exploring its practical clinical 9. Sarwal A, et al. J Ultrasound Med. 2015;34:1191–1200. 10. Puthucheary ZA, et al. Crit Care Med. 2015;43:1603–1611. https://doi.org/10.1016/j.jphys.2021.05.001 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 308 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Additional, mechanised upper limb self-rehabilitation in patients with subacute stroke is not more effective than basic stretching and active exercises in reducing upper limb impairment Synopsis Summary of: Rémy-Néris O, Le Jeannic A, Dion A, Médée B, Nowak E, Poiroux months after cessation of the intervention. Secondary outcomes were: É, Durand-Zaleski I; REM Investigative Team. Additional, Mechanized Upper severity of shoulder pain at rest and during active and passive movements, Limb Self-Rehabilitation in Patients With Subacute Stroke: The REM-AVC spasticity, functional status, upper limb function, quality of life, and Randomized Trial. Stroke. 2021;52(6):1938–1947. perception of their additional exercise intervention, measured on a 5-point Likert scale (higher scores were better). Results: Of the 215 randomised Question: In people in the subacute phase following stroke, do two 30- participants, 107 were allocated to the experimental group and 108 to the minute daily sessions of self-rehabilitation using a gravity-compensating control group. At the end of the 4-week period, data were available on 208 mechanical exoskeleton for game-based exercises improve upper limb participants. When adjusted for age and baseline measures at 4 weeks impairment compared with self-rehabilitation involving stretching and there was no difference in the mean change in Fugl-Meyer Assessment for basic active exercises? Design: Multicentre randomised trial with con- Upper Extremity score 1.62 (95% CI 20.78 to 4.02). There were no cealed allocation and blinded assessors. Setting: Twenty-one rehabilita- between-group differences in this score at any other time point. The only tion centres in France. Participants: Inclusion criteria were age 18 to 81 secondary outcome showing a difference was the perception regarding years, a diagnosis of middle cerebral artery stroke 3 weeks to 3 months ease of learning 0.34 (95% CI 0.06 to 0.62) and ease of practice 0.31 (95% CI previously, and a Fugl-Meyer Assessment for Upper Extremity score be- 0.02 to 0.61). Conclusion: In people with moderate-to-severe impairment tween 10 and 40 points. Exclusion criteria were pain in the affected of upper limb function shortly after a stroke, additional self-rehabilitation shoulder . 3 out of 10 on a visual analogue scale, a Boston Diagnostic performed using an exoskeleton was not more effective at reducing upper Aphasia Examination score  3 points, fatigue or visual impairment that limb impairment than basic stretching and active exercises. would prevent participation, and inability to sit independently. Interventions: Both groups received usual rehabilitation provided in each Provenance: Invited. Not peer reviewed. centre for 5 days per week. In addition, self-rehabilitation was performed twice daily for 30 minutes over a 4-week period. In the experimental Aline Scianni group, self-rehabilitation was performed using a gravity-compensating Department of Physical Therapy exoskeleton for games-based exercises. In the control group, self- Universidade Federal de Minas Gerais, Brazil rehabilitation comprised basic stretching and active exercises. Outcome measures: The primary outcome was the Fugl-Meyer Assessment for https://doi.org/10.1016/j.jphys.2021.08.002 Upper Extremity score, which was assessed at 30 days and 3, 6 and 12 Commentary The well-conducted REM-AVC trial confirms the existing literature synergies as measured with the Fugl-Meyer Assessment for Upper Ex- showing that upper limb robot training is equally as effective as usual tremity. However, this assumption is in contrast with findings from a care.1,2 A recent large phase-IV trial1 and meta-analysis of 44 trials2 suggest number of studies showing that intensity of practice and adaptive learning are the main drivers for improving upper limb capacity.6 that upper limb robot training may improve upper limb function by about 2 Provenance: Invited. Not peer reviewed. points on the Fugl-Meyer Assessment for Upper Extremity. This finding is irrespective of type of upper limb robot training that is applied.1,2 My first Gert Kwakkel concern is related to the heterogeneity introduced by the arbitrary Department of Rehabilitation Medicine, Amsterdam University Medical recruitment of subjects at a mean of 55 days (SD 22) after stroke in this Centre Amsterdam, the Netherlands REM-AVC trial. A recent observational study of 412 stroke subjects https://doi.org/10.1016/j.jphys.2021.08.001 showed that the proportional amount of spontaneous motor recovery References may range from , 10% for those with a low baseline score up to 90% in 1. Rodgers H, et al. Lancet. 2019;394:51–62. the first 10 weeks after stroke.3 The non-fixed timing of baseline as well 2. Veerbeek JM, et al. Neurorehabil Neural Repair. 2017;31:107–121. as post-intervention assessment and the lack of stratification on robust 3. van der Vliet R, et al. Ann Neurol. 2020;87:383–393. prognostic markers, such as voluntary finger extension,4 may easily result 4. Bernhardt J, et al. Neurorehabil Neural Repair. 2017;31:793–799. in underpowered trials in which . 450 participants per trial arm are 5. Winters C, et al. Trials. 2016;17:468. required.5 6. Winters C, et al. NeuroRehabilitation. 2018;43:19–30. A more fundamental concern is based on the equal dose of additional training of about 33 minutes per day applied in both the upper limb robot training group and the control group in the REM-AVC trial. Obviously, the authors did hypothesise in this trial that the difference in the type of training (upper limb robot training combined with virtual games versus self-rehabilitation involving stretching and basic active exercises) was sufficient to cause significant interaction effects in recovery of muscle 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 308 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Additional, mechanised upper limb self-rehabilitation in patients with subacute stroke is not more effective than basic stretching and active exercises in reducing upper limb impairment Synopsis Summary of: Rémy-Néris O, Le Jeannic A, Dion A, Médée B, Nowak E, Poiroux months after cessation of the intervention. Secondary outcomes were: É, Durand-Zaleski I; REM Investigative Team. Additional, Mechanized Upper severity of shoulder pain at rest and during active and passive movements, Limb Self-Rehabilitation in Patients With Subacute Stroke: The REM-AVC spasticity, functional status, upper limb function, quality of life, and Randomized Trial. Stroke. 2021;52(6):1938–1947. perception of their additional exercise intervention, measured on a 5-point Likert scale (higher scores were better). Results: Of the 215 randomised Question: In people in the subacute phase following stroke, do two 30- participants, 107 were allocated to the experimental group and 108 to the minute daily sessions of self-rehabilitation using a gravity-compensating control group. At the end of the 4-week period, data were available on 208 mechanical exoskeleton for game-based exercises improve upper limb participants. When adjusted for age and baseline measures at 4 weeks impairment compared with self-rehabilitation involving stretching and there was no difference in the mean change in Fugl-Meyer Assessment for basic active exercises? Design: Multicentre randomised trial with con- Upper Extremity score 1.62 (95% CI 20.78 to 4.02). There were no cealed allocation and blinded assessors. Setting: Twenty-one rehabilita- between-group differences in this score at any other time point. The only tion centres in France. Participants: Inclusion criteria were age 18 to 81 secondary outcome showing a difference was the perception regarding years, a diagnosis of middle cerebral artery stroke 3 weeks to 3 months ease of learning 0.34 (95% CI 0.06 to 0.62) and ease of practice 0.31 (95% CI previously, and a Fugl-Meyer Assessment for Upper Extremity score be- 0.02 to 0.61). Conclusion: In people with moderate-to-severe impairment tween 10 and 40 points. Exclusion criteria were pain in the affected of upper limb function shortly after a stroke, additional self-rehabilitation shoulder . 3 out of 10 on a visual analogue scale, a Boston Diagnostic performed using an exoskeleton was not more effective at reducing upper Aphasia Examination score  3 points, fatigue or visual impairment that limb impairment than basic stretching and active exercises. would prevent participation, and inability to sit independently. Interventions: Both groups received usual rehabilitation provided in each Provenance: Invited. Not peer reviewed. centre for 5 days per week. In addition, self-rehabilitation was performed twice daily for 30 minutes over a 4-week period. In the experimental Aline Scianni group, self-rehabilitation was performed using a gravity-compensating Department of Physical Therapy exoskeleton for games-based exercises. In the control group, self- Universidade Federal de Minas Gerais, Brazil rehabilitation comprised basic stretching and active exercises. Outcome measures: The primary outcome was the Fugl-Meyer Assessment for https://doi.org/10.1016/j.jphys.2021.08.002 Upper Extremity score, which was assessed at 30 days and 3, 6 and 12 Commentary The well-conducted REM-AVC trial confirms the existing literature synergies as measured with the Fugl-Meyer Assessment for Upper Ex- showing that upper limb robot training is equally as effective as usual tremity. However, this assumption is in contrast with findings from a care.1,2 A recent large phase-IV trial1 and meta-analysis of 44 trials2 suggest number of studies showing that intensity of practice and adaptive learning are the main drivers for improving upper limb capacity.6 that upper limb robot training may improve upper limb function by about 2 Provenance: Invited. Not peer reviewed. points on the Fugl-Meyer Assessment for Upper Extremity. This finding is irrespective of type of upper limb robot training that is applied.1,2 My first Gert Kwakkel concern is related to the heterogeneity introduced by the arbitrary Department of Rehabilitation Medicine, Amsterdam University Medical recruitment of subjects at a mean of 55 days (SD 22) after stroke in this Centre Amsterdam, the Netherlands REM-AVC trial. A recent observational study of 412 stroke subjects https://doi.org/10.1016/j.jphys.2021.08.001 showed that the proportional amount of spontaneous motor recovery References may range from , 10% for those with a low baseline score up to 90% in 1. Rodgers H, et al. Lancet. 2019;394:51–62. the first 10 weeks after stroke.3 The non-fixed timing of baseline as well 2. Veerbeek JM, et al. Neurorehabil Neural Repair. 2017;31:107–121. as post-intervention assessment and the lack of stratification on robust 3. van der Vliet R, et al. Ann Neurol. 2020;87:383–393. prognostic markers, such as voluntary finger extension,4 may easily result 4. Bernhardt J, et al. Neurorehabil Neural Repair. 2017;31:793–799. in underpowered trials in which . 450 participants per trial arm are 5. Winters C, et al. Trials. 2016;17:468. required.5 6. Winters C, et al. NeuroRehabilitation. 2018;43:19–30. A more fundamental concern is based on the equal dose of additional training of about 33 minutes per day applied in both the upper limb robot training group and the control group in the REM-AVC trial. Obviously, the authors did hypothesise in this trial that the difference in the type of training (upper limb robot training combined with virtual games versus self-rehabilitation involving stretching and basic active exercises) was sufficient to cause significant interaction effects in recovery of muscle 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 309 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: A self-directed, web-based exercise and physical activity program supported with text messages improves knee pain and function for people with knee osteoarthritis Synopsis Summary of: Nelligan RK, Hinman RS, Kasza J, Crofts SJC, Bennell KL. pain (average knee pain in the last week, scores 0 to 10 on numeric Effects of a self-directed web-based strengthening exercise and physical rating scale) and physical function (subscale of the Western Ontario and activity program supported by automated text messages for people with McMaster Universities Osteoarthritis Index, scores 0 to 68) from baseline knee osteoarthritis: a randomized clinical trial. JAMA Intern Med. to the end of the 24-week trial. Secondary outcomes were an additional 2021;181(6):776–785. pain measure, function in sport/recreation, quality of life, physical ac- tivity, self-efficacy, and satisfaction measured using appropriate stand- Question: Does a web-based, self-directed exercise regime and physical ardised assessment tools. Results: 180 participants completed the study. activity intervention supported by automated text messages improve At the end of the 24-week period the experimental group had greater knee pain and function in people with knee osteoarthritis? Design: improvements in overall pain ratings than the control group by 1.6 units Randomised controlled trial with concealed allocation with limited (95% CI 0.9 to 2.2). Physical function improved more in the experimental disclosure to blind participants, as all outcome assessments were patient than the control group by 5.2 units (95% CI 1.9 to 8.5). The groups reported. Setting: Online recruitment of participants from the commu- differed for most secondary outcomes in favour of the experimental nity throughout Australia. Participants: Inclusion criteria were age 45 group, except for physical activity and self-efficacy for function and years; clinical criteria for osteoarthritis; knee pain on most days for 3 exercise, where there were no between-group differences. Conclusion: A months, with average overall knee pain 4 on an 11-point numeric free-access digital intervention involving a self-directed, web-based rating scale in the previous week; and a mobile phone with text exercise and physical activity program supported by text messages messaging capabilities and home internet access. Key exclusion criteria improved knee pain and function in people with knee osteoarthritis. were scheduled joint replacement surgery, inflammatory arthritis or inability to speak English. Randomisation of 206 participants allocated Provenance: Invited. Not peer reviewed. 103 to an experimental group and 103 to a control group. Interventions: Both groups received access to a website containing information on knee Alicia Spittle pain, knee osteoarthritis and generic information on the importance of Department of Physiotherapy, University of Melbourne, Australia exercise and physical activity. In addition, the experimental group had access to a 24-week self-directed strengthening regimen and physical https://doi.org/10.1016/j.jphys.2021.08.004 activity program on the website, supported by automated text messages. Outcome measures: The primary outcomes were improvement in knee higher education,6–8 further work is also needed to find ways to reach those in most need of first-line osteoarthritis care. Commentary Provenance: Invited. Not peer reviewed. There are a number of factors behind the underutilisation of first-line care in knee osteoarthritis (ie, exercise and education); one important Pætur Mikal Holm factor is participation costs and lack of support.1 This randomised Research unit PROgrez controlled trial assessed an intervention of online structured exercise Department of Physiotherapy and Occupational Therapy and physical activity guidance along with text messaging support Næstved-Slagelse-Ringsted Hospitals strategies, which were based on the behaviour change wheel Region Zealand, Denmark framework, to address typical barriers to exercise and physical activity. Compared with a control group that received online education and https://doi.org/10.1016/j.jphys.2021.08.003 general advice about exercise and physical activity with osteoarthritis, the addition of this intervention resulted in greater relief in knee pain References and improvements in physical function and quality of life after 6 months. 1. Dobson F, et al. Am J Phys Med Rehabil. 2016;95:372–389. Mean between-group differences in pain, function and quality of life 2. Bellamy N, et al. J Rheumatol. 1992;19:451–457. were in the proximity of minimum clinically important differences.2,3 A 3. Angst F, et al. Arthritis Rheum. 2001;45:384–391. total of 57% of participants in the intervention group compared with 4. Pietrzak E, et al. Telemed J E Health. 2013;19:800–805. 27% in the control group improved overall, scoring better or much 5. Carnes D, et al. Clin J Pain. 2012;28:344–354. better on perceived effect after 6 months. 6. Bossen D, et al. J Med Internet Res. 2013;15:e257. 7. Allen KD, et al. Osteoarthritis and Cartilage. 2018;26:383–396. These findings suggest that a multicomponent approach is more 8. Eysenbach G. J Med Internet Res. 2007;9:e34. effective than a unidimensional approach for online interventions in facilitating exercise and physical activity behaviour in people with osteoarthritis.4–7 This intervention is easily scalable, unsupervised and free, and has the potential to be implemented at a population level. However, as online interventions seem to attract healthier people with 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 309 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: A self-directed, web-based exercise and physical activity program supported with text messages improves knee pain and function for people with knee osteoarthritis Synopsis Summary of: Nelligan RK, Hinman RS, Kasza J, Crofts SJC, Bennell KL. pain (average knee pain in the last week, scores 0 to 10 on numeric Effects of a self-directed web-based strengthening exercise and physical rating scale) and physical function (subscale of the Western Ontario and activity program supported by automated text messages for people with McMaster Universities Osteoarthritis Index, scores 0 to 68) from baseline knee osteoarthritis: a randomized clinical trial. JAMA Intern Med. to the end of the 24-week trial. Secondary outcomes were an additional 2021;181(6):776–785. pain measure, function in sport/recreation, quality of life, physical ac- tivity, self-efficacy, and satisfaction measured using appropriate stand- Question: Does a web-based, self-directed exercise regime and physical ardised assessment tools. Results: 180 participants completed the study. activity intervention supported by automated text messages improve At the end of the 24-week period the experimental group had greater knee pain and function in people with knee osteoarthritis? Design: improvements in overall pain ratings than the control group by 1.6 units Randomised controlled trial with concealed allocation with limited (95% CI 0.9 to 2.2). Physical function improved more in the experimental disclosure to blind participants, as all outcome assessments were patient than the control group by 5.2 units (95% CI 1.9 to 8.5). The groups reported. Setting: Online recruitment of participants from the commu- differed for most secondary outcomes in favour of the experimental nity throughout Australia. Participants: Inclusion criteria were age 45 group, except for physical activity and self-efficacy for function and years; clinical criteria for osteoarthritis; knee pain on most days for 3 exercise, where there were no between-group differences. Conclusion: A months, with average overall knee pain 4 on an 11-point numeric free-access digital intervention involving a self-directed, web-based rating scale in the previous week; and a mobile phone with text exercise and physical activity program supported by text messages messaging capabilities and home internet access. Key exclusion criteria improved knee pain and function in people with knee osteoarthritis. were scheduled joint replacement surgery, inflammatory arthritis or inability to speak English. Randomisation of 206 participants allocated Provenance: Invited. Not peer reviewed. 103 to an experimental group and 103 to a control group. Interventions: Both groups received access to a website containing information on knee Alicia Spittle pain, knee osteoarthritis and generic information on the importance of Department of Physiotherapy, University of Melbourne, Australia exercise and physical activity. In addition, the experimental group had access to a 24-week self-directed strengthening regimen and physical https://doi.org/10.1016/j.jphys.2021.08.004 activity program on the website, supported by automated text messages. Outcome measures: The primary outcomes were improvement in knee higher education,6–8 further work is also needed to find ways to reach those in most need of first-line osteoarthritis care. Commentary Provenance: Invited. Not peer reviewed. There are a number of factors behind the underutilisation of first-line care in knee osteoarthritis (ie, exercise and education); one important Pætur Mikal Holm factor is participation costs and lack of support.1 This randomised Research unit PROgrez controlled trial assessed an intervention of online structured exercise Department of Physiotherapy and Occupational Therapy and physical activity guidance along with text messaging support Næstved-Slagelse-Ringsted Hospitals strategies, which were based on the behaviour change wheel Region Zealand, Denmark framework, to address typical barriers to exercise and physical activity. Compared with a control group that received online education and https://doi.org/10.1016/j.jphys.2021.08.003 general advice about exercise and physical activity with osteoarthritis, the addition of this intervention resulted in greater relief in knee pain References and improvements in physical function and quality of life after 6 months. 1. Dobson F, et al. Am J Phys Med Rehabil. 2016;95:372–389. Mean between-group differences in pain, function and quality of life 2. Bellamy N, et al. J Rheumatol. 1992;19:451–457. were in the proximity of minimum clinically important differences.2,3 A 3. Angst F, et al. Arthritis Rheum. 2001;45:384–391. total of 57% of participants in the intervention group compared with 4. Pietrzak E, et al. Telemed J E Health. 2013;19:800–805. 27% in the control group improved overall, scoring better or much 5. Carnes D, et al. Clin J Pain. 2012;28:344–354. better on perceived effect after 6 months. 6. Bossen D, et al. J Med Internet Res. 2013;15:e257. 7. Allen KD, et al. Osteoarthritis and Cartilage. 2018;26:383–396. These findings suggest that a multicomponent approach is more 8. Eysenbach G. J Med Internet Res. 2007;9:e34. effective than a unidimensional approach for online interventions in facilitating exercise and physical activity behaviour in people with osteoarthritis.4–7 This intervention is easily scalable, unsupervised and free, and has the potential to be implemented at a population level. However, as online interventions seem to attract healthier people with 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 311 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Implicit motor learning is not superior to explicit motor learning for improving gait speed in chronic stroke Synopsis Summary of: Jie L-J, Kleynen M, Meijer K, Beurskens A, Braun S. Implicit measures were the modified Dynamic Gait Index, cognitive-motor dual-task and explicit motor learning interventions have similar effects on walking walking performance, Movement-Specific Reinvestment Scale, Stroke and speed in people after stroke: a randomized controlled trial. Phys Ther. Aphasia Quality of Life Scale-39, Global Perceived Effect scale, and a verbal 2021;101:1–10. protocol to assess the number of explicit rules used by participants while walking. Results: Seventy-nine participants started the intervention and 73 Question: Does implicit motor learning using analogy learning improve completed the study. There was no clear between-group difference for walking speed in chronic stroke more than explicit motor learning using change in gait speed at either the post-intervention assessment (MD 0.02 m/ verbal instructions? Design: Randomised controlled trial with concealed s, 95% CI 20.04 to 0.08) or at 1-month follow-up (MD 20.02 m/s, 95% CI 20.09 allocation and blinded outcome assessment. Setting: Home-based inter- to 0.05). There were no clear between-group differences for any secondary vention in the Netherlands. Participants: Individuals who were  6 months outcomes, with the exception that participants in the implicit group used after stroke, with gait speed , 1.0 m/s and ability to follow a three-step fewer explicit rules during walking after intervention compared with the command in Dutch. Key exclusion criteria were inability to walk 10 m, explicit group. Per-protocol analyses (n = 60) did not reveal any between- requiring manual assistance to walk on level surfaces, and other non-stroke group differences in treatment effect on gait speed. Conclusion: Implicit impairments that affected the gait pattern. Randomisation of 81 participants motor learning using analogies was not superior to explicit motor learning for allocated 39 to the implicit training group and 42 to the explicit training improving short-distance gait speed in adults with chronic gait impairment group. Interventions: For both groups, the intervention was delivered at due to stroke. home in nine 30-minute sessions over 3 weeks. The implicit training group participants received analogies meaningful to them that aimed to improve Provenance: Invited. Not peer reviewed. walking performance (eg, walk as if you are following footprints in the sand), whereas the explicit training group received detailed verbal instructions Prudence Plummer about how to alter aspects of their walking (eg, land with your heel first then Department of Physical Therapy, MGH Institute of Health Professions, roll through from heel to toe). Outcome measures: The primary outcome measure was 10-m gait speed, averaged over three trials, measured before USA and after intervention and at 1-month follow-up. Secondary outcome https://doi.org/10.1016/j.jphys.2021.09.002 1836-9553/© 2021 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). Commentary Implicit and explicit memory systems normally work together to versus explicit learning strategy difficult.5 Future work should quantify support motor learning,1 and dissociating these learning and memory the extent of damage in unique memory systems when testing different systems is notoriously difficult, even in highly controlled laboratory interventions designed to enhance motor learning. settings. Thus, Jie et al are to be commended for considering how the delivery of explicit instructions affected overground gait speed. The The critical next step to optimise learning and recovery after stroke is finding that gait speed improved regardless of explicit instructions to determine how individual patient characteristics (eg, cognitive capac- versus implicitly based analogies appears to be good news for therapists ity, lesion location, stroke severity) affect motor learning. who can chose to structure sessions in either format. Provenance: Invited. Not peer reviewed. However, several questions remain. The study used analogies to guide implicit learning. Yet, even in the laboratory where behaviour and tasks Lara A Boyd are tightly controlled, humans are very good at gaining explicit awareness Department of Physical Therapy & the Djavad Mowafaghian Centre for of behavioural regularities.2 As between-group explicit memory was not statistically tested (using the verbal protocol data) it is unclear how much Brain Health, University of British Columbia, Vancouver, Canada overlap occurred between memory systems (or if the explicit group was truly aware of the task ‘rules’). https://doi.org/10.1016/j.jphys.2021.09.003 Additionally, learning of continuous versus discrete motor skills can be References differently affected by explicit instructions.2,3 This may result from the difficulty in applying an explicit ‘rule’ to an evolving movement and 1. Bond KM, et al. J Neurophysiol. 2015;113:3836–3849. from the challenge of updating ongoing motor plans. Finally, the 2. Hadjiosif AM, et al. Eur J Neurosci. 2021;53:499–503. implicit memory system is distributed across the brain, and damage to 3. Boyd L, et al. J Neurol Phys Ther. 2006;30:46–57. different regions creates unique deficits.4 While the heterogeneous 4. Boyd LA, et al. Learn Memory. 2004;11:388–396. population of individuals with stroke reflects clinical reality, it likely 5. Morgan OP, et al. Cerebellum. 2021;20:222–245. made conclusions surrounding who would benefit from an implicit 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 311 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Implicit motor learning is not superior to explicit motor learning for improving gait speed in chronic stroke Synopsis Summary of: Jie L-J, Kleynen M, Meijer K, Beurskens A, Braun S. Implicit measures were the modified Dynamic Gait Index, cognitive-motor dual-task and explicit motor learning interventions have similar effects on walking walking performance, Movement-Specific Reinvestment Scale, Stroke and speed in people after stroke: a randomized controlled trial. Phys Ther. Aphasia Quality of Life Scale-39, Global Perceived Effect scale, and a verbal 2021;101:1–10. protocol to assess the number of explicit rules used by participants while walking. Results: Seventy-nine participants started the intervention and 73 Question: Does implicit motor learning using analogy learning improve completed the study. There was no clear between-group difference for walking speed in chronic stroke more than explicit motor learning using change in gait speed at either the post-intervention assessment (MD 0.02 m/ verbal instructions? Design: Randomised controlled trial with concealed s, 95% CI 20.04 to 0.08) or at 1-month follow-up (MD 20.02 m/s, 95% CI 20.09 allocation and blinded outcome assessment. Setting: Home-based inter- to 0.05). There were no clear between-group differences for any secondary vention in the Netherlands. Participants: Individuals who were  6 months outcomes, with the exception that participants in the implicit group used after stroke, with gait speed , 1.0 m/s and ability to follow a three-step fewer explicit rules during walking after intervention compared with the command in Dutch. Key exclusion criteria were inability to walk 10 m, explicit group. Per-protocol analyses (n = 60) did not reveal any between- requiring manual assistance to walk on level surfaces, and other non-stroke group differences in treatment effect on gait speed. Conclusion: Implicit impairments that affected the gait pattern. Randomisation of 81 participants motor learning using analogies was not superior to explicit motor learning for allocated 39 to the implicit training group and 42 to the explicit training improving short-distance gait speed in adults with chronic gait impairment group. Interventions: For both groups, the intervention was delivered at due to stroke. home in nine 30-minute sessions over 3 weeks. The implicit training group participants received analogies meaningful to them that aimed to improve Provenance: Invited. Not peer reviewed. walking performance (eg, walk as if you are following footprints in the sand), whereas the explicit training group received detailed verbal instructions Prudence Plummer about how to alter aspects of their walking (eg, land with your heel first then Department of Physical Therapy, MGH Institute of Health Professions, roll through from heel to toe). Outcome measures: The primary outcome measure was 10-m gait speed, averaged over three trials, measured before USA and after intervention and at 1-month follow-up. Secondary outcome https://doi.org/10.1016/j.jphys.2021.09.002 1836-9553/© 2021 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). Commentary Implicit and explicit memory systems normally work together to versus explicit learning strategy difficult.5 Future work should quantify support motor learning,1 and dissociating these learning and memory the extent of damage in unique memory systems when testing different systems is notoriously difficult, even in highly controlled laboratory interventions designed to enhance motor learning. settings. Thus, Jie et al are to be commended for considering how the delivery of explicit instructions affected overground gait speed. The The critical next step to optimise learning and recovery after stroke is finding that gait speed improved regardless of explicit instructions to determine how individual patient characteristics (eg, cognitive capac- versus implicitly based analogies appears to be good news for therapists ity, lesion location, stroke severity) affect motor learning. who can chose to structure sessions in either format. Provenance: Invited. Not peer reviewed. However, several questions remain. The study used analogies to guide implicit learning. Yet, even in the laboratory where behaviour and tasks Lara A Boyd are tightly controlled, humans are very good at gaining explicit awareness Department of Physical Therapy & the Djavad Mowafaghian Centre for of behavioural regularities.2 As between-group explicit memory was not statistically tested (using the verbal protocol data) it is unclear how much Brain Health, University of British Columbia, Vancouver, Canada overlap occurred between memory systems (or if the explicit group was truly aware of the task ‘rules’). https://doi.org/10.1016/j.jphys.2021.09.003 Additionally, learning of continuous versus discrete motor skills can be References differently affected by explicit instructions.2,3 This may result from the difficulty in applying an explicit ‘rule’ to an evolving movement and 1. Bond KM, et al. J Neurophysiol. 2015;113:3836–3849. from the challenge of updating ongoing motor plans. Finally, the 2. Hadjiosif AM, et al. Eur J Neurosci. 2021;53:499–503. implicit memory system is distributed across the brain, and damage to 3. Boyd L, et al. J Neurol Phys Ther. 2006;30:46–57. different regions creates unique deficits.4 While the heterogeneous 4. Boyd LA, et al. Learn Memory. 2004;11:388–396. population of individuals with stroke reflects clinical reality, it likely 5. Morgan OP, et al. Cerebellum. 2021;20:222–245. made conclusions surrounding who would benefit from an implicit 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 310 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Stable supportive shoes improved knee pain more than flat flexible shoes in people with moderate to severe radiographic medial knee osteoarthritis Synopsis Summary of: Paterson KL, Bennell KL, Campbell PK, Metcalf BR, Wrigley physical function via the Western Ontario and McMaster Universities TV, Kasza J, Hinman RS. The effect of flat flexible versus stable supportive Osteoarthritis Index function subscale (0 = no dysfunction, 68 = maximum shoes on knee osteoarthritis symptoms: a randomized trial. Ann Intern dysfunction). Secondary outcome measures were Knee Injury and Osteo- Med. 2021;174:462–471. arthritis Outcome Score subscales, pain in back/hips/knees/feet/ankles, health-related quality of life, physical activity and global changes in pain Question: Are flat flexible shoes superior to stable supportive shoes in and physical function. Results: In total, 161 (98%) participants completed improving walking knee pain and physical dysfunction in people with the study. There was no evidence found at 6 months that flat flexible shoes moderate to severe radiographic medial knee osteoarthritis? Design: Su- were superior to stable supportive shoes regarding any primary or sec- periority randomised controlled trial with concealed allocation and blinded ondary outcome. There was evidence showing a between-group difference outcome assessment. Setting: Community participants from Melbourne, in change in pain favouring stable supportive shoes (1.1 units, 95% CI 0.5 to Australia. Participants: Adults aged  50 years with knee pain on most 1.8), but not function (2.3 units, 95% CI –0.9 to 5.5). Fewer participants days of the past month; knee pain during walking in the past week of  4 reported adverse events with stable supportive shoes (n = 12, 15%) on an 11-point numerical rating scale; and moderate to severe radiographic compared with flat flexible shoes (n = 26, 32%) (risk difference –0.17, 95% CI tibiofemoral osteoarthritis (Kellgren-Lawrence grade 3 to 4). Main exclu- –0.30 to –0.05). Conclusion: Flat flexible shoes were not superior to stable sion criteria were: lateral  medial joint space narrowing, recent (past 6 supportive shoes. Stable supportive footwear resulted in greater reductions months) or planned (next 6 months) knee surgery, and/or currently using in walking knee pain over 6 months and may be a useful self-management shoe orthoses or customised shoes. Randomisation of 164 participants strategy in this subgroup of patients with knee osteoarthritis. allocated 82 to flat flexible footwear and 82 to stable supportive footwear. Interventions: Flat flexible shoes had heel height , 15 mm, shoe pitch , 10 Provenance: Invited. Not peer reviewed. mm, no arch support, minimal sole rigidity and weighed  200 grams. Stable supportive shoes had heel height . 30 mm, shoe pitch . 10 mm, Nina Østerås arch support, rigid sole and weighed . 300 grams. Participants chose two Division of Rheumatology and Research, Diakonhjemmet Hospital, Oslo, different pairs within their allocated group and were advised to increase shoe wear by 1 hour/day until wearing the shoes as much as possible ( 6 Norway hours/day) for 6 months. Outcome measures: Primary outcomes were 6-month change in: average walking pain over the previous week via an https://doi.org/10.1016/j.jphys.2021.08.006 11-point numerical rating scale (0 = no pain, 10 = worst pain possible); and Commentary A recent overview of international guideline recommendations on footwear designed to reduce medial knee loading. However, this contra- nonpharmacological and nonsurgical interventions for osteoarthritis re- dicts their statement in the discussion that it is unlikely that their findings ported conflicting advice on shoes for knee osteoarthritis.1 This timely, would differ in the subgroup with mild knee osteoarthritis. high-quality randomised controlled trial addressed this issue by comparing two types of shoes for knee osteoarthritis patients. The authors Overall, this high-quality randomised controlled trial brings us a step hypothesised that flat flexible shoes would lead to greater improvements closer to answering the question of which shoes are most suitable in in knee pain and physical function compared to stable supportive shoes reducing pain in knee osteoarthritis. over 6 months. Interestingly, contrary to what they expected, stable supportive footwear led to greater reductions in knee pain during Provenance: Invited. Not peer reviewed. walking. Sabine E Kloprogge, Ingrid A Szilagyi and Sita MA Bierma-Zeinstra The study was well conducted with high internal validity. However, Department of General Practice, Erasmus MC, University Medical Center, there were some limitations. First, the authors concluded that stable supportive footwear led to greater reductions in knee pain, but the dif- Rotterdam, The Netherlands ferences with the flat flexible shoes did not meet the pre-specified min- imum clinically important difference.2 Second, the generalisability to https://doi.org/10.1016/j.jphys.2021.08.005 clinical practice is limited, as guidelines state that you do not need radiography for non-surgical treatment of knee osteoarthritis.3 References Approximately half of the eligible patients with no or mild radiographic or non-medial osteoarthritis were excluded. The authors justified these 1. Bierma-Zeinstra S, et al. Best Prac Res Clin Rheumatol. 2020;34:101564. inclusion criteria because people with moderate to severe knee osteoar- 2. Chan LS. Am J Public Health. 2013;103:e24–e25. thritis are more likely than those with mild disease to benefit from 3. Sakellariou G, et al. Ann Rheum Dis. 2017;76:1484–1494. 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 310 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Stable supportive shoes improved knee pain more than flat flexible shoes in people with moderate to severe radiographic medial knee osteoarthritis Synopsis Summary of: Paterson KL, Bennell KL, Campbell PK, Metcalf BR, Wrigley physical function via the Western Ontario and McMaster Universities TV, Kasza J, Hinman RS. The effect of flat flexible versus stable supportive Osteoarthritis Index function subscale (0 = no dysfunction, 68 = maximum shoes on knee osteoarthritis symptoms: a randomized trial. Ann Intern dysfunction). Secondary outcome measures were Knee Injury and Osteo- Med. 2021;174:462–471. arthritis Outcome Score subscales, pain in back/hips/knees/feet/ankles, health-related quality of life, physical activity and global changes in pain Question: Are flat flexible shoes superior to stable supportive shoes in and physical function. Results: In total, 161 (98%) participants completed improving walking knee pain and physical dysfunction in people with the study. There was no evidence found at 6 months that flat flexible shoes moderate to severe radiographic medial knee osteoarthritis? Design: Su- were superior to stable supportive shoes regarding any primary or sec- periority randomised controlled trial with concealed allocation and blinded ondary outcome. There was evidence showing a between-group difference outcome assessment. Setting: Community participants from Melbourne, in change in pain favouring stable supportive shoes (1.1 units, 95% CI 0.5 to Australia. Participants: Adults aged  50 years with knee pain on most 1.8), but not function (2.3 units, 95% CI –0.9 to 5.5). Fewer participants days of the past month; knee pain during walking in the past week of  4 reported adverse events with stable supportive shoes (n = 12, 15%) on an 11-point numerical rating scale; and moderate to severe radiographic compared with flat flexible shoes (n = 26, 32%) (risk difference –0.17, 95% CI tibiofemoral osteoarthritis (Kellgren-Lawrence grade 3 to 4). Main exclu- –0.30 to –0.05). Conclusion: Flat flexible shoes were not superior to stable sion criteria were: lateral  medial joint space narrowing, recent (past 6 supportive shoes. Stable supportive footwear resulted in greater reductions months) or planned (next 6 months) knee surgery, and/or currently using in walking knee pain over 6 months and may be a useful self-management shoe orthoses or customised shoes. Randomisation of 164 participants strategy in this subgroup of patients with knee osteoarthritis. allocated 82 to flat flexible footwear and 82 to stable supportive footwear. Interventions: Flat flexible shoes had heel height , 15 mm, shoe pitch , 10 Provenance: Invited. Not peer reviewed. mm, no arch support, minimal sole rigidity and weighed  200 grams. Stable supportive shoes had heel height . 30 mm, shoe pitch . 10 mm, Nina Østerås arch support, rigid sole and weighed . 300 grams. Participants chose two Division of Rheumatology and Research, Diakonhjemmet Hospital, Oslo, different pairs within their allocated group and were advised to increase shoe wear by 1 hour/day until wearing the shoes as much as possible ( 6 Norway hours/day) for 6 months. Outcome measures: Primary outcomes were 6-month change in: average walking pain over the previous week via an https://doi.org/10.1016/j.jphys.2021.08.006 11-point numerical rating scale (0 = no pain, 10 = worst pain possible); and Commentary A recent overview of international guideline recommendations on footwear designed to reduce medial knee loading. However, this contra- nonpharmacological and nonsurgical interventions for osteoarthritis re- dicts their statement in the discussion that it is unlikely that their findings ported conflicting advice on shoes for knee osteoarthritis.1 This timely, would differ in the subgroup with mild knee osteoarthritis. high-quality randomised controlled trial addressed this issue by comparing two types of shoes for knee osteoarthritis patients. The authors Overall, this high-quality randomised controlled trial brings us a step hypothesised that flat flexible shoes would lead to greater improvements closer to answering the question of which shoes are most suitable in in knee pain and physical function compared to stable supportive shoes reducing pain in knee osteoarthritis. over 6 months. Interestingly, contrary to what they expected, stable supportive footwear led to greater reductions in knee pain during Provenance: Invited. Not peer reviewed. walking. Sabine E Kloprogge, Ingrid A Szilagyi and Sita MA Bierma-Zeinstra The study was well conducted with high internal validity. However, Department of General Practice, Erasmus MC, University Medical Center, there were some limitations. First, the authors concluded that stable supportive footwear led to greater reductions in knee pain, but the dif- Rotterdam, The Netherlands ferences with the flat flexible shoes did not meet the pre-specified min- imum clinically important difference.2 Second, the generalisability to https://doi.org/10.1016/j.jphys.2021.08.005 clinical practice is limited, as guidelines state that you do not need radiography for non-surgical treatment of knee osteoarthritis.3 References Approximately half of the eligible patients with no or mild radiographic or non-medial osteoarthritis were excluded. The authors justified these 1. Bierma-Zeinstra S, et al. Best Prac Res Clin Rheumatol. 2020;34:101564. inclusion criteria because people with moderate to severe knee osteoar- 2. Chan LS. Am J Public Health. 2013;103:e24–e25. thritis are more likely than those with mild disease to benefit from 3. Sakellariou G, et al. Ann Rheum Dis. 2017;76:1484–1494. 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Journal of Physiotherapy 67 (2021) 242 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Expressions of interest for the Journal of Physiotherapy Editorial Fellowship Expressions of interest are now invited for the next Editorial Fellowship term, which will run for two years from January 2022 to December 2023. The Editorial Fellowship (one post available) will enable an emerging Australian physiotherapy researcher to gain important exposure to an international Editorial Board. This Fellowship will allow an early career researcher to develop skills in the key Journal of Physiotherapy portfolios, including Critically Appraised Papers and Clinimetrics. By rotating through each portfolio, there will be opportunities to receive mentoring from the Scientific Editor and Editorial Board members across all stages of the publication process. The position will also have a communications focus, contributing to our social media presence. A tangible outcome of the Fellowship will be a co-authored editorial or appraisal paper in the Journal. The successful applicant will be expected to attend quarterly Editorial Board meetings during the appointed term (as a non-voting member) and to prepare a brief report at the end of the Fellowship. Please note that this is an unpaid Fellowship. To be eligible to apply, applicants must: 1. be within 5 years of PhD award (taking career disruption into account) 2. be a financial member of the Australian Physiotherapy Association at the time of application. To be considered, applicants must submit: 1. a cover letter (maximum 2 pages) summarising: (a) what they hope to gain from the Editorial Fellowship; (b) what they would uniquely bring to the Journal of Physiotherapy; and (c) their specific areas of interest with respect to Editorial Fellowship rotations 2. a brief CV (maximum 3 pages) that also addresses the eligibility criteria above Expressions of interest will be assessed by the Editorial Board, with a focus on written communication skills, research track record (relative to opportunity), and research dissemination experience. Short-listed applicants will be asked to share their idea(s) for broadening journal reach and enhancing reader engagement; for example, through the preparation of a sample tweet, infographic, podcast outline or plain language research summary. Short-listed applicants may be interviewed. Expressions of interest close Friday, 12 November 2021 and should be directed to Marko Stechiwskyj at [email protected] https://doi.org/10.1016/j.jphys.2021.08.013 1836-9553/

Journal of Physiotherapy 67 (2021) 235–237 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Editorial Innovative physiotherapy clinical education in response to the COVID-19 pandemic with a clinical research placement model Amabile Dario a, Milena Simic b a Work Integrated Learning, The University of Sydney, Sydney, Australia; b Discipline of Physiotherapy, The University of Sydney, Sydney, Australia Coronavirus Disease 2019 (COVID-19) is challenging the health- near-completion physiotherapy students via clinical research place- care system and affecting current and future generations of physio- ments. It discusses how placements were co-designed with re- therapists. Physiotherapists in the workforce have been rapidly searchers and clinicians, and how the model was piloted in two adapting to new models of care under significant pressure.1–3 Phys- research trials to build students’ clinical skills and graduate qualities. iotherapy students have also been impacted profoundly by the global Strengths and barriers of clinical research placements have been pandemic, as most face-to-face teaching and learning activities have suggested by researchers, clinicians, academics and students, and been restricted to reduce the rapid spread of the virus. One aspect of these aspects are discussed in this Editorial. physiotherapy education that has been severely impacted by the COVID-19 pandemic is clinical education.4 Clinical education is a core Developing clinical research placements component in the curriculum of undergraduate and graduate-entry physiotherapy degrees and a requirement for registration in many Recognising the sudden and dramatic challenges that COVID-19 countries.5 Prior to the pandemic, clinical education for physio- presented to timely graduation of physiotherapy students, a therapy students was mostly delivered face-to-face in hospitals, pri- detailed consultation was undertaken by Work Integrated Learning vate practices and aged-care facilities, with remote activities such as (WIL) academics in charge of clinical education. The main goal of the telehealth rarely encouraged.4 This Editorial presents alternative so- consultation was to evaluate: the number of physiotherapy place- lutions to overcome barriers in traditional models and assist clinical ments needed to be filled, the student clinical education re- education during and beyond the pandemic. quirements to satisfy course accreditation, the timing of clinical placement blocks, the tasks required to assist students’ clinical ex- The impact of COVID-19 on physiotherapy clinical education was periences during a research clinical placement, and the student evident in Australia in March 2020, with restrictions on face-to-face upskill required prior to commencement of the research clinical teaching and learning activities. Almost half of traditional physio- placement. As clinical educational models rely heavily on successful therapy placements were paused or cancelled at the University of partnerships with internal and external stakeholders,7,8 WIL aca- Sydney. In 2021, the impact of the pandemic remains, and managing demics explored several opportunities for partnerships with health- clinical placement shortages has become part of the educational care providers and researchers, including novel collaborations to context in Australia and internationally. The reasons for placement deliver high-quality evidence-based care in clinical research settings. cancellations are multifactorial and include restrictions on face-to- We initially consulted with all research theme leaders across the face consultations and redeployment of physiotherapists to COVID- University of Sydney School of Health Sciences and collated infor- 19-related work.1,5 Other contributing factors may include insufficient mation about past successfully implemented research-integrated personal protective equipment and concerns that students could be placements across other health disciplines (eg, Bachelor of Applied infected and transmit the virus unknowingly.6 Placement shortages Science - Exercise Physiology). Due to the nature of the pandemic, it have presented a substantial challenge for progressing students was identified that telehealth-based clinical trials in physiotherapy through their degree, particularly through their final year and into the would be targeted. To ensure an appropriate learning experience for future healthcare workforce. A shortage of new graduates is a na- physiotherapy students, the placement model was co-designed by tional threat, as timely graduation and registration are critical to joint efforts from WIL academics and researchers, considering the workforce sustainability,5 especially during the pandemic when research context, accreditation requirements, educational practices physiotherapists have several roles in managing patients admitted to and students’ learning needs.8 hospitals with confirmed or suspected COVID-19.2 Piloting clinical research placements within trials In response to COVID-19, rapid innovations in clinical education have been developed to provide authentic learning experiences for The clinical research placement was piloted by integrating phys- physiotherapy students to develop their clinical skills. One of the iotherapy students into two telehealth trials led by physiotherapy educational solutions that has been implemented successfully to researchers. The trials from the Musculoskeletal Research Group at provide real-life experiences of evidence-based patient management the Sydney School of Health Sciences investigated physiotherapist- for physiotherapy students, while adhering to accreditation re- led interventions for people with chronic musculoskeletal condi- quirements,5 is ‘clinical research placements’. In this model, clinical tions (knee osteoarthritis and low back pain). These trials were the education is delivered in the context of research, whereby students few unaffected by the social distancing measures implemented by are integrated in research projects that involve the delivery of local governments during the pandemic. Both trials involved the evidence-based care. This Editorial provides an overview of the strategies explored to identify adequate learning opportunities for https://doi.org/10.1016/j.jphys.2021.08.008 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

236 Editorial Table 1 Examples of learning and teaching opportunities for students and clinician-researchers involved in the clinical research placement pilot. Activities Student learning opportunities Added value for research teams Engaging in telehealth Students were exposed to safe evidence-based practice, communication While learning, students assisted with background research, preparation consultations styles, behaviour change strategies, various physical therapies and for therapy and assessment sessions, delivery of therapy or aspects of Simulated telehealth consultations interventions, exercise prescription and advice, goal-setting and therapy and follow-up consolidation activities. Screening of potential management planning of musculoskeletal clients. participants Clinical notes and Students practised skills learnt during placement with peers and Throughout clinical education, clinician-researchers mentored students monitoring of participants Quality improvement clinician-researchers, including communication, clinical reasoning, and promoted knowledge translation. project (eg, literature review and presentation) implementation of therapy and management plans. Students participated in screening for red flags, diagnosis evaluation and Students assisted with screening and quality control of participant entry triage of patients. into trials. Students liaised regarding management plans for participants, follow-up Students assisted therapists to maximise individualised therapy and and reference to baseline data. documentation. Students received additional training on the conditions being studied, Students prepared material that can be used for knowledge translation research methodology, literature review and research communication and presentations, generating rich discussions on the topic and ideas for skills. Students gained experience presenting to an audience of future clinical practice and research. researchers and clinicians. delivery of telehealth interventions focusing on education, strength efficiently and develop skills in research communication. This model training, pain management strategies, physical activity prescription may be particularly relevant to provide learning opportunities for and aspects of cognitive behavioural therapy. Students engaged in physiotherapy students in several areas that are common barriers to most of the activities and, after participant consent was obtained, evidence-based practice, such as confidence to identify and critically they had opportunities to observe and practise their skills with a appraise research. variety of patient presentations. The educational model that was used was shared care,9 where experienced physiotherapists led clinical Observations about clinical research placements assessments and key aspects of therapeutic delivery, while student primary roles included support of clinicians through various aspects To date, 21 physiotherapy students have completed a clinical of healthcare delivery and research. This was identified as the most research placement designed to address placement shortages and appropriate educational approach for the pilot, to ensure adequate progress students at the University of Sydney. As part of clinical learning experiences for students while leaving the original study placement quality control, we gathered general information on the design and aims unaffected. acceptability and perceived value to students completing these placements via debriefings with the WIL academics, self-reflection Similar to a traditional placement, a full-time 5-week program forms and unit of study survey. Overall, students reported valuable provided students with the opportunity to be clinically involved in learning experiences related to the combination of clinical and managing patients with musculoskeletal conditions. Students research activities and enjoyed being part of a clinical research team. engaged in a 100% online placement, which was designed to be safe They found that immersion in clinical research was effective in pro- and offered a variety of learning opportunities to enhance student moting their learning and expanding their clinical skills. We observed skills, knowledge and employability in telehealth care and evidence- that students often spoke enthusiastically about their placement based practice (Table 1). This placement model was unique in that the experiences and reported a good balance between learning about students’ education was delivered by clinicians and researchers. The research while optimising clinical knowledge and practical skills. learning experiences were designed to expand students’ under- Students suggested that the placement may assist some in contem- standing of the latest evidence but to also offer wider insights into the plating participation in clinical research and research training after process of how evidence is developed and applied in musculoskeletal graduation (eg, PhD). A summary of the suggested benefits and bar- health. The placement required students to have both telehealth riers to clinical research placements noted by the authors is listed in clinical days, where students engaged with telehealth physiothera- Box 1. pists during their consultations and assisted with therapy delivery (eg, exercise prescription) and clinical notes, as well as research days, Clinicians and researchers reported high levels of satisfaction in where students worked with the study project manager on research- working together to educate physiotherapy students and suggested based tasks for a quality improvement project. On completion of the that this educational model provides mutual benefits (Box 1). Cli- pilot, a registered physiotherapist assessed the five physiotherapy nicians and researchers observed that students gained and applied students involved in the project, and they met the learning outcomes evidence-based physiotherapy skills while assisting with clinical according to the Assessment of Physiotherapy Practice.10 and research tasks. Students learned from and supported clinical researchers with a variety of tasks including recruitment, prepara- Embedding clinical research placements across healthcare tion for assessment/therapy sessions, writing clinical notes and settings background research. With the quality improvement projects in the pilot, students assisted the research team by developing educational Following the successful pilot, acceptability of clinical research materials that will be used in knowledge translation and placements among academics, clinicians, researchers and students presentations. has been growing. Clinical research placements have now been embedded in tertiary and community healthcare settings across Clinicians and researchers valued the opportunity to teach, Sydney, including a research institute embedded in these settings. inspire and contribute to the future workforce by providing clinical The Institute for Musculoskeletal Health is a research collaboration and research education; however, potential barriers were identified between Sydney Local Health District and the University of Sydney in the implementation of this educational model. Clinical education and is offering ongoing clinical research placement opportunities for often needed to be tailored to the individual student’s knowledge physiotherapy students. As part of this placement, physiotherapy and skill level. To provide adequate student support, training and students engage in clinical trials and evidence-based training with time commitments were required from researchers, clinicians and leading musculoskeletal researchers and participate in traditional WIL academics implementing the placement. As with any new clinical activities with physiotherapists from the Local Health District. clinical educator, the research team comprising clinicians and re- Students complete the online Physiotherapy Evidence Database searchers worked with WIL academics to ensure that adequate (PEDro) Scale training program to understand and appraise trials learning opportunities were available for students (eg, placement manual, orientation session and supervision strategies). Another

Editorial 237 Box 1. Benefits and barriers of clinical research placements. Benefits for students  Experience how research is implemented in clinical practice  Gain and apply skills associated with the presentation of research to an audience of clinicians and Benefits for clinician- researchers researchers Barriers for students  Develop confidence with how to find and critically appraise evidence and to link evidence to practice  Gain a better understanding of research methods and how research is developed Barriers for clinician- researchers  Satisfaction educating the next generation of physiotherapists  Assistance with clinical activities (eg, preparation for therapy and assessment sessions, delivery of therapy or aspects of therapy and writing clinical case notes for documentation)  Assistance with clinical research activities (eg, recruitment, screening, data collection and literature review)  Difficulty finding, understanding and interpreting evidence from published papers prior to receiving the additional training from the clinician-researchers/researchers  Challenge in interpreting de-identified data from assessment timepoints (eg, identification of the baseline pain levels from the raw data of a questionnaire)  Difficulty understanding research study protocols  Some lack of clarity about what is expected from students in research tasks  Lack of practice with clinical education  Limited time to design and implement educational activities and supervision  Concerns with research ethical constraints when involving students  Need to provide induction and training for students every 5-week block  Variability in recruitment rate can affect students’ activities barrier raised by researchers was related to research ethics re- Ethics approval: Nil. quirements: for each research project, researchers need to identify Competing interests: The authors declare that they have no appropriate activities that could be conducted with students and competing interests. consult with ethics committees for any potential changes to Source(s) of support: MS is supported by the Sydney Research research activities. Accelerator (SOAR) fellowship. Acknowledgements: We thank each of the highly committed What does the future hold? clinicians and researchers that promptly engaged in the education of the future physiotherapy workforce. We are grateful to the support The impact of COVID-19 in health sectors and education is from the Musculoskeletal Research Group at the Sydney School of extending well into 2021 and beyond. This challenging time has Health Sciences and Institute for Musculoskeletal Health for collab- created several opportunities to re-think clinical education for orating with Workplace Integrated Learning to design and implement physiotherapy students. Unquestionably, innovations in clinical edu- clinical research placements during the COVID-19 pandemic. cation are necessary to address placement shortages and ensure Provenance: Not invited. Peer reviewed. timely graduation for physiotherapy students to enter the healthcare Correspondence: Amabile Borges Dario, Work Integrated workforce. Clinical research placements are one potential Learning, Sydney School of Health Sciences, Faculty of Medicine and solution—assisting students in developing clinical skills while un- Health, The University of Sydney, Sydney, Australia. Email: derstanding how research is developed and applied in authentic [email protected] healthcare settings. From an educational perspective, a clinical research placement is a suitable placement model for maintaining References access to a range of opportunities for students to build their clinical skills and promote evidence-based practice. We are now expanding 1. Haines KJ, et al. J Physiother. 2020;66:67–69. partnerships with the physiotherapy research community to inte- 2. Thomas P, et al. J Physiother. 2020;66:73–82. grate students into clinical research projects and promote research 3. Quigley A, et al. Physiother Can. 2021;71:1–2. capacity. We encourage clinical researchers to consider integrating 4. Kay J, et al. Int J Work-Integr Learn. 2020;21:491–503. clinical education in their projects. This will assist in promoting op- 5. Australian Government Department of Health. National principles for clinical edu- portunities for evidence-based practice and ensure that emerging physiotherapy graduates have applied knowledge and experience in cation during COVID-19. 2020. https://www.adc.org.au/sites/default/files/Media_ research methodology. Libraries/National_principles_for_clinical_education_during_the_COVID19_pandemic_ FINAL_17.04.20.pdf. Accessed 16 August, 2021. The impact of COVID-19 has forced innovations in clinical educa- 6. Rose S. JAMA. 2020;323:2131–2132. tion that may assist future graduates to value clinical research and 7. Fleming J, et al. Int J Work-Integr Learn. 2018;19:321–335. reduce the existing gap between evidence and practice in physio- 8. Nisbet G, et al. Med Educ. 2021;55:45–54. therapy.11 Future clinical research occurring in professional practice 9. Clinical Education Managers of Australia and New Zealand. Physiotherapy student areas should consider integrating clinical education at the inception clinical placements: A guide for private practice providers. Compiled by members of of research design and funding applications, to ensure that this the physiotherapy CEMANZ committee; 2018: https://australian.physio/sites/default/ research experience and evidence-based practice is standard practice files/Physiotherapy_student_clinical_placements.pdf. Accessed 16 August, 2021. in physiotherapy education in Australia and internationally. 10. Reubenson A, et al. J Physiother. 2020;66:113–119. 11. Zadro J, et al. BMJ Open. 2019;9:e032329. Websites PEDro Scale training program. https://training.pedro.org.au/

Journal of Physiotherapy 67 (2021) 240–241 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Editorial Knee osteoarthritis Anthony J Goff a,b, Mark R Elkins c,d a Singapore Institute of Technology, Singapore; b La Trobe Sport and Exercise Medicine Research Centre, La Trobe University, Melbourne, Australia; c Editor, Journal of Physiotherapy; d Faculty of Medicine and Health, University of Sydney, Sydney, Australia This Editorial introduces another of Journal of Physiotherapy’s A qualitative study examined the experiences of physiotherapists article collections.1–3 These are collections of papers in a specific field of research, published in the Journal of Physiotherapy within the past delivering care for people with knee OA and how these experiences aligned with a national clinical guideline.18 Physiotherapists reported decade and curated to alert readers to important findings and research trends in that field, while highlighing avenues for further that their management of knee OA was mostly consistent with the research. This latest article collection examines physiotherapy quality care standard. Nevertheless, the authors concluded that research into knee osteoarthritis (OA). ‘Knee’ and ‘osteoarthritis’ were physiotherapists could improve their management by increasing the two of the top 20 search terms in an analysis of the searches con- psychosocial focus of care, offering longer-term reviews, and being ducted of the Physiotherapy Evidence Database (PEDro).4 more proactive with advice and/or referral regarding weight loss, Up to 25% of people over the age of 50 years have knee OA,5,6 with pain medications and knee surgery.18 many progressing to knee replacement surgery.7 Current clinical Diet-induced weight loss improves function as a standalone practice guidelines recommend knee replacement surgery only after intervention and improves pain when combined with exercise ther- apy for people with knee OA.19 These findings are similar to those of first-line and second-line management options have been exhaus- patient education and support guideline recommendations that pa- ted.8,9 Such non-surgical management options include patient edu- tient education, exercise therapy and weight loss (when appropriate) cation, exercise therapy, weight management and advice about should be provided in combination as first-line management for knee pharmacological management of symptoms.8,9 OA.8,9,11–14 Whilst physiotherapists regularly provide patient educa- tion and exercise therapy, they less frequently provide weight loss Most of the recommended non-surgical interventions have been support for people with knee OA.18,20 The reasons for this are multifactorial; however, research from the Journal of Physiotherapy included in clinical practice guidelines because of robust evidence; provides some key insights, with Teo et al identifying that physio- however, education has often been recommended based on evidence therapists can perceive weight loss as outside their professional role18 related to OA of other joints or other forms of arthritis.10 To address and Setchell et al identifying that many physiotherapists demonstrate this, a recent systematic review pooled the available evidence about weight stigma.21 Combined, these findings may impact the provision education for knee OA.10 It found that patient education may reduce of guideline-recommended first-line management of knee OA and should be explored further in clinical research. pain and improve function compared with usual care, although these Although the benefits of exercise are well established, people with effects may not be large enough to be clinically important in isolation. knee OA do not always adhere to a formal exercise regimen or maintain adequate levels of physical activity in the longer term.22 However, combining patient education with exercise therapy should be People with knee OA reported that lack of motivation, time, phys- encouraged, given the statistically superior and clinically important ical environment and monitoring were barriers to exercise and gen- improvements in function compared with patient education alone.10 eral physical activity.22 These barriers seemed similar, regardless of the levels of supervision, individualisation and progression that were Cryotherapy is another non-surgical intervention with unclear evidence in knee OA. Some guidelines recommend it,11,12 whereas provided when exercise was commenced. Instead, the presence and others do not.8,13,14 A recent systematic review summarised the quality of a therapeutic alliance with a physiotherapist facilitated available evidence and determined that in people with symptomatic adherence to exercise and general physical activity. Another strategy that might assist with adherence to exercises is the use of an app with knee OA, any beneficial effect of the short-term application of cryo- therapy on pain is so small that most would not consider it to be remote support to prescribe and monitor the formal home exercise worthwhile.15 The effects of short-term cryotherapy on function and program. Compared to merely having the exercise program pre- quality of life in people with symptomatic knee OA were unclear. scribed on paper, patients who had a home exercise program for their musculoskeletal pain prescribed on an app with remote support Conflicting advice exists within clinical practice guidelines about adhered better to their home exercise program.23 The randomised the use of manual therapies, including manipulation, mobilisation trial that established this beneficial effect on adherence did not have a and massage, in the management of knee OA. Local16 and interna- precise enough estimate to prove that the effect on adherence is tional11 guidelines conditionally recommend the use of manual clinically worthwhile. Nevertheless, this should not discourage the techniques but only as an adjunct to first-line management, whereas use of the app for home exercise programs for knee OA because the other international guidelines either conditionally recommend that app is freely available, has high user satisfaction, permits adherence manual techniques are not used12 or have not included any recom- monitoring and is quick and easy to use.23 mendations about manual therapies due to insufficient evidence.8,13 A systematic review of the effects of massage17 identified three trials in It may be particularly appropriate to consider use of the app with knee OA. The pooled short-term effects on knee pain and function remote support23 for people who have progressed to surgical knee were each estimated to be clinically worthwhile, but the confidence intervals spanned both worthwhile and trivial effects. Overall, the certainty of this evidence was low, partly because only two of the three trials contributed to each meta-analysis. https://doi.org/10.1016/j.jphys.2021.08.009 1836-9553/© 2021 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

Editorial 241 replacement, given that they also struggle to maintain satisfactory Competing interests: Nil. exercise adherence. Patients with lower limb orthopaedic conditions Source(s) of support: Nil. in inpatient rehabilitation are relatively inactive and do not meet Acknowledgements: Nil. current physical activity guidelines. Given the importance of physical Provenance: Invited. Not peer reviewed. activity for general health and functional improvements following Correspondence: Mark R Elkins, Centre for Education & Workforce hospitalisation, it is important to develop methods to decrease Development, Sydney Local Health District, Sydney, Australia. Email: sedentary behaviour and increase physical activity levels in rehabil- [email protected] itation.24 After total knee arthroplasty, 42% of people were not active enough to maintain their health and fitness.25 Physiotherapists References should encourage people with a total knee arthroplasty to undertake the recommended exercise regimens to maintain health and fitness. 1. Bonnevie T, et al. J Physiother. 2020;66:3–4. Males, more educated participants and respondents living with 2. Dennett A, et al. J Physiother. 2020;66:70–72. family were found to have higher odds of meeting the health, fitness 3. Hwang R, et al. J Physiother. 2020;66:193–195. and both recommendations;25 therefore, particular attention should 4. Stevens ML, et al. Methods Inf Med. 2016;55:333–339. be given to those people with characteristics known to be associated 5. Litwic A, et al. Br Med Bull. 2013;105:185–199. with poor adherence to the recommendations. 6. Cross M, et al. Ann Rheum Dis. 2014;73:1323–1330. 7. Ackerman IN, et al. BMC Musculoskelet Disord. 2019;20:90. Despite general symptomatic improvement following knee 8. McAlindon TE, et al. Osteoarth Cartil. 2014;22:363–388. replacement, up to one in five people remain dissatisfied with out- 9. Australian Commission on Safety and Quality in Health Care. Osteoarthritis of the comes,26 and can experience persistent muscle weakness, ongoing functional difficulties and pain when compared with healthy age- Knee Clinical Care Standard; 2017. https://www.safetyandquality.gov.au/sites/defa matched controls.27 A large randomised controlled trial sought to ult/files/migrated/Osteoarthritis-of-the-knee-Clinical-Care-Standard.pdf. Accessed determine whether outcomes related to these deficiencies could be 12 August, 2021. improved by incorporating hip abductor strengthening exercises into 10. Goff AJ, et al. J Physiother. 2021;67:177–189. a 6-week rehabilitation program.28 However, similar improvements 11. Chae KJ, et al. National Institute for Health and Care Excellence. Osteoarthritis: Care in muscle strength, functional performance and patient-reported and Management. December 2014. outcomes were observed whether specific hip-strengthening exer- 12. Kolasinski SL, et al. Arthritis Care Res. 2020;72:149–162. cises were incorporated or general functional exercises were 13. Fernandes L, et al. Ann Rheum Dis. 2013;72:1125–1135. continued instead as part of a postoperative rehabilitation program 14. Brosseau L, et al. Clin Rehabil. 2017;31:596–611. for participants after knee replacement.28 15. Dantas LO, et al. J Physiother. 2019;65:215–221. 16. The Royal Australian College of General Practitioners. Guideline for the management In summary, this online article collection includes a range of of knee and hip osteoarthritis. 2nd ed. Melbourne: RACGP; 2018. Available from important developments in the physiotherapy management of knee https://www.racgp.org.au/download/Documents/Guidelines/Musculoskeletal/guide OA. The papers in the collection also highlight some important unan- line-for-the-management-of-knee-and-hip-oa-2nd-edition.pdf swered questions. For example, while education and exercise seem to 17. Bervoets DC, et al. J Physiother. 2015;61:106–116. be effective in the rehabilitation of knee OA, are there more effective 18. Teo PL, et al. J Physiother. 2020;66:256–265. approaches in the provision of patient education and self-management 19. Hall M, et al. Semin Arthritis Rheum. 2019;48:765–777. (eg, theory-based, co-designed)? Also, are the exercises that we 20. Spitaels D, et al. Musculoskelet Sci Pract. 2017;27:112–123. currently use ideal or do we need to include other exercises, in- 21. Setchell J, et al. J Physiother. 2014;60:157–162. terventions or dosages (such as aerobic exercise or general increases in 22. Fransen M, et al. Br J Sports Med. 2015;49:1554–1557. physical activity)? Finally, how can physiotherapists be better sup- 23. Lambert TE, et al. J Physiother. 2017;63:161–167. ported to provide weight loss for people with knee OA? These ques- 24. Peiris CL, et al. J Physiother. 2013;59:39–44. tions should be prioritised in future research with a focus on facilitating 25. Groen J-W, et al. J Physiother. 2012;58:113–116. and improving longer-term behaviour change and patient outcomes. 26. Bourne RB, et al. Clin Orthop Relat Res. 2010;468:57–63. 27. Schache MB, et al. Knee. 2014;21:12–20. 28. Schache MB, et al. J Physiother. 2019;65:136–143. Websites PEDro www.pedro.org.au

Journal of Physiotherapy 67 (2021) 243–251 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Invited Topical Review Physiotherapy management of Down syndrome Nora Shields Department of Physiotherapy, Podiatry and Prosthetics and Orthotics, La Trobe University, Melbourne, Australia KEY WORDS [Shields N (2021) Physiotherapy management of Down syndrome. Journal of Physiotherapy 67:243–251] © 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under Down syndrome, Exercise Meta-analysis the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Physical therapy Introduction syndrome, which can be inherited, an additional chromosome 21 is attached to another chromosome. Mosaic Down syndrome or partial Over the past century, the life expectancy of people born with trisomy 21 is the result of a mutation during mitosis, where the Down syndrome has increased from 9 to 60 years,1 with data indi- embryo has a combination of typical cells, as well as cells with a third cating that it could match the general population within a genera- copy of chromosome 21. People with mosaic Down syndrome usually tion.2 This rapid increase in survival is largely due to earlier surgical have fewer clinical features than those with other types of Down correction of heart defects, better treatment of infections and syndrome, but this depends on how early in development the mu- improved general healthcare.3 This success, however, is tempered by tation occurs. the substantially higher risk and early onset of several chronic health conditions in adults with Down syndrome, particularly cognitive Burden of Down syndrome decline commencing in their early 40s. Down syndrome has whole-of-genome and epigenetic effects, Evidence-based clinical guidelines for the medical care of adults with consequences to the structure and function of the nervous, with Down syndrome were recently published, with one strong cardiovascular, musculoskeletal and endocrine systems. The primary recommendation that screening for Alzheimer-type dementia start clinical feature of Down syndrome is intellectual disability, which is after the age of 40 years.4 By comparison, there are no known clinical usually moderate but can range from mild to severe. Central nervous practice guidelines to support physiotherapy management.5 The system structural differences include a smaller cerebrum, cerebellum physiotherapy profession has much to offer people with Down syn- and brain stem. Adults with Down syndrome are at ultra-high risk of drome: a wide breadth of practice spanning early infancy through to experiencing early cognitive decline, with a cumulative risk of de- old age, unique contributions as part of a multidisciplinary team in mentia of 45% by 55 years and 80% by 65 years9 compared with 20 to managing chronic health conditions and, specifically, expertise in 35% by 75 years in the general population. Common structural dif- exercise and physical activity. Therefore, it is timely to review the ferences in the cardiovascular system are congenital heart defects, available evidence to guide physiotherapists in their management of affecting 40 to 55% of infants with Down syndrome.10 Although the Down syndrome. This review focuses on the most recent evidence limited available evidence suggests a reduced risk of atherosclerotic (published since 2000) from randomised controlled trials involving cardiovascular disease,4 cardiovascular functioning is usually people with Down syndrome across their lifespan. It provides an compromised in people with Down syndrome who have very low evaluation of interventions that are either currently within the scope cardiorespiratory fitness,11 altered physiology (such as autonomic of, or could become part of, physiotherapy practice. dysfunction),12 and are at higher risk of stroke across all age groups than the general population.13 Respiratory illness is the most com- What is Down syndrome? mon reason for hospitalisation of children,14 and pneumonia is a leading cause of death in older people with Down syndrome.15 Down syndrome (or trisomy 21) is the most common genetic Obstructive sleep apnoea is also more common and more severe in cause of intellectual disability,6 occurring in an estimated 1 in 800 adolescents and adults with Down syndrome than in the general births worldwide.6 Approximately 11,000 people with Down syn- population.16 drome live in Australia7 and 250,000 in the USA.8 Down syndrome is caused by the over-expression of normal genetic material, usually an Down syndrome is associated with extensive musculoskeletal extra chromosome 21. Trisomy 21 occurs by meiotic nondisjunction, sequelae, including muscle weakness, hypermobility, ligamentous when the egg or sperm carries an extra copy of chromosome 21.6 The laxity and skeletal deformities. Musculoskeletal conditions such as risk of trisomy 21 increases with parental age. In translocation Down atlantoaxial instability, scoliosis, foot deformities, and hip and https://doi.org/10.1016/j.jphys.2021.08.016 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

244 Shields: Physiotherapy management of Down syndrome patellar instability are common. Prevalence estimates of atlan- gravity; an inability to initiate weight shift; ineffective postural con- toaxial instability from population-based studies are 4 to 8%,17,18 trol; the tendency to become stuck in a position; and difficulty although less than 1 to 2% of people with Down syndrome developing fine motor skills. In many countries, infants with Down develop symptoms of spinal cord compression,19 and permanent or syndrome would expect to receive early intervention support from a sudden damage to the spinal cord rarely occurs without previous physiotherapist. Despite this, six trials (n = 144) supporting evidence- neurological symptoms.20 Annual health screening to check for based early intervention practice by physiotherapists have been signs and symptoms of cervical myelopathy using targeted history published; these trials investigated parent-delivered treadmill and physical examination are recommended instead of routine training and massage therapy. cervical spine x-rays in asymptomatic people with Down syn- drome.4 Prevalence estimates of scoliosis rise from 2% in children18 Parent-delivered treadmill training to 7% in adolescents and young adults,17 and increase after thora- Treadmill training aims to promote earlier independent walking cotomy for the treatment of congenital heart disease.21 The risk of major complications after surgical correction of scoliosis is greater among infants with Down syndrome and the development of a more among those with Down syndrome than those with idiopathic typical walking pattern. Achieving the fundamental skill of walking is scoliosis,22 likely due to more concurrent comorbidities in those desirable, as it provides opportunities to interact with the environ- with Down syndrome. Prevalence estimates for hip and patellar ment, facilitating motor, social and cognitive development. Three instability among children with Down syndrome are 1 to 7% and 1 trials (n = 77),33–42 completed by one research group, found imme- to 4%, respectively. These conditions are usually due to bony diate and longer-term benefits for infants with Down syndrome from anomalies23,24 and can interfere with walking. Young people with home-based, parent-delivered treadmill training in terms of devel- Down syndrome have a high incidence of foot deformities, such as opmental milestone attainment, movement efficacy and physical flat feet (76%),25 and many complain of foot problems that interfere activity levels. with daily life.17 Recent research shows that children with Down syndrome have shorter and wider feet with greater girth,26 which An initial trial39 (n = 30) found that infants with Down syndrome can make shoe fitting problematic. who commenced treadmill training in addition to their usual care when able to sit independently at around 10 months walked on The endocrine system is adversely impacted by Down syndrome, average 3 months earlier than those who received usual care only with increased incidence of osteoporosis, thyroid disease, diabetes (MD 101 days, 95% CI 18 to 184). A subsequent trial (n = 30), reported and obesity. Six studies (n = 796),4 rated as poor quality by recent across seven articles,33–35,38,40–42 found that infants with Down syn- clinical guidelines, have reported wide-ranging prevalence estimates drome completing high-intensity treadmill training, incorporating for osteoporosis (1 to 45%) among adults with Down syndrome. progression via belt speed, duration and resistance, also achieved Higher rates of osteoporosis in this cohort are likely related to lower motor milestones earlier,38 displayed more typical walking pat- bone mineral density, early menopause, thyroid disease, low physical terns,34 more advanced maturation of joint kinematics,41 earlier activity and muscle weakness. However, there is some evidence that adoption of mature strategies to walk over obstacles42 and spent triplication of certain genes on chromosome 21 itself confers a risk.15 more time in higher-intensity physical activity33 than those receiving Prevalence estimates for diabetes are also higher among adults with lower-intensity training. A third trial (n = 17)36,37 reported that Down syndrome compared with the general population across all age supramalleolar orthoses may have a detrimental effect on gross groups: 3.5% versus 0.7% for young adults and 5.5% versus 2.7% for motor skills development of infants with Down syndrome. Although adults aged  30 years.27 The prevalence of hypothyroidism is high the effect on time to onset of walking was unclear, the infants who (27% across 19 cohort and case-control studies),28 placing people with had worn orthotics in addition to treadmill training, from the time Down syndrome at higher risk of depression.29 Indeed, population- they could pull themselves to standing at around 20 months, had based data indicate that about one-third of young people with lower gross motor function measure (GMFM) total scores and lower Down syndrome have a mental health condition17 such as anxiety standing and walking scores (D subscale) and running and jumping (22%) or depression (11%). Being overweight and obese is another scores (E subscale) 1 month after walking onset37 compared with major problem in adolescents (61%)30 and adults (72%)31 with Down those who did not wear orthotics. The effect on the amount of hand syndrome. These rates of mental health conditions and obesity are support used by infants when standing upright, as a proxy for substantially higher than in the general population. Other common perceived stability, was unclear.36 Therefore, orthotics should not be co-occurring impairments among people with Down syndrome are prescribed for infants with Down syndrome prior to the onset of vision impairment, hearing impairment, speech and language dis- walking. Future studies are needed to determine if orthotics confer a abilities and behavioural conditions.4 beneficial or detrimental effect on young children with Down syn- drome when prescribed after the onset of walking. Physiotherapy management of Down syndrome Despite the favourable evidence, a recent US-based survey of A physiotherapist’s role when working with a person with Down current practice43 found that only 6.5% of physiotherapists imple- syndrome depends on life-stage and usually relates to facilitating mented treadmill training, with almost a third indicating that this physical activity. Depending on individual presentation, early in life was because of the requirement for specialist equipment (a custom- the focus is on optimising motor skills and minimising development ised treadmill operating at 0.2 m/s). This intervention also requires of abnormal compensatory movement patterns. During adolescence substantial parental effort, time commitment and training, and and young adulthood, the focus is on maximising physical and mental ongoing therapist support. While some families may like the program health. In adulthood, the focus is on maintaining function, slowing structure, knowing what to do, how to do it and for how long, it may physical deterioration due to early ageing and delaying the onset of be impractical for others implementing this training at home for 8 Alzheimer’s disease. minutes a day, 5 days per week for approximately 10 months. In- dications of difficulty adhering to the intended training protocol have been reported;37,40 however, data on how families felt about the intervention are unavailable. Early intervention Massage therapy Typically, infants with Down syndrome learn to walk but are Massage therapy is not typically prescribed for children with delayed in attaining this milestone. The probability that a child with Down syndrome will walk by 24 months is 40%, by 30 months it is developmental motor disorders, as it is a passive intervention 74% and by 36 months it is 92%.32 Delayed and abnormal motor requiring minimal physical demand from the recipient.44 However, development can include reduced movement, especially against there is some evidence that massage therapy can support develop- mental growth in preterm infants,45 and early intervention treat- ments for children at risk of, or with, developmental motor disorders

Invited Topical Review 245 often incorporate sensory elements such as tactile stimulation.46 Study WMD (95% CI) Three trials44,47,48 (n = 67) have reported positive short-term effects Millar Random Varela on development (global, motor, visual-motor, language, social) in Rosety-Rodriguez Boer & Moss (continuous) favour of massage therapy as an adjunct to early intervention for infants aged 4 to 8 months47 and young children (mean age 2 years) Total with Down syndrome.44,48 Massage therapy was delivered in two trials47,48 by trained staff weekly and at home by (trained) parents –10 –5 0 5 10 daily for 10 to 15 minutes for 1 month47 or 5 months,48 and in one Favours no training (m/kg/min) Favours aerobic training trial exclusively by a massage therapist for 30 minutes, twice a week for 2 months.44 Only the latter trial44 controlled for attention to ac- count for the increase in contact, communication and dedicated time spent with the infant with Down syndrome. Summary of trials relating to early intervention Figure 1. Mean difference (95% CI) in effect of continuous aerobic exercise on relative Early developmental intervention is frequently used by phys- VO2peak (ml/kg/min), estimated by pooling data from four studies (n = 78). WMD = weighted mean difference. iotherapists to improve overall functional outcomes for infants at risk of developmental delay and is supported by growing evidence. Exercise However, most early intervention trials systematically exclude infants with chromosomal conditions, intellectual disability and/or Aerobic training congenital heart defects. As shown, the evidence about early People with Down syndrome have very low cardiorespiratory intervention specifically for infants with Down syndrome is sparse, but (with the caveat that these trials lack methodological fitness (relative VO2peak); the average relative VO2peak of adoles- rigour) they support parent-delivered interventions with a pri- cents and young adults with Down syndrome is equivalent to that of a mary focus (repetition of movement or sensory stimulation) and 60-year-old with heart disease.55 Low cardiorespiratory fitness daily application to promote earlier developmental milestone negatively impacts participation in daily and recreational activities of attainment. people with Down syndrome.56 It is also likely associated with the high rates of secondary health conditions in people with Down syn- Task-specific training interventions drome (eg, stroke, cancer), given the strong associations between relative VO2peak and chronic disease in the general population.57,58 The gap in gross motor skills abilities between children with and Aerobic training (including walking, swimming, cycling, rowing and without Down syndrome widens with age.49 Task-specific training exergaming) increases relative VO2peak in the general population involves massed practice of participant-relevant, context-specific when implemented at sufficient intensity, frequency and time.59 It is tasks, where the intervention focuses on the skills needed for a the exercise intervention most studied in Down syndrome, with 20 task.50 The effect of task-specific training on gross motor abilities trials60–83 investigating outcomes related to cardiovascular fitness, among children with Down syndrome has been investigated in two body size, functional activities, low-grade inflammation and cogni- trials49,51 (n = 73); both showed benefits when implemented tion. There is substantial heterogeneity within these trials in terms of during early adolescence. A 5-day program for learning to ride a participant age (range 4 to 63 years), program duration (6 to 36 two-wheel bicycle,49 in a summer camp setting and comprising weeks) and training intensity (low to high intensity) but all trials intensive individualised instruction for 75 minutes per day, incorporated supervision in the intervention design. enabled 19 of 36 (56%) participants (mean age 12 years) to achieve independent cycling over 9 m. Participants commenced training on A Cochrane review,84 published in 2010, was inconclusive about adapted bicycles and progressed to their own standard two-wheel the effects of aerobic training on cardiovascular fitness outcomes for bicycle by the last day of the program. Compared to a wait-list adults with Down syndrome. The addition of two subsequent tri- control group, at 12 months those who completed the program als62,66 of continuous aerobic training to their forest plots shows spent less time sedentary, spent more time in moderate to vigorous similar inconclusive findings for relative VO2peak (Figure 1) and physical activity and had lower subcutaneous fat.49 Similarly, nine pulmonary ventilation, but there was a between-group difference for 45-minute sessions of throwing instruction over 3 weeks improved maximum test time (time to exhaustion) in favour of continuous throwing accuracy, but not coordination, compared with everyday aerobic training compared with no exercise, based on data from two activities in adolescents with Down syndrome (mean age 13 trials62,67 (Table 1); for a more detailed forest plot, see Figure 2 on the years).51 eAddenda. A single trial62 reported a between-group difference in favour of interval training compared with continuous aerobic exercise Summary of trials relating to task-specific training for relative VO2peak, pulmonary ventilation and time to exhaustion. Most children with Down syndrome eventually learn a basic When data from five trials that implemented any exercise training (ie, interval or continuous aerobic and/or strengthening training) were repertoire of motor skills49 and continue to develop motor profi- combined, between-group differences in favour of exercise were ciency into adolescence.52 Although evidence-based motor in- found for relative VO2peak (Figure 3) and pulmonary ventilation; for terventions involving task specificity are frequently applied by more a detailed forest plot, see Figure 4 on the eAddenda. The lower physiotherapists in adult neurological rehabilitation, the emphasis on cardiorespiratory fitness of people with Down syndrome may be motor development in clinical programs for children and adolescents partially explained by lower muscle quantity and quality.85 This may with Down syndrome usually declines once they can walk. Based on explain why programs including aerobic and strengthening exercise two trials, task-specific training appears beneficial for children and have a positive effect on cardiorespiratory fitness, when aerobic adolescents with Down syndrome to learn and develop proficiency in training alone have not. motor skills at doses (6.25 to 6.75 hours) lower than children with cerebral palsy require to achieve upper limb individual goals (14 to 25 Meta-analyses show that aerobic training has positive effects on hours) or general function goals (30 to 40 hours).53 Continued focus physical function (6-minute walk distance, sit-to-stand test, Timed by clinicians on motor skills beyond walking is especially important Up and Go), waist circumference and body mass index (BMI) for those with Down syndrome, given the consequences for their compared with no exercise (Table 1). One trial80 also reported participation in education and recreational settings, but also because between-group differences in favour of swim training (three times a motor ability better predicts functional limitations in this population week for 50 minutes, in groups of up to eight, for 36 weeks) than cognitive performance.54 compared with recreational water games (twice a week) on waist circumference, BMI, percentage body fat and some skinfold

246 Shields: Physiotherapy management of Down syndrome Table 1 Trials (n) Participants (n) MD (95% CI) Certainty Summary of findings. Ex Con Outcome Continuous aerobic training versus no exercise 4 41 37 0.84 (21.38 to 3.06) lowa,b 3 VO2peak, (ml/kg/min) 2 30 28 6.46 (20.81 to 13.73) lowa,b Pulmonary ventilation, (l/min) 6 Time to exhaustion, (min) 4 21 24 2.23 (1.11 to 3.35) lowa,b Body weight, (kg) 4 Percentage body fat, (%) 5 100 104 21.1 (22.4 to 0.1) lowa,b Waist circumference, (cm) 5 Body mass index 5 44 46 21 (23 to 2) lowa,b Timed Up and Go, (s) 3 Six-minute walk distance, (m) 2 44 46 22.7 (25.3 to 20.1) lowa,b Sit-to-stand, (n) Hand-grip strength, (kg) 95 94 21.8 (23.3 to 20.3) very lowa,b,c 60 58 21.7 (22.4 to 20.9) very lowa,b 52 56 51 (26 to 75) lowa,b 32 35 1.8 (0.4 to 3.1) very lowa,b,c 25 29 1.2 (22.6 to 4.9) lowa,b Any exercise training versus no exercise 5 71 59 2.93 (1.25 to 4.62) lowa,b 3 30 28 9.67 (1.49 to 17.85) lowa,b VO2peak, (ml/kg/min) Pulmonary ventilation, (l/min) Con = control, ex = exercise. a Downgraded due to limitations of studies: , 75% of studies were rated at low risk of bias overall. b Downgraded due to imprecision: there were very large confidence intervals with the higher end indicating appreciable benefit and the lower end indicating either a little effect or worse outcome. c Downgraded due to inconsistency: there was statistical heterogeneity (I2 values . 40%). measurements (suprailiac, triceps) in 45 adolescents with Down with 24 weeks of sensory integration, neurodevelopmental therapy syndrome (mean age 14 years). The Cochrane review84 reported data and perceptual motor activities.76 A third trial79 reported between- group differences in overall balance stability in favour of exercising from a single trial showing no effect of aerobic training on two using a ‘suspension system’ compared with treadmill walking and balance training in young children with Down syndrome (4 to 9 anthropometric outcomes (weight, percent body fat) for people with years). In adults with Down syndrome, a between-group difference in Down syndrome. The addition of five subsequent trials61,62,75,77,78 on favour of 8 weeks of swim training compared with no training for body weight, and three subsequent trials62,66,77 on percentage body dynamic balance (walking on a balance beam) was reported61 in one trial but a second trial found no difference when exergaming was fat to their forest plot shows similar findings (Figures 5 and 6). For compared with usual activities.77 more detailed forest plots, see Figures 7 and 8 on the eAddenda. A meta-analysis of two trials62,77 also found no effect on handgrip The effect of aerobic training on a range of other outcomes has been investigated in at least one trial. Four months of aerobic training plus a strength test when compared with no exercise (Table 1). calcium supplement had a greater effect on femoral neck bone density than either aerobic training alone, or calcium supplement alone in one More recent trials have investigated the effects of aerobic training trial73 of 48 children with Down syndrome (mean age 9 years). Compared to no training, small positive changes in muscle strength on low-grade systematic inflammation and cognitive function. One have been reported after exergaming, swim training and treadmill trial,64–66 reported across three articles and involving 20 premeno- training in adolescents,75 adults61 and older adults63 with Down syndrome. Ten weeks of Nordic walking for adults with Down syn- pausal women with Down syndrome (mean age 25 years), found that drome (mean age 31 years) resulted in improvements in spatiotem- poral gait parameters compared with no training. One trial72 (n = 29) aerobic training had positive effects on low-grade systematic in children with Down syndrome (mean age 9 years) found no dif- ferences in pulmonary function tests for those who completed rowing inflammation. Reported between-group differences after 10 weeks of training (12 weeks) compared with positioning, breathing exercises, aerobic exercise, three times per week, for 30 to 40 minutes at 55 to postural drainage and incentive spirometry. Two trials61,77 reported between-group differences for aerobic capacity (modified shuttle 65% peak heart rate indicated reduced levels of pro-inflammatory tests) in favour of 8 weeks of exergaming or swim training compared cytokines (TNF-a, IL6),65 some acute phase proteins (CRP and fibrin- with usual activities or no training. No between-group differences ogen but not a1-antitrypsin),65 and leptin but not adiponectin64 were found for gait speed, physical activity levels, or wellbeing in a immediately after training; however, these changes were mostly feasibility trial78,86 involving 16 young adults with Down syndrome (mean age 21 years) allocated to 8 weeks of 150 minutes of walking not retained at follow-up 3 months later. Three small trials have re- per week compared with sedentary social activities. ported mixed results on the effect of aerobic training on cognitive Progressive resistance training function. One trial (n = 12)81 found no between-group differences in Muscle strength in the upper and lower limbs is up to 50% less in short-term working memory or selective attention for exergaming people with Down syndrome compared with those without disability,87,88 negatively impacting the ability to perform everyday compared with regular activities in older adults with Down syndrome activities.56 Progressive resistance training is regarded as the best (mean age 50 years). Another trial70,71 (n = 27), reported across two way to improve muscle strength in almost all populations, when completed with sufficient intensity and progression of load.59 Four articles, found no differences in cognitive function between groups trials89–93 (n = 151) have investigated the effects of this type of training in adolescents and young adults with Down syndrome, exercising on 2 days compared with 1 day per week for 12 weeks in implementing similar exercises, but with differences in intensity, duration, group size (individual versus small group), and qualification adults with Down syndrome (mean age 28 years) in a program of the supervisor. Three trials91–93 of moderate-to-vigorous intensity delivered by videoconference.70 A third trial (n = 34)68,69 involving young adults with Down syndrome (mean age 18 years) reported between-group differences after 8 weeks of training in favour of assisted cycling compared with voluntary cycling for cognitive plan- ning,68 some executive function outcomes (reaction time and response inhibition) but not others (set shifting and language fluency)69 and for manual dexterity (which was shown to be strongly associated with cognitive planning ability and verbal working memory in adolescents with Down syndrome).68 Five trials have reported effects of aerobic training on balance outcomes in children and adolescents (three trials, n = 165) and adults (two trials, n = 53) with Down syndrome. Two trials,76,82 involving children with Down syndrome, found positive changes in balance in favour of exergaming compared with 6 weeks of strengthening and walking activities82 but no difference compared

Invited Topical Review 247 Study WMD (95% CI) Study WMD (95% CI) Millar Random Varela Random Varela Lin & Wuang Rimmer Shields & Taylor Rosety-Rodriguez Boer & Moss (continuous) Boer & Moss (continuous) Silva Boer Total Total –10 –5 0 5 10 –15 –7.5 0 7.5 15 Favours no exercise (ml/kg/min) Favours exercise Favours aerobic training (kg) Favours no exercise Figure 3. Mean difference (95% CI) in effect of any exercise on relative VO2peak (ml/kg/ min), estimated by pooling data from five studies (n = 130). Figure 5. Mean difference (95% CI) in effect of aerobic exercise on body weight (kg), WMD = weighted mean difference. estimated by pooling data from six studies (n = 204). WMD = weighted mean difference. training (60 to 80% of 1RM) led to increases in muscle strength and in favour of exercise on lumbar spine and hip bone mineral content, lower limb function (measured using the timed stair climb) but not and in lumbar spine bone mineral density. This trial101 also reported upper limb function (measured using grocery shelving task) improvements in physical fitness in favour of conditioning exercise compared with no exercise immediately after a 10-week program.94 but no between-group differences in physical activity at 12 months. Changes in leg but not arm muscle strength were maintained at the 3-month follow-up.93 A between-group difference in physical activity Inspiratory muscle training levels also favoured progressive resistance training at the 6-month Respiratory disorders are a common cause of illness and death in follow-up. Compared to no training, low-intensity to moderate- intensity progressive resistance training (40 to 65% of 8RM) led to a children and adults with Down syndrome.6 Upper, lower and general reduction in low-grade systemic inflammation, indicated by respiratory tract conditions and infections account for 40% of all between-group differences in leptin, TNF-a and IL-6 levels,90 and to hospital admissions in children with Down syndrome.14 Those with reduced susceptibility to infection, indicated by between-group dif- congenital heart disease are at an even higher risk of hospitalisation ferences in IgA levels,89 in men with Down syndrome. for respiratory infections than those without.104 Inspiratory muscle training105 and incentive spirometry106 are two interventions tar- Progressive resistance training was incorporated into the exercise geting pulmonary function to have been investigated in trials protocols of an additional four trials95–98 (n = 136). The largest of involving children with Down syndrome (n = 50). Positive effects on these trials,98 involving 52 adults with Down syndrome (mean age 39 pulmonary function and respiratory muscle strength were reported years), found between-group differences in cardiovascular fitness, in one trial105 involving 16 children with Down syndrome (aged 11 muscle strength and body weight in favour of 12 weeks of combined years), half of whom trained 5 days per week for 4 weeks with an training (30 to 45 minutes of continuous aerobic training at 50 to 70% inspiratory muscle training device set at 40% maximal inspiratory V02peak, 15 to 20 minutes of progressive resistance training at 70% of pressure compared with a control group training at 0% maximal 1RM; 3 days per week) compared with no exercise. Two trials95,96 inspiratory pressure. However, the addition of incentive spirometry involving adolescents with Down syndrome reported improvement to an oromotor exercise program had no effect on either pulmonary in leg muscle strength and balance after either 6 weeks of a combined or oromotor function in a trial106 of 34 children with Down syndrome low-intensity to moderate-intensity progressive resistance and bal- (mean age 8 years). One trial83 investigating respiratory aspects of ance training compared with usual activity or 12 weeks of lower limb speech production reported a between-group difference in favour of isokinetic training and balance/isotonic strengthening program swim training (12 weeks, three times a week, for 60 minutes) for compared with a balance/isotonic strengthening program only. One maximum phonation duration, but not initiation volume or expired trial97 found that a 12-week program of positioning, breathing exer- mean airflow, in adolescents with Down syndrome (n = 28) compared cises and incentive spirometry in 30 children with Down syndrome with no training. (mean age 12 years) had positive effects on pulmonary function (forced expiratory volume in 1 second and maximum voluntary Balance training ventilation, but not for forced vital capacity and peak expiratory flow Limitations to balance and postural control among children and rate) compared with a low-intensity lower limb strengthening program. adolescents with Down syndrome are well documented107 and show the slowest development longitudinally.52 The primary impairment Combined aerobic and non-progressive strengthening exercise programs in the postural control system is reduced muscle tone. Related People with Down syndrome have low bone mineral density, secondary issues are insufficiency of muscular co-contractions, insufficiency of balance reactions, reduced proprioception and which confers a higher risk of osteoporosis in adulthood.15 Exercise is hypermobility. As a result, children with Down syndrome experience important for bone formation during childhood and to maintain bone problems in achieving and maintaining posture and movement, and mass in adulthood, as a mechanical stimulus to encourage bone inadequate development of their motor abilities.108 Six trials60,109–113 modelling and re-modelling.99 Three trials100–103 (n = 100) have (n = 187) have investigated the effect of exercise on balance in chil- studied the effects of ‘conditioning’ exercise programs (comprising dren and adolescents with Down syndrome. In three trials,60,109,110 non-progressive strengthening and aerobic training two or three (n = 91) core stability exercise training for 8 weeks was found to times weekly for 21 to 52 weeks) on bone or muscle mass compared have beneficial effects on static balance compared with either no with no exercise in adolescents and young adults with Down syn- training110 or in addition to strength, balance and postural control drome. One trial102 reported between-group differences for total lean exercises60,109 in children with Down syndrome (mean ages 4 and 9 mass and lower limb lean mass in favour of conditioning exercise. No years). One of these trials60 also reported between-group differences effect on bone mineral content100,103 or bone mineral density100 was for functional balance (Berg balance scale). Between-group differ- reported in two trials implementing a 6-month program, but a larger ences were also reported in static and dynamic balance after 6 weeks trial101 implementing a 12-month program reported positive effects

248 Shields: Physiotherapy management of Down syndrome Study WMD (95% CI) Summary of trials relating to health education and behaviour change Varela Random interventions Rosety-Rodriguez Boer & Moss (continuous) Healthy eating and physical activity are complex health behav- Silva iours impacted by capability, motivation and opportunity. Having an intellectual disability and limitations in literacy further impact these Total health behaviours. Based on two trials, health education and behav- iour change interventions, which physiotherapists could implement –10 –5 0 5 10 as part of a multidisciplinary team, have benefits for adults with Favours aerobic training (%) Favours no training Down syndrome in improving attitudes towards exercise, participa- tion in physical activity and weight loss. Indeed, the level of weight Figure 6. Mean difference (95% CI) in effect of aerobic exercise on percentage body fat loss achieved with a group-based intervention (24 hours total dura- (%), estimated by pooling data from four studies (n = 90). tion per four to five participants) delivered by a multidisciplinary WMD = weighted mean difference. team (dietitian and therapeutic recreation specialist) was commen- surate with that achieved by an individual intervention (3.5 hours of strengthening and balance exercise compared with strengthening total duration) delivered by a primary care practitioner to members of and aerobic exercise in young children with Down syndrome (mean the general population. age 4 years).113 However, another trial112 found that 6 weeks of vestibular stimulation activities had a superior effect on functional Whole-body vibration balance compared with strengthening and balance exercise in chil- dren with Down syndrome aged 6 to 9 years. The only trial111 Whole-body vibration training exposes the entire body to me- involving adolescents with Down syndrome (mean age 17 years) chanical vibrations as an individual stands (static) on a platform that found no between-group difference in static balance between those oscillates at a particular frequency and amplitude.116 Based on who completed balance exercises for 45 minutes, twice a week, for 12 piezoelectric theory, the interaction of the mechanical vibrations with weeks, compared with no training. A systematic review107 of 11 the structures of the body stimulates bone formation.117 Two tri- randomised and non-randomised clinical trials reached similar als95,118 of whole-body vibration implemented the same protocol of conclusions. standing in a squat position on a vertical platform for 15 to 20 mi- nutes, three times a week, for 6 months, while a third119 imple- Summary of trials relating to exercise mented a shorter 3-month program. Compared with either no Exercise (aerobic training, progressive resistance training, com- training118,120,121 or as an adjunct to exercise,95,119 these trials (n = 90) found short-term positive effects on static standing balance in chil- bined programs, balance training) appears to have similar effects on dren and adolescents with Down syndrome,118 on lower limb muscle people with Down syndrome as the general population. Provided that strength in children with Down syndrome95,119 and subtotal bone physiotherapists implement exercise in accordance with the recom- mineral density (ie, total body except the head) and subtotal bone mended guidelines (ie, sufficient dose and intensity), it can improve mineral content in adolescents with Down syndrome.121 No short- cardiovascular fitness, muscle strength and reduce activity limita- terms effects were found on either body fat in children119 or ado- tions. Uncertainty remains about the specific role of exercise in lescents120 or on lean body mass in adolescents.120 The longer-term addressing the issue of cognitive decline in people with Down syn- effects of whole-body vibration training have not been tested. drome and, more generally, the long-term benefits of exercise on the prevention of chronic diseases, particularly those associated with Summary of trials relating to whole-body vibration low-grade inflammation. Although the primary aim of applying whole-body vibration is to Health education and behaviour change interventions stimulate bone formation, only one trial, with methodological limi- tations, found positive between-group changes in subtotal bone High levels of obesity and low levels of physical activity are key mineral content (3%) and bone mineral density (1.6%) for this passive issues throughout life for people with Down syndrome. Health edu- intervention after 20 weeks. Further, these effects are lower than cation and behavioural change interventions play an important role those found for an (active) exercise intervention, reported from a in addressing these problems in the general population. However, similar-sized trial with similar methodological limitations, albeit over only two trials114,115 (n = 74) have investigated the benefits of these a longer duration (1 year). Given the additional benefits of exercise on interventions for adults with Down syndrome. Compared with no muscle mass and physical fitness, it seems prudent for physiothera- intervention, a 12-week health promotion program, comprising 1 pists to choose to implement exercise interventions over whole-body hour of exercise and 1 hour of health education delivered three times vibration. per week, improved attitudes towards exercise and psychological wellbeing, but not community integration or depression, immediately Future directions for research and practice after the intervention among 53 adults with Down syndrome (mean age 40 years).115 The second trial114 involving 21 participants with A recent consensus initiative involving invited scientific experts BMI  85th percentile (mean age 20 years) reported between-group identified research gaps relating to almost every aspect of Down differences in body weight (MD 23.2 kg, 95% CI 21 to 25.5 kg) syndrome (cognition, behaviour, communication, sleep, the various sustained at 1 year (MD 23.6 kg, 95% CI 21.4 to 25.9 kg) in favour of body systems and community engagement), including general parent training in behavioural strategies (diet and activity moni- research needs (research training, research inclusion, open access toring, modification of stimulus control conditions, goal setting and data and robust trial designs).122 From a physiotherapy perspective, positive reinforcement) in addition to a 6-month nutrition and ac- the most pressing health concern that people with Down syndrome tivity education program (16 sessions of 90-minutes duration each). face is their ultra-high risk of cognitive decline with an onset in their Between-group differences in favour of the experimental group were early 40s. Only recently have trials, albeit with methodological limi- also reported for moderate to vigorous physical activity levels tations, investigated the effect of exercise on cognitive function in immediately post intervention (25 minutes, 95% CI 6 to 43 minutes) people with Down syndrome; although evidence from animal123,124 but not at 1 year (15 minutes, 95% CI 23 to 34 minutes). and longitudinal studies125 suggests that regular, long-term, moder- ate-intensity to high-intensity exercise has the potential to reduce their risk of cognitive decline. Rigorous trials are urgently needed to determine if exercise is effective in maintaining cognitive function and delaying the onset of dementia in adults with Down syn- drome.125 As the meta-analyses presented as part of this review show,

Invited Topical Review 249 there is preliminary evidence that exercise has positive short-term 19. Tomlinson C, Campbell A, Hurley A, Fenton E, Heron N. Sport preparticipation effects on physical function and may have positive short-term ef- screening for asymptomatic atlantoaxial instability in patients with Down syn- fects on cognitive function. Therefore, new trials of long-term exer- drome. Clin J Sport Med. 2020;30:293–295. cise interventions are urgently needed to better understand if exercise is effective and cost-effective in reducing the ultra-high risk 20. Lagan N, Huggard D, Mc Grane F, Leahy TR, Franklin O, Roche E, et al. Multiorgan of cognitive decline in people with Down syndrome and, if effective, involvement and management in children with Down syndrome. Acta Paediatr. whether a dose-response relationship exists. Aligned with this is the 2020;109:1096–1111. need for studies to understand how best to support the long-term implementation of exercise in clinical and community practice, 21. Milbrandt TA, Johnston CE. Down syndrome and scoliosis: a review of a 50-year given the needs of many people with Down syndrome for supervision experience at one institution. Spine. 2005;30:2051–2055. to facilitate exercise, particularly at higher intensities. 22. Chung AS, Renfree S, Lockwood DB, Karlen J, Belthur M. Syndromic scoliosis: na- A second area of need, within both research and practice, is the tional trends in surgical management and inpatient hospital outcomes: a 12-year effect of interventions implemented by physiotherapists on partici- analysis. Spine. 2019;44:1564–1570. pation outcomes. As shown, most research has focused on changes in body structure (eg, weight, waist circumference and percentage body 23. Maranho D, Fuchs K, Kim Y-J, Novais EN. Hip instability in patients with Down fat), body function (eg, muscle strength, cardiovascular fitness and syndrome. J Am Acad Orthop Surg. 2018;26:455–462. balance) and activities (eg, 6-minute walk distance and sit-to-stand). What is missing are data related to participation: involvement in life 24. Dugdale TW, Renshaw T. Instability of the patellofemoral joint in Down syndrome. situations, describing attendance (being there) and involvement J Bone Joint Surg Am. 1986;68:405–413. (experience of participation). Participation is arguably the most important outcome for people with Down syndrome, and physio- 25. Lim PQ, Shields N, Nikolopoulos N, Barrett JT, Evans AM, Taylor NF, et al. The as- therapy research and practice need to adapt to better reflect this. sociation of foot structure and footwear fit with disability in children and ado- lescents with Down syndrome. J Foot Ankle Res. 2015;8:1–10. eAddenda: Figures 2, 4, 7 and 8 can be found online at https://doi. org/10.1016/j.jphys.2021.08.016. 26. Hassan NM, Buldt AK, Shields N, Landorf KB, Menz HB, Munteanu SE. Differences in foot dimensions between children and adolescents with and without Down syn- Ethics approval: Nil. drome. Disabil Rehabil. 2021:1–8. Competing interests: Nil. Source(s) of support: Nil. 27. Alexander M, Petri H, Ding Y, Wandel C, Khwaja O, Foskett N. Morbidity and Acknowledgements: The author thanks Natalie Pearse, Senior medication in a large population of individuals with Down syndrome compared to Library Research Advisor at La Trobe University, for her assistance in the general population. Dev Med Child Neurol. 2016;58:246–254. completing the database searches that supported this review. Provenance: Invited. Peer reviewed. 28. Capone GT, Chicoine B, Bulova P, Stephens M, Hart S, Crissman B, et al. Co- Correspondence: Nora Shields, Department of Physiotherapy, occurring medical conditions in adults with Down syndrome: a systematic review Podiatry and Prosthetics and Orthotics, La Trobe University, Mel- toward the development of health care guidelines. Am J Med Genet A. bourne, Australia. Email: [email protected] 2018;176:116–133. References 29. Walker J, Dosen A, Buitelaar J, Janzing J. Depression in Down syndrome: a review of the literature. Res Dev Disabil. 2011;32:1432–1440. 1. Carfì A, Brandi V, Zampino G, Mari D, Onder G. Care of adults with Down syn- drome: gaps and needs. Eur J Intern Med. 2015;26:375–376. 30. Maïano C, Hue O, Morin AJ, Moullec G. Prevalence of overweight and obesity among children and adolescents with intellectual disabilities: a systematic review 2. Bittles A, Glasson E. Clinical, social, and ethical implications of changing life ex- and meta-analysis. Obes Rev. 2016;17:599–611. pectancy in Down syndrome. Dev Med Child Neurol. 2004;46:282–286. 31. Stancliffe R, Lakin K, Larson S, Engler J, Bershadsky J, Taub S, et al. Overweight and 3. Glasson E, Dye D, Bittles AH. The triple challenges associated with age-related obesity among adults with intellectual disabilities who use services in 20 US states. comorbidities in Down syndrome. J Intellect Disabil Res. 2014;58:393–398. Am J Intellect Dev Disabil. 2011;116:401–418. 4. Tsou AY, Bulova P, Capone G, Chicoine B, Gelaro B, Harville TO, et al. Medical care of 32. Palisano RJ, Walter SD, Russell DJ, Rosenbaum PL, Gémus M, Galuppi BE, et al. Gross adults with Down syndrome: a clinical guideline. JAMA. 2020;324:1543–1556. motor function of children with Down syndrome: creation of motor growth curves. Arch Phys Med Rehabil. 2001;82:494–500. 5. Ruiz-González L, Lucena-Antón D, Salazar A, Martín-Valero R, Moral-Munoz J. Physical therapy in Down syndrome: systematic review and meta-analysis. 33. Angulo-Barroso R, Burghardt AR, Lloyd M, Ulrich DA. Physical activity in infants J Intellect Disabil Res. 2019;63:1041–1067. with Down syndrome receiving a treadmill intervention. Infant Behav Dev. 2008;31:255–269. 6. Bull MJ. Down syndrome. N Engl J Med. 2020;382:2344–2352. 7. National Disability Insurance Agency. NDIS Quarterly Report to disability ministers 34. Angulo-Barroso RM, Wu J, Ulrich DA. Long-term effect of different treadmill in- terventions on gait development in new walkers with Down syndrome. Gait 31 March 2021. 2021. Canberra. Posture. 2008;27:231–238. 8. Presson AP, Partyka G, Jensen KM, Devine OJ, Rasmussen SA, McCabe LL, et al. 35. Lloyd M, Burghardt A, Ulrich DA, Angulo-Barroso R. Physical activity and walking Current estimate of Down syndrome population prevalence in the United States. onset in infants with Down syndrome. Adapt Phys Active Q. 2010;27:1–16. J Pediatr. 2013;163:1163–1168. 9. McCarron M, McCallion P, Reilly E, Mulryan N. A prospective 14-year longitudinal 36. Looper J, Ulrich D. Does orthotic use affect upper extremity support during upright follow-up of dementia in persons with Down syndrome. J Intellect Disabil Res. play in infants with Down syndrome? Pediatr Phys Ther. 2011;23:70–77. 2014;58:61–70. 10. Santoro SL, Steffensen EH. Congenital heart disease in Down syndrome–a review of 37. Looper J, Ulrich DA. Effect of treadmill training and supramalleolar orthosis use on temporal changes. J Congenit Cardiol. 2021;5:1–14. motor skill development in infants with Down syndrome: a randomized clinical 11. Roizen N, Patterson D. Down’s syndrome. Lancet. 2003;361:1281–1289. trial. Phys Ther. 2010;90:382–390. 12. Hilgenkamp TI, Wee SO, Schroeder EC, Baynard T, Fernhall B. Peripheral blood flow regulation in response to sympathetic stimulation in individuals with Down syn- 38. Ulrich DA, Lloyd MC, Tiernan CW, Looper JE, Angulo-Barroso RM. Effects of in- drome. Artery Res. 2018;24:16–21. tensity of treadmill training on developmental outcomes and stepping in infants 13. Sobey CG, Judkins CP, Sundararajan V, Phan TG, Drummond GR, Srikanth VK. Risk with Down syndrome: a randomized trial. Phys Ther. 2008;88:114–122. of major cardiovascular events in people with Down syndrome. PLoS One. 2015;10: e0137093. 39. Ulrich DA, Ulrich BD, Angulo-Kinzler RM, Yun J. Treadmill training of infants with 14. Fitzgerald P, Leonard H, Pikora TJ, Bourke J, Hammond G. Hospital admissions in Down syndrome: evidence-based developmental outcomes. Pediatrics. children with Down syndrome: experience of a population-based cohort followed 2001;108:E84. from birth. PLoS One. 2013;8:e70401. 15. Torr J, Strydom A, Patti P, Jokinen N. Aging in Down syndrome: morbidity and 40. Wu J, Looper J, Ulrich BD, Ulrich DA, Angulo-Barroso RM. Exploring effects of mortality. J Policy Pract Intellect Disabil. 2010;7:70–81. different treadmill interventions on walking onset and gait patterns in infants with 16. Lal C, White DR, Joseph JE, van Bakergem K, LaRosa A. Sleep-disordered breathing Down syndrome. Dev Med Child Neurol. 2007;49:839–845. in Down syndrome. Chest. 2015;147:570–579. 17. Pikora TJ, Bourke J, Bathgate K, Foley K-R, Lennox N, Leonard H. Health conditions 41. Wu J, Looper J, Ulrich DA, Angulo-Barroso RM. Effects of various treadmill in- and their impact among adolescents and young adults with Down syndrome. PLoS terventions on the development of joint kinematics in infants with Down syn- One. 2014;9:e96868. drome. Phys Ther. 2010;90:1265–1276. 18. Thomas K, Bourke J, Girdler S, Bebbington A, Jacoby P, Leonard H. Variation over time in medical conditions and health service utilization of children with Down 42. Wu J, Ulrich DA, Looper J, Tiernan CW, Angulo-Barroso RM. Strategy adoption syndrome. J Pediatr. 2011;158:194–200.e1. and locomotor adjustment in obstacle clearance of newly walking toddlers with Down syndrome after different treadmill interventions. Exp Brain Res. 2008;186:261–272. 43. Johnson R, Looper J, Fiss A. Current Trends in pediatric physical therapy practice for children with Down syndrome. Pediatr Phys Ther. 2021;33:74–81. 44. Hernandez-Reif M, Field T, Largie S, Mora D, Bornstein J, Waldman R. Children with Down syndrome improved in motor functioning and muscle tone following mas- sage therapy. Early Child Dev Care. 2006;176:395–410. 45. Field TM, Schanberg SM, Scafidi F, Bauer CR, Vega-Lahr N, Garcia R, et al. Tactile/kinesthetic stimulation effects on preterm neonates. Pediatrics. 1986;77:654–658. 46. Blauw-Hospers CH, Hadders-Algra M. A systematic review of the effects of early intervention on motor development. Dev Med Child Neurol. 2005;47: 421–432. 47. Pinero-Pinto E, Benítez-Lugo M-L, Chillón-Martínez R, Rebollo-Salas M, Bellido- Fernández L-M, Jiménez-Rejano J-J. Effects of massage therapy on the development of babies born with Down syndrome. Evid Based Complement Altern Med. 2020. Article ID 4912625. 48. Silva LM, Schalock M, Garberg J, Smith CL. Qigong massage for motor skills in young children with cerebral palsy and Down syndrome. Am J Occup Ther. 2012;66:348–355. 49. Ulrich DA, Burghardt AR, Lloyd M, Tiernan C, Hornyak JE. Physical activity benefits of learning to ride a two-wheel bicycle for children with Down syndrome: a ran- domized trial. Phys Ther. 2011;91:1463–1477. 50. Hubbard IJ, Parsons MW, Neilson C, Carey LM. Task-specific training: evidence for and translation to clinical practice. Occup Ther Int. 2009;16:175–189.

250 Shields: Physiotherapy management of Down syndrome 51. Noghondar NV, Sohrabi M, Taheri HR, Kobravi HR, Khodashenas E. The effect of 83. Casey AF, Emes C. The effects of swim training on respiratory aspects of speech pro- training on variability and accuracy of overhand throwing in children with Down duction in adolescents with down syndrome. Adapt Phys Active Q. 2011;28:326–341. syndrome. Int J Dev Disabil. 2019. https://doi.org/10.1080/20473869.2019.1574388. 84. Andriolo RB, El Dib R, Ramos L, Atallah ÁN, da Silva EM. Aerobic exercise training 52. Jobling A. Motor development in school-aged children with Down syndrome: a programmes for improving physical and psychosocial health in adults with Down longitudinal perspective. Int J Disabil Dev Educ. 1998;45:283–293. syndrome. Cochrane Database Syst Rev. 2005. 53. Jackman M, Lannin N, Galea C, Sakzewski L, Miller L, Novak I. What is the threshold 85. Beck V, Baynard T, Lefferts E, Hibner B, Fernhall B, Hilgenkamp TI. Anthropometry dose of upper limb training for children with cerebral palsy to improve function? A does not fully explain low fitness among adults with Down syndrome. J Intellect systematic review. Aust Occup Ther J. 2020;67:269–280. Disabil Res. 2021;65:373–379. 54. Volman MJ, Visser JJ, Lensvelt-Mulders GJ. Functional status in 5 to 7-year-old 86. https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=363104. children with Down syndrome in relation to motor ability and performance mental 87. Croce RV, Pitetti KH, Horvat M, Miller J. Peak torque, average power, and ability. Disabil Rehabil. 2007;29:25–31. hamstring/quadriceps ratios in nondisabled adults and adults with mental retar- 55. Baynard T, Pitetti KH, Guerra M, Unnithan VB, Fernhall B. Age-related changes in dation. Arch Phys Med Rehabil. 1996;77:369–372. aerobic capacity in individuals with mental retardation: a 20-yr review. Med Sci 88. Pitetti KH, Climstein M, Mays MJ, Barrett PJ. Isokinetic arm and leg strength of Sports Exerc. 2008;40:1984–1989. adults with Down syndrome: a comparative study. Arch Phys Med Rehabil. 1992;73:847–850. 56. Cowley PM, Ploutz-Snyder LL, Baynard T, Heffernan KS, Young JS, Hsu S, et al. The 89. Fornieles G, Rosety MA, Elosegui S, Rosety JM, Alvero-Cruz JR, Garcia N, et al. effect of progressive resistance training on leg strength, aerobic capacity and Salivary testosterone and immunoglobulin A were increased by resistance training functional tasks of daily living in persons with Down syndrome. Disabil Rehabil. in adults with Down syndrome. Braz J Med Biol Res. 2014;47:345–348. 2011;33:2229–2236. 90. Rosety-Rodriguez M, Camacho A, Rosety I, Fornieles G, Rosety MA, Diaz AJ, et al. Resistance circuit training reduced inflammatory cytokines in a cohort of male 57. Schmid D, Leitzmann MF. Cardiorespiratory fitness as predictor of cancer mortal- adults with Down syndrome. Med Sci Monit. 2013;19:949–953. ity: a systematic review and meta-analysis. Ann Oncol. 2015;26:272–278. 91. Shields N, Taylor NF. A student-led progressive resistance training program in- creases lower limb muscle strength in adolescents with Down syndrome: a 58. Wang Y, Li F, Cheng Y, Gu L, Xie Z. Cardiorespiratory fitness as a quantitative randomised controlled trial. J Physiother. 2010;56:187–193. predictor of the risk of stroke: a dose–response meta-analysis. J Neurol. 92. Shields N, Taylor NF, Dodd KJ. Effects of a community-based progressive resistance 2020;267:491–501. training program on muscle performance and physical function in adults with Down syndrome: a randomized controlled trial. Arch Phys Med Rehabil. 59. Liguori G, American College of Sports Medicine. ACSM’s guidelines for exercise 2008;89:1215–1220. testing and prescription. Philadelphia, USA: Lippincott Williams & Wilkins; 93. Shields N, Taylor NF, Wee E, Wollersheim D, O’Shea SD, Fernhall B. A community- 2020. based strength training programme increases muscle strength and physical ac- tivity in young people with Down syndrome: a randomised controlled trial. Res Dev 60. Alsakhawi RS, Elshafey MA. Effect of core stability exercises and treadmill training Disabil. 2013;34:4385–4394. on balance in children with Down syndrome: randomized controlled trial. Adv 94. Wentz EE, Looper J, Menear KS, Rohadia D, Shields N. Promoting participation in Ther. 2019;36:2364–2373. physical activity in children and adolescents with Down syndrome. Phys Ther. 2021;101:pzab032. 61. Boer PH. The effect of 8 weeks of freestyle swim training on the functional fitness 95. Eid MA. Effect of whole-body vibration training on standing balance and muscle of adults with Down syndrome. J Intellect Disabil Res. 2020;64:770–781. strength in children with Down syndrome. Am J Phys Med Rehabil. 2015;4:633– 643. 62. Boer P, Moss S. Effect of continuous aerobic vs. interval training on selected 96. Gupta S, Rao BK, Kumaran SD. Effect of strength and balance training in children anthropometrical, physiological and functional parameters of adults with Down with Down’s syndrome: a randomized controlled trial. Clin Rehabil. 2011;25:425– syndrome. J Intellect Disabil Res. 2016;60:322–334. 432. 97. Hussein ZA. Strength training versus chest physical therapy on pulmonary func- 63. Carmeli E, Kessel S, Coleman R, Ayalon M. Effects of a treadmill walking program tions in children with Down syndrome. Egypt J Med Hum Genet. 2017;18:35–39. on muscle strength and balance in elderly people with Down syndrome. J Geront A: 98. Rimmer JH, Heller T, Wang E, Valerio I. Improvements in physical fitness in adults Biol Sci Med Sci. 2002;57:M106–M110. with Down syndrome. Am J Ment Retard. 2004;109:165–174. 99. Specker B, Thiex NW, Sudhagoni RG. Does exercise influence pediatric bone? A 64. Ordonez FJ, Fornieles-Gonzalez G, Camacho A, Rosety MA, Rosety I, Diaz AJ, et al. systematic review. Clin Orthop Relat Res. 2015;473:3658–3672. Anti-inflammatory effect of exercise, via reduced leptin levels, in obese women with Down syndrome. Int J Sport Nutr Exerc Metab. 2013;23:239–244. 100. Dilek G, Öztürk C, Hepgüler S, Özkınay F, Dilek M. The effect of exercise on bone mineral density in patients with Down syndrome. J Pediatr Res. 2018;5:1. 65. Ordonez FJ, Rosety MA, Camacho A, Rosety I, Diaz AJ, Fornieles G, et al. Aerobic training improved low-grade inflammation in obese women with intellectual 101. Ferry B, Gavris M, Tifrea C, Serbanoiu S, Pop AC, Bembea M, et al. The bone tissue of disability. J Intellect Disabil Res. 2014;58:583–590. children and adolescents with Down syndrome is sensitive to mechanical stress in certain skeletal locations: a 1-year physical training program study. Res Dev Disabil. 66. Rosety-Rodriguez M, Diaz AJ, Rosety I, Rosety MA, Camacho A, Fornieles G, et al. 2014;35:2077–2084. Exercise reduced inflammation: but for how long after training? J Intellect Disabil Res. 2014;58:874–879. 102. González-Agüero A, Vicente-Rodríguez G, Gómez-Cabello A, Ara I, Moreno LA, Casajús JA. A combined training intervention programme increases lean mass in 67. Varela AM, Sardinha LB, Pitetti KH. Effects of an aerobic rowing training regimen in youths with Down syndrome. Res Dev Disabil. 2011;32:2383–2388. young adults with Down syndrome. Am J Ment Retard. 2001;106:135–144. 103. González-Agüero A, Vicente-Rodríguez G, Gómez-Cabello A, Ara I, Moreno LA, 68. Holzapfel SD, Ringenbach SD, Mulvey GM, Sandoval-Menendez AM, Cook MR, Casajus JA. A 21-week bone deposition promoting exercise programme increases Ganger RO, et al. Improvements in manual dexterity relate to improvements in bone mass in young people with Down syndrome. Dev Med Child Neurol. cognitive planning after assisted cycling therapy (ACT) in adolescents with down 2012;54:552–556. syndrome. Res Dev Disabil. 2015;45-46:261–270. 104. So S, Urbano R, Hodapp R. Hospitalizations of infants and young children with 69. Ringenbach SD, Holzapfel SD, Mulvey GM, Jimenez A, Benson A, Richter M. The Down syndrome: evidence from inpatient person-records from a statewide effects of assisted cycling therapy (ACT) and voluntary cycling on reaction time and administrative database. J Intellect Disabil Res. 2007;51:1030–1038. measures of executive function in adolescents with Down syndrome. J Intellect Disabil Res. 2016;60:1073–1085. 105. Vural M, Özdal M, Pancar Z. Effects of inspiratory muscle training on respiratory functions and respiratory muscle strength in Down syndrome: a preliminary study. 70. Ptomey LT, Szabo AN, Willis EA, Gorczyca AM, Greene JL, Danon JC, et al. Changes in Isokinet Exerc Sci. 2019;27:283–288. cognitive function after a 12-week exercise intervention in adults with Down syndrome. Disabil Health J. 2018;11:486–490. 106. Ibrahim AF, Salem EE, Gomaa NE, Abdelazeim FH. The effect of incentive spirom- eter training on oromotor and pulmonary functions in children with Down’s 71. Ptomey LT, Szabo AN, Willis EA, Greene JL, Danon JC, Washburn RA, et al. Remote syndrome. J Taibah Univ Med Sci. 2019;14:405–411. exercise for adults with Down syndrome. Transl J Am Coll Sports Med. 2018;3:60. 107. Maïano C, Hue O, Lepage G, Morin AJS, Tracey D, Moullec G. Do exercise in- 72. El Kafy EM, Helal OF. Effect of rowing on pulmonary functions in children with terventions improve balance for children and adolescents with Down syndrome? A Down syndrome. Pediatr Phys Ther. 2014;26:437–445. systematic review. Phys Ther. 2019;99:507–518. 73. Reza SM, Rasool H, Mansour S, Abdollah H. Effects of calcium and training on the 108. Lauteslager P, Vermeer A, Helders P. Disturbances in the motor behaviour of chil- development of bone density in children with Down syndrome. Res Dev Disabil. dren with Down’s syndrome: the need for a theoretical framework. Physiotherapy. 2013;34:4304–4309. 1998;84:5–13. 74. Skiba A, Marchewka J, Skiba A, Podsiadło S, Sulowska I, Chwała W, et al. Evaluation 109. Aly SM, Abonour AA. Effect of core stability exercise on postural stability in children of the effectiveness of Nordic walking training in improving the gait of persons with Down syndrome. Int J Med Res Health Sci. 2018;5:213–222. with Down syndrome. BioMed Res Int. 2019. Article ID 6353292. 110. Ghaeeni S, Bahari Z, Khazaei AA. Effect of core stability training on static balance of 75. Lin HC, Wuang YP. Strength and agility training in adolescents with Down syn- the children with Down syndrome. Phys Treat. 2015;5:49–54. drome: a randomized controlled trial. Res Dev Disabil. 2012;33:2236–2244. 111. Jankowicz-Szymanska A, Mikolajczyk E, Wojtanowski W. The effect of the degree 76. Wuang YP, Chiang CS, Su CY, Wang CC. Effectiveness of virtual reality using Wii of disability on nutritional status and flat feet in adolescents with Down syndrome. gaming technology in children with Down syndrome. Res Dev Disabil. Res Dev Disabil. 2013;34:3686–3690. 2011;32:312–321. 112. Kamatchi K, Balachandar V, Kaviraja N. Comparative evaluation of weight bearing 77. Silva V, Campos C, Sa A, Cavadas M, Pinto J, Simoes P, et al. Wii-based exercise exercise and vestibular stimulation on balance in children with Down syndrome. program to improve physical fitness, motor proficiency and functional mobility in Int J Pharma Bio Sci. 2018;9:227–234. adults with Down syndrome. J Intellect Disabil Res. 2017;61:755–765. 113. Rahman SAA, Shaheen A. Efficacy of weight bearing exercises on balance in 78. Shields N, Taylor NF. The feasibility of a physical activity program for young adults children with Down syndrome. Egypt J Neurol Psychiatr Neurosurg. 2010;47: with Down syndrome: A phase II randomised controlled trial. J Intellect Dev Disabil. 37–42. 2015;40:115–125. 114. Curtin C, Bandini LG, Must A, Gleason J, Lividini K, Phillips S, et al. Parent support 79. El-Meniawy GH, Kamal HM, Elshemy SA. Role of treadmill training versus sus- pension therapy on balance in children with Down syndrome. Egypt J Med Hum Genet. 2012;13:37–43. 80. Suarez-Villadat B, Luna-Oliva L, Acebes C, Villagra A. The effect of swimming program on body composition levels in adolescents with Down syndrome. Res Dev Disabil. 2020;102:103643. 81. Perrot A, Maillot P, Le Foulon A, Rebillat A-S. Effect of exergaming on physical fitness, functional mobility, and cognitive functioning in adults with Down syn- drome. Am J Intellect Dev Disabil. 2021;126:34–44. 82. Rahman SA, Rahman A. Efficacy of virtual reality-based therapy on balance in children with Down syndrome. World Appl Sci J. 2010;10:254–261.

Invited Topical Review 251 improves weight loss in adolescents and young adults with Down syndrome. 121. Matute-Llorente A, Gonzalez-Aguero A, Gomez-Cabello A, Olmedillas H, Vicente- J Pediatr. 2013;163:1402–1408. Rodriguez G, Casajus JA. Effect of whole body vibration training on bone mineral density and bone quality in adolescents with Down syndrome: a randomized 115. Heller T, Hsieh K, Rimmer JH. Attitudinal and psychosocial outcomes of a fitness controlled trial. Osteoporos Int. 2015;26:2449–2459. and health education program on adults with down syndrome. Am J Ment Retard. 2004;109:175–185. 122. Hendrix JA, Amon A, Abbeduto L, Agiovlasitis S, Alsaied T, Anderson HA, et al. Opportunities, barriers, and recommendations in Down syndrome research. Transl 116. Bidonde J, Busch AJ, van der Spuy I, Tupper S, Kim SY, Boden C. Whole body vi- Sci Rare Dis. 2020;5:99–129. bration exercise training for fibromyalgia. Cochrane Database Syst Rev. 2017. 123. Kida E, Rabe A, Walus M, Albertini G, Golabek AA. Long-term running alleviates 117. Weber-Rajek M, Mieszkowski J, Niespodzinski B, Ciechanowska K. Whole-body some behavioral and molecular abnormalities in Down syndrome mouse model vibration exercise in postmenopausal osteoporosis. Menopause Rev. 2015;14:41. Ts65Dn. Exp Neurol. 2013;240:178–189. 118. Villarroya MA, Gonzalez-Aguero A, Moros T, Gomez-Trullen E, Casajus JA. Effects of 124. Walus M, Kida E, Rabe A, Albertini G, Golabek AA. Widespread cerebellar tran- whole body vibration training on balance in adolescents with and without Down scriptome changes in Ts65Dn Down syndrome mouse model after lifelong running. syndrome. Res Dev Disabil. 2013;34:3057–3065. Behav Brain Res. 2016;296:35–46. 119. Emara HA. Effects of whole body vibration on body composition and muscle 125. Pape SE, Baksh RA, Startin C, Hamburg S, Hithersay R, Strydom A. The As- strength of children with Down syndrome. Int J Ther Rehabil Res. 2016;5:1–8. sociation between Physical Activity and CAMDEX-DS Changes Prior to the Onset of Alzheimer’s Disease in Down Syndrome. J Clin Med. 2021;10: 120. Gonzalez-Aguero A, Matute-Llorente A, Gomez-Cabello A, Casajus JA, Vicente- 1882. Rodriguez G. Effects of whole body vibration training on body composition in adolescents with Down syndrome. Res Dev Disabil. 2013;34:1426–1433.

Journal of Physiotherapy 67 (2021) 315–318 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Research Note: Transporting causal effects from randomised trials to target populations to improve external validity Randomised trials can reliably estimate average treatment effects estimate a contrastb of the average potential outcome had all in- from a sample that is drawn from a population.1 When they are well dividuals in a population been treated versus the average potential conducted, trials can rigorously mitigate threats to internal validity outcome had all individuals in a population not been treated. Pro- and enable credible claims to be made about the causal effects of vided that there are minimal losses to follow-up and adequate interventions. However, investigators and clinicians have noted dif- adherence,7 the intention-to-treat effect can be interpreted as the ferences between patients seen in practice and participants enrolled expected average causal effect of the treatment in a population that is in trials.2,3 These concerns may cast doubt over the external validity resembled by the trial sample (ie, the trial population). Unfortunately, of trials and limit the ability of trial findings to inform policy and for some trials, it is not guaranteed that the trial population reflects a practice with confidence.4 Recent discussions have suggested that for target population the investigator wishes to make inferences about or trials to reliably inform decisions about intervening on populations, the population where they wish to implement the findings. The re- they should not only generate internally valid estimates of treatment lationships between the target population, trial population and trial effects but also enable reliable inferences about relevant target pop- sample are illustrated in Figure 2. ulations.3,5 This research note discusses how average treatment ef- fects from a randomised trial can be extended to clinically relevant In practice, eligibility criteria are used to restrict the entire pop- target populations when trial samples may not represent target ulation to some trial-eligible population. It may appear that the populations. eligibility criteria fully define the target population of interest. But that is not true. Trial eligibility criteria only serve to place boundaries Most randomised trials cannot enrol simple random samples of a around a relevant subset of the population, and it is expected that target population samples drawn from the trial-eligible population will vary with respect to individual characteristics. Also, applying reasonable eligi- Under most clinical settings it is unfeasible to enrol the entire bility criteria may still exclude individuals who would have been target population into a randomised trial. Even the largest and most considered suitable for the treatment being tested in the trial. pragmatic physiotherapy trials typically recruit a fraction of the target Therefore, eligibility criteria are just the investigators’ intention to population. Therefore, depending on the extent to which there is partially define a target population. Making inferences about pop- selective invitation and consent into the trial, the trial sample will be ulations requires further understanding about how individuals in a more or less resemblant of the target population. Selective recruit- trial sample compare to those in the target population. ment is common: some hospitals do not participate in trials, clini- cians may forget to extend invitations, participation may be limited to Selective trial participation alone rarely causes problems for selected geographical regions, and language barriers could exclude making inferences about target populations, but when coupled certain populations. Rothwell claims that across medical disciplines, with effect heterogeneity, external validity can be compromised less than 10% of patients who would have been considered eligible for a relevant trial end up being recruited.3 In principle, comprehensive It may be easy to criticise the generalisability of a randomised trial recruitment strategies can overcome selective invitation; however, just because the trial sample is not randomly drawn from the target selective consent is certainly inevitable. All individuals, guardians and population. However, non-random sampling would not compromise attorneys have the right to refuse consent. This guarantees that the external validityc if treatment effects are homogenous.d Only when trial sample will never be a perfect random sample of the target causes of the outcome in the target population are unequally repre- population.a Figure 1 illustrates key sampling processes that would sented in the trial sample, the trial may have limited capacity to make typically occur for a randomised trial. valid inferences about the target population. For example, if a falls prevention strategy for people with Parkinson’s disease increases the Randomised trials can make inferences about the population it rate of falling for those who have freezing of gait,9 and more people represents, but that population may not reflect a clinically freeze in the target population compared to the trial sample, we relevant target population might expect the average treatment effect from the trial to differ from the expected treatment effect in the target population. Most randomised trials aim to estimate treatment effects that b Typically, a ‘contrast’ refers to a difference or ratio of potential outcome means, apply to some population.1,6 More precisely, randomised trials aim to probabilities or hazards. a The population represented by the trial sample will always be different to the target c Broadly, any variation in treatment composition, treatment mechanisms, outcome population with respect to time. Because trial findings are typically applied to a target measure, setting and individual characteristics between the trial sample and target population after the trial has closed, the source population that produced the trial population can influence external validity. To limit the scope of this Research Note, it is sample is historical and could be different to the target population where the findings assumed that all factors are consistent across the trial sample and target population are to be implemented.5,8 except for individual characteristics. d The absence of evidence for treatment effect heterogeneity does not imply that treatment effects are homogenous. https://doi.org/10.1016/j.jphys.2021.08.007 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

316 Appraisal Research Note Figure 1. Selective sampling process from a population to a trial sample. In principle, there is a subset of the entire population that can be labelled the target population. From the target population investigators define a trial-eligible population. Trial-eligible individuals are then invited (or self-enrol), then a subset of the invited individuals provide consent to participate. Those who consent make up the trial sample. At each stage of selection there is a chance for the excluded subsets (in red) to differ from the included subsets (in blue), which would result in a trial sample that does not resemble the target population. Figure 2. The relationship between the target population, trial sample and trial population. The trial sample that is drawn from the entire population can make valid inferences about a population it represents (the trial population). But when there is selective sampling (Figure 1), the trial population may not reflect the target population with respect to patient characteristics (indicated by the colours and percentages). This fictitious example only displays the distributions of one characteristic. In reality, there will be more characteristics and their combinations to consider. Claims about the external validity of trials should always be made effect in the UK population with Parkinson’s. However, if there was with reference to a target population interest in applying the findings to care home residents with Parkin- son’s in New Zealand where there is a higher prevalence of freezing, it is It is not meaningful to talk about the external validity of a trial unlikely that the trial will provide reliable inferences about the New without reference to a target population that is of scientific or practical Zealand care home population. This shows that we cannot just claim interest.5 Because the fundamental goal of the randomised trial is to that the Parkinson’s trial is externally valid or invalid. These statements make inferences about a target population, a claim or assessment about must be made with careful reference to the target population of interest. the external validity of a trial should be made in reference to a target Arguably, where possible, trials should report baseline characteristics of population. This means that a trial could be externally valid for one a target population alongside the baseline characteristics of the trial target population but not another. Take for example the Parkinson’s falls sample (ie, Table 1 of the trial report) so that its external validity can be prevention trial conducted in the UK’s NHS9 and assume that it sampled assessed descriptively. participants at random so that all causes of the outcome were equally distributed in the trial and target populations. There is a good chance The following sections aim to provide a broad overview of that this trial would allow valid inferences about the average treatment emerging methods that can extend the findings of a possibly non- representative trial to a target population.

Appraisal Research Note 317 Generalisability and transportability analyses The first assumption, conditional exchangeability for trial selec- tion, requires that there are no unmeasured causes of outcomes that By combining data from a randomised trial and a registry or are unequally distributed between the trial and target popula- cohort that can be considered representative of a target populatione tion.10,11 For example, if the Exeter stem that was tested in WHiTE-3 recent developments have produced rigorous frameworks8,10 and offered better quality of life for people who took bone medication, methods11–13 for estimating the average treatment effect in a target and there were more bone medication users in the target popula- population. The intuition behind this approach is to utilise the rigour tion than the trial, this difference should be adjusted in the analysis. of randomised trials to ensure internal validity, and the breadth of This assumption is often challenging to satisfy because the inves- registries and cohorts to improve external validity. Informally, these tigator will usually need to rely on subject matter knowledge to methods can be used to answer the question: ‘what would be the identify plausible causes of outcomes. Selection diagrams10 can be average treatment effect had we conducted the same trial in the useful for identifying a sufficient set of variables that would relax entire target population?’. this assumption (Figure 3). In recent discussions, the term generalisability has been used to The second assumption, positivity of trial participation, requires describe the objective of learning about a population of all trial- that all individuals in the target population have a positive probability eligible individuals, and transportability as the objective of learning of being selected into the trial. For example, say 2% of population about a broader target population who are not necessarily part of the represented by the UK National Hip Fracture Database had scoliosis trial-eligible population.14 The methods outlined below generally and we believe that having scoliosis might cause the outcome (thus is apply to both generalisability and transportability, with a few ex- required to achieve conditional exchangeability for trial selection). If ceptions that are noted as footnotes. Note there are also design ad- nobody in the trial had scoliosis, positivity could be violated because aptations to the randomised trial (eg, two-stage preference trials) that the 2% of people in the target population with scoliosis are not rep- can also enhance generalisability,15 but they are not the focus of this resented in the trial (ie, have zero probability of being selected into paper. The following sections describe analytical methods that can the trial). In most cases, this type of violation can be avoided by extend trial findings when design adaptations are unfeasible, un- carefully limiting the target population to individuals who would be ethical or uneconomical. considered eligible for the trial.g To illustrate key concepts, we use an example of a generalisability Estimating average causal effects in the target population analysis that extended the WHiTE-3 trial (n = 958) to a subset of the UK’s National Hip Fracture Database who would have been consid- If all the above conditions are met, the average treatment effect in ered eligible for the WHiTE-3 trial (n = 190,894).16 The WHiTE-3 trial a clearly defined target population can be estimated. The analytical compared the effects of two hip implants (modern Exeter hemi- methods can be broadly classified into three types: weighting by trial arthroplasty versus the traditional Thompson monoblock) on quality participation,11,12 outcome modelling14,18 and their combination.14 As of life and length of hospital stay in a convenience sample of par- the weighting method is perhaps the most widely used approach, this ticipants aged . 60 years with displaced intracapsular fractures of the section will focus on that approach, and readers are referred to the hip. references for the other approaches. Study design and data requirements The basic principle behind the weighting method is to weight individuals in the trial so that they closely resemble the target pop- Dahabreh et al outlined two major study designs for extending ulation with respect to variables that affect the outcome. The first trial findings:17 a nested trial design, where a randomised trial is step involves fitting a model for trial membership based on a set of embedded in a cohort or registry, and a non-nested trial design, variables collected in the trial and target population. This model is where data from a randomised trial are combined with data from a used to obtain predicted probabilities of trial membership that are separate cohort or registry. In both designs, it is assumed that the used to calculate sampling weightsh for all individuals in the trial.11 cohort or registry is either a census or a simple random sample of the Intuitively, individuals in the trial who have characteristics that target population. As most physiotherapy trials are not nested in were over-sampled get assigned a small weight, and individuals who cohorts or registries, the natural design choice would be the non- have characteristics that were under-sampled get assigned a large nested trial design. weight. Provided that the trial participation model is well specified, applying the sampling weights to all trial participants makes the trial In both designs, individual participant data on treatment, partic- sample more like the target population. Finally, the potential out- ipant characteristics (plausible causes of outcome) and outcome are comes for the treatment and control groups can be estimated in the required from the trial, and individual participant data on participant weighted trial sample to obtain the average treatment effect for the characteristics are required from the cohort or registry. In most ap- target population.11,12 plications, particularly in the non-nested case, measures on partici- pant characteristics will need to be harmonised across the trial and The generalisability analysis of the WHiTE-3 trial showed that target population datasets. there were differences between the trial sample and target popula- tion in plausible causes of the outcome such as the use of walking Assumptions under which randomised trials can be extended to aids and pre-fracture living status. Despite these differences, the target populations average treatment effect in the target population was comparable with the average treatment effect from the trial. For quality of life, the To obtain an unbiased estimate of the average treatment effect for estimate for the target population was 0.05 (95% CI –0.01 to 0.11) and a target population, a set of unverifiable assumptions must be met. estimate from the trial was 0.06 (95% CI –0.01 to 0.12). Similarly, for The following key assumptions are required in addition to the stan- length of hospital stay (in days), the estimate for the target popula- dard assumptions for obtaining an unbiased average treatment effect tion was –1.14 (95% CI –2.35 to 0.08), whereas the estimate from the from a randomised trial alone.f trial was –0.70 (95% CI –1.90 to 0.51). Provided that the measured covariates were sufficient to achieve conditional exchangeability of e A registry or cohort may not be a random sample, unless it is a census of the target population. Therefore, we often make an extra assumption that the registry or cohort g If the goal is to estimate an average treatment effect in a broader population of is a simple random sample of the target population the investigator is concerned individuals who are not necessarily trial-eligible (ie, transportability), investigators about. may intentionally breach the positivity assumption to allow extrapolation to a wider target population. f These assumptions include consistency (well-defined interventions), positivity of treatment assignment and exchangeability of treatment assignment (no con- h The sampling weights are calculated as the inverse probability of trial membership founding).1 In most cases these assumptions will be satisfied by design in a rand- or the odds of trial membership, depending on whether the inferential goal is omised trial. As noted in footnote ‘c’, we restrict our discussion to cases where the generalisability or transportability, respectively.12 treatment mechanism is consistent between the trial and target populations.

318 Appraisal Research Note Figure 3. Selection diagrams. Selection diagrams are Directed Acyclic Graphs (DAGs) that represent the trial population, combined with selection nodes (- S) that indicate dif- ferences between the trial and target populations. Please refer to a previous Research Note on causal inference (Herbert 2020) for a gentle introduction to DAGs.6 To estimate the effect of a treatment on an outcome in the target population, all selection nodes must be made independent of the outcome. On the left panel, pre-fracture mobility is a cause of the outcome (length of hospital stay) and could be differently distributed in the trial and target population, as indicated by the selection node. To make the selection node independent of the outcome, pre-fracture mobility would need to be adjusted. On the right panel, pre-fracture mobility is a cause of the outcome, and age is a cause of the outcome and pre- fracture mobility. As there is no selection node going into age, we assume age is equally distributed between the trial and target population. Here we want to adjust for pre-fracture mobility to separate the selection node and outcome. But doing so induces a correlation (dotted line) between selection and age because pre-fracture mobility is a collider of selection and age. This opens a back-door path from selection to the outcome via age. Therefore, the estimate of the treatment effect should adjust for mobility and age to achieve conditional exchangeability for trial selection. trial selection, these results provide some assurance that the WHiTE- common data models may create opportunities for robust general- 3 findings would apply to the wider UK population of individuals who isability and transportability analyses. would have been eligible for the trial. Competing interests: Nil. The transported causal effect will often include the effect of Sources of Support: Nil. treatment and the effects of trial participation Acknowledgements: We thank Dr Aidan Cashin for earlier com- ments and proofing of this paper. Lastly, it is important to acknowledge that participating in a Provenance: Invited. Peer reviewed. randomised trial may have effects on the outcome that are not Correspondence: Hopin Lee, Centre for Statistics in Medicine, mediated through the actual treatment being tested. For example, Nuffield Department of Orthopaedics Rheumatology and Musculo- participating in the WHiTE-3 trial may have increased the frequency skeletal Sciences (NDORMS), University of Oxford, Oxford, UK. Email: and quality of postoperative exercises because participants in both [email protected] arms were closely monitored by physiotherapists during the trial. Furthermore, especially for self-reported outcomes, there may be Hopin Leea,b and Sarah E Lamba,c Hawthorne effects from trial participation.19 As it can be challenging aCentre for Statistics in Medicine, Nuffield Department of Orthopaedics to partition the effect of treatment from the effects of participating in a trial, when we generalise or transport trials to target populations, Rheumatology and Musculoskeletal Sciences (NDORMS), we are extending the effect of the treatment and also the possible University of Oxford, Oxford, UK effects of trial participation.20 bSchool of Medicine and Public Health, University of Newcastle, Concluding remarks Newcastle, Australia Well conducted randomised trials can provide robust estimates of cCollege of Medicine and Health, University of Exeter Medical School, average treatment effects. But they are more useful when results are Exeter, UK externally valid and applicable to relevant target populations. The increasing availability of routinely collected data and recent meth- References odological developments have strengthened the ability to improve the external validity of randomised trials. Generalisability and 1. Rubin DB. J Educ Psychol. 1974;66:688–701. transportability analyses could help remove barriers to imple- 2. Kennedy-Martin T, et al. Trials. 2015;16:495. mentation by assuring clinicians and decision-makers about the 3. Rothwell PM. Lancet. 2005;365:82–93. external validity of selected trials to their populations of interest. 4. Huebschmann AG, et al. Annu Rev Public Health. 2019;40:45–63. They could also identify trial findings that are unsuitable for imple- 5. Westreich D, et al. Am J Epidemiol. 2019;188:438–443. mentation, based on their lack of generalisability or transportability. 6. Herbert RD. J Physiother. 2020;66:273–277. Trials could provide greater return on investment if they can be 7. Kasza J. J Physiother. 2021;67:147–149. transported to multiple settings to guide implementation and 8. Bareinboim E, Pearl J. J Causal Inference. 2013;1:107–134. resource allocation. 9. Ashburn A, et al. Health Technol Assess. 2019;23:1–150. 10. Pearl J, Bareinboim E. Transportability of Causal and Statistical Relations: A Formal These are exciting prospects, but they come with challenges. We should not forget that these methods rely on strong assumptions that Approach. In: 2011 IEEE 11th International Conference on Data Mining Workshops. must be reasonably satisfied to permit valid claims about general- IEEE; 2011:540–547. isability and transportability. Relatedly, the quality of registry data 11. Dahabreh IJ, et al. Biometrics. 2019;75:685–694. and their overlap with trial data must be sufficient and well-aligned 12. Westreich D, et al. Am J Epidemiol. 2017;186:1010–1014. so that conditional exchangeability for selection assumption can be 13. Stuart EA, et al. Prev Sci. 2015;16:475–485. met. Finally, harmonisation of data across trials and registries through 14. Dahabreh IJ, et al. Stat Med. 2020;39:1999–2014. 15. Marcus SM, et al. Psychol Methods. 2012;17:244–254. 16. Lee H, et al. J Clin Epidemiol. 2021;131:141–151. 17. Dahabreh IJ, et al. http://arxiv.org/abs/1905.07764. Accessed 9 March, 2021. 18. Kern HL, et al. J Res Educ Eff. 2016;9:103–127. 19. McCarney R, et al. BMC Med Res Methodol. 2007;7:30. 20. Dahabreh IJ, Hernán MA. Eur J Epidemiol. 2019;34:719–722.

Journal of Physiotherapy 67 (2021) 271–283 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research Respiratory physiotherapy interventions focused on exercise training and enhancing physical activity levels in people with chronic obstructive pulmonary disease are likely to be cost-effective: a systematic review Glenn Leemans a, Jan Taeymans b,c, Paul Van Royen d, Dirk Vissers e a Department of Rehabilitation Sciences and Physiotherapy, University of Antwerp, Antwerp, Belgium; b Department of Movement and Sport Sciences, Vrije Universiteit Brussel, Brussels, Belgium; c Department of Health Professions, Bern University of Applied Sciences, Bern, Switzerland; d Department of Family Medicine and Population Health, University of Antwerp, Antwerp, Belgium; e Department of Rehabilitation Sciences and Physiotherapy, University of Antwerp, Antwerp, Belgium KEY WORDS ABSTRACT Systematic review Question: What is the cost-effectiveness of respiratory physiotherapy interventions for people with chronic Healthcare economics and organisations obstructive pulmonary disease? Design: Systematic review of full economic evaluations alongside clinical Physical therapy modalities trials published between 1997 and 2021. Reviewers independently screened studies for inclusion, extracted Chronic obstructive pulmonary disease data and assessed methodological quality. Participants: People with chronic obstructive pulmonary disease. Intervention: Respiratory physiotherapy interventions as defined in the respiratory physiotherapy curricu- lum of the European Respiratory Society. Outcome measures: Costs expressed in monetary units, effect sizes expressed in terms of disease-specific quality of life (QOL), quality-adjusted life years (QALYs) or monetary units. Results: This review included 11 randomised trials with 3,261 participants. The interventions were pulmonary rehabilitation, airway clearance techniques, an integrated disease-management program and an early assisted discharge program, including inpatient respiratory physiotherapy. Meta-analysis was consid- ered irrelevant due to the extensive heterogeneity of the reported interventions. A total of 45 incremental cost-effectiveness ratios (ICERs) were extracted. Regardless of the economic perspectives, 67% of all QOL- related ICERs and 71% of all QALY-related ICERs were situated in the north-east or south-east quadrants of the cost-effectiveness plane. Six studies could be seen as cost-effective when compared with a specified cost- effectiveness threshold per QALY gained. Conclusion: Respiratory physiotherapy interventions focusing on exercise training in combination with enhancing physical activity levels are likely to be cost-effective in terms of costs per unit QOL gained and QALYs. Some uncertainty still exists on the various estimates of cost- effectiveness due to differences in the content and intensity of the type of interventions, outcome measures and comparators. Registration: PROSPERO CRD42018088699. [Leemans G, Taeymans J, Van Royen P, Vissers D (2021) Respiratory physiotherapy interventions focused on exercise training and enhancing physical activity levels in people with chronic obstructive pulmonary disease are likely to be cost- effective: a systematic review. Journal of Physiotherapy 67:271–283] © 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Background accounted for 42% of the total direct cost of treating respiratory dis- eases.4 For the United States, the cost of COPD in 2010 was projected Chronic obstructive pulmonary disease (COPD) is a common, often preventable and treatable disease, characterised by persistent respi- to be approximately $50 billion, which included $20 billion in indirect ratory symptoms and airflow limitation. Significant comorbidities costs and $30 billion in direct healthcare expenditures.5 Key cost- may have an impact on morbidity and mortality.1 Furthermore, COPD has a number of intrapulmonary and extrapulmonary components driving factors for direct medical costs are inpatient hospitalisation whose dynamic interactions along time are not linear, and not all of and medication due to exacerbations.6 In order to release this pres- these components are present in all individuals at any given time point.2 Besides this complex and heterogeneous nature, COPD is also sure on healthcare budgets in the future while still increasing the one of the most prevalent chronic respiratory diseases worldwide and associated with a significant social and economic burden.3 In the patients’ quality of life, it is expected that COPD treatments will European Union alone, the direct cost of COPD amounted to 6% of the increasingly be tailored to individual patients’ needs to accommodate total annual healthcare budget in 2011 (380 billion Euros) and the complexity and heterogeneity of this disease.7,8 Individualised approaches for patients with COPD are not new in non-pharmacological therapies. Pulmonary rehabilitation (PR), for example, has been defined as a comprehensive intervention based on a thorough patient assessment followed by patient-tailored therapies https://doi.org/10.1016/j.jphys.2021.08.018 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

272 Leemans et al: Cost-effectiveness of respiratory physiotherapy in COPD designed to improve the physical and psychological condition of Box 1. Inclusion criteria. patients with chronic respiratory disease and to promote the long- term adherence to health-enhancing behaviours.9 While PR is Design considered to be one of the most cost-effective therapies for  All types of full economic evaluations: individuals with COPD,10 comparing various estimates of cost- effectiveness is complicated, and further research in terms of costs cost-effectiveness analyses per disease-specific quality of life (QOL) and costs per quality- cost-utility analyses adjusted life years (QALYs) gained by PR is needed to reach definite cost-benefit analyses conclusions.9 This also seems to be the case for other respiratory  The economic data were collected alongside data from a single physiotherapy interventions such as airway clearance techniques. prospective clinical trial Participants This systematic review aimed to carry out a critical appraisal and  People with chronic obstructive pulmonary disease synthesis of health economic evaluations that investigated the cost- Intervention effectiveness of respiratory physiotherapy interventions in patients  Any skills as described in the modules of the European with COPD. Respiratory Society respiratory physiotherapy curriculum Outcomes Therefore, the research question for this systematic review was:  Measures of cost-effectiveness, cost-benefit and cost-utility Comparators What is the cost-effectiveness of respiratory physiotherapy  Standard of care interventions for people with COPD?  Any other respiratory physiotherapy treatment Method performed the search, the screening process and the initial inclusion of studies based on title and abstract. Full texts of relevant studies This systematic review was carried out and reported following the were consulted for definitive inclusion by two independent re- Preferred Reporting Items for Systematic Reviews and Meta-Analyses searchers (GL, DV). A consensus discussion between all researchers (PRISMA) guidelines.11 The five-step approach for preparing a sys- took place after title and abstract screening, as well as after full-text tematic review of economic evaluations by van Mastrigt et al was also consultation. used where appropriate.12 A search of the Cochrane Library revealed no similar published literature reviews. Assessment of characteristics of studies Identification and selection of studies Based on Wijnen et al,15 the following data were extracted from the included studies: study design, characteristics of study partici- A comprehensive literature search was performed in March 2020. pants, details of the intervention/comparator, and measurement and The five selected sources were a combination of several general da- valuation of effects and costs. When studies referred to data sets tabases (Medline/PubMed, Web of Science, Wiley Online Library), a previously published in other manuscripts, information was extracted specific database for economic evaluations of healthcare in- from those articles to make a complete data extraction. In the case of terventions (NHS EED) and a physiotherapy-specific database missing data, the study authors were contacted. Two independent (PEDro). In order to increase the sensitivity of the search, references researchers (GL, DV) extracted data on study characteristics following from included articles were checked (backward citation tracking). All the criteria in Box 1 and entered the data in a prepared digital form. A databases were searched with the following search terms: ‘COPD’, consensus discussion took place at the end of the data extraction ‘therapy’, ‘economics’ and ‘costs’. Boolean operators were used to process with the whole research team. combine search terms. Details of the search strategy are presented in Appendix 1 on the eAddenda. For PubMed, MeSH terms were used Quality when available. The searches were adapted and repeated across all The standardised JBI Critical Appraisal Checklist for Economic databases. Search results were stored in reference management softwarea and duplicates were removed. Studies that performed a full Evaluation,16 based on the guidelines developed by Drummond health economic evaluation involving people with COPD were et al,13 was used to assess the methodological quality and validity of included when two or more alternative interventions were the relevant economic evaluations. All 11 items of this checklist can compared, and both costs and effects (or benefits) of the compared be rated as ‘yes’, ‘no’ (inadequate methodology), or ‘unclear’ (insuf- treatment were taken into account.13 ficient information) and ‘not applicable’. All included articles were scored by two researchers independently (GL, DV). Discrepancies Interventions were defined as any skill performed by a respiratory were discussed in a consensus meeting. physiotherapist as described in the mandatory modules of the res- piratory physiotherapy curriculum of the European Respiratory So- Study design and setting ciety.14 Multicomponent interventions such as disease management The following details were extracted: study design, time horizon, programs, which focused on different health outcomes, were included if physiotherapy interventions as previously defined were a type of economic evaluation, country, perspective of the economic component of the rehabilitation program. The compared treatment analyses (healthcare, societal or third-party payer) and reported was either standard of care or any other respiratory treatment. The reference year with currency used. outcomes of the included studies were costs, expressed in monetary units, and effect sizes expressed in terms of natural units (eg, change Participants in lung function), disease-specific QOL, healthy years (eg, QALYs) or in monetary units. Box 1 summarises the inclusion criteria for studies in To describe the participants, the following data were extracted: this systematic review. major inclusion criteria, sample size, age and forced expiratory vol- ume in 1 second (FEV1) in percent predicted. Regarding study design, all types of full economic evaluations were eligible, namely: cost-effectiveness analyses (CEA), cost-utility Intervention analyses (CUA) or cost-benefit analyses if the economic data were collected alongside data from a single prospective clinical trial. Partial The details extracted about the interventions from each included and model-based economic evaluations were excluded. Publications study were: type of the (multicomponent) intervention, specific in- in languages other than English, German and French were excluded. terventions related to respiratory physiotherapy and (where re- Neither the publication date nor timeframe of the economic analyses, ported) the number of weeks, sessions or hours of one or more the time horizon, was specified. Two independent researchers

Research 273 interventions. Data about the control group were: description of provided an incremental analysis of cost and consequences. The usual care, type of (multicomponent) intervention and (where re- majority of the studies reported clinical effectiveness (nine studies), ported) the number of weeks, sessions or hours of one or more reflected on issues of concern to users in the study results (eg, interventions. decision-makers) (10 studies), discussed transferability of how to generalise results to other settings with similar characteristics (eight Outcome measures studies), adjusted costs for differential timing (eight studies) and conducted sensitivity analysis regarding uncertainty in estimates of The outcome measures considered by this review are listed in costs methods (nine studies). More specifically, the following sensi- Box 1. Detailed breakdowns of the costs per cost item and the tivity analysis methods were used: probabilistic (nine studies), uni- outcome parameters used for measuring effectiveness were extracted variate (six studies) and scenario analysis (one study). from the included studies. Design, setting and participants Data analysis Eleven economic evaluations alongside a clinical trial were Heterogeneity is well recognised in the content, healthcare pro- included in the final analysis. All were published between 1997 and vider and organisational aspects of respiratory physiotherapy in- 2020. All studies were randomised controlled trials, of which two terventions.14,17 Taking this heterogeneity into account, it was not used cluster-randomisation. All cluster-randomised trials took clus- plausible to pool effects and, hence, pooling was not undertaken. tering into account for the statistical analysis.19,20 Two trials were Therefore, this analysis remained purely descriptive and studies were designed to show equivalence rather than superiority between con- qualitatively analysed. A summary of incremental costs, incremental trol and intervention.21,22 In light of the research question, we effects and incremental cost-effectiveness ratios (ICERs) was tabu- reversed the control (usual care, including inpatient respiratory lated. Where applicable, 95% confidence intervals were reported as physiotherapy) and intervention group (early assisted discharge by well. ICERs were given with their location on the cost-effectiveness nurses) of Goossens et al23 in comparison with the original publica- plane. The cost-effectiveness plane presents the effectiveness of the tion. Studies were carried out in The Netherlands (four studies), intervention on the x-axis and the total costs on the y-axis and United Kingdom (four studies), Australia (one study), Ireland (one consists of four quadrants. ICERs in the south-east quadrant indicate study) and Canada (one study). Time horizons ranged from 3 months that the intervention compared to the alternative is more effective to 2 years; eight studies used time horizons between 1 and 2 years. and less expensive. ICERs in the south-west quadrant indicate that Included studies used the healthcare perspective alone (six studies), the intervention is less effective and less expensive. In the north-west societal perspective alone (one study), both healthcare and societal quadrant, the intervention is less effective and more expensive, while perspectives (three studies) and the societal and third-party payer’s ICERs in the north-east quadrant of the plane indicate more effective perspective (one study). Studies consisted of CEA alone (one study), but also more expensive interventions as compared to the alternative. CUA alone (two studies) and both CEA and CUA (eight studies). In In this situation, the cost-effectiveness depends on the willingness to most studies, inclusion criteria were stable COPD diagnosed accord- pay, which is defined as the maximum amount of money an indi- ing to GOLD guidelines1 and a smoking history of  10 pack-years. vidual is willing to pay to avoid or reduce a specific health problem or Only Cross et al22 and Goossens et al23 recruited people with COPD to gain a specific health benefit.13 during an acute exacerbation. The number of participants across the included studies ranged from 89 to 1,086, with an aggregate total (at Nonetheless, to enhance comparability of included studies all time of randomisation) of 3,261. All trials recruited a similar number study currencies for the ICERs were converted to 2019 Euros by the of participants for both study groups. Participants’ mean age at Campbell and Cochrane Economics Methods Group (CCEMG) and the recruitment was 67.1 years (range 49.6 to 71), while mean FEV1 was Evidence for Policy and Practice Information and Coordinating Centre 51% predicted (range 35 to 68). Groups were broadly comparable at (EPPI-Centre) cost converterb.18 This free web-based tool automati- baseline, albeit with some differences: Boland et al20 reported a cally adjusts estimates for costs and price year, taking purchasing higher percentage of males in the usual care group (51 versus 57%) power parities between countries into account. and the control group of Burns et al24 had on average a higher EuroQol five dimensional (EQ-5D) utility (0.7 versus 0.6) than the Results intervention group. Flow of studies through the review Interventions After removing 100 duplicates, the database search yielded 435 Reported interventions were PR programs (eight studies), airway records, from which 59 articles were identified to review for eligi- clearance techniques (ACTs) (one study), an integrated disease man- bility. These were obtained in full text and assessed, resulting in the agement program (one study) and one study on the effect of early inclusion of 11 studies in this review. The most frequent reasons for assisted discharge by nurses compared with usual care, including exclusion were intervention (eg, no involvement of respiratory inpatient physiotherapy interventions. The majority of the in- physiotherapist as healthcare provider during intervention), outcome terventions that were studied were compared against usual care (eg, no information about cost-effectiveness) and study design (eg, (seven studies). Four interventions were compared against another reviews, abstracts and economic models). A full overview of the flow intervention: a home-based PR versus centre-based PR program,21 is provided in Figure 1. Study details of all included studies are pre- ACTs with manual percussion versus ACTs without percussion,22 PR sented in Table 1. in hospital versus community settings with and without telephone follow-up,25 and a self-management program versus a community- Characteristics of included studies based exercise program within that self-management program.26 Using the syllabus items of the European Respiratory Society core Quality curriculum in respiratory physiotherapy27 as a framework to report The results of the critical appraisal of the studies are presented in the different interventions, it was observed that 82% of the in- terventions were exercise training in combination with enhancing Table 2. Five of 11 studies fulfilled all 11 criteria on the checklist. All 11 physical activity levels (see Table 3). studies: posed a clear research question; provided a clear description of the interventions and comparators; identified and measured all The reported content, timing and organisational aspects differed important costs and outcomes and valued those credibly; and greatly between the exercise programs. First, the intervention