Journal of Physiotherapy 64 (2018) 142–158 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Research Additional weekend allied health services reduce length of stay in subacute rehabilitation wards but their effectiveness and cost-effectiveness are unclear in acute general medical and surgical hospital wards: a systematic review Mitchell N Sarkies a, Jennifer White a, Kate Henderson b, Romi Haas a, John Bowles c Evidence Translation in Allied Health (EviTAH) Group1 a Department of Physiotherapy, Monash University; b Physiotherapy Department, Monash Health; c Allied Health Research Unit, Monash University, Melbourne, Australia KEY WORDS ABSTRACT Systematic review Question: Are additional weekend allied health services effective and cost-effective for acute general Meta-analysis medical and surgical wards, and subacute rehabilitation hospital wards? Design: Systematic review and Allied health meta-analysis of studies published between January 2000 and May 2017. Two reviewers independently Physical therapy screened studies for inclusion, extracted data, and assessed methodological quality. Meta-analyses were Weekend conducted for relative measures of effect estimates. Participants: Patients admitted to acute general medical and surgical wards, and subacute rehabilitation wards. Intervention: All services delivered by allied health professionals during weekends (Saturday and/or Sunday). This study limited allied health professions to: occupational therapy, physiotherapy, social work, speech pathology, dietetics, art therapy, chiropractic, exercise physiology, music therapy, oral health (not dentistry), osteopathy, podiatry, psychology, and allied health assistants. Outcome measures: Hospital length of stay, hospital re-admission, adverse events, discharge destination, functional independence, health-related quality of life, and cost of hospital care. Results: Nineteen articles (20 studies) were identified, comprising 10 randomised and 10 non-randomised trials. Physiotherapy was the most commonly investigated profession. A meta-analysis of randomised, controlled trials showed that providing additional weekend allied health services in subacute rehabilitation wards reduced hospital length of stay by 2.35 days (95% CI 0.45 to 4.24, I2 = 0%), and may be a cost-effective way to improve function (SMD 0.09, 95% CI –0.01 to 0.19, I2 = 0%), and health-related quality of life (SMD 0.10, 95% CI –0.01 to 0.20, I2 = 0%). For acute general medical and surgical hospital wards, it was unclear whether the weekend allied health service model provided in the two identified randomised trials led to significant changes in measured outcomes. Conclusion: The benefit of providing additional allied health services is clearer in subacute rehabilitation settings than for acute general medical and surgical wards in hospitals. Registration: PROSPERO CRD76771. [Sarkies MN, White J, Henderson K, Haas R, Bowles J, Evidence Translation in Allied Health (EviTAH) Group (2018) Additional weekend allied health services reduce length of stay in subacute rehabilitation wards but their effectiveness and cost-effectiveness are unclear in acute general medical and surgical hospital wards: a systematic review. Journal of Physiotherapy 64: 142–158] © 2018 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 1 The EviTAH Group: Terry P Haines, School of Primary and Allied Health, Monash Allied health services, medicine and nursing are considered to University; Meg E Morris, La Trobe Centre for Sport and Exercise Medicine Research, comprise three pillars of the healthcare system.1 Allied DF3_I1]$T[health La Trobe University; Leeanne Carey, Occupational Therapy, School of Allied Health, professionals provide diagnostic and therapeutic services across La Trobe University, and the Florey Institute of Neuroscience and Mental Health, different settings,2and represent a large proportion of the healthcare Neurorehabilitation and Recovery, Melbourne Brain Centre; Nicholas F Taylor, La workforce internationally.3,4 Allied health is often organised and Trobe Centre for Sport and Exercise Medicine Research, La Trobe University; Anne E Holland, Department of Rehabilitation, Nutrition and Sport, La Trobe University; managed in professional groups, including physiotherapy, psychol- Anne Bardoel, Department of Management and Marketing, Swinburne University; Cylie Williams, Allied Health Department, Peninsula Health; Lisa O’Brien, ogy, occupational therapy, speech pathology, dietetics, podiatry, and Department of Occupational Therapy, Monash University; Elizabeth H Skinner, Allied Health Research Unit, Monash University; Kathleen Philip, Department of social work, within an over-riding inter-professional comprehensive Health and Human Services; Kelly-Ann Bowles, Department of Community care model.5,6 Emergency Health and Paramedic Practice, Monash University, Melbourne, Australia. The routine provision of weekend allied health services is variable across hospitals both in Australia and worldwide. For example, a survey of tertiary care hospitals in Canada reported that https://doi.org/10.1016/j.jphys.2018.05.004 1836-9553/© 2018 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/).
Research 143 97% of facilities provided routine weekend physiotherapy services, Box 1. Inclusion criteria. with diversity in the amount and focus of service delivery.7 In Australia, 61% of hospitals routinely provided physiotherapy on Design Saturdays, and 45% on Sundays, with more provided in metropoli- Randomised, controlled trials tan and acute hospitals than regional and subacute hospitals.8 Non-randomised, controlled trials Aside from physiotherapy, comparatively little is known about the Observational studies practices of other allied health professions. The evidence about the Participants effects of providing these services during weekends is unclear. In-patients on a general medical or surgical wards or a Providing earlier,9,10 additional,11–13 or higher intensity14 allied health services can improve health outcomes. However, it is subacute rehabilitation ward of a hospital unclear whether these benefits occur when weekend allied health Intervention staffing models are used to deliver additional services on a routine Additional allied health service delivered at the weekend basis.15 The provision of allied health services on weekends incurs Outcome measures more cost and logistical difficulty than during traditional business Hospital length of stay hours, with uncertainty around the experience of staff, appropri- Hospital re-admission ateness of referrals, and whether the mix of professions achieves Adverse events the intended benefits.16 Discharge destination Functional independence The aim of this review was to synthesise the available evidence Health-related quality of life examining the effectiveness and cost-effectiveness of providing Cost of hospital care additional weekend allied health services to patients on acute Comparisons general medical and surgical hospital wards, and subacute Additional allied health services versus usual allied health rehabilitation hospital wards. services only Therefore, the research question for this systematic review was: considered to be different to those on inpatient rehabilitation and Are additional weekend allied health services effective and geriatric evaluation and management wards. cost-effective for acute general medical and surgical wards, and subacute rehabilitation hospital wards? Interventions focused upon in this review included all services delivered by allied health professionals during weekends (Satur- Methods day and/or Sunday). This study limited allied health professions to: occupational therapy, physiotherapy, social work, speech patholo- This systematic review was conducted and reported in gy, dietetics, art therapy, chiropractic, exercise physiology, music accordance with the Preferred Reporting Items for Systematic therapy, oral health (not dentistry), osteopathy, podiatry, psychol- Reviews and Meta-Analysis (PRISMA) guidelines.17 ogy, and allied health assistants.19 An allied health service not delivered by an allied health professional or allied health assistant Identification and selection of studies (eg, nursing staff or self-directed) was not eligible. Weekends were defined as complementary to the traditional workweek, as per the Ovid MEDLINE (all fields), PubMed (all fields), CINAHL country the study was performed in. Studies that reported data (keyword, title, CINAHL subject headings, abstract, and instru- relating to the provision of additional allied health services as part mentation fields), Cochrane library (title, abstract, keywords), and of changing timing of commencement, intensity, frequency or Scopus (title, abstract, keywords) were searched for articles duration with a weekend component were included, but only if published between 1 January 2000 to 5 May 2017 to retrieve data relating specifically to weekend services with appropriate contemporary literature. Terms relevant to the population and controls could be extracted. intervention were combined and results were limited to English- language publications. See Appendix 1 on the eAddenda for the full Two reviewers (MS and JW) screened titles and abstracts search strategy. Electronic database searches were supplemented independently against the above criteria. Studies determined to be by cross-checking the reference list of included articles and potentially eligible were retrieved for full-text review. Two relevant systematic reviews identified during the screening reviewers (MS and JW) independently assessed the full-text process. Publication lists from key authors in the field were also articles to ascertain eligibility for inclusion. Where there was any hand searched to identify additional studies. A web-based disagreement during the screening, a third independent reviewer applicationa[D]F32$_IT was used for reference management.18 (KH) was consulted. Authors of studies whose full-text article could not be retrieved were contacted. In the cases of non- The inclusion criteria for the review are presented in Box 1. For response, these articles were excluded. the purposes of this review, acute general medical and surgical wards included: general medical, general surgical, medical Assessment of characteristics of studies assessment unit, orthopaedic, vascular, plastics, ear nose and throat, thoracic, respiratory, coronary care unit, renal, rheumatol- Data were extracted using a customised pro-forma, which was ogy, neurology (including stroke units), infectious diseases, developed and piloted for this review. One (JW) and either of two colorectal, endocrine, urological, and gastroenterology. Excluded other reviewers (KH or JB) independently extracted data relating to acute wards were emergency department, intensive care unit, high the study details, design, setting, population, intervention, out- dependency unit, burns, spinal, maternity, paediatrics, mental comes, and results for all included studies. Discrepancies in health, and palliative care. These wards were excluded because the extracted data were resolved by discussion. Where agreement allied health role was considered to be potentially different in could not be reached, a fourth independent reviewer (MS) was these settings compared to acute general medical and surgical consulted. wards. For the purposes of this study, subacute rehabilitation wards included inpatient rehabilitation (both mixed and condi- Quality tion-specific wards), and geriatric evaluation and management Two of three reviewers (JW, KH or JB) independently assessed wards. Excluded subacute wards comprised mental health and psychiatric, chronic and long-term care, alternative level of care, the risk of bias for randomised, controlled trials using the Cochrane and extended care patients. The goals of care on these wards were Collaboration’s tool for assessing risk of bias,20 and the Newcastle- Ottawa Quality Assessment Scale for observational studies.21 Any discrepancy in the assessments of risk of bias was resolved by
144 Sarkies et al: Additional weekend allied health services in hospital discussion and, if necessary, consultation of a fourth independent Two of the studies identified and included in this review were reviewer (RH). stepped-wedge cluster randomised trials, conducted in succession at the same research location (hospitals and wards). These were Participants the only randomised trials identified in the acute setting. Data To describe the participants in the study, the following data available from this study were used at the participant level (for continuous outcomes) and cluster level (for proportion outcomes) were extracted from the published report: sample size, age, rather than summative data. This was done to incorporate the country, and ward (acute or subacute). dependency of observations gathered from the same wards between the two trials. Weighted mean difference (WMD) was Intervention used for continuous outcomes, and risk difference (RD) for The details extracted about the allied health service from each proportional outcomes. included study were: the number of hospitals, wards and/or beds Results being covered by the service, where reported; the number of additional hours of service; and the number and professions of the Flow of studies through the review allied health staff providing the service. Data about the control intervention were the number of hours of weekday service, where A total of 3413 titles were identified, with 3405 from the reported, and the number and professions of the allied health staff electronic search strategy and eight from hand searching publica- providing the weekday service. tion lists of prominent authors. Duplicates (n = 293) were removed using Endnote (n = 224) and Covidence (n = 69), resulting in Outcome measures 3120 titles remaining for screening (Figure 1). After title and The outcomes considered by this review are listed in Box abstract screening, 72 records were considered potentially eligible for inclusion. These were obtained in full text and assessed, 1. These were intended to cover all patient and health service resulting in the inclusion of 19 articles (n = 20 studies) in this outcomes that might be reported in the included studies. review.25–43 (See Appendix 2 on the eAddenda for a list of the excluded full-text articles.) Data analysis Characteristics of included studies Analysis was performed using Stata 13 softwareb. Relative measures of effect estimates were pooled according to study Quality setting and design (separate for acute and subacute settings, and The Cochrane Collaboration tool for assessing risk of bias in the randomised and non-randomised study designs). Random effect meta-analysis accounted for differences in populations, interven- included randomised, controlled trials is presented in tions and outcomes across studies, and was performed where data Figure 2. Eight studies reported adequate methods of random were available for similar outcomes evaluated in more than one sequence generation.25,27–33 All studies reported adequate alloca- study. tion concealment; however, blinding of participants and personnel was not possible in any included studies. Hospital length of stay A majority of analyses used summative, study-level data. was unable to be blinded in any included studies,25–33 although the Weighted mean difference (WMD) effect size estimates were used stepped-wedge, cluster-randomised trials limited this as a poten- for continuous outcomes, where measurements were reported in tial source of bias.25 Risk of bias for selective reporting was the same units (eg, length of stay and cost). Where measures used identified in two studies,27,32 with another study reporting further different scales, standardised mean difference (SMD) was used to outcomes to be reported in other publications.25 No other potential estimate effect size for function and health-related quality of life sources of bias were identified for any of the included studies. outcomes. The effect size for SMD was interpreted according to Cohen’s d, with 0.2 considered small, 0.5 moderate, and 0.8 a large For non-randomised, controlled trials assessed using the effect size.22 Risk ratios (RR) were used for dichotomous outcomes Newcastle-Ottawa quality assessment scale (Table 1), eight studies (eg, number of adverse events, patients discharged home, and reported adequate methods to control for potential risk of delayed discharge). selection bias,34–36,38,39,41–43 with two studies potentially at risk of selection bias.37,40 There was adequate reporting of methods to Authors were contacted to request additional data for studies control for potential risk of comparability bias for all studies. Seven not reporting sufficient outcome data for inclusion in the meta- studies reported adequate methods to control for potential analysis. Data were pooled from subgroups to estimate the total outcome or exposure risk of bias,34,36,38,39,41–43 with three studies population effect size where data were only reported according to potentially at risk of outcome bias.35,37,40 subgroups within an individual trial. The proportion of patients discharged to supported residential aged-care facilities reported in Design, setting and participants subgroups according to level of care were summed to create a total Ten randomised, controlled trials (published in nine articles) proportion of patients discharged to a supported residential aged care facility. When two control groups were used for non- evaluated the effect of providing allied health services during randomised, controlled trials, the first mentioned control group weekends. Two studies were performed in acute general medical was selected as the comparator. This applied only to observational and surgical hospital ward settings.25,44 Eight were within studies where multiple time periods/cohorts were reported within subacute rehabilitation hospital wards, including mixed patient the study. If study results were reported as medians and range or groups,26–30 orthopaedic,31 and stroke rehabilitation.32,33 Ten non- interquartile range, and the mean and standard deviation could not randomised, controlled trials (10 articles) were also identified. be obtained, means and standard deviation were estimated using Seven were cohort studies performed in acute general medical and the methods of Wan et al.23 A sensitivity analysis was undertaken surgical hospital wards, including orthopaedic,34–37,43 rheumatol- to assess the effect of excluding studies where data were estimated ogy,38 and stroke.39 Two cohort studies were performed in mixed due to inadequate reporting. rehabilitation subacute hospital wards,40,41 and one cross-section- al study was conducted in a mixed rehabilitation setting.42 The Heterogeneity in study results was represented using the majority of studies were performed in Australia (n = 14),25–35, I-squared statistic (I2), with values > 50% considered substan- 40,41,43 with one study each in England,38 Canada,42 Japan,39 tial.24 An iterative approach was used to explore possible Singapore,36 and Scotland.37 Study details are presented in Table 2. explanations for heterogeneity by subgrouping studies according to variables such as allied health profession, patient population, and potential sources of bias. A formal meta-regression was not planned due to the anticipated low number of studies for inclusion in the meta-analysis.
re_1)TD$FIG]ugi[( Research 145 Records identified through Additional records identified database searching (n = 3405) through other sources (n = 8) Records after duplicates removed (n = 3120) Records screened (n = 3120) Records excluded (n = 3048) Potentially relevant articles Full-text articles excluded (n = 53) retrieved for evaluation of full • duplicate publication (n = 13) text (n = 72) • no appropriate control (n = 20) • ineligible outcome measures (n = Included in synthesis (n = 19 articles) 7) (n = 20 studies) • therapy not delivered by an allied Included in meta-analysis (n = 16 studies) health professional (n = 7) • therapy provided in an ineligible setting (n = 2) • commentary (n = 1) • non-English language (n = 1) • outpatient setting (n = 1) • systematic review (n = 1) Figure 1. Flow of studies through the review. Interventions weekend allied health) service with a ‘new stakeholder driven’ weekend model including Friday, Saturday, Sunday and Monday Provision of physiotherapy during weekends was the most services.25 In this ‘new stakeholder driven service’, weekend services examined allied health service (n = 9),26,32–38,40 followed by were distributed across Friday, Saturday, Sunday and Monday. Further details of intervention and control conditions are presented in Table 2. combinations of: physiotherapy and occupational therapy (n = 7);27–31,39,42 physiotherapy, occupational therapy, and social Effect of additional weekend allied health services work (n = 1);41 and physiotherapy, occupational therapy, speech pathology, dietetics, and social work (n = 3).25,43 The numerical results of all meta-analyses, subgroup analyses and sensitivity analyses are presented in Tables 3, 4, 5 and Ten studies compared a Monday to Friday allied health service with 6. Streamlined forest plots are presented in the paper, and detailed a model that incorporated Saturday and Sunday services,25,32–35, forest plots are presented in Appendix 3 on the eAddenda. 37,38,42,43 eight compared Monday to Friday with a Saturday service,26–31,40,41 and one compared a Monday to Saturday service w[(Fgure_2)TD$IG]i ith a Sunday service.39 One study compared a Monday to Friday (no Random sequence generation Effect on hospital length of stay in the acute setting Allocation concealment Blinding of participants Randomised trials Blinding of personnel Blinding of outcome assessors Two acute, stepped-wedge, randomised, controlled trials Incomplete outcome data (n = 27 508) were identified.25F_3I]D[T$ Meta-analysis of individual partici- Selective reporting pant level data in these trials demonstrated no difference between Other sources of bias Risk of bias intervention and control conditions for hospital length of stay low (WMD 0.08 days, 95% CI –0.15 to 0.32, I2 = 99%), as shown in unclear high Table 1 Risk of bias of the nine included cohort studies and one included cross-sectional Haines 2017 25 study, assessed using the Newcastle-Ottawa quality assessment scale for Brusco 2007 26 observational studies. Brusco 2014 29 Brusco 2014 28 Cohort study Selection Comparability Outcome Brusco 2015 30 English 2014 33 Boxall 2004 35 **** * ** English 2015 32 David 2003 38 **** * *** Haas 2017 43 **** * *** Peiris 2012 31 Kinoshita 2017 39 **** * *** Peiris 2013 27 Maidment 2014 34 **** * *** Pengus 2015 37 *** * ** Figure 2. Risk of bias of the included randomised trials, assessed using the Cochrane Pua 2011 36 **** * *** Risk of Bias tool. Caruana 2016 40 *** * ** Hakkennes 2015 41 **** * *** Cross-sectional study Selection Comparability Exposure DiSotto-Monastero 2012 42 **** * *** The higher the number of stars in each column indicates lower risk of bias. A different scale version was used for cross-sectional studies, where ‘outcome’ was replaced with ‘exposure’.
Table 2 Detailed characteristics of the included studies. Study Setting Participants a Con Haines et al 2017 25 Australia No allied hea Study A Mixed general medical and surgical weekends Acute (hospitals = 2; wards = 12; n = 14 834) RCT Con: n = 6796, age = 61 (20) No allied hea Exp: n = 8038, age = 59 (21) weekends Haines et al 2017 25 Australia Study B Acute Hospital 1, 6 inpatient wards: orthopaedic 5-day weekd surgery, stroke, thoracic/vascular/general physiotherap RCT surgery & medical, general medicine, head/ patient) neck/plastics, and surgical PT 1 Brusco et al 2007 26 Australia RCT Sub-acute Hospital 2, 6 inpatient wards: medical (2 wards), infectious diseases/respiratory, plastics/ENT/head/neck surgery, general surgery/colorectal/breast/endocrine/urology, and general surgery/vascular/thoracic/upper gastrointestinal Mixed general medical and surgical (hospitals = 2; wards = 11; n = 12 674) Con: n = 6869, age = 60 (21) Exp: n = 5805, age = 60 (20) Hospital 1, 6 inpatient wards: orthopaedic surgery, stroke, thoracic/vascular/general surgery & medical, general medicine, head/ neck/plastics, and surgical Hospital 2, 5 inpatient wards: medical (2 wards), infectious diseases/respiratory, plastics/ENT/head/neck surgery, general surgery/colorectal/breast/endocrine/urology, and general surgery/vascular/thoracic/upper gastrointestinal Mixed rehabilitation (wards = 2; n = 262) Con: n = 132, age = 77 (13) Exp: n = 130, age = 77 (13)
ntrol Intervention Outcome 146 Sarkies et al: Additional weekend allied health services in hospital alth services on Usual care allied health services on weekends Primary Hospital 1 (hrs/hospital) Hospital length of stay Length of stay longer than expected length Sat: PT 8, OT 3, SP 3.5, DT 2, SW 1, AHA 4 Sun: PT 11, OT 3, SP 3, DT 2, SW 1, AHA 4 of stay Hospital readmission Hospital 2 (hrs/hospital) Adverse events Sat: PT 3.25, OT 3.5 Secondary Sun: PT 3.25 Compliments and complaints Discharge destination Occasions of allied health service Cost of hospital admission Clinical exceptions Staff absenteeism alth services on Newly developed stakeholder-driven weekend Primary allied health service Hospital length of stay Length of stay longer than expected length Hospital 1 (hrs/hospital) Fri: PT 8, OT 4, AHA 4 of stay Sat: PT 3.5, OT 2, SP 3.5, DT 1, SW 1 Hospital readmission Sun: PT 7, OT 2, SP 3, DT 1, SW 1, AHA 4 Adverse events Mon: PT 4, OT 4 Secondary Compliments and complaints Hospital 2 (hrs/hospital) Discharge destination Sat: ICU PT 1, IRS 4, SP 2 Occasions of allied health service Sun: ICU PT 1 Cost of hospital admission Clinical exceptions Staff absenteeism day Additional Saturday physiotherapy (hrs/patient) Primary py (daily hrs/ PT 1 Hospital length of stay Physiotherapy length of stay Secondary EuroQol Functional Independence Measure Functional reach 10-m Walk Test Timed Up and Go test Motor Assessment Scale Knee and hip range of motion Discharge destination Adverse events Follow-up therapy
Table 2 (Continued ) Setting Participants a Con Study Australia Brusco et al 2014 29 Sub-acute Mixed rehabilitation (hospitals = 2; beds = 90; 5-day weekda RCT n = 996) physiotherapy Brusco et al 2014B 28 Australia Con: n = 500, age = 74 (13) occupational RCT Sub-acute Exp: n = 496, age = 75 (13) hrs/patient) PT 1, OT 1 Brusco et al 2015 30 Mixed rehabilitation (hospitals = 2; beds = 90; RCT n = 137) 5-day weekda English et al 2014 33 Con: n = 63, age = 61 (13) physiotherapy RCT Exp: n = 74, age = 63 (12) occupational hours per pat English et al 2015 32 Australia Mixed rehabilitation (hospitals = 2; beds = 90; PT 1, OT 1 RCT Sub-acute n = 996) Con: n = 500, age = 74 (13) 5-day weekda Australia Exp: n = 496, age = 75 (13) physiotherapy Sub-acute occupational Stroke rehabilitation (centres = 5; n = 21) hrs/patient) Con: n = 10, age = N/S PT 1, OT 1 Exp: n = 11, age = N/S Usual care ph Australia Stroke rehabilitation (centres = 5; n = 190) (recruitment Sub-acute Con: n = 94, age = 68 (13) Exp: n = 96, age = 72 (12) Individual the per week (3/5 Individual or between 1 to week (2/5 sit Weekend the patients (2/5 Usual care ph (recruitment Individual the per week (3/5 Individual or between 1 to week (2/5 sit Weekend the patients (2/5
ntrol Intervention Outcome ay Additional Saturday physiotherapy and occupational therapy (daily hrs/patient) Health service and therapy utilisation py and PT 1, OT 1 (30-day follow-up) therapy (daily Additional Saturday physiotherapy and Cost of inpatient rehabilitation ay occupational therapy (daily hrs/patient) (30-day follow-up) py and PT 1, OT 1 Primary therapy (daily Return to work tient) Average hours worked Paid income ay Additional Saturday physiotherapy and Secondary Research py and occupational therapy (daily hrs/patient) Functional Independence Measure PT 1, OT 1 EuroQol-5D therapy (daily Hospital length of stay 7-day physiotherapy (daily duration/patient) hysiotherapy PT (matched to preceding week) Health service and therapy utilisation sites) (12-month follow-up) 7-day physiotherapy (daily duration/patient) erapy 5 days PT (matched to preceding week) Cost of inpatient rehabilitation (12-month 5 sites) follow-up) group therapy o 4 days per Therapy duration tes) Reasons for shortened therapy erapy for some Reasons for non-attendance Activity across day sites) Activity during therapy Activity outside therapy hysiotherapy Activity location sites) Activity with people present erapy 5-days Primary 5 sites) Six-minute walk test Secondary group therapy Walking speed o 4 days per Functional ambulation category tes) Functional Independence Measure erapy for some Wold Motor Function test Stroke Impact Scale physical subscale sites) Hospital length of stay Assessment of quality of life Adverse events Resource utilisation 147
Table 2 (Continued ) Setting Participants a Con Study Australia Peiris et al 2012 31 Sub-acute Mixed rehabilitation (hospitals = 2; beds = 90; 5-day weekd RCT n = 105) physiotherap Australia Con: n = 54, age = 73 (13) occupational Peiris et al 2013 27 Sub-acute Exp: n = 51, age = 75 (12) hrs/patient) RCT PT 1, OT 1 Mixed rehabilitation (hospitals = 2; beds = 90; Boxall et al 2004 35 n = 996) 5-day weekd Cohort Con: n = 500, age = 74 (13) physiotherap Exp: n = 496, age = 75 (13) occupational David et al 2003 38 hrs/patient) Cohort PT 1, OT 1 Haas et al 2017 43 Cohort Australia Orthopaedic (wards = 1; n = 240) 5-day weekd Acute Con: n = 120, age = 68 physiotherap Exp: n = 120, age = 68 England 5-day weekd Acute Rheumatology (wards = 1; beds = 28; n = 361) physiotherap Australia Con: n = 146, age = N/S PT 2, AHA 0.3 Acute Exp: n = 215, age = N/S 5-day weekd Orthopaedic (wards = 2; beds = 48; n = 276) health servic Con: n = 146, age = 69 (10) physiotherap Exp: n = 130, age = 68 (11) occupational speech patho and social wo
ntrol Intervention Outcome 148 Sarkies et al: Additional weekend allied health services in hospital day Additional Saturday physiotherapy and Primary py and occupational therapy (daily hrs/patient) Steps per day l therapy (daily PT 1, OT 1 Time spent upright Secondary day Additional Saturday physiotherapy and Time spent inactive py and occupational therapy (daily hrs/patient) Activity completed in therapy l therapy (daily PT 1, OT 1 Primary day 7-day physiotherapy service Functional Independence Measure py EuroQol-5D Hospital length of stay day Additional weekend physiotherapy (total EFT) Secondary py (total EFT) PT 0.2, AHA 0.4 Personal Care-Participation Assessment and 3 Saturday and Sunday allied health: Resource Tool physiotherapy, occupational therapy, speech 10-m Walk Test day allied pathology, dietetics, and social work Timed Up and Go test ce: Modified Motor Assessment Scale py, Discharge destination l therapy, Follow-up therapy ology, dietetics, Adverse events ork Hospital length of stay Days to independent transfers Days to independent mobility Reasons for delayed discharge Discharge destination Pre-admission clinic attendance Number of rheumatology admissions Hospital length of stay Day of admission and discharge Staff cost and utilisation Hospital length of stay Adverse events Discharge destination Time to first postoperative transfer Staff profession assisting first postoperative transfer Physiotherapy session rate Time to first postoperative physiotherapy session Modified Barthel Index De Morton Mobility Index EuroQol-5D utility EuroQol-5D VAS Patient satisfaction Pain Opioid use
Table 2 (Continued ) Setting Participants a Con Study Japan 5-day or 6-da Kinoshita et al 2017 39 Acute Stroke (hospitals = 14; n = 3072) physiotherapy Cohort Con: n = 1997, age = 73 (13) b occupational Australia Exp: n = 1075, age = 74 (12) b Maidment et al 2014 34 Acute 5-day physiot Cohort Orthopaedic (ward = 1; n = 145) service Scotland Con: n = 59, age = 72 (8) b Pengas et al 2015 37 Acute Exp: n = 86, age = 69 (7) b 5-day weekda Cohort physiotherapy Singapore Orthopaedic (n = 792) Pua et al 2011 36D$]IF[2_T Acute Con: n = 600, age = N/S 6-day Monda Cohort Exp: n = 192, age = N/S physiotherapy Australia PT 1 Caruana et al 2016 40 Sub-acute Orthopaedic (n = 155) Cohort Con: n = 82, age = 65 (7) 5-day Monda Exp: n = 73, age = 66 (8) physiotherapy occupational Mixed rehabilitation (ward = 1; beds = 40; speech patho n = 270) dietetics Con: n = 108, age = 75 (4) Exp: n = 162, age = 78 (12) DiSotto-Monastero et al 2012 42 Canada Mixed rehabilitation (n = 3500) 5-day physiot Cross-sectional Sub-acute Con: n = 1692, age = 72 (13) occupational Exp: n = 1808, age = 72 (14) Hakkennes et al 2015 41 Australia Mixed rehabilitation (hospital = 1; beds = 100; 5-day weekda Cohort Sub-acute n = 976) Con: n = 499, age = 78 (10) b Exp: n = 477, age = 78 (12) b AHA = allied health assistant, AIN = assistant in nursing, Con = control group, DT = dietician, Exp = experimental group, ICU = OT = occupational therapist, PT = physiotherapist, RCT = randomised, controlled trial, SP = speech pathologist, SW = social a Age is presented as mean (SD) in years. b Converted from median (range/inter-quartile range) to mean (SD).
ntrol Intervention Outcome ay 7-day physiotherapy and occupational therapy Modified Rankin Scale py and Hospital length of stay Adverse events therapy Discharge destination therapy 7-day physiotherapy service Physiotherapy length of stay Hospital length of stay Number of physiotherapy sessions Reasons for delayed discharge ay Additional Saturday and Sunday physiotherapy Days to mobilising with two sticks py provided by an allied health assistant (daily Hospital length of stay hours) Time to achieve 90 knee flexion AHA 3 Range of motion ay to Saturday Additional Sunday physiotherapy Hospital length of stay py (daily OOS) Passive range of motion D_3[FT$]-I knee Straight leg raise Independent mobility ay to Friday Additional Saturday physiotherapy (daily hours) Primary Research py, PT and AIN 4 Hospital length of stay Secondary therapy, Functional Independence Measure ology, and Timed Up and Go test 10-m Walk Test Functional reach Step test Feet together eyes closed Balance Outcome Measure for Elder Rehabilitation therapy and 7-day physiotherapy and occupational therapy Functional Independence Measure therapy Number of admissions and discharges Hospital length of stay Rehabilitation workload ay therapy Additional Saturday physiotherapy, Functional Independence Measure occupational therapy, and social work (daily Hospital length of stay hrs/hospital) Number of admissions on Saturday PT 20, OT 16, SW 8, AHA 6 = Intensive Care Unit, IRS = Immediate Response Service, n = number of patients, N/S = not stated, OOS = occasions of service, l worker. 149
150 Sarkies et al: Additional weekend allied health services in hospital Table 3 Randomised, controlled trial meta-analysis investigating the effect of weekend allied health on acute general medical and surgical wards. Outcome Weighted mean difference Risk difference (95% CI) I2 (95% CI) I2 Hospital length of stay (days) –0.08 (–0.32 to 0.15) 99% 0.00 (–0.02 to 0.02) 78% Hospital length of stay (log transformed) –0.02 (–0.05 to 0.02) 88% 0.01 (0.00 to 0.02) 43% Proportion of patients staying longer than expected 0.00 (–0.01 to 0.01) 83% Unplanned hospital readmissions within 28 days –118 (–511 to 274) 99% 0.00 (0.00 to 0.01) 31% Proportion of patients with any adverse event Proportion of patients discharged to aged care Cost to healthcare system per admission (AUD) Total effect size for continuous outcomes calculated from participant level data, and total effect size for proportion outcomes were calculated from cluster level data. Positive values favour intervention. Negative values favour control. No subgroup or sensitivity analysis were performed because only two randomised trials in the acute setting were identified. Figure 3a. When patient diagnosis was taken into account, there Non-randomised studies was no difference in the proportion of patients whose hospital This was consistent with the results of one non-randomised length of stay was longer than their expected length of stay between intervention and control conditions using cluster-level trial that also demonstrated no between-group difference in data (RD 0.00 days; 95% CI –0.02 to 0.02, I2 = 79%), as shown in unplanned hospital re-admission within 6 weeks of discharge.39 Figure 4. High levels of heterogeneity in the study results were examined in post-hoc exploratory analysis by Haines et al,25 Effect on hospital readmissions in the subacute setting suggesting that there was a significant change in hospital length of stay outcomes between control conditions, but when intervention Meta-analysis was unable to be performed for this outcome in conditions were compared, there was no significant difference either randomised or non-randomised trials in the subacute between patient hospital length of stay outcomes.25 setting. However, one subacute randomised, controlled trial reported no difference between intervention and control condi- Non-randomised studies and subgroup analyses tions for hospital re-admission within 30 days of discharge.29 These meta-analyses results were somewhat concordant with Effect on adverse events in the acute setting those involving non-randomised studies. Meta-analysis (n = 4676) of six acute non-randomised, controlled trials34–39,43 showed no Randomised trials effect of providing weekend allied health services on hospital Meta-analysis (n = 27 508) of cluster-level data from two acute, length of stay (WMD 0.24 days, 95% CI –0.17 to 0.66, I2 = 96%), as shown in Figure 3b. Heterogeneity levels were reduced when stepped-wedge, randomised, controlled trials25 indicated no subgroup analysis of total hip arthroplasty patients was performed statistically significant difference in the proportion of patients (WMD 0.08 days, 95% CI –0.12 to 0.29, I2 = 24%), which was possible experiencing adverse events (n = 2464) for the events measured using data from three studies34,35,37 (Table 4). (falls, pressure injuries, pulmonary embolism, deep vein throm- bosis, rapid response medical team call, transfer to intensive care, Effect on hospital length of stay in the subacute setting transfer to high dependency unit, and death) between those receiving and not receiving weekend allied health services (RD Randomised trials 0.00, 95% CI –0.01 to 0.01, I2 = 83%), as shown in Figure 6a. High Meta-analysis (n = 1437) of three randomised, controlled levels of heterogeneity in the study results were examined in post- hoc exploratory analysis by Haines et al,25 suggesting that there trials26,27,32 conducted in the subacute setting indicated that was a significant change in patients experiencing any adverse weekend allied health services reduced subacute hospital length of event between control conditions; however, when intervention stay (WMD 2.35 days, 95% CI 0.45 to 4.24, I2 = 0.0%), as shown in conditions were compared, there was no significant difference.25 Figure 3c. Non-randomised studies and subgroup analyses Non-randomised studies and subgroup analyses The results from randomised trials were concordant with those The results from the randomised trials in the subacute setting from two acute non-randomised, controlled trials,39,43 for which were somewhat concordant with non-randomised studies. A meta-analysis (n = 3348) showed no effect of weekend allied meta-analysis (n = 5012) of three subacute non-randomised, health on the number of adverse events (n = 135) for the events controlled trials40–42 showed a trend towards reduced hospital measured (falls, pressure injuries, pulmonary embolism, deep vein length of stay in favour of weekend allied health provision (WMD thrombosis, rapid response medical team calls, transfer to 0.49 days, 95% CI –0.87 to 1.85, I2 = 83%), as shown in Figure 3d. intensive care or high dependency unit, and deaths) between However, high levels of heterogeneity in study results were those receiving and not receiving weekend allied health services observed between studies. (RR 1.18, 95% CI 0.51 to 2.73, I2 = 78%), as shown in Figure 6b. However, high levels of heterogeneity in study results were Effect on hospital readmissions in the acute setting observed between studies. Randomised trials Effect on adverse events in the subacute setting Meta-analysis (n = 27 508) of cluster-level data from two acute, Randomised trials stepped-wedge, randomised, controlled trials25 showed no signifi- Meta-analysis (n = 1437) of three subacute randomised, cant difference in the proportion of patients who had an unplanned hospital re-admission within 28 days of hospital discharge controlled trials26,27,32 indicated no difference between week- between groups with available and unavailable weekend allied end and no weekend allied health for the number of adverse health services (RD 0.01, 95% CI 0.00 to 0.02, I2 = 43%), as shown in events (n = 303) for the adverse events measured (falls, skin Figure 5. tears, infections, re-admission to acute service, and death) between those receiving and not receiving weekend allied health services (RR 1.13, 95% CI 0.92 to 1.39, I2 = 0%), as shown in Figure 6c.
Table 4 Non-randomised, controlled trial meta-analyses, subgroup meta-analyses, and sensitivity analyses investigating the ef Outcome Total effect size (95% CI) I2D$9ITF]_2[ Sub-group analyses (9 Orthopaedic Orthopaedic total Orthopaedic to knee arthroplasty hip arthroplas Hospital length of stay, 0.24 0.51 0.49 0.08 WMD (days) (–0.17 to 0.66) (0.26 to 0.76) (0.11 to 0.87) (–0.12 to 0.29 Adverse events, RR 96% 84% 90% 24% 1.18 0.74 No studies No studies Patients discharged (0.51 to 2.73) (0.39 to 1.41) home, RR 78% N/Aa No studies No studies 1.19 1.22 Functional (1.03 to 1.38) (0.87 to 1.72) 0.39 0.31 independence, SMD 59% 80% (0.15 to 0.64) (0.12 to 0.51 0.19 0.19 (–0.12 to 0.50) (–0.12 to 0.50) 0% 0% 77% 77% N/A = not applicable, PT = physiotherapist, SMD = standardised mean difference, WMD = weighted mean difference. Positive values favour intervention. Negative values favour control. a I2 could not be calculated where data from only one study were available.
ffect of weekend allied health services on acute general medical and surgical hospital wards. 95% CI) 2_I9D$TF]I[ 2 Sensitivity analyses (95% CI) I2IFDT_03[]$ otal Neurological Studies Studies with Studies with Studies where sty with PT service identified selection identified outcome data were estimated only (no other bias removed bias removed due to inadequate professions) reporting removed 9) –0.50 0.66 0.22 0.18 –0.04 Research 1) (–0.74 to –0.26) (0.49 to 0.83) (–0.27 to 0.72) (–0.33 to 0.69) (–1.45 to 1.38) N/Aa 68% 96% 97% 49% No studies 1.18 1.18 1.75 1.75 (0.51 to 2.73) (0.51 to 2.73) (1.11 to 2.75) 1.04 78% 78% N/Aa (1.11 to 2.75) (0.86 to 1.25) 1.19 1.27 1.19 N/Aa (1.03 to 1.38) (1.03 to 1.57) (1.03 to 1.38) N/Aa 59% 66% 59% 1.17 0.34 0.09 –0.10 0.09 (0.19 to 0.49) (–0.33 to 0.52) (–0.35 to 0.15) (–0.33 to 0.52) (1.08 to 1.38) 71% N/Aa 71% N/Aa 0% No studies 151
Table 5 Randomised, controlled trial meta-analyses, subgroup meta-analyses, and sensitivity analyses investigating the effect o Outcome Total effect size Sub-group analyses (95% CI) I2 (95% CI) I2 Mixed Neurological Studies with i rehabilitation rehabilitation PT service only ge (no other professions) Hospital length of stay, 2.35 2.33 3.00 3.19 WMD (days) (0.45 to 4.24) (0.42 to 4.25) (–0.32 to 6.70) (–11.12 to 17.12) 0% 0% N/Aa 0% Adverse events, RR 1.13 1.18 0.77 1.01 (0.92 to 1.39) (0.95 to 1.46) (0.70 to 1.46) (0.38 to 1.53) 0% 0% N/Aa 0% Functional independence, 0.09 0.09 0.07 0.06 SMD (–0.01 to 0.19) (–0.02 to 0.20) (–0.13 to 0.24) (–0.22 to 0.37) 0% 0% N/Aa 0% Functional walking 3.36 5.07 0.12 5.14 speed, SMD (0.56 to 6.15) (–4.89 to 15.04) (–4.71 to 14.98) (–0.17 to 0.42) 100% 100% N/Aa 10 0% Functional stroke 0.03 –1.05 0.07 –1.03 outcomes, SMD (–0.20 to 0.26) (–4.95 to 2.86) (–5.04 to 2.97) (–0.26 to 0.40) 0% 43% N/Aa 45% Functional mobility, 0.06 0.01 0.27 0.16 SMD (–0.09 to 0.21) (–0.13 to 0.14) (–0.08 to 0.38) (–0.03 to 0.57) 18% 0% N/Aa 14% Quality of life, SMD 0.10 0.12 –0.03 0.08 (–0.01 to 0.20) (0.00 to 0.23) (–0.11 to 0.27) (–0.33 to 0.28) 0% 0% N/Aa 0% N/A = not applicable, PT = physiotherapist, SMD = standardised mean difference, WMD = weighted mean difference. Positive values favour intervention. Negative values favour control. a I2 could not be calculated where data from only one study were available.
of weekend allied health services on subacute rehabilitation wards. 152 Sarkies et al: Additional weekend allied health services in hospital Sensitivity analyses (95% CI) I2 Studies with Studies with Studies with Studies with Studies where identified bias for identified bias for identified bias identified bias data estimated random sequence blinding of outcome for incomplete outcome due to inadequate eneration removed assessment removed data removed for selective reporting removed reporting removed 2.03 2.03 2.33 2.35 (–0.20 to 4.25) (–0.20 to 4.25) (0.42 to 4.25) 3.20 (0.45 to 4.24) (–0.43 to 6.83) 0% 0% 0% 0% N/Aa 1.08 1.08 1.18 1.13 (0.75 to 1.45) (0.75 to 1.45) (0.95 to 1.46) 1.12 (0.92 to 1.40) (0.73 to 1.73) 30% 0% 0% 0% N/Aa 0.10 0.10 0.09 0.09 (–0.01 to 0.21) (–0.01 to 0.21) (–0.02 to 0.20) 0.04 (0.01 to 0.19) (–0.20 to 0.29) 0% 0% 0% 0% N/Aa 0.03 0.03 5.07 3.36 (–0.11 to 0.16) (–0.11 to 0.16) (–4.89 to 15.04) 10.17 (0.56 to 6.15) (9.26 to 11.08) 0% 0% 100% 100% N/Aa 0.03 0.03 –1.05 0.03 (–0.20 to 0.26) (–0.20 to 0.26) (–4.95 to 2.86) –4.83 (–0.20 to 0.26) (–11.98 to 2.31) 0% 0% 43% 0% N/Aa 0.10 0.10 0.01 0.06 (–0.15 to 0.36) (–0.15 to 0.36) (–0.13 to 0.14) 0.03 (–0.09 to 0.21) (–0.27 to 0.34) 59% 59% 0% 0% N/Aa 0.09 0.09 0.12 0.10 (–0.03 to 0.20) (–0.03 to 0.20) (–0.01 to 0.23) 0.15 (–0.01 to 0.20) (–0.10 to 0.39) 0% 0% 0% 0% N/Aa
Research 153 Table 6 Non-randomised, controlled trial meta-analyses and sensitivity analyses investigating the effect of weekend allied health services on subacute rehabilitation wards. Outcome Total effect size Sensitivity analyses (95% CI) I2 (95% CI) I2 Studies with Studies with identified Studies with identified Studies where data PT service only selection bias removed outcome bias removed were estimated due to (no other professions) inadequate reporting removed Hospital length 0.49 1.70 0.18 0.18 1.11 of stay, WMD (days) (–0.87 to 1.85) (–0.52 to 3.92) (–1.29 to 1.64) (–1.29 to 1.64) (0.22 to 2.00) Patients discharged 83% N/Aa 88% 88% 0% to aged care, RR 1.00 No studies 1.00 1.00 1.00 (0.75 to 1.34) (0.75 to 1.34) (0.75 to 1.34) (0.75 to 1.34) Functional 0% –0.21 0% 0% 0% independence, SMD 0.05 (–0.45 to 0.04) 0.15 0.15 –0.06 (–0.17 to 0.28) (–0.11 to 0.42) (–0.11 to 0.42) (–0.28 to 0.16) 89% N/Aa 93% 93% 69% N/A = not applicable, PT = physiotherapist, SMD = standardised mean difference, WMD = weighted mean difference. Positive values favour intervention. Negative values favour control. a I2 could not be calculated where data from only one study were available. Effect on discharge destination in the acute setting 0.31, 95% CI 0.12 to 0.51, I2 = 0%), total knee arthroplasty patients (SMD 0.39, 95% CI 0.15 to 0.64, I2 = 0%), or studies examining only Randomised trials physiotherapy (no other professions) was performed (SMD 0.34, Meta-analysis (n = 27 508) of cluster-level data from two acute, 95% CI 0.19 to 0.49, I2 = 0%), as shown in Table 5. stepped-wedge, randomised, controlled trials25 indicated no differ- Effect on functional independence in the subacute setting ence in the proportion of patients discharged to aged care between those receiving and not receiving weekend allied health services (RD Randomised trials 0.00, 95% CI –0.00 to 0.01; I2 = 31%), as shown in Figure 7a. The Functional Independence Measure was pooled and cate- Non-randomised studies and subgroup analyses gorised as a functional activity of daily living outcome measured in This result was somewhat discordant with those from three three subacute randomised, controlled trials.26,27,32 The Timed Up and Go test, and the Wolf Motor Function Test were transformed to acute non-randomised studies35,39,43 (n = 3588), where a meta- conform to the same effect direction. These were then pooled and analysis showed that patients receiving weekend allied health may categorised as functional mobility outcomes measured in three have been more likely to be discharged home to private residence subacute randomised, controlled trials.26,27,32 Walking speed and 10- from hospital (RR 1.19, 95% CI 1.03 to 1.38, I2 = 59%), as shown in Metre Walk Test were transformed to conform to the same effect Figure 7b. However, high levels of heterogeneity in study results direction. These were then pooled and categorised as functional were observed between studies. walking speed outcomes measured in three subacute randomised, controlled trials.26,27,32 The Motor Assessment Scale and the physical Effect on discharge destination in the subacute setting dimension of the Stroke Impact Scale were transformed to conform to the same effect direction. These were then pooled and categorised Randomised trials as stroke-specific functional outcomes in three subacute non- Meta-analysis was not possible for randomised trials in the randomised, controlled trials.26,27,32 Measurement time points were grouped prior to inpatient discharge. subacute setting. However, one subacute randomised, controlled trial reported no difference between intervention and control Meta-analysis (n = 1437) showed a trend towards improved conditions for the number of patients discharged to aged care.26 functional activities of daily living taken prior to hospital discharge in favour of weekend allied health service provision, although this Non-randomised studies and subgroup analyses was not statistically significant (SMD 0.09, 95% CI –0.01 to 0.19, This result was concordant with those from two subacute non- I2 = 0%), as shown in Figure 8b. No difference was identified between intervention and control conditions in meta-analyses of randomised studies41,42 (n = 4476), for which meta-analysis functional mobility (n = 335), walking speed (n = 438), and stroke- showed no effect of weekend allied health on the relative risk of specific outcome measures (n = 210). patients being discharged to aged care (RR 1.00, 95% CI 0.75 to 1.34, I2 = 0%), as shown in Figure 7c. Effect on functional independence in the acute setting Non-randomised studies and subgroup analyses The Functional Independence Measure was pooled and cate- Randomised trials Meta-analysis was not possible for this outcome in randomised gorised as a functional activity of daily living outcome measure in three subacute non-randomised, controlled trials.40–42 Results from trials in the acute setting. meta-analysis of randomised, controlled trials were somewhat concordant with meta-analysis of three subacute non-randomised, Non-randomised studies and subgroup analyses controlled trials (n = 4746), which showed no statistically significant Functional independence with activities of daily living were difference between weekend and no weekend allied health for functional activities of daily living outcome measures (SMD 0.05, 95% measured in three acute non-randomised, controlled trials.35,37,43 The CI –0.17 to 0.28, I2 = 89%), as shown in Figure 8c. However, high levels Barthel Index, ‘days to mobilising with two sticks’, and ‘days to of heterogeneity in study results were observed between studies. independent mobility’ were transformed to conform to the same effect direction. These were then pooled and categorised as functional Effect on quality of life in the acute setting activities of daily living outcomes. Meta-analysis of three acute non- randomised studies (n = 1201) showed no difference in functional Meta-analysis was not possible for this outcome in randomised independence between those who had weekend allied health services or non-randomised trials in the acute setting. No randomised trials available and those who did not (SMD 0.19, 95% CI –0.12 to 0.50, reported health-related quality of life in the acute setting, and one I2 = 77%), as shown in Figure 8a. Heterogeneity levels were reduced non-randomised trial demonstrated no difference in quality of life when subgroup analysis of only total hip arthroplasty patients (SMD
(Figure_3)TD$IG][154 Sarkies et al: Additional weekend allied health services in hospital (a) WMD (95% CI) (c) WMD (95% CI) Random Random Study Area Haines A25 –10 –5 0 5 10 Study –20 –10 0 10 20 Haines B25 Favours con Favours exp Favours con Favours exp Mixed Total Peiris27 Brusco26 (b) WMD (95% CI) Subtotal Random Area Stroke Study English32 Subtotal Orthopaedic Total Maidment34 Boxall35 (d) WMD (95% CI) Pua36 Random Pengas37 Study Haas43 Caurana40 Subtotal DiSotto-Monastero42 Hakkennes41 Stroke Kinoshita39 Total Subtotal Total –10 –5 0 5 10 –10 –5 0 5 10 Favours con Favours exp Favours con Favours exp Figure 3. Weighted mean difference (95% CI) in the effect of weekend allied health on hospital length of stay: (a) acute randomised, controlled trials, calculated from [(Figure_4)TD$IG]participant-level data; (b) acute non-randomised, controlled trials; (c) subacute randomised, controlled trials; (d) subacute non-randomised, controlled trials. Study RD (95% CI) at 4 days postoperatively between weekend and no-weekend Random allied health groups.43 Haines A25 Effect on quality of life in the subacute setting Haines B25 In three subacute randomised, controlled trials, the EuroQol Five Total Dimensions questionnaire and Assessment of Quality of Life were pooled and categorised as health-related quality of life outcome –0.04 –0.02 0 0.02 0.04 measures taken prior to discharge.26,27,32 Meta-analysis (n = 1423) Favours con Favours exp indicated a trend towards improved health-related quality of life in favour of weekend allied health service provision (SMD 0.10, 95% CI Figure 4. Risk difference (95% CI) in acute randomised, controlled trials of the effect –0.01 to 0.20, I2 = 0%), as shown in Figure 9. Although this result was of weekend allied health on proportion of patients whose hospital length of stay not statistically significant, it indicated possible improved efficiency, was longer then their expected length of stay. as improvements in quality of life may have been achieved during a shorter length of hospital stay. GIF$DT)5_eru[(ig] Study RD (95% CI) Effect on cost of hospital care in the acute setting Haines A25 Random Haines B25 Randomised trials Meta-analysis (n = 27 508) of individual participant level data from Total two acute, stepped-wedge, randomised, controlled trials25 indicated –0.03 0 0.03 no difference in patient cost to the healthcare system per hospital Favours con Favours exp admission between those with weekend allied health services available and unavailable (WMD AUD118, 95% CI –274.5 to 510.5, Figure 5. Risk difference (95% CI) in acute randomised, controlled trials of the effect I2 = 99%), as shown in Figure 10. Clinical costing data were captured of weekend allied health on hospital readmission. using routinely applied hospital data collection and resource allocation procedures, largely driven by hospital length of stay and procedures performed, which do not take into consideration cost
GIF$DT)6_er[(igu] Researc7)TD_er[u(F]GI$igh 155 (a) RD (95% CI) (a) RD (95% CI) Random Random Study Study Haines A25 Haines A25 Haines B25 Haines B25 Total Total –0.05 0 0.05 –0.0169 0 0.0169 Favours con Favours exp Favours con Favours exp (b) RR (95% CI) (b) RR (95% CI) Random Random Area Area Study Study Orthopaedic Orthopaedic Haas43 Boxall35 Subtotal Haas43 Subtotal Stroke Kinoshita39 Stroke Subtotal Kinoshita39 Total Subtotal Total 0.1 1 10 0.1 1 10 Favours con Favours exp Favours con Favours exp (c) (c) Area RR (95% CI) Area RR (95% CI) Study Random Study Random Mixed Mixed Peiris27 Brusco26 DiSotto-Monastero42 Subtotal Hakkennes41 Stroke English32 Subtotal Subtotal Total Total 0.1 1 10 0.1 1 10 Favours con Favours exp Favours con Favours exp Figure 6. Effect of weekend allied health on adverse events, calculated as risk Figure 7. Effect of weekend allied health on discharge destination, calculated as risk difference (95% CI) in: (a) acute randomised, controlled trials, calculated from difference (95% CI) in: (a) acute randomised, controlled trials, calculated from cluster-level data, and calculated as relative risk (95% CI); (b) acute non- cluster-level data, and calculated as relative risk (95% CI); (b) acute non- randomised, controlled trials; and (c) subacute randomised, controlled trials. randomised, controlled trials; and (c) subacute non-randomised, controlled trials. relative to patient diagnosis. High levels of heterogeneity in the study randomised, controlled trial reported economic outcomes at 3- results were explained in exploratory analysis by Haines et al,25 month, 6-month and 12-month follow-up.29,30 At 3-month follow-up indicating that there was a significant change in total cost favouring post hospital discharge, there was a mean cost saving of AUD1673 the original weekend allied health service delivery model, although favouring weekend allied health service provision. An incremental these outcomes did not account for differences in patient diagnosis cost utility ratio saving of AUD41 825 per quality of life year gained, categories between phases.25 and an incremental cost-effectiveness ratio found a saving of AUD16 003 in achieving a minimal clinically important difference Effect on cost of hospital care in the subacute setting in functional independence for the group receiving additional weekend allied health service provision. If willingness to pay per Meta-analysis was not possible for this outcome in randomised or quality of life year gained or minimal clinically important difference non-randomised trials in the subacute setting. However, one subacute in functional independence was AUD0, the probability of cost- effectiveness would be 96 and 95%, respectively. A sensitivity analysis removing the Saturday penalty rate salary loading of 50% did not alter the results of the primary analysis. At 6-month follow-up there was a mean cost saving of AUD6445 favouring weekend allied health service provision; however, this saving was no longer significant at 12-month follow-up. The incremental cost-effectiveness
1[(Fig]GI$DT)8_eru 56 FI$DT)9_erugGi([]Sarkies et al: Additional weekend allied health services in hospital (a) SMD (95% CI) Area SMD (95% CI) Random Study Random Study Boxall35 –1.0 –0.5 0 0.5 1.0 Mixed Pengas37 Favours con Favours exp Peiris27 Hass43 Brusco26 Subtotal Total Stroke English32 Subtotal Total (b) –4 –2 0 2 4 Area SMD (95% CI) Favours con Favours exp Study Random Figure 9. Standardised mean difference (95% CI) in subacute randomised, controlled Mixed trials of the effect of weekend allied health on health-related quality of life. Peiris27 Brusco26 $IGe[(FigD]urT)_10 Subtotal Study WMD (95% CI) Stroke Haines A25 Random English32 Haines B25 Total Subtotal –4000 –2000 0 2000 4000 Total Favours con (AUD) Favours exp –1.0 –0.5 0 0.5 1.0 Favours con Favours exp Figure 10. Weighted mean difference (95% CI) in the effect of weekend allied health on cost in acute randomised, controlled trials, calculated from participant-level data. (c) SMD (95% CI) Random nursing did not produce favourable results,46 and there is a paucity Study of research on medical staffing models in subacute rehabilitation Caruana40 settings. DiSotto-Monastero42 Hakkennes41 The evidence generated by this review may assist resource allocation decisions by healthcare policymakers and managers Total considering implementing out-of-hours and 7-day healthcare service models. The identified benefits include improved patient –1.0 –0.5 0 0.5 1.0 function and health-related quality of life during shorter hospital Favours con Favours exp length of stay, which indicate that increased rehabilitation efficiency is achievable. Further, the reduced hospital length of stay may also Figure 8. Standardised mean difference (95% CI) in the effect of weekend allied provide incentives for investment in additional weekend allied health on functional independence in: (a) acute non-randomised, controlled trials; health service provision because reductions in length of stay in (b) subacute randomised, controlled trials; and (c) subacute non-randomised, subacute rehabilitation hospitals can improve patient flow in acute controlled trials. services and contribute to improved cost-effectiveness.47]D_F[3TI$5 ratio found a saving of AUD41 825 (95% CI –2817 to 74 620) per quality In contrast, there was insufficient evidence to conclude that of life year gained for the weekend allied health group at 6 months. routinely using a weekend allied health model to provide additional services in the acute general medical and surgical ward setting would Discussion achieve intended benefits for all patients. It is possible that the difference in outcomes between acute and subacute ward types is due This meta-analysis supported previous syntheses showing that to variability in the focus of allied health activities in these settings. For allied health rehabilitation therapy services improve patient example, physiotherapy services for chronic obstructive pulmonary outcomes.12,15,45F43T[ID]$_ This is the first review to show that benefits disease in the acute hospital ward setting may focus on preventing can be achieved in the subacute rehabilitation setting by providing respiratory failure by providing non-invasive ventilation,48whereas in additional allied health services on weekends. The same benefits the subacute rehabilitation ward setting, these services may empha- have not been demonstrated by additional weekend nursing and sise therapies specifically aimed at improving functional indepen- medical staffing in the subacute rehabilitation setting. A study dence.49 Therefore, an evaluation of providing weekend allied health examining the effect of additional rehabilitation provided by services in either the acute or subacute ward setting should reflect these outcomes of interest. Another possible explanation for unclear results in the acute setting is that providing additional allied health services in isolation may not change outcomes, due to limited medical and nursing staffing. A recent study in the acute general medical setting incorporated a multi-disciplinary 7-day medical, nursing, and allied health staffing model, which reported improvements in hospital length of stay, reduced general medical bed occupancy days, and improved weekend discharges.50 However, comprehensive 7-day
Research 157 medical, nursing, and allied health service provision may not improve risk of bias for outcomes unable to be blinded such as hospital outcomes in all circumstances.51 Unclear results in the acute setting length of stay, where allied health personnel could delay or could also be explained by previously examined logistical difficulties expedite hospital discharge for either the intervention or control in implementing weekend allied health service models. This may be groups, thereby affecting the hospital length of stay outcome data. due to the employment of less experienced staff during weekends to The use of a stepped-wedge, randomised, controlled trial design in reduce cost;16 employing staff who work outside of the organisation the acute setting may have reduced this risk of bias, as service during normal business hours;16 reduced medical and nursing staffing demands would be prohibitive to interference across extended over the weekend;52–54 unavailability of services from community- control and intervention periods. Also, use of outcome measures, based organisations needed to facilitate discharge on weekends; and such as hospital length of stay and cost to the health service per patient expectation of rest on the weekend.55 Findings in the acute admission, was potentially problematic in some of the included general medical and surgical ward setting should not be extrapolated studies. It was not always clear whether the data for hospital to weekday, after-hours, on-call, or 7-days a week service provision length of stay and cost were relative or absolute because patient models, and only applies to relevant wards, as defined in this review. cohorts varied over time. Changes in these outcome variables could However, it should also be noted that only two randomised, controlled have resulted from changing cohorts, as well as improvements trials (one article) from a single country have been published from interventions. One way to control for this would be to evaluating acute weekend allied health services;25 considering consider outcomes relative to diagnoses using coding data, for weekend service variability, it is possible that other untested targeted example hospital length of stay relative to expected length of stay. models of weekend allied health service could provide benefit. Future studies could consider whether alternative models have the potential Providing additional allied health services to patients in to deliver cost-effective weekend allied health services. subacute rehabilitation wards on the weekend reduced hospital length of stay, and may be an economically efficient way to It is believed that this is the first systematic review examining the improve functional independence and health-related quality of effectiveness of providing additional allied health services to acute and life. However, for acute general medical and surgical hospital subacute hospital wards during weekends. It is also the first meta- wards, the impact of weekend allied health services is unclear. This analysis to include a stepped-wedge, randomised, controlled trial design. synthesis of evidence to date suggests that the benefits of routinely This design is valuable because it provides a statistically powerful, providing additional allied health services on subacute rehabilita- methodologically rigorous approach for evaluating weekend allied tion wards is clearer than in the acute general medical and surgical health services where provision is considered usual care. The ward setting. Future studies are required to further investigate the comprehensive search strategy resulted in 19 articles (20 studies) being effect of targeted weekend allied health services. However, the identified, with 38 732 participants, which is the largest and most goals of care for weekend allied health service provision may differ contemporaneous review to be conducted on this topic area. The pooling between acute and subacute hospital ward settings, and should of results according to study designs and setting may assist the therefore be reflected in the outcomes evaluated in future studies. understanding of how non-randomised trials can be compared with the results from randomised studies, and applied to clinical areas of interest. What was already known on this topic: The provision of However, a number of excluded studies exploring interventions that weekend allied health services in hospitals is variable, with include a weekend allied health service component must be differences in the amount and focus of service delivery. Pro- acknowledged.13,56–58 In most cases these were excluded, as the effect viding earlier, additional or high-intensity allied health services of weekend therapy was unable to be isolated from concurrent weekday is generally beneficial, but the specific effect of additional interventions within the trials (eg,increased weekday, after-hours,oron- weekend services is unclear. call services in addition to a weekend service). Future research should What this study adds: Providing additional weekend allied focus on implementing additional allied health through different models health services in subacute rehabilitation wards significantly such as increasing frequency, intensity, time, and type of services during reduced hospital length of stay, and may be a cost-effective normal departmental operating hours; 7-day service provision; way of improving function and quality of life. The effect in acute evaluating outreach models; on-call services; comprehensive care general medical and surgical wards remains unclear despite compared to discipline-specific models; and after-hours provision. substantial data. Despite the large number of trials that were identified overall, Footnotes: a Covidence, Melbourne, Australia. b StataCorp LP, the relatively small number of studies with randomised and non- College Station, USA. randomised designs in acute and subacute settings limited the generalisability of these findings. This was particularly evident for eAddenda: Appendices 1, 2, 3 can be found online at https://doi. the acute randomised, controlled trials, where there were only two org/10.1016/j.jphys.2018.05.004. studies. However, a large number of participants were included in the acute randomised trial meta-analysis (n = 27 508), and there Ethics approval: Not applicable. were relatively narrow confidence intervals for effect size Competing interests: Nil. estimates and the overall meta-analysis. Heterogeneity in the Sources of support: This systematic review was funded by a results of randomised studies in the acute setting for hospital partnership grant from the National Health and Medical Research length of stay and adverse event outcomes was examined via Council (NHMRC) Australia (APP1114210), and the Victorian exploratory analysis in the original study, suggesting that there Department of Health and Human Services. was a significant change in outcomes between control conditions, Acknowledgements: We wish to thank Monash University, and when intervention conditions were compared, there was no Monash Health, and the Victorian Department of Health and significant difference in outcomes.25 However, this post-hoc Human Services for providing the support for this project. analysis was hypothesis generating, suggesting the need for Provenance: Not invited. Peer reviewed. additional research in this setting. Data relating to the frequency, Correspondence: Mitchell Sarkies, Department of Physiotherapy, intensity, time and type of additional weekend allied health service Monash University, Melbourne, Australia. Email: mitchell.sarkies@ provision were unable to be extracted due to limitations in monash.edu reporting. Therefore, the results of this meta-analysis should be interpreted as applying to staffing models, rather than specific References allied health assessments and particular interventions. 1. Philip K. Allied health: untapped potential in the Australian health system. Aust Due to the nature of providing additional allied health services Health Rev. 2015;39:244–247. over weekends, blinding of participants and allied health person- nel was not possible. Blinding of outcome assessors was used to 2. Turnbull C, Grimmer-Somers K, Kumar S, May E, Law D, Ashworth E. Allied, reduce risk of bias in some studies; however, there was a potential scientific and complementary health professionals: a new model for Australian allied health. Aust Health Rev. 2009;33:27–37.
158 Sarkies et al: Additional weekend allied health services in hospital 3. Dorning H, Bardsley M. Focus on: Allied health professionals. Can we measure quality 32. English C, Bernhardt J, Crotty M, Esterman A, Segal L, Hillier S. Circuit class therapy of care?The Health Foundation and the Nuffield Trust; 2014. or seven-day week therapy for increasing rehabilitation intensity of therapy after stroke (CIRCIT): a randomized controlled trial. Int J Stroke. 2015;10:594–602. 4. Levit L, Patlak M. Ensuring quality cancer care through the oncology workforce: sustaining care in the 21st century: workshop summary. Washington DC: National 33. English C, Bernhardt J, Hillier S. Circuit class therapy and 7-day-week therapy Academies Press; 2009. increase physiotherapy time, but not patient activity: early results from the CIRCIT trial. Stroke. 2014;45:3002–3007. 5. Soh SE, Morris ME, Watts JJ, McGinley JL, Iansek R. Health-related quality of life in people with Parkinson’s Disease. Aust Health Rev. 2016;40:613–618. 34. Maidment ZL, Hordacre BG, Barr CJ. Effect of weekend physiotherapy provision on physiotherapy and hospital length of stay after total knee and total hip replace- 6. Morris ME, Murphy AT, Watts JJ, Jolley D, Campbell D, Soh SE, et al. The health ment. Aust Health Rev. 2014;38:265–270. profile of people living with Parkinson’s Disease managed in a comprehensive care setting. J Aging Sci. 2015;3:1–7. 35. Boxall A-M, Sayers A, Caplan GA. A cohort study of 7 day a week physiotherapy on an acute orthopaedic ward. J Orthop Nurs. 2004;8:96–102. 7. Campbell L, Bunston R, Colangelo S, Kim D, Nargi J, Hill K, et al. The provision of weekend physiotherapy services in tertiary-care hospitals in Canada. Physiother 36. Pua YH, Ong PH, Chong HC, Lo NN. Sunday physiotherapy reduces inpatient stay in Can. 2010;62:347–354. knee arthroplasty: a retrospective cohort study. Arch Phys Med Rehabil. 2011;92:880–885. 8. Shaw KD, Taylor NF, Brusco NK. Physiotherapy services provided outside of business hours in Australian hospitals: a national survey. Physiother Res Int. 37. Pengas IP, Khan WS, Bennett CA, Rankin KS. Impact of weekend physiotherapy 2013;18:115–123. service on the cost effectiveness of elective orthopaedic hip and knee arthroplasty. Open Orthop J. 2015;9:515–519. 9. Haas R, Sarkies M, Bowles K-A, O'Brien L, Haines T. Early commencement of physical therapy in the acute phase following elective lower limb arthroplasty produces 38. David C, Price N, Price T, Sheeran T, Mulherin D. Impact of weekend physiotherapy favorable outcomes: a systematic review and meta-analysis examining allied delivery on the throughput of rheumatology inpatients: Feasibility study. Physio- health service models. Osteoarthritis Cartilage. 2016;24:1667–1681. therapy. 2003;89:25–29. 10. Schweickert WD, Pohlman MC, Pohlman AS, Nigos C, Pawlik AJ, Esbrook CL, et al. 39. Kinoshita S, Momosaki R, Kakuda W, Okamoto T, Abo M. Association between Early physical and occupational therapy in mechanically ventilated, critically ill 7 days per week rehabilitation and functional recovery of patients with acute patients: a randomised controlled trial. Lancet. 2009;373(9678):1874–1882. stroke: a retrospective cohort study based on the Japan Rehabilitation Database. Arch Phys Med Rehabil. 2017;98:701–706. 11. Beder J. Evaluation research on social work interventions: a study on the impact of social worker staffing. Soc Work Health Care. 2008;47:1–13. 40. Caruana EL, Kuys SS, Clarke J, Bauer SG. A pragmatic implementation of a 6-day physiotherapy service in a mixed inpatient rehabilitation unit. Disabil Rehabil. 12. Peiris CL, Taylor NF, Shields N. Extra physical therapy reduces patient length of stay and 2016;1–6. improves functional outcomes and quality of life in people with acute or subacute conditions: a systematic review. Arch Phys Med Rehabil. 2012;92:1490–1500. 41. Hakkennes S, Lindner C, Reid J. Implementing an inpatient rehabilitation Saturday service is associated with improved patient outcomes and facilitates patient flow 13. Mills E, Hume V, Stiller K. Increased allied health services to general and acute across the health care continuum. Disabil Rehabil. 2015;37:721–727. medical units decreases length of stay: comparison with a historical cohort. Aust Health Rev. Published early online March 30, 2017: https://doi.org/10.1071/ 42. DiSotto-Monastero M, Chen X, Fisch S, Donaghy S, Gomez M. Efficacy of 7 days per AH16220 [Accessed May 6, 2018]. week inpatient admissions and rehabilitation therapy. Arch Phys Med Rehabil. 2012;93:2165–2169. 14. Cifu DX, Kreutzer JS, Kolakowsky-Hayner SA, Marwitz JH, Englander J. The rela- tionship between therapy intensity and rehabilitative outcomes after traumatic 43. Haas R, O’Brien L, Bowles KA, Haines T. Effectiveness of a pragmatic weekend brain injury: a multicenter analysis. Arch Phys Med Rehabil. 2003;84:1441–1448. physical therapy service on short-term outcomes following hip and knee arthro- plasty. Clin Rehabil. 2018. http://dx.doi.org/10.1177/0123456789123456. Pre-pub- 15. Scrivener K, Jones T, Schurr K, Graham PL, Dean CM. After-hours or weekend lished June 5. rehabilitation improves outcomes and increases physical activity but does not affect length of stay: a systematic review. J Physiother. 2015;61:61–67. 44. Haines TP, O’Brien L, Mitchell D, Bowles KA, Haas R, Markham D, et al. Study protocol for two randomized controlled trials examining the effectiveness and 16. Mitchell D, O’Brien L, Bardoel A, Haines T. Challenges, uncertainties and perceived safety of current weekend allied health services and a new stakeholder-driven benefits of providing weekend allied health services—a managers’ perspective. model for acute medical/surgical patients versus no weekend allied health ser- BMC Health Serv Res. 2017;17:118. vices. Trials. 2015;16:1. 17. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JPA, et al. The 45. Kwakkel G, van Peppen R, Wagenaar RC, Dauphinee SW, Richards C, Ashburn A, et PRISMA statement for reporting systematic reviews and meta-analyses of studies al. Effects of augmented exercise therapy time after stroke a meta-analysis. Stroke. that evaluate health care interventions: explanation and elaboration. Ann Intern 2004;35:2529–2539. Med. 2009;151:W65–W94. 46. Davidson I, Hillier VF, Waters K, Walton T, Booth J. A study to assess the effect of 18. Covidence. https://www.covidence.org/. [Accessed November 18, 2016]. nursing interventions at the weekend for people with stroke. Clin Rehabil. 19. Allied Health: Therapy and Science Disciplines Position Paper. Department of 2005;19:126–137. Health and Human Services, Victoria, Australia: 2016. https://www2.health.vic. 47. Poulos CJ, Eagar K. Determining appropriateness for rehabilitation or other sub- gov.au/about/publications/policiesandguidelines/allied-health-categories- acute care: is there a role for utilisation review? Aust N Z Health Policy. 2007;4:3. position-paper [Accessed May 6, 2018]. 20. Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD, et al. The Cochrane 48. Keenan SP, Sinuff T, Burns KE, Muscedere J, Kutsogiannis J, Mehta S, et al. Clinical Collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343: practice guidelines for the use of noninvasive positive-pressure ventilation and d5928. noninvasive continuous positive airway pressure in the acute care setting. Can Med 21. Wells G, Shea B, O'Connell D, J Peterson, V Welch, M Losos, et al. Newcastle-Ottawa Assoc J. 2011;183:E195–E214. quality assessment scale cohort studies. 2014. http://www.ohri.ca/programs/ clinical_epidemiology/oxford.asp [Accessed May 6, 2018]. 49. Puhan MA, Scharplatz M, Troosters T, Steurer J. Respiratory rehabilitation after 22. Cohen J. Statistical power analysis for the behavioral sciences. Hilsdale, NJ, USA: acute exacerbation of COPD may reduce risk for readmission and mortality–a Lawrence Earlbaum Associates; 1988. systematic review. Respir Res. 2005;6:54. 23. Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res 50. Gilfillan C, Newnham E, Nagappan R, Evans J, Compton J. A 7-day team-based Methodol. 2014;14:135. model of care in general medicine: implementation and outcomes at 12 months. 24. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta- Intern Med J. 2016;46:79–85. analyses. BMJ. 2003;327(7414):557–560. 25. Haines T, Bowles K, Mitchell D, Haas R, Markham D, Plumb S, et al. Impact of 51. Gan HW, Wong DJN, Dean BJF, Hall AS. Do expanded seven-day NHS services disinvestment from weekend allied health services across acute medical and improve clinical outcomes? Analysis of comparative institutional performance surgical wards: 2 stepped-wedge cluster randomised controlled trials. PLoS from the NHS Services, Seven Days a Week project 2013-2016. BMC Health Serv Med. 2017;14:e1002412. Res. 2017;17:552. 26. Brusco NK, Shields N, Taylor NF, Paratz J. A Saturday physiotherapy service may decrease length of stay in patients undergoing rehabilitation in hospital: a ran- 52. Chow KM, Szeto CC. Impact of enforcing the Labour Ordinance, with 1-in-7-day off domised controlled trial. Taust J Physiother. 2007;53:75–81. for hospital doctors, on weekend hospital discharge rate. J Public Health. 27. Peiris CL, Shields N, Brusco NK, Watts JJ, Taylor NF. Additional Saturday rehabilita- 2005;27:189–191. tion improves functional independence and quality of life and reduces length of stay: a randomized controlled trial. BMC Med. 2013;11:1. 53. Varnava AM, Sedgwick JEC, Deaner A, Ranjadayalan K, Timmis AD. Restricted 28. Brusco NK, Watts JJ, Shields N, Chan SP, Taylor NF. Does additional acute phase weekend service inappropriately delays discharge after acute myocardial infarc- inpatient rehabilitation help people return to work? A subgroup analysis from a tion. Heart. 2002;87:216–219. randomized controlled trial. Clin Rehabil. 2014;28:754–761. 29. Brusco NK, Watts JJ, Shields N, Taylor NF. Are weekend inpatient rehabilitation 54. Schilling PL, Campbell Jr DA, Englesbe MJ, Davis MM. A comparison of in-hospital services value for money? An economic evaluation alongside a randomized mortality risk conferred by high hospital occupancy, differences in nurse controlled trial with a 30 day follow up. BMC Med. 2014;12:89. staffing levels, weekend admission, and seasonal influenza. Med Care. 2010;48: 30. Brusco NK, Watts JJ, Shields N, Taylor NF. Is cost effectiveness sustained after 224–232. weekend inpatient rehabilitation? 12 month follow up from a randomized con- trolled trial. BMC Health Serv Res. 2015;15:1. 55. Peiris CL, Taylor NF, Shields N. Patients value patient-therapist interactions more 31. Peiris CL, Taylor NF, Shields N. Additional Saturday allied health services increase than the amount or content of therapy during inpatient rehabilitation: a qualitative habitual physical activity among patients receiving inpatient rehabilitation for study. J Physiother. 2012;58:261–268. lower limb orthopedic conditions: a randomized controlled trial. Arch Phys Med Rehabil. 2012;93:1365–1370. 56. Babu AS, Noone MS, Haneef M, Samuel P. The effects of ‘on-call/out of hours' physical therapy in acute exacerbations of chronic obstructive pulmonary disease: a randomized controlled trial. Clin Rehabil. 2010;24:802–809. 57. Said CM, Morris ME, Woodward M, Churilov L, Bernhardt J. Enhancing physical activity in older adults receiving hospital based rehabilitation: a phase II feasibility study. BMC Geriatr. 2012;12:26. 58. Duncan C, Hudson M, Heck C. The impact of increased weekend physiotherapy service provision in critical care: a mixed methods study. Physiother Theory Pract. 2015;31:547–555.
Journal of Physiotherapy 64 (2018) 199 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Appraisal of Clinical Practice Guideline: American Academy of Orthopaedic Surgeons Clinical Practice Guideline on the Management of Osteoarthritis of the Hip Date of latest update: March 2017. Date of next update: Not stated. Description: This guideline is published in an 854-page document and Patient group: Adult patients (age 18 years) with osteoarthritis of provides evidence for the management of osteoarthritis of the hip in the hip. Intended audience: This guideline is intended to be used by adult patients. The full guideline includes recommendations, detailed orthopaedic surgeons and physicians managing adult patients with evidence tables, and discussion of implementing the guideline recom- osteoarthritis of the hip. Additional versions: Not stated. Additional mendations. The guideline summarises the evidence for the use of risk documents include appendices and a systematic review. Expert assessment tools and potential risk factors in patients with symptom- working group: The current guideline was developed by the atic osteoarthritis of the hip undergoing total hip arthroplasty, such as American Academy of Orthopaedic Surgeons (AAOS) Management of obesity, patient age, mental health disorders and tobacco use. Evidence Osteoarthritis of the Hip Guideline Multidisciplinary Clinician for management of hip osteoarthritis is outlined, beginning with non- Guideline Development Group. steroidal anti-inflammatory drugs, glucosamine sulfate and intraartic- ular injectables. Evidence is also discussed for various anaesthetic Funded by: The Clinical Practice Guideline was funded exclusively by types, tranexamic acid and surgical approaches. Three of the recom- the American Academy of Orthopaedic Surgeons, who received no mendations pertain specifically to physiotherapy. One of them funding from outside commercial sources to support the development addresses physiotherapy as a conservative treatment, while the others of this document. Consultation with: Members of the AAOS Board of cover physiotherapy as a pre-operative and postoperative intervention. Directors, members of the Council on Research and Quality, members of the Board of Councillors, and members of the Board of Specialty Provenance: Invited. Not peer reviewed. Societies. Approved by: The American Academy of Orthopaedic Surgeons (AAOS) Location: The guideline and additional documents Ze Lua and Joy C MacDermidb are available at: aSt Joseph’s Health Care https://www.guideline.gov/summaries/summary/50737/american- bWestern University, Canada academy-of-orthopaedicsurgeons-clinical-practice-guideline-on- management-of-osteoarthritis-of-the-hip https://doi.org/10.1016/j.jphys.2018.02.016 Appraisal of Clinical Practice Guideline: 2018 Guidelines for the Early Management of Patients with Acute Ischemic Stroke Date of latest update: January 2018. Date of next update: Not initiated within 2 weeks of an acute arterial ischaemic stroke. These indicated. Patient group: Adult patients with acute arterial guidelines cover over 200 statements including the categories of ischaemic stroke. Intended audience: The guidelines are intended prehospital care, emergency evaluation and treatment, general sup- for prehospital care providers, physicians, allied health professionals portive care, in-hospital management (general care and treatment of and hospital administrators. Additional versions: These guidelines acute complications) and secondary prevention strategies, which are supersede the 2013 ‘Guidelines for the Early Management of Patients often commenced during the first 2 weeks of hospitalisation. The with Acute Ischemic Stroke’. Expert working group: Members of the recommendations regarding rehabilitation have the most direct rel- writing group were appointed by the American Heart Association evance to the physiotherapy profession. The guidelines are based on Stroke Council’s Scientific Statements Oversight Committee, repre- two systematic reviews that were published in conjunction with the senting various areas of medical expertise. Funded by: No explicit guidelines. The structure allows for ease of reading, with each rec- statement about funding was provided. Consultation with: ommendation being graded on the strength and quality of supportive Endorsed by the American Association of Neurological Surgeons evidence using the American College of Cardiology/American Heart and Congress of Neurological Surgeons, as well as the Society for Association 2015 Class of Recommendations and Level of Evidence. Academic Emergency Medicine. Approved by: The American Heart All new or revised recommendations are accompanied by a short Association Science Advisory and Coordinating Committee and the description and supplemental tables summarising the key studies American Heart Association Executive Committee. Location: The supporting these recommendations. guidelines and additional documents are available at: http://stroke. ahajournals.org/content/49/3/e46 Provenance: Invited. Not peer reviewed. Natasha Pocovi Description: These guidelines are published as a 66-page compre- hensive set of recommendations concerning the process of care for Macquarie University, Australia adult patients with acute arterial ischaemic stroke. The guidelines have been restricted to adult patients and secondary prevention https://doi.org/10.1016/j.jphys.2018.04.001 1836-9553/© 2018 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 64 (2018) 199 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Appraisal of Clinical Practice Guideline: American Academy of Orthopaedic Surgeons Clinical Practice Guideline on the Management of Osteoarthritis of the Hip Date of latest update: March 2017. Date of next update: Not stated. Description: This guideline is published in an 854-page document and Patient group: Adult patients (age 18 years) with osteoarthritis of provides evidence for the management of osteoarthritis of the hip in the hip. Intended audience: This guideline is intended to be used by adult patients. The full guideline includes recommendations, detailed orthopaedic surgeons and physicians managing adult patients with evidence tables, and discussion of implementing the guideline recom- osteoarthritis of the hip. Additional versions: Not stated. Additional mendations. The guideline summarises the evidence for the use of risk documents include appendices and a systematic review. Expert assessment tools and potential risk factors in patients with symptom- working group: The current guideline was developed by the atic osteoarthritis of the hip undergoing total hip arthroplasty, such as American Academy of Orthopaedic Surgeons (AAOS) Management of obesity, patient age, mental health disorders and tobacco use. Evidence Osteoarthritis of the Hip Guideline Multidisciplinary Clinician for management of hip osteoarthritis is outlined, beginning with non- Guideline Development Group. steroidal anti-inflammatory drugs, glucosamine sulfate and intraartic- ular injectables. Evidence is also discussed for various anaesthetic Funded by: The Clinical Practice Guideline was funded exclusively by types, tranexamic acid and surgical approaches. Three of the recom- the American Academy of Orthopaedic Surgeons, who received no mendations pertain specifically to physiotherapy. One of them funding from outside commercial sources to support the development addresses physiotherapy as a conservative treatment, while the others of this document. Consultation with: Members of the AAOS Board of cover physiotherapy as a pre-operative and postoperative intervention. Directors, members of the Council on Research and Quality, members of the Board of Councillors, and members of the Board of Specialty Provenance: Invited. Not peer reviewed. Societies. Approved by: The American Academy of Orthopaedic Surgeons (AAOS) Location: The guideline and additional documents Ze Lua and Joy C MacDermidb are available at: aSt Joseph’s Health Care https://www.guideline.gov/summaries/summary/50737/american- bWestern University, Canada academy-of-orthopaedicsurgeons-clinical-practice-guideline-on- management-of-osteoarthritis-of-the-hip https://doi.org/10.1016/j.jphys.2018.02.016 Appraisal of Clinical Practice Guideline: 2018 Guidelines for the Early Management of Patients with Acute Ischemic Stroke Date of latest update: January 2018. Date of next update: Not initiated within 2 weeks of an acute arterial ischaemic stroke. These indicated. Patient group: Adult patients with acute arterial guidelines cover over 200 statements including the categories of ischaemic stroke. Intended audience: The guidelines are intended prehospital care, emergency evaluation and treatment, general sup- for prehospital care providers, physicians, allied health professionals portive care, in-hospital management (general care and treatment of and hospital administrators. Additional versions: These guidelines acute complications) and secondary prevention strategies, which are supersede the 2013 ‘Guidelines for the Early Management of Patients often commenced during the first 2 weeks of hospitalisation. The with Acute Ischemic Stroke’. Expert working group: Members of the recommendations regarding rehabilitation have the most direct rel- writing group were appointed by the American Heart Association evance to the physiotherapy profession. The guidelines are based on Stroke Council’s Scientific Statements Oversight Committee, repre- two systematic reviews that were published in conjunction with the senting various areas of medical expertise. Funded by: No explicit guidelines. The structure allows for ease of reading, with each rec- statement about funding was provided. Consultation with: ommendation being graded on the strength and quality of supportive Endorsed by the American Association of Neurological Surgeons evidence using the American College of Cardiology/American Heart and Congress of Neurological Surgeons, as well as the Society for Association 2015 Class of Recommendations and Level of Evidence. Academic Emergency Medicine. Approved by: The American Heart All new or revised recommendations are accompanied by a short Association Science Advisory and Coordinating Committee and the description and supplemental tables summarising the key studies American Heart Association Executive Committee. Location: The supporting these recommendations. guidelines and additional documents are available at: http://stroke. ahajournals.org/content/49/3/e46 Provenance: Invited. Not peer reviewed. Natasha Pocovi Description: These guidelines are published as a 66-page compre- hensive set of recommendations concerning the process of care for Macquarie University, Australia adult patients with acute arterial ischaemic stroke. The guidelines have been restricted to adult patients and secondary prevention https://doi.org/10.1016/j.jphys.2018.04.001 1836-9553/© 2018 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 64 (2018) 140 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Call for applications for membership of the Editorial Board The Editorial Board currently consists of 12 members: eight local and four international. Applications are invited to fill the following Editorial Board vacancies beginning in 2019: four local. All incumbents are entitled to re-apply in the current round. Editorial Board members are given portfolios with substantial responsibilities. This might involve, for example, soliciting submissions and editing contributions for one of the journal’s ‘Appraisal’ sections. Potential applicants who are not prepared to take on portfolio responsibilities should not apply. The initial term of office commences on 1 January 2019 and expires on 31 December 2021. Editorial Board members are entitled to renominate for a further two successive terms. Knowledge and skills required: 1. broad understanding of research methods 2. extensive experience in publication of research 3. excellent written communication skills 4. good working knowledge of the physiotherapy profession and an interest in its future. To be eligible to apply, Australian applicants must: 5. hold a PhD 6. be a physiotherapist registered in Australia 7. be a financial member of the Australian Physiotherapy Association (APA). Responsibilities: contribute to the establishment of policies that guide the publication of the journal participate in the activities of the Editorial Board as a voting member manage one of the journal portfolios attend regular Editorial Board teleconferences and a two-day face-to-face meeting annually meet and liaise with other members of the Editorial Board and the journal Editor as required undertake specific tasks from time to time to promote the standing of the journal. To be considered, physiotherapists applying for positions must submit: 1. a cover letter addressing the numbered criteria, above 2. a brief CV, which includes a clear explanation of the impact of any career interruption(s) over the last 5 years. Applicants will be assessed against the knowledge and skills listed above and potentially against other criteria. Applications close 5.00 pm AEST, Friday 28 September 2018 and should be directed to Marko Stechiwskyj at [email protected] https://doi.org/10.1016/j.jphys.2018.05.007 1836-9553/
Journal of Physiotherapy 64 (2018) 141 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Call for applications for the role of Scientific Editor The role of Scientific Editor requires a high level of expertise and entails considerable responsibility. Applications are invited to fill the Scientific Editor vacancy beginning in 2019. The incumbent is entitled to re-apply. The initial term of office commences on 1 January 2019 and expires on 31 December 2021. The Scientific Editor is entitled to nominate for up to four successive terms. An honorarium is provided. On completion of the term of office, the Scientific Editor is required to serve an additional year (without an honorarium) as a member of the Editorial Board to assist the smooth transition of the role. Knowledge and skills required: 1. broad understanding of research methods 2. extensive experience in publication of research 3. excellent written communication skills 4. good working knowledge of the physiotherapy profession and an interest in its future. To be eligible to apply, Australian applicants must: 5. hold a PhD 6. be a physiotherapist registered in Australia 7. be a financial member of the Australian Physiotherapy Association (APA). Responsibilities manage the peer-review process identify potential reviewers make decisions about the acceptability of manuscripts edit manuscript content closely review copyedited proofs for accuracy respond to enquiries about the suitability of a manuscript for submission deal with correspondence regarding published papers write or solicit editorials solicit expert topical reviews liaise with the Editorial Board and International Advisory Board attend regular Editorial Board teleconferences and a two-day face-to-face meeting annually maintain the Journal’s presence on social media. To be considered, physiotherapists applying for positions must submit: 1. a cover letter addressing the numbered criteria, above 2. a brief CV, which includes a clear explanation of the impact of any career interruption(s) over the last 5 years. Applicants will be assessed against the knowledge and skills listed above and potentially against other criteria. Applications close 5.00 pm AEST, Friday 28 September 2018 and should be directed to Marko Stechiwskyj at [email protected] https://doi.org/10.1016/j.jphys.2018.05.008 1836-9553/
Journal of Physiotherapy 64 (2018) 198 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Clinimetrics: Hospital Anxiety and Depression Scale Summary The Hospital Anxiety and Depression Scale (HADS) was is required to administer the HADS. The HADS is available in 115 lan- developed to assess depression and anxiety in patients with guages, including English.2 illness and in the general population.1 It is a self-rating patient- reported outcome measure (PROM) for both hospital and Validity, reliability and responsiveness: Several studies have con- community settings.2 Fourteen items are equally divided in two firmed the two-factor structure,6,7 but others have implicated some subscales: anxiety (HADS-A) and depression (HADS-D). The HADS- caution in the use of the two subscales as independent measures.7,8 A includes items such as tension, worry, fear, panic, difficulties in relaxing, and restlessness. The HADS-D includes items predomi- The internal consistency for the subscales is high (Cronbach’s alphas nantly measuring anhedonia (not experiencing joy). Respondents ranging from 0.68 to 0.93 for the HADS-A and from 0.67 to 0.90 for the indicate how they currently feel. Responses are rated on a 4-point HADS-D).3,6 Studies found high test-retest correlations (r > 0.80).9 Com- Likert scale and range from 0 to 3, with higher scores indicating pared to other commonly used (but often longer) depression and anxiety higher severity. Anchor points for the Likert items vary depending on the item (eg, ‘I can sit still and feel relaxed’ scores as 0 and ‘not at measures (eg, the Beck Depression Inventory, the General Health Ques- all’ scores as 3).1 The ratings of the 14 items are summed to yield a total score (0 to 42), or for each subscale separately (0 to 21), with tionnaire), validity is adequate, with correlations ranging between 0.60 special attention to reversed items. The HADS is used in many (good) and 0.80 (very good).3,4,6 The questionnaire’s ability to discrimi- different populations, including medical populations and healthy nate is moderate to high, with sensitivities and specificities 0.80, which able-bodied individuals (aged 16 to 65 years).3,4 is similar to other self-rating screening tools.6,10 Accessibility and feasibility: The HADS is copyrighted and available The following cut-offs are recommended: 0 to 7 for normal or from: [email protected]. The HADS is an easy to use and short questionnaire,3 with an accompanying test manual.5 It no anxiety/depression, 8 to 10 for mild, 11 to 14 for moderate, and takes around 2 to 5 minutes to fill in the questionnaire and no training 12 to 21 for severe anxiety/depression.5 For specific populations, alternate cut-offs have been proposed. There is some evidence that the HADS is sensitive to change and responsive to interventions.8 Normative UK data included a higher median anxiety score in women (6, IQR 4 to 9) than in men (5, IQR 2 to 8). The median depression score for both women and men was 3 (IQR 1 to 6).11 Commentary Emmylou Beekmana and Arianne Verhagenb aResearch Centre for Autonomy and Participation of Persons with a The HADS is a questionnaire that is widely used to quickly and Chronic Illness, Zuyd University of Applied Sciences, The Netherlands easily detect possible psychological distress in patients. Assessing anxiety and depression are essential for clinicians who treat bSchool of Physiotherapy, Graduate School of Health, University of patients with long-term conditions. Also in physiotherapy practice, Technology Sydney, Australia existing anxiety or depression in patients might influence treatment and its effectiveness, and thereby influence treatment References outcomes. 1. Zigmond AS, et al. Acta Psychiatra Scand. 1983;67(6):361–370. The HADS appears to be a valid and reliable instrument for use 2. GL Assessment. https://www.gl-assessment.co.uk/products/hospital-anxiety-and- in clinical practice and research. Nevertheless, the HADS does not adequately detect the presence of specific anxiety and depression depression-scale-hads. disorders and is not a diagnostic tool, but rather helps identify 3. Smarr KL, et al. Arthritis Care Res. 2011;63(Suppl 11):S454–S466. general symptoms.4 4. Julian LJ. Arthritis Care Res. 2011;63(Suppl 11):S467–S472. 5. Snaith RP, et al. The Hospital Anxiety and Depression Scale manual. Windsor, UK: Somatic items relating to emotional and physical disorders are excluded from the HADS. Therefore, the HADS may be particularly Nfer-Nelson; 1994. useful when studying the cognitive processes associated with 6. Bjelland I, et al. J Psychosom Res. 2002;52(2):69–77. depressive symptoms and anxiety, since it is free of physical 7. Iani L, et al. Health Qual Life Outcomes. 2014;12:84. symptoms such as insomnia and weight loss.3 8. Cosco TD, et al. J Psychosom Res. 2012;72(3):180–184. 9. Roberts SB, et al. Br J Health Psychol. 2001;6(Part 4):373–383. Provenance: Invited. Not peer reviewed. 10. Brennan C, et al. J Psychosom Res. 2010;69(4):371–378. 11. Breeman S, et al. Qual Life Res. 2015;24(2):391–398. https://doi.org/10.1016/j.jphys.2018.04.003 1836-9553/© 2018 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 64 (2018) 197 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Clinimetrics: Modified Canadian Aerobic Fitness Test Test description: The modified Canadian Aerobic Fitness Test diastolic blood pressures return below 144 mmHg and 94 mmHg, (mCAFT) is an eight-stage stepping test that was developed respectively.1 to estimate a client’s maximum oxygen uptake (VO2-max Equipment: The following equipment is required: heart mlÁkgÀ1ÁminÀ1) during maximum intensity exercise.1 Based on rate monitor, a double 20.3-cm step stool, timer/stopwatch, stethoscope, sphygmomanometer, mCAFT music, music player, a client’s age and gender, an initial 3-minute stepping stage (1, 2, 3, and data collection worksheet. or 4) is determined. Furthermore, the client’s ceiling post-exercise heart rate (ie, 85% of maximum heart rate) is calculated using Calculations: Estimated VO2-max (mlÁkgÀ1ÁminÀ1) is: the formula: 17.2 + (1.29Â oxygen cost at the final stage)À (0.09 Â weight in kg) 85% HRmax = 0.85 (220 À age).1 À (0.18 Â age in years).1 Once the client’s initial stepping stage and ceiling post-exercise Clinical measurement properties: The mCAFT demonstrated heart rate are obtained, the mCAFT can begin with the client relative and absolute test-retest reliability (ICC = 0.98 and stepping on a double 20.3-cm step stool at a predetermined SEM = 4.08) in a sample of able-bodied healthy participants, cadence set by the mCAFT music (STEP-STEP-UP, STEP-STEP- respectively.2 The mCAFT has shown to be highly correlated with DOWN).1 Multiple 3-minute stepping stages are carried out to maximal treadmill testing in estimation of VO2-max levels, attain a heart rate measure that equals or exceeds the calculated demonstrating criterion validity of r = 0.88.3 The sensitivity and ceiling post-exercise heart rate. Once the ceiling post-exercise specificity of the mCAFT in categorising an individual whose heart rate is attained, the mCAFT is complete.1 Upon completion, measured maximal treadmill VO2-max was excellent were 81 and the client is required to walk at a low intensity for 3 minutes, and 72%, respectively.4 In categorising an individual whose measured then sit rested for another 3 minutes. This is to ensure that the maximal treadmill VO2-max was poor, sensitivity and specificity of client’s heart rate returns below 99 beats/minute, and systolic and mCAFT were 100 and 50%, respectively.4 Commentary Goris Nazari Western University, Canada The mCAFT is a field-based sub-maximal fitness test that follows standardised procedures to provide an estimate of VO2-max levels. References Furthermore, it is quick to administer and requires minimal equipment. The mCAFT testing provides an easy and inexpensive 1. The Canadian physical activity, fitness and lifestyle appraisal. 2nd ed. Ottawa: Canadian alternative to maximal testing to quantify a change in an individual’s Society for Exercise Physiology; 1998. cardiorespiratory fitness levels, and has wide application in research involving able-bodied healthy individuals. However, the reliability, 2. Weller & Corey. Med Sci Sports Exerc. 1998; 30:1530–1536. validity and responsiveness properties of the mCAFT in clinical 3. Weller. et al. Can J Appl Physiol. 1994;20:211–221. populations are lacking. Measurement studies aiming to establish 4. Weller. et al. Can J Appl Physiol. 1993;18(2):175–188. such properties could be of great use to clinicians in order to facilitate mCAFT testing in clinical settings. Provenance: Invited. Not peer reviewed. https://doi.org/10.1016/j.jphys.2018.02.017 1836-9553/© 2018 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 64 (2018) 200–201 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Correspondence: Time-based versus repetition-based sit-to-stand measures: choice of metrics matters In both healthy and clinical older populations,1–4 the sit-to-stand Research implications (STS) test is a common and popular test of functional performance. STS performance is quantified by two metrics: time-based STS metric Given that test-retest errors in repetition-based STS metric are is the time taken to complete a specified number of STS repetitions; reportedly homoscedastic,2 the nonlinear association suggests that repetition-based STS metric is the number of STS repetitions the corresponding test-retest errors in time-based metric are completed in a specified period of time. Physiologically, when the heteroscedastic. The intuition here is that if the repetition-based specified number of STS repetitions is low (< 10 repetitions) and the STS measures fluctuated at a constant level, the corresponding specified period of time is short (<30 seconds), both time-based and fluctuations on the time scale would increase at decreasing values on repetition-based STS metrics assess lower extremity muscle the repetition-based metric. To illustrate this point, we generated a strength and give relatively the same information.3 Mathematically, simulated dataset for both test and retest repetition-based STS data time-based and repetition-based STS metrics are equivalent: less which, by design, showed no systematic bias and heteroscedasticity. time taken to perform one STS repetition equates to more repetitions However, heteroscedasticity appeared when the repetition-based completed in a specified period of time. However, the association is metric was converted into time-based metric (Figure 2). Reviewing not linear because one metric is the inverse of the other: the literature, one study reported heteroscedasticity in the time-based STS metric but postulated only biological reasons for metric ¼ SpecifRieedpertietpioentitimonetrnicumber: the findings.5 Among studies that did not rigorously test for data T ime heteroscedasticity, three studies reported greater reliability in repetition-based than in time-based STS metric.6–8 Another study This nonlinear association (graphically shown in Figure 1) has reported improved absolute reliability for the time-based metric research and clinical implications that are not well reported or when participants performed the STS test more rapidly.9 If data studied. These are described below. Figure 1. Time-based sit-to-stand (STS) metric is a nonlinear function of repetition- Figure 2. Bland–Altman plots showing the 95% limits of agreement for the based metric. In 30 seconds, increasing the number of STS repetitions performed from simulated homoscedastic sit-to-stand (STS) repetition-based data (top panel) and 7 to 10 improves the corresponding five times-STS performance by 6.4 seconds (from the corresponding heteroscedastic STS time-based data (bottom panel). This figure 21.4 to 15 seconds). However, increasing the number of STS repetitions from 10 to was generated from https://sgh-physio.shinyapps.io/STSmetrics/ 13 only improves the corresponding five times-STS performance by 3.5 seconds. This figure was generated from https://sgh-physio.shinyapps.io/STSmetrics/ https://doi.org/10.1016/j.jphys.2018.05.010 1836-9553/© 2018 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/).
Appraisal Correspondence 201 heteroscedasticity in the time-based metric were present but not repetition metric into equivalent values on the time metric. It is accounted for, (spuriously) large absolute errors could arise. hoped that this web application and letter will stimulate ‘nonlinear thinking’ in readers and help them to make better decisions. Clinical implications Yong-Hao Puaa, Julian Thumboob,c and Ross Alan Clarkd At the individual level, to detect important improvement or aDepartment of Physiotherapy, Singapore General Hospital, Singapore deterioration in STS performance, if the minimal clinically bDepartment of Rheumatology and Immunology, Singapore General important difference on the repetition-based metric is a constant,2 the corresponding change scores on the time-based metric are Hospital, Singapore non-constant and asymmetrical (Figure 1). With this asymmetry, cProgram in Health Services and Systems Research, Duke-NUS Medical patients may appear to be getting decreasing returns when they improve the time-based metric but constant returns when they School, Singapore improve the repetition-based metric. Thus, it is logical to speculate dResearch Health Institute, University of the Sunshine Coast, Sunshine that patients may be more motivated to improve their STS performance on a repetition-based metric. To our knowledge, no Coast, Australia clinical studies have examined this intriguing question; however, emerging research in business management10,11 has shown that References the choice of different – but equivalent – performance metrics can influence employee motivation. 1. Jordre B, et al. J Geriatr Phys Ther. 2013;36:47–50. 2. Bennell K, et al. Arthritis Care Res (Hoboken). 2011;63(Suppl 11):S350–S370. In conclusion, while we have focused on the STS test, the 3. Vaidya T, et al. Respir Med. 2017;128:70–77. implications that we have described may apply to other time-based 4. Paul SS, et al. Five-repetition sit-to-stand. J Physiother. 2014;60:168. and rate-based physical performance measures. A web application, 5. Schwenk M, et al. Physiol Meas. 2012;33:1931–1946. available at https://sgh-physio.shinyapps.io/STSmetrics/, allows 6. Segura-Orti E, et al. Phys Ther. 2011;91:1244–1252. readers to: simulate test-retest STS data; visualise the nonlinear 7. Mikkelsen LR, et al. Clin Rehabil. 2015;29:165–174. associations between time-based and repetition-based STS metrics; 8. Petersen C, et al. J Geriatr Phys Ther. 2017;40:223–226. and convert minimal clinically important difference values on the 9. Regterschot GR, et al. Gait Posture. 2014;40:220–224. 10. De Langhe B, et al. Harv Bus Rev. 2017;(May–June):11. 11. Stangl T, et al. Manuf Serv Oper Manag. 2017;19:472–488.
Journal of Physiotherapy 64 (2018) 194 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Critically appraised paper: Preoperative physiotherapy education halved postoperative pulmonary complications in patients after upper abdominal surgery Synopsis Summary of: Boden I, Skinner EH, Browning L, Reeve J, Anderson L, pulmonary complications, individualised risk assessment, education about Hill C, et al. Preoperative physiotherapy for the prevention of respiratory the impact of surgery, instructions on breathing exercises, and memory cues. complications after upper abdominal surgery: pragmatic, double blinded, Outcome measures: The primary outcome was presence of postoperative multicentre randomised controlled trial. BMJ. 2018;360:j5916. pulmonary complications within 14 postoperative days assessed daily using the Melbourne Group Scale. Secondary outcomes were: hospital-acquired Question: Does preoperative physiotherapy reduce postoperative pulmo- pneumonia, length of hospital stay, use of intensive care services, hospital nary complications in adults after elective, major, open upper abdominal costs, health-related quality of life (at 6 weeks), and all-cause mortality (at surgery? Design: Randomised, controlled trial with concealed allocation and 12 months). Results: A total of 432 participants completed the trial. The blinding of participants and assessors. Setting: Three tertiary public hospi- incidence of postoperative pulmonary complications was halved (adjusted tals in Australia and New Zealand. Participants: Adults ( 18 years) within hazard ratio 0.48, 95% CI 0.30 to 0.75) in the intervention group compared 6 weeks of elective upper abdominal surgery requiring general anaesthesia, with the control group, with an absolute risk reduction of 15% (95% CI 7 to 22) a minimum overnight hospital stay, an incision 5 cm above the umbilicus, and a number needed to treat of seven (95% CI 5 to 14). No significant and attendance at an outpatient preadmission clinic. Key exclusion criteria differences in other secondary outcomes were detected. Conclusion: In were: current hospital inpatients, organ transplant recipients, abdominal adults undergoing upper abdominal surgery, adding a single preoperative hernia repairs, being unable to walk 1 minute, or unable to attend the physiotherapy session, which comprised education and breathing exercise preoperative physiotherapy session. Randomisation of 441 participants coaching, to a standardised assessment and delivery of information on allocated 222 to the intervention group and 219 to the control group. postoperative physiotherapy via a booklet halved the incidence of postop- Interventions: Both groups attended the preadmission clinic and received erative pulmonary complications. preoperative physiotherapy comprising a standardised assessment and a booklet containing information about postoperative pulmonary complica- Provenance: Invited. Not peer reviewed. tions, potential prevention with early ambulation and breathing exercises, and a prescription for deep breathing exercises starting immediately after Nora Shields surgery. In addition, the intervention group received a 30-minute education School of Allied Health, La Trobe University, Australia and breathing exercise coaching session with a physiotherapist immediately after the preoperative session. This included: details on risk of postoperative https://doi.org/10.1016/j.jphys.2018.04.008 1836-9553/© 2018 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 The positive effects of breathing exercises on the recovery of wounded conducting the preoperative session may influence the intervention soldiers was reported in 1915.1 The need to prevent postoperative outcome. This study conveyed an important message that preoperative pulmonary complications was well noted in the 1950s, and the role of physiotherapy should target not only optimal breathing technique, but chest physiotherapy in the prevention of postoperative pulmonary also empower the patient to conduct self-directed breathing techniques complications was first acknowledged in 1954.2 The association between as early as practicable after surgery. The appreciation that prevention of postoperative pulmonary complications, mortality and length of hospital postoperative pulmonary complications relies upon the principle of stay has led to a multitude of reports on a variety of interventions, and maintenance of alveoli and airway patency, rather than the re-expansion been the pathophysiological basis for the prevention of postoperative of collapsed alveoli, is easily understood but may be clinically neglected. pulmonary complications after thoracic and abdominal surgery. The fidelity of the communication between the physiotherapist and the The Lung Infection Prevention Post Surgery Major Abdominal with patient affects the outcome of the intervention. It is our role as Pre-Operative Physiotherapy (LIPPSMAck-POP) trial is the first multi- physiotherapists to deliver optimal care to our patients. This study centre, international, randomised, controlled trial to investigate the effect provides objective evidence that the calibre of preoperative respiratory of the ‘quality’ of preoperative physiotherapy intervention on postopera- instructions and immediacy of postoperative implementation directly tive pulmonary complications. This well-designed study by Boden and impacts postsurgical patient recovery. colleagues addressed many of the factors that confound postoperative pulmonary complication data, and demonstrated that a control group Provenance: Invited. Not peer reviewed. Alice Jones that received an information booklet provided preoperatively, together with postoperative physiotherapy management, had a postoperative Discipline of Physiotherapy, Faculty of Health Sciences, The University pulmonary complication rate of 27%. However, the addition of a 30-minute preoperative physiotherapy consultation, clearly of Sydney, Sydney, Australia targeting breathing and coughing technique, with an emphasis on breathing technique immediately after regaining consciousness, reduced References the incidence of postoperative pulmonary complications by 50%. Further, this study also suggested that the experience of the physiotherapist 1. MacMahon C. Lancet. 1915;186:769–770. 2. Thoren L. Acta Chir Scand. 1954;107:193–205. https://doi.org/10.1016/j.jphys.2018.04.006 1836-9553/© 2018 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 64 (2018) 195–196 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Commentary described and delivered directly to patients while in hospital. The effect could be explained by the fact that most patients were well Surgery is a stressor that causes large physiological changes educated and seen within 6 weeks before surgery. Another ranging from tissue trauma, immobility, and systemic effects, to consideration is that most patients (around 70%) had cancer and it psychological distress. After upper abdominal surgery, patients is possible that these patients were motivated to take greater care of present with respiratory changes, including atelectasis, diaphragm their health postoperatively. In many developing countries, it might dysfunction and reduced lung volumesD[2.TFI_]$ This leads to postoperative be difficult to implement this type of intervention because of a lack of hypoventilationFI,_3DT[]$ which is associated with a high risk of pulmonary routine preadmission clinics. However, this trial will encourage complications. Preoperative physiotherapy interventions, includ- clinicians to re-consider their preoperative care. ing exercise training1 and breathing therapy strategies,2 aim to increase alveolar stability and mobilise secretions. These inter- Provenance: Invited. Not peer reviewed. ventions usually take around 20 minutes to complete. Several Celso RF Carvalho protocols having been tested but none have been found to reduce postoperative pulmonary complications.3 Department of Physical Therapy, School of Medicine, University of São Paulo, São Paulo, Brazil I commend the authors for this well designed and reported trial that satisfied all possible criteria of the PEDro scale, and recruited a References large number of patients. The intervention included patienteducation about postoperative pulmonary complications and training patients 1. Pouwels S, et al. Int J Surg. 2014;12:134–140. how to perform breathing exercises. The intervention was provided 2. Carvalho CRF, et al. Braz J Phys Ther. 2011;15:160–165. either face to face or by telephone. F_]T4ID[$T he simple intervention halved the 3. Lunardi AC, et al. Chest. 2015;148:1003–1010. incidence of postoperative pulmonary complications, including pneumonia, within 14 postoperative hospital days and the effect https://doi.org/10.1016/j.jphys.2018.04.004 was maintained a[5_]TD$FI t 12 months. It seems amazing how such a simple intervention was more effective than other interventions previously Commentary postoperative pulmonary complications along with coaching on simple breathing exercises would have the profound impact of The trial results point towards an effective intervention in which a halving complications in the upper abdominal surgery cohort. single 30-minute preoperative coaching session from a physiother- Implementation challenges could occur if decision-makers were not apist on breathing exercises can halve the incidence of postoperative fully convinced by or committed to the results. Anecdotally, pulmonary complications after upper abdominal surgery. However, physiotherapy preoperative assessment and education of those there were several imbalances in the randomised groups that might undergoing ‘at-risk’ surgery, such as cardiac, lung lobectomy and explain some of the effect. The intervention group:[$D23I]F_T was younger; had upper abdominal surgery, was prevalent in the later part of the 20th lower American Society of Anaesthesiology scores; had fewer century2$DI_F[T5] but funding for preoperative education, in the absence of respiratory, diabetes and cardiac co-morbidities; had fewer current evidence, has long been directed elsewhere. This trial is generating smokers and had a lower pack-year history. Additionally, the much conversation, but needs replicating across jurisdictions before intervention group reportedly had higher preoperative handgrip the water cooler chatter can settle and practice change ensues. strength and estimated VO2mTD$4[I]F_ ax. Further, less upper gastrointestinal/ hepatobiliary surgeries were performed in the intervention group. Provenance: Invited. Not peer reviewed. Are any of these large enough to create an unbalanced risk profile Shane Patman between intervention and control groups, and therefore cast doubt on the results? Probably not in isolation, but collectively? The School of Physiotherapy, The University of Notre Dame Australia, investigators undertook adjustments to their results for some Perth, Australia baseline variables considered to potentially affect the primary outcome, but it is speculative whether this was sufficient. What is References known is that presently, preadmission education by physiothera- pists for those undergoing upper abdominal surgery is not usual care 1. Patman S, et al. Arch Physiother. 2017;7:11. in Australian and New Zealand hospitals.1 Therefore, for clinical 2. Reeve J, et al. N Z J Physiother. 2016;44:33–D$48]IF[_T 9. practice change to occur, results have to pass the ‘water cooler test’. It intuitively seems too good to be true that such a minimal-risk https://doi.org/10.1016/j.jphys.2018.04.007 preoperative intervention of ‘shock and awe’ education on risks of Commentary (PEDro 9/10) demonstrating that a single preoperative physiothera- py session halves respiratory complications after major abdominal Strong claims require strong evidence. The Lung Infection surgery. Previous trials have reported very large effects: a 70 to 80% Prevention Post Surgery Major Abdominal with Pre-Operative Physiotherapy (LIPPSMAck-POP) is the latest and most robust trial 1836-9553/© 2018 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).
196 Appraisal reduction.1 Is this too good to be true? Maybe if based on a single trial, While physiotherapy in preoperative clinics is still common but three positive trials across four countries with a combined practice in Europe, Australian physiotherapists have disinvested sample size of close to 1000 provide D$IF][4_Tvery strong evidence. from this highly effective therapy over the past 20 years. Considering the mounting evidence also supporting preoperative In our trial, group imbalances were present at baseline. exercise training,3 it is time for physiotherapists to turn back the Mathematical modelling indicates that imbalances are statistically clock and get back into preop! likely (>70% likelihood) to occur when >50 variables are reported. Provenance: Invited. Not peer reviewed. In accordance with recommendations,2$DTIF][_3 it is important to specify a priori key variables known to influence postoperative pulmonary Ianthe Boden complications and to adjust results for these if required. Rather than on behalf of the LIPPSMAck POP investigators speculation, this is a deliberate measure to ensure that results are a Physiotherapy Department, Allied Health Services, Tasmanian Health true reflection of independent treatment effects. Additionally, the Services, North Launceston General Hospital, Australia imbalances are small in absolute numbers; at most, baseline covariate balance differed by eight people. Considering the large References effect size, robust trial methods, and detailed statistical analysis, it is unlikely that these small imbalances affected the overall outcome. 1. Reeve J, et al. N Z J Physiother. 2016;44:33–39. 2. Pocock S, et al. Stat Med. 2002;21:2917–2930. LIPPSMAck-POP confirmed the findings of previous trials: 3. Barberan-Garcia A, et al. Ann Surg. 2018;267:50–56. preoperative education, in addition to early ambulation, prevents postoperative pulmonary complications.1 Most importantly, tim- https://doi.org/10.1016/j.jphys.2018.04.005 ing is vital. Patients should commence breathing exercises immediately after surgery, not a day or two later as per usual physiotherapy service in Australia. Preadmission clinics are the opportune time to train patients on these breathing exercises.
Journal of Physiotherapy 64 (2018) 195–196 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Commentary described and delivered directly to patients while in hospital. The effect could be explained by the fact that most patients were well Surgery is a stressor that causes large physiological changes educated and seen within 6 weeks before surgery. Another ranging from tissue trauma, immobility, and systemic effects, to consideration is that most patients (around 70%) had cancer and it psychological distress. After upper abdominal surgery, patients is possible that these patients were motivated to take greater care of present with respiratory changes, including atelectasis, diaphragm their health postoperatively. In many developing countries, it might dysfunction and reduced lung volumesD[2.TFI_]$ This leads to postoperative be difficult to implement this type of intervention because of a lack of hypoventilationFI,_3DT[]$ which is associated with a high risk of pulmonary routine preadmission clinics. However, this trial will encourage complications. Preoperative physiotherapy interventions, includ- clinicians to re-consider their preoperative care. ing exercise training1 and breathing therapy strategies,2 aim to increase alveolar stability and mobilise secretions. These inter- Provenance: Invited. Not peer reviewed. ventions usually take around 20 minutes to complete. Several Celso RF Carvalho protocols having been tested but none have been found to reduce postoperative pulmonary complications.3 Department of Physical Therapy, School of Medicine, University of São Paulo, São Paulo, Brazil I commend the authors for this well designed and reported trial that satisfied all possible criteria of the PEDro scale, and recruited a References large number of patients. The intervention included patienteducation about postoperative pulmonary complications and training patients 1. Pouwels S, et al. Int J Surg. 2014;12:134–140. how to perform breathing exercises. The intervention was provided 2. Carvalho CRF, et al. Braz J Phys Ther. 2011;15:160–165. either face to face or by telephone. F_]T4ID[$T he simple intervention halved the 3. Lunardi AC, et al. Chest. 2015;148:1003–1010. incidence of postoperative pulmonary complications, including pneumonia, within 14 postoperative hospital days and the effect https://doi.org/10.1016/j.jphys.2018.04.004 was maintained a[5_]TD$FI t 12 months. It seems amazing how such a simple intervention was more effective than other interventions previously Commentary postoperative pulmonary complications along with coaching on simple breathing exercises would have the profound impact of The trial results point towards an effective intervention in which a halving complications in the upper abdominal surgery cohort. single 30-minute preoperative coaching session from a physiother- Implementation challenges could occur if decision-makers were not apist on breathing exercises can halve the incidence of postoperative fully convinced by or committed to the results. Anecdotally, pulmonary complications after upper abdominal surgery. However, physiotherapy preoperative assessment and education of those there were several imbalances in the randomised groups that might undergoing ‘at-risk’ surgery, such as cardiac, lung lobectomy and explain some of the effect. The intervention group:[$D23I]F_T was younger; had upper abdominal surgery, was prevalent in the later part of the 20th lower American Society of Anaesthesiology scores; had fewer century2$DI_F[T5] but funding for preoperative education, in the absence of respiratory, diabetes and cardiac co-morbidities; had fewer current evidence, has long been directed elsewhere. This trial is generating smokers and had a lower pack-year history. Additionally, the much conversation, but needs replicating across jurisdictions before intervention group reportedly had higher preoperative handgrip the water cooler chatter can settle and practice change ensues. strength and estimated VO2mTD$4[I]F_ ax. Further, less upper gastrointestinal/ hepatobiliary surgeries were performed in the intervention group. Provenance: Invited. Not peer reviewed. Are any of these large enough to create an unbalanced risk profile Shane Patman between intervention and control groups, and therefore cast doubt on the results? Probably not in isolation, but collectively? The School of Physiotherapy, The University of Notre Dame Australia, investigators undertook adjustments to their results for some Perth, Australia baseline variables considered to potentially affect the primary outcome, but it is speculative whether this was sufficient. What is References known is that presently, preadmission education by physiothera- pists for those undergoing upper abdominal surgery is not usual care 1. Patman S, et al. Arch Physiother. 2017;7:11. in Australian and New Zealand hospitals.1 Therefore, for clinical 2. Reeve J, et al. N Z J Physiother. 2016;44:33–D$48]IF[_T 9. practice change to occur, results have to pass the ‘water cooler test’. It intuitively seems too good to be true that such a minimal-risk https://doi.org/10.1016/j.jphys.2018.04.007 preoperative intervention of ‘shock and awe’ education on risks of Commentary (PEDro 9/10) demonstrating that a single preoperative physiothera- py session halves respiratory complications after major abdominal Strong claims require strong evidence. The Lung Infection surgery. Previous trials have reported very large effects: a 70 to 80% Prevention Post Surgery Major Abdominal with Pre-Operative Physiotherapy (LIPPSMAck-POP) is the latest and most robust trial 1836-9553/© 2018 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/).
Journal of Physiotherapy 64 (2018) 194 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Critically appraised paper: Preoperative physiotherapy education halved postoperative pulmonary complications in patients after upper abdominal surgery Synopsis Summary of: Boden I, Skinner EH, Browning L, Reeve J, Anderson L, pulmonary complications, individualised risk assessment, education about Hill C, et al. Preoperative physiotherapy for the prevention of respiratory the impact of surgery, instructions on breathing exercises, and memory cues. complications after upper abdominal surgery: pragmatic, double blinded, Outcome measures: The primary outcome was presence of postoperative multicentre randomised controlled trial. BMJ. 2018;360:j5916. pulmonary complications within 14 postoperative days assessed daily using the Melbourne Group Scale. Secondary outcomes were: hospital-acquired Question: Does preoperative physiotherapy reduce postoperative pulmo- pneumonia, length of hospital stay, use of intensive care services, hospital nary complications in adults after elective, major, open upper abdominal costs, health-related quality of life (at 6 weeks), and all-cause mortality (at surgery? Design: Randomised, controlled trial with concealed allocation and 12 months). Results: A total of 432 participants completed the trial. The blinding of participants and assessors. Setting: Three tertiary public hospi- incidence of postoperative pulmonary complications was halved (adjusted tals in Australia and New Zealand. Participants: Adults ( 18 years) within hazard ratio 0.48, 95% CI 0.30 to 0.75) in the intervention group compared 6 weeks of elective upper abdominal surgery requiring general anaesthesia, with the control group, with an absolute risk reduction of 15% (95% CI 7 to 22) a minimum overnight hospital stay, an incision 5 cm above the umbilicus, and a number needed to treat of seven (95% CI 5 to 14). No significant and attendance at an outpatient preadmission clinic. Key exclusion criteria differences in other secondary outcomes were detected. Conclusion: In were: current hospital inpatients, organ transplant recipients, abdominal adults undergoing upper abdominal surgery, adding a single preoperative hernia repairs, being unable to walk 1 minute, or unable to attend the physiotherapy session, which comprised education and breathing exercise preoperative physiotherapy session. Randomisation of 441 participants coaching, to a standardised assessment and delivery of information on allocated 222 to the intervention group and 219 to the control group. postoperative physiotherapy via a booklet halved the incidence of postop- Interventions: Both groups attended the preadmission clinic and received erative pulmonary complications. preoperative physiotherapy comprising a standardised assessment and a booklet containing information about postoperative pulmonary complica- Provenance: Invited. Not peer reviewed. tions, potential prevention with early ambulation and breathing exercises, and a prescription for deep breathing exercises starting immediately after Nora Shields surgery. In addition, the intervention group received a 30-minute education School of Allied Health, La Trobe University, Australia and breathing exercise coaching session with a physiotherapist immediately after the preoperative session. This included: details on risk of postoperative https://doi.org/10.1016/j.jphys.2018.04.008 1836-9553/© 2018 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 The positive effects of breathing exercises on the recovery of wounded conducting the preoperative session may influence the intervention soldiers was reported in 1915.1 The need to prevent postoperative outcome. This study conveyed an important message that preoperative pulmonary complications was well noted in the 1950s, and the role of physiotherapy should target not only optimal breathing technique, but chest physiotherapy in the prevention of postoperative pulmonary also empower the patient to conduct self-directed breathing techniques complications was first acknowledged in 1954.2 The association between as early as practicable after surgery. The appreciation that prevention of postoperative pulmonary complications, mortality and length of hospital postoperative pulmonary complications relies upon the principle of stay has led to a multitude of reports on a variety of interventions, and maintenance of alveoli and airway patency, rather than the re-expansion been the pathophysiological basis for the prevention of postoperative of collapsed alveoli, is easily understood but may be clinically neglected. pulmonary complications after thoracic and abdominal surgery. The fidelity of the communication between the physiotherapist and the The Lung Infection Prevention Post Surgery Major Abdominal with patient affects the outcome of the intervention. It is our role as Pre-Operative Physiotherapy (LIPPSMAck-POP) trial is the first multi- physiotherapists to deliver optimal care to our patients. This study centre, international, randomised, controlled trial to investigate the effect provides objective evidence that the calibre of preoperative respiratory of the ‘quality’ of preoperative physiotherapy intervention on postopera- instructions and immediacy of postoperative implementation directly tive pulmonary complications. This well-designed study by Boden and impacts postsurgical patient recovery. colleagues addressed many of the factors that confound postoperative pulmonary complication data, and demonstrated that a control group Provenance: Invited. Not peer reviewed. Alice Jones that received an information booklet provided preoperatively, together with postoperative physiotherapy management, had a postoperative Discipline of Physiotherapy, Faculty of Health Sciences, The University pulmonary complication rate of 27%. However, the addition of a 30-minute preoperative physiotherapy consultation, clearly of Sydney, Sydney, Australia targeting breathing and coughing technique, with an emphasis on breathing technique immediately after regaining consciousness, reduced References the incidence of postoperative pulmonary complications by 50%. Further, this study also suggested that the experience of the physiotherapist 1. MacMahon C. Lancet. 1915;186:769–770. 2. Thoren L. Acta Chir Scand. 1954;107:193–205. https://doi.org/10.1016/j.jphys.2018.04.006 1836-9553/© 2018 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 64 (2018) 195–196 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Commentary described and delivered directly to patients while in hospital. The effect could be explained by the fact that most patients were well Surgery is a stressor that causes large physiological changes educated and seen within 6 weeks before surgery. Another ranging from tissue trauma, immobility, and systemic effects, to consideration is that most patients (around 70%) had cancer and it psychological distress. After upper abdominal surgery, patients is possible that these patients were motivated to take greater care of present with respiratory changes, including atelectasis, diaphragm their health postoperatively. In many developing countries, it might dysfunction and reduced lung volumesD[2.TFI_]$ This leads to postoperative be difficult to implement this type of intervention because of a lack of hypoventilationFI,_3DT[]$ which is associated with a high risk of pulmonary routine preadmission clinics. However, this trial will encourage complications. Preoperative physiotherapy interventions, includ- clinicians to re-consider their preoperative care. ing exercise training1 and breathing therapy strategies,2 aim to increase alveolar stability and mobilise secretions. These inter- Provenance: Invited. Not peer reviewed. ventions usually take around 20 minutes to complete. Several Celso RF Carvalho protocols having been tested but none have been found to reduce postoperative pulmonary complications.3 Department of Physical Therapy, School of Medicine, University of São Paulo, São Paulo, Brazil I commend the authors for this well designed and reported trial that satisfied all possible criteria of the PEDro scale, and recruited a References large number of patients. The intervention included patienteducation about postoperative pulmonary complications and training patients 1. Pouwels S, et al. Int J Surg. 2014;12:134–140. how to perform breathing exercises. The intervention was provided 2. Carvalho CRF, et al. Braz J Phys Ther. 2011;15:160–165. either face to face or by telephone. F_]T4ID[$T he simple intervention halved the 3. Lunardi AC, et al. Chest. 2015;148:1003–1010. incidence of postoperative pulmonary complications, including pneumonia, within 14 postoperative hospital days and the effect https://doi.org/10.1016/j.jphys.2018.04.004 was maintained a[5_]TD$FI t 12 months. It seems amazing how such a simple intervention was more effective than other interventions previously Commentary postoperative pulmonary complications along with coaching on simple breathing exercises would have the profound impact of The trial results point towards an effective intervention in which a halving complications in the upper abdominal surgery cohort. single 30-minute preoperative coaching session from a physiother- Implementation challenges could occur if decision-makers were not apist on breathing exercises can halve the incidence of postoperative fully convinced by or committed to the results. Anecdotally, pulmonary complications after upper abdominal surgery. However, physiotherapy preoperative assessment and education of those there were several imbalances in the randomised groups that might undergoing ‘at-risk’ surgery, such as cardiac, lung lobectomy and explain some of the effect. The intervention group:[$D23I]F_T was younger; had upper abdominal surgery, was prevalent in the later part of the 20th lower American Society of Anaesthesiology scores; had fewer century2$DI_F[T5] but funding for preoperative education, in the absence of respiratory, diabetes and cardiac co-morbidities; had fewer current evidence, has long been directed elsewhere. This trial is generating smokers and had a lower pack-year history. Additionally, the much conversation, but needs replicating across jurisdictions before intervention group reportedly had higher preoperative handgrip the water cooler chatter can settle and practice change ensues. strength and estimated VO2mTD$4[I]F_ ax. Further, less upper gastrointestinal/ hepatobiliary surgeries were performed in the intervention group. Provenance: Invited. Not peer reviewed. Are any of these large enough to create an unbalanced risk profile Shane Patman between intervention and control groups, and therefore cast doubt on the results? Probably not in isolation, but collectively? The School of Physiotherapy, The University of Notre Dame Australia, investigators undertook adjustments to their results for some Perth, Australia baseline variables considered to potentially affect the primary outcome, but it is speculative whether this was sufficient. What is References known is that presently, preadmission education by physiothera- pists for those undergoing upper abdominal surgery is not usual care 1. Patman S, et al. Arch Physiother. 2017;7:11. in Australian and New Zealand hospitals.1 Therefore, for clinical 2. Reeve J, et al. N Z J Physiother. 2016;44:33–D$48]IF[_T 9. practice change to occur, results have to pass the ‘water cooler test’. It intuitively seems too good to be true that such a minimal-risk https://doi.org/10.1016/j.jphys.2018.04.007 preoperative intervention of ‘shock and awe’ education on risks of Commentary (PEDro 9/10) demonstrating that a single preoperative physiothera- py session halves respiratory complications after major abdominal Strong claims require strong evidence. The Lung Infection surgery. Previous trials have reported very large effects: a 70 to 80% Prevention Post Surgery Major Abdominal with Pre-Operative Physiotherapy (LIPPSMAck-POP) is the latest and most robust trial 1836-9553/© 2018 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/).
Journal of Physiotherapy 64 (2018) 172–177 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Research Downton Fall Risk Index during hospitalisation is associated with fall-related injuries after discharge: a longitudinal observational study Mahnaz Mojtaba a, Hassan Alinaghizadeh b,c, Elisabeth Rydwik d,e a Jakobsberg’s Geriatric Clinic; b Academic Primary Care Center, Stockholm County Council, Stockholm; c Department of Medical Sciences, Occupational and Environmental Medicine, Uppsala University, Uppsala; d FOU nu Research and Development Center for the Elderly, Stockholm County Council, Stockholm; e Department of Neurobiology, Care Sciences and Society, Division of Physiotherapy, Karolinska Insitutet, Huddinge, Sweden KEY WORDS ABSTRACT Elderly Questions: Among older people who are hospitalised, what is the predictive validity of the Downton Fall Previous falls Risk Index (DFRI) in relation to fall-related injury after discharge? What is the predictive validity of the Unsafe gait DFRI among males and females in this setting? Design: Prospective, longitudinal, observational study. Medication Participants: All hospital admissions during 2012 at three geriatric clinics in the Stockholm County Gender Council were monitored. Patients aged > 65 years who did not die during the admission and who lived in the Stockholm County Council region were included. Outcome measures: The DFRI consists of five modules: previous falls, medication, sensory deficits, mental state, and gait. Three or more points indicate an increased fall risk. Data on DFRI, health status and medications were collected prior to discharge. Data regarding fall-related injuries were collected up to 6 months after discharge. Poisson multivariate regression analyses were conducted to evaluate the association between DFRI and fall-related injuries. Results: In total, 6650 patients were analysed. The cut-off 3 points in the DFRI was significantly associated with fall-related injury when confounding variables were controlled for (IRR 1.94, 95% CI 1.60 to 2.38). Among individual modules, only previous falls (IRR 2.58, 95% CI 2.22 to 3.01) and unsafe gait (IRR 1.79, 95% CI 1.53 to 2.09) were associated with fall-related injuries. Stratified analyses showed a higher risk ratio for men compared to women regarding the DFRI, but the test for an interaction effect was not significant. Conclusion: The risk of post-discharge fall-related injury is increased among older hospitalised people with an increased fall risk, according to the DFRI, especially those who had previous falls or unsafe gait. Although the DFRI tool is predictive, previous falls and gait are the measures that are most worthy of focus. [Mojtaba M, Alinaghizadeh H, Rydwik E (2018) Downton Fall Risk Index during hospitalisation is associated with fall-related injuries after discharge: a longitudinal observational study. Journal of Physiotherapy 64: 172–177] © 2018 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 frailty are risk factors for women, while older age, depression and standing balance are risk factors for men.9 Older people who are discharged from hospital are at risk of falling. Two studies have shown that 13 to 15% fall within 4 weeks Fall prevention actions during hospitalisation might include the of discharge.1,2 The risk of a fall after discharge is associated with impairment in activities of daily living, decreased mobility, use of use of a fall screening instrument and frequent fall-risk meetings assistive devices, and pre-admission falls.1,2 Falling once or more with an interdisciplinary approach.10 Several fall screening during hospitalisation is also associated with increased likelihood instruments are available.11 The Downton Fall Risk Index (DFRI) of falls after discharge.3 is a validated screening instrument used in a Swedish quality Risk factors for fall-related injuries (eg, fractures) are similar to risk factors for falls.4,5 However, additional risk factors are register and is recommended to be used in hospitals, geriatric important in regard to fall-related injuries: low bone mineral density, reaction time, amount of soft tissue padding, previous clinics, care of the elderly, and primary care in Sweden. Several fractures, the characteristics of the fall (such as falling backwards or sideways), and the energy of the fall.6 Women have a higher risk studies have shown that the DFRI can predict falls in residential of falls than men, and this has been shown to be related to higher care,12 among community-dwellers,13 and in hospitals.14 However, gait variability during dual-task activities.7 They are also more likely to be multiple fallers than men.8 Risk factors for falls also it is believed that no studies have evaluated the DFRI in relation to differ between men and women. For example, incontinence and clinically relevant outcomes such as fall-related injuries after hospital discharge. The DFRI has been shown to have higher sensitivity than other instruments,11 although several studies have also shown low specificity.11–14 In addition, results from several studies have indicated that the nursing staff’s clinical assessment might be as accurate as existing screening tools.15,16 Since the existing literature shows diverse results and no studies have https://doi.org/10.1016/j.jphys.2018.05.005 1836-9553/© 2018 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/).
Research 173 evaluated the DFRI in a hospital setting related to fall-related Independent variable injuries after discharge or investigated differences between men The DFRI was used to assess risk of falls.12,13 Fall risk was and women, it requires further evaluation. Thus, the primary aim of this study was to evaluate the predictive validity of DFRI in assessed at admission by the responsible nurse and registered in relation to post-discharge fall-related injury and the secondary aim was to investigate this separately in men and women. the medical records. The instrument consists of five modules: previous falls, medication, sensory deficits, mental state, and gait. Therefore, the research questions for this prospective, longi- This results in 11 different risk factors, which are summarised into tudinal, observational study were: a score between 0 and 11. Scores 3 points indicate an increased fall risk.12,13 1. Among older people who are hospitalised, what is the predictive Potentially confounding variables validity of the DFRI in relation to fall-related injury after The following information was extracted from the medical discharge? records from each participant’s last hospital stay in 2012. Age, 2. What is the predictive validity of the DFRI among males and gender, and number of diagnoses were recorded. The number and females in this setting? type of medications were also extracted from the medication record. Method Mini Nutritional Assessment was used to measure nutritional Design status.17 The instrument is based on scores between 0 and 14 points, where 0 to 7 points is regarded as malnutrition, 8 to This was a multi-centre, prospective, longitudinal, observa- 11 points is regarded as at risk of malnutrition, and 12 to tional study conducted in Stockholm County, Sweden. All hospital 14 points is regarded as normal nutritional status. In the admissions in 2012 were monitored at three geriatric clinics in regression analyses, this variable was dichotomised as 1 for Stockholm County Council. If a participant had more than one malnutrition or 0 for at risk of malnutrition or normal hospital admission during the data collection period, the last nutritional status. Body mass index was used to classify admission was used in the analysis. Participants were followed for underweight (< 18.5), overweight (25.0 to 29.9) and obesity 6 months after discharge from hospital to observe for fall-related ( 30.0). Information about body mass index was extracted from injury. the Mini Nutritional Assessment. The variable was dichotomised and a cut-off for underweight < 23 kg/m2 was used.18 Participants Haemoglobin was used to classify anaemia. This variable was All admitted patients were screened for eligibility. To be eligible dichotomised using a cut-off for anaemia set at < 115 g/l.19 Data on for inclusion, patients were required to be aged 65 years. blood pressure were dichotomised according Klein et al,19 where a low blood pressure can increase the risk of falling. The cut-off for Data collection low blood pressure was set to < 120 mmHg for systolic and < 80 mmHg for diastolic.20 Data were retrieved from two different data sources and then merged into one dataset for analyses. Baseline characteristics C-reactive protein was used to estimate inflammatory levels. collected during each participant’s last admission to the geriatric The variable was dichotomised and a level of > 10 g/l was used as a clinics during 2012 were retrieved from the medical records. cut-off indicating a high inflammatory level.21 Follow-up data on fall-related injuries (such as fractures, contusions and fall accidents) were retrieved from the Stockholm Data analysis County Council Health Care Consumption database within 6 months after discharge from the geriatric clinic. All healthcare Demographic and background factors at baseline were pre- providers within the Stockholm County Council – including both sented as mean, standard deviation and frequencies. Between- hospitals and outpatient clinics (primary care) – were obliged to group comparison was carried out by t-test for continuous data and report the data digitally. The Stockholm County Council Health Chi-squared test for nominal data, where statistical assumptions Care Consumption database covers 99% of all care in Stockholm. were not violated. When normality assumptions for continuous The first clinically relevant diagnosis of a fall-related injury that data were not met, the Mann-Whitney test was used to compare was found was recorded as an event. All observations with no differences between groups. For between-group analysis in event followed were censored at 6 months. Data regarding relation to the DFRI, the Kruskal-Wallis equality-of-populations falls could not be retrieved from this data source, so data on rank test was used. non-injurious falls were unobtainable. Poisson regression was used to assess how strongly the Dependent variable outcome of fall-related injury was associated with the DFRI and The dependent variable was fall-related injury. The Interna- its individual modules. The analysis treated fall-related injury dichotomously and therefore did not allow for the possibility of tional Classification of Diseases (ICD)-10 codes (S, T and W) based more than one injury per participant. Due to multicollinearity, all on the World Health Organization’s International Classification of of the DFRI modules were used as exposures and a separate model Diseases were used to extract information about diagnoses related was applied for the full DFRI (dichotomised based on the cut- to fall-related injuries from the Stockholm County Council Health off 3 points) as well as each module. Apart from the potential Care Consumption database. Specific information about the ICD confounding variables described above, an interaction term with codes used is presented in Appendix 1 (see eAddenda). gender and the DFRI module was added in each model. For those models where a significant association was found, bootstrapping A total of 1146 participants had a fall-related injury. Among with 1000 replications and Jackknife estimation were applied.22 these participants, 117 sustained a fracture, 37 sustained a contusion, 145 had a fall accident and 847 had some combination Complementary analyses were also conducted among specific of the three; the most common combination was fracture and fall medications that have been shown to be associated with an accident (n = 630). increased risk of falls and fall-related injuries.23 The following medications were chosen for further analyses: diuretics, antihyper- tensives (not diuretics), antiparkinsonian drugs, neuroleptics, antidepressants, tramadol, Propavan, benzodiazepine (long-lasting), and Nozinan. All statistical analyses were performed using SAS softwarea. A 5% significance level was chosen to reject the null hypothesis.
174 Mojtaba et al: Fall risk and fall-related injury Admissions to one of the three Table 1 study geriatric clinics Demographic and clinical characteristics of all participants at baseline, also (n = 10062) categorised according to no fall injury and fall injury within 6 months after discharge. Excluded (n = 72) • aged < 65 years (n = 72) Risk factors All No fall injury Fall injury (n = 6650) (n = 5504) (n = 1146) Participants enrolled in the study (n = 6650) Women, n (%)a 4277 (64) 3478 (63) 799 (70) 2026 (37) 347 (30) Participants analysed Men, n (%) 2373 (36) (n = 6650) 84 (7) 84 (8) Age, mean (SD) 84 (7) 4 (2) 4 (2) Figure 1. Flow of participants through the study. 4 (2) 4 (2) Diagnoses (n), mean (SD) 4 (2) 232 (4) 43 (4) Results Medications (n), mean (SD) 4 (2) 1544 (28) 294 (26) Flow of participants through the study 2811 (51) 606 (53) missing n (%) 275 (4) 1020 (19) 229 (20) During the data collection period, 10 062 admissions occurred at three participating geriatric clinics in Stockholm County Council. Mini Nutritional Assessment (points), n (%) 129 (2) 17 (1) Several patients had more than one admission. Therefore, the 10 062 hospital admissions resulted in 6650 participants who were 0 to 7 1833 (28) 2906 (53) 621 (54) included and analysed in the study. See Figure 1. 2352 (43) 491 (43) 8 to 11 3417 (51) Characteristics of the study participants 246 (4) 34 (3) 12 to 14 1249 (19) A description of the baseline characteristics of all participants 1761 (32) 429 (37) are shown in Table 1. The characteristics of the subgroups of missing 146 (2) 2842 (52) 564 (49) participants who did and did not sustain a fall-related injury after Body Mass Index, n (%)b 901 (16) 153 (13) discharge are also presented in Table 1. Of the 6650 subjects, 1146 (17%) sustained a fall-related injury during the 6-month follow-up. < 23 3527 (53) 1432 (26) 257 (22) The mean age in both groups with and without fall-related injury 3847 (70) 856 (75) was 84 years. Women were significantly more likely to sustain a 23 2843 (43) fall-related injury than men. There were significant differences 225 (4) 33 (3) between those with and without fall-related injury regarding missing 280 (4) previous falls, gait and fall risk according to DFRI ( 3 points), as Haemoglobin (g/l), n (%)a 3908 (71) 823 (72) presented in Table 2. Among those who sustained a fall-related 1370 (25) 290 (25) injury, 83% had a high fall risk according to DFRI. Of those who did 115 2190 (33) not sustain a fall-related injury, 22% were correctly classified as 226 (4) 33 (3) having no fall risk (Table 2). The corresponding proportion for men > 115 3406 (51) and women was similar. Of those who sustained a fall-related 698 (61) 3169 (58) injury, mean days to sustain the injury were 59 days (SD 53, 95% CI missing 1054 (16) 256 (22) 1289 (23) 56 to 62) (Figure 2). 192 (17) 1046 (19) Systolic blood pressure (mmHg), n (%)c The regression analyses showed a significant association between the cut-off 3 points according to the DFRI and fall- 120 1689 (25) related injury when confounding variables were controlled for. When analysing the individual modules of the DFRI, only previous > 120 4703 (71) falls and unsafe gait were associated with fall-related injuries (Table 3). missing 258 (4) Since there was no significant association between the Diastolic blood pressure (mmHg), n (%) medication module and fall-related injury, despite previously shown associations, complementary analyses were conducted. As 80 4731 (71) seen in Table 3, there was a low but significant association between not using diuretics and fall-related injuries, when > 80 1660 (25) controlling for confounding variables. There was no significant interaction between gender and fall risk according to DFRI missing 259 (4) ( 3 points). There was also no significant interaction between gender and either of the two significant DFRI modules in the final C-reactive protein (g/l), n (%) models. However, due to the significant main effect (DFRI module and gender), these differences are visualised in Figure 3. On the 10 3867 (58) other hand, the association between not using diuretics and fall- related injury was only significant among women (IRR 1.40, 95% > 10 1545 (23) CI 1.16 to 1.70). missing 1238 (19) a p < 0.001, indicating a significant difference between women and men in the prevalence of fall injury. b Determined according to the Mini Nutritional Assessment. c p < 0.05, indicating a significant difference between no fall injury and fall injury. Discussion This study showed a significant association between fall-related injuries and a high fall risk according to the DFRI. Similarly, a significant association was shown between fall-related injuries and either the previous falls or unsafe gait modules. These results are in line with a recently published study that showed that DFRI independently predicted fall-related injuries24 however, the populations partly differed. The current results are also in line with a study by Vassallo et al,25 which showed that previous falls and unsafe gait are associated with fall-related injuries. Similar to that study,25 clinical characteristics between subjects with or without fall-related injuries did not differ in most of the variables in the current study (Table 1). These results suggested that identifying patients at risk of injury is difficult and multi-faceted. This difficulty is emphasised by the paradox shown in a study by Aranda-Gallardo et al,26 in which there was an increased odds ratio for fall-related injury among those classified as low risk according to the DFRI. This highlights the problematic features of relying too much on screening, where patients screened with a low risk are not further investigated, and therefore no preventive actions are taken.26 The current results showed that 17% of the sample sustained an injurious fall within 6 months of discharge. This is somewhat different compared to another study showing that 11% sustained an injurious fall after discharge.3 The difference might be explained by a difference in length of follow-up and whether the injury required hospital care or not. To our knowledge, this is the first study to conduct stratified analyses by gender regarding fall risk screening. The current results indicated a higher incidence risk ratio for fall-related injuries in men according to the DFRI; however, the difference was not significant (shown by the non-significant interaction effect on
Research 175 Table 2 who were regarded as frail had a higher risk for falls compared to Results of screening with the Downton Fall Risk Index and its individual modules in women who were regarded as frail.27 relation to fall injuries (n = 6650). Somewhat opposite to other studies, this study showed no Risk factors No fall injury Fall injury association with polypharmacy,28 or between antidepressants and (n = 5504) (n = 1146) psychotropic drugs with fall-related injury.23,29 However, this Previous falls, n (%)a 2571 (47) 307 (27) might partly be explained by different outcomes (falls instead of no 2785 (51) 818 (71) fall-related injuries,23 and hip fractures alone).29 Using hyperten- yes missing 148 (2) 21 (2) sive drugs was not related to fall-related injuries – neither in the multivariate analyses nor in the stratified analysis – which is in line Medication, n (%) 957 (17) 207 (18) with other studies.30,31 Surprisingly, this study showed that none 2143 (39) 448 (39) tranquilisers/sedatives 1670 (30) 349 (30) women who did not use diuretics had an increased risk of fall- diuretics 103 (9) antihypertensive (other than diuretics) 463 (8) related injuries. This is both in line with and somewhat opposite to antiparkinsonian drugs 108 (2) 17 (1) other studies. Thorell et al29 showed no association between antidepressants 14 (< 1) 2 (< 1) missing 149 (3) 20 (2) diuretics and hypertensives with hip fracture. Another study Sensory deficits, n (%) 1449 (26) 298 (26) showed a higher prevalence of diuretics among non-fallers none 2675 (49) 561 (49) visual impairment 1122 (20) 250 (22) compared to fallers, as well as among non-injured fallers compared hearing impairment to injured fallers.25 A Cochrane review showed that thiazide limb impairment 111 (2) 16 (1) diuretics reduced the risk of hip fracture;31 however, the missing 147 (3) 21 (2) conclusion was based solely on observational studies solely. One Mental state, n (%) 4111 (75) 838 (73) orientated 1250 (23) 287 (25) clinical explanation could be that women suffering from dizziness confused (cognitively impaired) missing 143 (3) 21 (2) or orthostatic hypotension might have discontinued with the Gait, n (%)a 1997 (36) 293 (26) drugs. Again, different outcomes and stratification might explain normal (safe with or without walking aids) 3364 (61) 833 (73) the different results, but this also highlights the complexity of the unsafe (with or without walking aids) missing 143 (3) 20 (2) area being studied. DFRI fall risk (1 to 11), n (%)a 1201 (22) 172 (15) Mental state was not associated with fall-related injuries, which 1 to 2 (no fall risk) 4144 (75) 950 (83) is not in line with a previous review.32 However, the data on 3 to 11 (fall risk) missing 159 (3) 24 (2) cognition in this study relate to the mental state module in DFRI, a p < 0.001, indicating a significant difference on Chi-squared test. which makes comparison difficult. Unfortunately, data on cogni- tion measured with the Mini Mental State Examination were only gender). However, data regarding sensitivity and specificity were similar in men and women (data not shown), so this problem available for a small portion of the sample. Therefore, no remains. A recent study investigated differences in risk factors for falls between women and men, and found that some risk factors supplementary analyses were conducted. were the same, while others (such as incontinence and frailty) were significant only in women.9 Another study showed that men The DFRI showed high sensitivity but low specificity in relation to fall-related injuries (Table 2). This is partly in line with previous studies that showed similar results, but in relation to falls.8–10 Olsson Möller et al13 investigated other cut-offs in frail communi- ty-dwelling older adults in relation to falls, and concluded that the predictive validity was low regardless of cut-off. Lower cut-offs showed higher sensitivity and lower specificity and higher cut-offs the opposite.13 120 Mean = 59 SD = 53 N = 1146 100 Participants with fall-related injury (n) 80 60 40 20 0 12345 6 0 Duration of follow-up since hospital discharge (months) Figure 2. Distribution of observation with fall-related injuries at 6-month follow-up. The dashed line shows mean time to fall-related injury (59 days).
176 Mojtaba et al: Fall risk and fall-related injury Table 3 Observed incident rate ratios (95% CI) for fall injury depending on risk factors at the 6-month follow-up. Explanatory variables Observed IRR (Normal-based 95% CI)a Crude model Model A Model B Downton 3 1.57 (1.32 to 1.86) 1.59 (1.34 to 1.88) 1.94 (1.58 to 2.37) Previous falls 2.36 (2.07 to 2.70) 2.37 (2.07 to 2.72) 2.53 (2.15 to 2.96) Gait 1.65 (1.44 to 1.89) 1.67 (1.44 to 1.93) 1.78 (1.51 to 2.11) Not using diuretics 1.28 (1.12 to 1.46) 1.28 (1.12 to 1.46) 1.34 (1.13 to 1.58) Not using antihypertensives 1.13 (1.00 to 1.27) 1.13 (1.00 to 1.28) 1.07 (0.92 to 1.24) Downton < 3, no previous falls, safe gait, using diuretics and using antihypertensive are reference categories. Model A was adjusted for age and gender. Model B was adjusted for age, gender, gender*explanatory variable, number of diagnoses, number of medicines, Mini Nutritional Assessment, body mass index, haemoglobin, systolic and diastolic blood pressure, and C-reactive protein. a Bootstrap with 1000 replications and Jackknife estimation. 4 Incidence Rate Ratio (95% CI) 3 2 1 0 DFRI Unsafe Previous DFRI Unsafe Previous gait falls gait falls Men Women Figure 3. Stratified incidence rate ratio by sex for fall injury according Downton Fall Risk Index when controlling for confounding variables. The results in this and previous studies suggest that screening risk of fall-related injuries if the patient had a risk for falls tools should only be used as indicators of risks and should be according to the DFRI, especially if they had previous falls and followed by further evaluation of the individual’s risk profile for unsafe gait. Although there was similarity in results with many falls and fall-related injuries.33 It seems as though previous falls, previous studies, these results could be further validated in future unsafe gait and maybe also previous fractures6 are important risk studies. The results suggest that identifying patients at risk of factors and warrant further individualised evaluations in clinical injury is problematic and multi-faceted, and highlight the settings. The National Institute for Health and Care Excellence complexity of the area and the need for a broad, individualised (NICE) guidelines suggest a thorough examination. This should be assessment with an inter-disciplinary approach. However, a fall- followed by a multidisciplinary, person-centred approach regard- risk assessment should be conducted before discharge, focusing on ing fall prevention, with a specific focus on unsafe gait.16,33 There is previous falls and gait instability. In order to decrease the risk of substantial evidence that exercise can prevent risk factors of falls fall-related injuries after discharge, the results of the assessments such as unsafe gait.33 should then influence the level of care provided. This study had some limitations. There were no data regarding What was already known on this topic: After discharge causes for admission at baseline. Since this is a database study, from hospital, older people are at risk of falling. Several risk there was no information about fall characteristics, presence of factors for these falls and fall-related injuries have been iden- osteoporosis, current health status, and other factors associated tified. Some of these risk factors differ between men and with the fall-related injury. In addition, there was only information women. The Downton Fall Risk Index incorporates some of about those who sought care for the fall-related injury, and the these risk factors to predict falls in the community, residential number of falls not requiring care was unknown. However, care and hospitals. strengths of the study included the large sample size and the What this study adds: Among hospitalised older people, the 99% coverage of all care in Stockholm by the Stockholm County Downton Fall Risk Index predicts fall-related injury after dis- Council Health Care Consumption database. charge. Previous falls and gait were the items on the Downton Fall Risk Index that were most predictive. In conclusion, this study provides new insights into the DFRI and fall-related injuries. No previous studies have evaluated the individual modules of the DFRI. This study showed an increased
Research 177 Footnotes: a SAS 9.4, SAS Institute Inc, Cary, NC, USA. 13. Olsson Möller U, Kristensson J, Midlöv P, Ekdahl C, Jakobsson U. Predictive validity eAddenda: Appendix 1 can be found online at https://doi.org/ and cut-off scores in four diagnostic tests for falls – A study in frail older people at 10.1016/j.jphys.2018.05.005 home. Phys Occup Ther Geriatr. 2012;30:189–201. Ethics approval: Approved by the Regional Ethical Review Board in Stockholm, Sweden (DNR: 2013/1620-31/2). Since data in this 14. Haines TP, Hill K, Walsh W, Osbome R. Design-related bias in hospital fall risk study were only based on database information, Ethics Approval screening tool predictive accuracy evaluations: Systematic review and meta-anal- permitted informed consents not to be collected. ysis. J Gerontol A Biol Sci Med Sci. 2007;6:664–672. Competing interest: The authors declare no conflict of interest. Source of support: This research did not receive any specific 15. Meyer G, Köpke S, Haastert B, Mühlhauser I. Comparison of a fall risk assessment grant from funding agencies in the public, commercial, or not-for- tool with nurses’ judgement alone: a cluster-randomised controlled trial. Age profit sectors. Ageing. 2009;38:417–423. Acknowledgements: Nil. Provenance: Not invited. Peer reviewed. 16. National Institute for Health and Care Excellence, NICE Falls in older people, 2015. Correspondence: Elisabeth Rydwik, Department of Neurobiolo- https://www.nice.org.uk/guidance/qs86 [accessed 16/05/2016]. gy, Care Sciences and Society, Division of Physiotherapy, Karolinska Insitutet, Huddinge, Sweden. Email: [email protected] 17. Vallas B, Villars H, Abellan G, Soto ME. Overview of the MNA – Its history and challenges. J Nutr Health Aging. 2006;10:456–463. References 18. Al Snih S, Ottenbacher K, Markides K, Kuo Y-F, Eschbach K, Goodwin JS. The effect of 1. Mahoney JE, Palta M, Johnson J, Jalaluddin M, Gray S, Park S, et al. Temporal obesity on disability vs. mortality in older Americans. Arch Intern Med. 2007;167:774– association between hospitalization and rate of falls after discharge. Arch Intern 780. Med. 2000;160:2788–2795. 19. Nilsson-Ehle H, Jagenburg R, Landahl S, Svanborg A. Blood haemoglobin declines in 2. Mahoney J, Sager M, Dunham NC, Johnson J. Risk of falls after hospital discharge. J the elderly: implications for reference intervals from age 70 to 88. Eur J Haematol. Am Geriatr Soc. 1994;42:269–274. 2000;65:297–305. 3. Davenport RD, Vaidean GD, Jones CB, Chandler AM, Kessler LA, Mion LC, et al. Falls 20. Klein D, Nagel G, Kleiner Ulmer H, Rehberger B, Concin H, Rapp K. Blood pressure following discharge after an in-hospital fall. BMC Geriatr. 2009;9:53. and falls in community-dwelling people aged 60 years and older in the VHM&PP cohort. BMC Geriatr. 2013;13:50. 4. Oliver D, Daly F, Martin FC, McMurdo MET. Risk factors and risk assessment tools for falls in hospital in-patients: A systematic review. Age Ageing. 2004;33:122–130. 21. Heeschen C, Hamm C, Bruemmer J, Simoons ML, CAPTURE Investigators. Predictive value of C-reactive protein and troponin T in patients with unsafe angina: a 5. Ensrud KE, Ewing SK, Taylor BC, Fibk HA, Stone KL, Cauley JA, et al. Frailty and risk of comparative analysis. J Am Coll Cardiol. 2000;35:1535–1542. falls, fracture, and mortality in older women: the study of osteoporotic fractures. Study of Osteoporotic Fractures Research Group. J Gerontol A Biol Sci Med Sci. 22. Mooney CZ, Duval RD. Bootstrapping: A Nonparametric Approach to Statistical 2007;62:744–751. Inference. Newbury Park, CA: Sage; 1993. 6. Ensrud K. Epidemiology of fracture risk with advancing age. J Gerontol A Biol Sci Med 23. Titler MG, Shever LL, Kanak MF, Picone DM, Qin R. Factors associated with falls Sci. 2013;68:1236–1242. during hospitalization in an older adult population. Res Theory Nurs Pract. 2011;25:127–152. 7. Johansson J, Nordström A, Nordström P. Greater fall risk in elderly women than in men is associated with increased gait variability during multitasking. J Am Med Dir 24. Nilsson M, Eriksson J, Larsson B, Odén A, Johansson H, Lorentzon M. Fall risk Assoc. 2016;17(6):535–540. assessment predicts fall-related injury, hip fracture, and head injury in older adults. J Am Geriatr Soc. 2016;64:2242–2250. 8. Morris M, Osborne D, Hill K, Kendig H, Lundgren-Lindquist B, Browning C, et al. Predisposing factors for occasional and multiple falls in older Australians who live 25. Vassallo M, Vignaraja R, Sharma JC, Briggs R, Allen S. The relationship of falls to at home. Aust J Physiother. 2004;50:153–159. injury among hospital in-patients. Int J Clin Pract. 2005;59:17–20. 9. Gale CR, Cooper C, Sayer AA. Prevalence and risk factors for falls in older men and 26. Aranda-Gallardo M, Morales-Asencio JM, Canca-Sanchez JC, Toribio-Montero JC. women: The English Longitudinal Study of Ageing. Age Ageing. 2016;45:789–794. Circumstances and causes of falls by patients at a Spanish acute care hospital. J Eval Clin Pract. 2014;20:631–637. 10. Vassallo M, Vignaraja R, Sharma JC, Hallam H, Binns K, Briggs R, et al. The effect of changing practice on fall prevention in a rehabilitative hospital. J Am Geriatr Soc. 27. Kojima G. Frailty as a predictor of future falls among community-dwelling older 2004;52:335–339. people: A systematic review and meta-analysis. J Am Med Dir Assoc. 2015;16:1027– 1033. 11. Vassallo M, Stockdale R, Sharma JC, Briggs R, Allen S. A comparative study of the use of four fall risk assessment tools on acute medical wards. J Am Geriatr Soc. 28. Laflamme L, Monárrez-Espino J, Johnell K, Elling B, Möller J. Type, number or both? 2005;53:1034–1038. A population-based matched case-control study on the risk of fall-related injuries among older people and number of medications beyond fall-inducing drugs. PLoS 12. Rosendahl E, Lundin-Olsson L, Kallin K, Jensen J, Gustafson Y, Nyberg L. Prediction One. 2015;10:e0123390. of falls among older people in residential care facilities by the Downton index. Aging Clin Exp Res. 2003;15:142–147. 29. Thorell K, Ranstad K, Midlöv P, Borgquist L, Halling A. Is use of fall risk-increasing drugs in an elderly population associated with an increased risk of hip fracture, after adjustment for multimorbidity level: a cohort study. BMC Geriatr. 2014;14:131. 30. Zia A, Kamaruzzaman SB, Myint PK, Tan MP. The association of antihypertensives with postural blood pressure and falls among seniors residing in the community: a case-control study. Eur J Clin Invest. 2015;45:1069–1076. 31. Aung K, Htay T. Thiazide diuretics and the risk of hip fracture. Cochrane Database Syst Rev. 2011;10:CD005185. 32. Muir SW, Gopaul K, Montero Odasso MM. The role of cognitive impairment in fall risk among older adults: a systematic review and meta-analysis. Age Ageing. 2012;41:299–308. 33. Sherrington C, Tiedemann A. Physiotherapy in the prevention of falls in older people. J Physiother. 2015;61:54–60.
Journal of Physiotherapy 64 (2018) 159–165 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Research Exercise programs may be effective in preventing a new episode of neck pain: a systematic review and meta-analysis Tarcisio F de Campos a1[_F,TD$I]8 Chris G Maher b, Daniel Steffens c,d]D,FI91_T[$ Joel T Fuller a, Mark J Hancock a a Department of Health Professions, Macquarie University; b The University of Sydney School of Public Health, F$D[IT]_1F2 aculty of Medicine and Health, The University of Sydney; c Surgical Outcomes Research Centre (SOuRCe), Royal Prince Alfred Hospital, Sydney, Australia; D$[IF_dT]23 Sydney Medical School, The University of Sydney KEY WORDS ABSTRACT Neck pain Question: What is the effectiveness of interventions that aim to prevent a new episode of neck pain? Prevention Design: Systematic review and meta-analysis of randomised, controlled trials. Participants: People Randomised controlled trial without neck pain at study entry. Intervention: Any intervention aiming to prevent a future episode of Systematic review neck pain. Outcome measures: New episode of neck pain. Results: Five trials including a total of Meta-analysis 3852 individuals met the inclusion criteria. The pooled results from two randomised, controlled trials (500 participants) found moderate-quality evidence that exercise reduces the risk of a new episode of neck pain (OR 0.32, 95% CI 0.12 to 0.86). One of the meta-analysed trials included some co-interventions with the exercise. There was low-quality evidence from three randomised, controlled trials (3352 participants) that ergonomic programs do not reduce the risk of a new neck pain episode (OR 1.00, 95% CI 0.74 to 1.35). Conclusion: This review found moderate-quality evidence supporting the effectiveness of an exercise program for reducing the risk of a new episode of neck pain. There is a need for high-quality randomised, controlled trials evaluating interventions to prevent new episodes of neck pain. Registration: PROSPERO CRD42017055174. [de Campos TF, Maher CG, Steffens D, Fuller JT, Hancock MJ (2018) Exercise programs may be effective in preventing a new episode of neck pain: a systematic review. Journal of Physiotherapy 64: 159–165] © 2018 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 Therefore, the research question for this systematic review was: Neck pain is one of the most significant health problems What is the effectiveness of interventions that aim to prevent a worldwide.1FDI$0_2][T It has been ranked the fourth leading cause of years new episode of neck pain? lived with disability, according to the Global Burden of Disease Study.2 Mean lifetime prevalence is estimated to be 48.5% and is Method expected to increase due to the ageing population.2,3$[24_]FIDT The natural course of an episode of neck pain is favourable;4 however, This systematic review adhered to the statement for reporting recurrence rates are reported to be high,5 which contributes to systematic reviews and meta-analyses of studies that evaluate the high global social and economic burden. The Global Burden of healthcare interventions (PRISMA).13 Disease studies1,2 and Task Forces6 worldwide have called for prevention strategies for neck and back pain. Recent clinical Identification and selection of studies practice guidelines for neck pain lack recommendations for prevention.7 Consequently, a comprehensive, high-quality system- A comprehensive search of five electronic databases (MEDLINE via atic review of the literature is required to examine the effective- Ovid, EMBASE via Ovid, CINAHL, Physiotherapy Evidence Database ness of prevention strategies for neck pain. (PEDro), and The Cochrane Central Register of Controlled Trials (CENTRAL) via The Cochrane Library) was conducted from the earliest A number of systematic reviews that examined the effective- records published to 27 April, 2018. A sensitive search strategy was ness of interventions for preventing neck pain have been used based on the recommendations of the Cochrane Back and Neck published.8–12 However, these systematic reviews have important Group14 for ‘randomised controlled trials’ and ‘neck pain’, combined limitations. Some were published > 10 years ago,8,9 some did not with search terms for ‘prevention’. The detailed search strategy for publish a pre-specified study protocol,10,12 some included non- each database is presented in Appendix 1 (see eAddenda for randomised studies,10–12 and some included studies recruiting Appendix 1). In addition, reference lists of relevant reviews and symptomatic participants at study entry.9,11 There has been no included randomised, controlled trials were manually searched for systematic review investigating strategies for prevention of neck additional randomised, controlled trials, and citation tracking of all pain including only randomised, controlled trials (randomised, included trials was performed. Non-English language studies were controlled trials) and asymptomatic participants at baseline. https://doi.org/10.1016/j.jphys.2018.05.003 1836-9553/© 2018 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/).
160 de Campos et al: Interventions to prevent neck pain Box 1. Inclusion criteria. Outcome measures To be eligible for inclusion, trials had to report an outcome Design Randomised, controlled trials measure of a new episode of neck pain (eg, number of participants Participants experiencing a new episode of neck pain, or number of participants People not meeting the study’s definition of an episode of taking sick leave due to a new episode of neck pain), or a measure of neck pain or disability over the follow-up period (pain or disability neck pain at study entry measures at a single point in time did not satisfy this criterion). Intervention Any intervention aiming to prevent a new episode of neck Data extraction and analysis pain Data for each included trial were extracted by two independent Outcome measures reviewers (TFC and MJH or JTF) using a standardised data A new episode of neck pain extraction form and discrepancies were resolved by discussion A new episode of neck pain leading to care seeking, with a third author (CGM). Extracted data included the character- istics of the trial (eg, demographic characteristics of the activity limitation or work loss participants, description of the interventions, duration of treat- Measures of pain or disability over the follow-up period ment, and description of the outcomes) and outcome data. Comparisons Whenever possible, raw outcome data (number of participants The intervention group must be compared to no having a new episode of neck pain and total number of participants) in both the intervention group and control group intervention/placebo or minimal intervention were extracted. Treatment effect estimates were calculated using Studies investigating the additional benefit of a treatment methods recommended in the Cochrane Handbook for Systematic Review of Interventions.18 Attempts were made to contact authors (eg, exercise + education versus exercise alone) of included trials to clarify any relevant information or request additional data, when required. included if an appropriate translation could be obtained; otherwise, they were noted but excluded from analyses. The overall quality of evidence was assessed for each intervention contrast and rated as high, moderate, low, or very low, as Randomised, controlled trials assessing the effectiveness of recommended by the Grading of Recommendations Assessment, prevention strategies for neck pain were included if they met the Development and Evaluation (GRADE) system.19 The GRADE inclusion criteria listed in Box 1. A three-stage screening process classification was downgraded one level per study flaw, from high was used to select relevant randomised, controlled trials for this quality, if any of the following flaws were present: design limitation review. In the first stage, one reviewer (TFC) screened all titles for (more than a quarter of participants from studies with high risk of eligibility and excluded clearly irrelevant studies. In the second bias, PEDro score < 7); inconsistency of results (substantial hetero- stage, each study title and abstract was independently evaluated geneity, I2F2D]$I_[6T > 50%); and imprecision (based on a threshold by two reviewers (TFC and DS or JTF). In the third stage, the full text of < 400 participants for each pooled outcome, and also observation for each potentially eligible study was retrieved and assessed of the 95% CIs in cases of dichotomous outcomes). This review did not against the eligibility criteria by two independent reviewers (TFC consider the indirectness criterion because the eligibility criteria and DS or JTF). In cases of disagreement, a third reviewer (MJH or ensured a specific population with relevant outcomes. In addition, CGM) was consulted. the review did not assess publication bias due to insufficient study numbers. Two reviewers (TFC and MJH or DS or JTF) independently Assessment of characteristics of studies performed GRADE assessments for each treatment contrast. Risk of bias Trials considered homogeneous were grouped into the same Risk of bias was assessed using the PEDro Scale15,16]D[F25$T_I by prevention strategy category. Odds ratios (ORs) and 95% CIs were calculated and a random-effects model was used to pool estimates downloading the available scores from the PEDro database. If a using commercial meta-analysis softwarea. For randomised, con- study had not been rated on the website, two experienced PEDro trolled trials that did not report the sample size at the end of the raters scored the study. The total score on the PEDro scale is the follow-up period, the OR (95% CI) was calculated using the baseline addition of ‘yes’ (criterion is clearly satisfied) responses for Items sample size. Outcome data on short-term follow-up ( 12 months) 2 to 11 (Item 1 is not used for calculation of the total PEDro scale and long-term follow-up (> 12 months) were assessed. Statistical score because it is more related to external validity) and range from heterogeneity was assessed visually and using the I2 statistic. 0 (high risk of bias) to 10 (low risk of bias). There is evidence that the PEDro scale total score has acceptably high reliability and Results validity15,16 and Rasch analysis has confirmed that it can be used as a continuous scale.17 Flow of studies through the review Participants Overall, the comprehensive database search strategy identified Randomised, controlled trials were included if the participants 12 725 records. After screening articles by title and abstract, 114 potentially eligible studies were identified, and their full texts did not have neck pain at study entry or did not meet all of the were retrieved. In total, five trials (3852 participants) met the study’s criteria for an episode of neck pain at baseline. For example, inclusion criteria and were included in the review.20–24 The if a small proportion of participants had mild neck pain at study included studies were three randomised, controlled trials20,22,24 entry but all were working, and the study outcome was a new and two cluster-randomised, controlled trials.21,23 An outline of the episode of work absence due to neck pain, then the study would be screening and reviewing process can be seen in Figure 1. considered eligible. Characteristics of studies Intervention To be eligible for inclusion, trials had to evaluate an Risk of bias Risk of bias scores for four of the randomised, controlled intervention aiming to prevent a future episode of neck pain. The experimental group had to be compared to a group that trials20,21,23,24 were found on the PEDro database website. The fifth received no intervention, sham intervention or minimal inter- vention. Randomised, controlled trials investigating multimodal interventions were also included.
IF$DT]G1_erugi([) Research 161 Records identified from databases (n = 12 725) Additional records identified through other sources • MEDLINE search (n = 1447) (n = 8) • Embase search (n = 8077) • CINAHL search (n = 1430) • Cochrane Library search (n = 1715) • PEDro search (n = 56) 1755 duplicates removed 10 978 records screened Records excluded (Title and Abstract) (n = 10 864) Full-text articles assessed for eligibility (n = 114) Full-text articles excluded (n = 109) • not a randomised trial (n = 34) • ineligible participants (n = 58) • no eligible outcome (n = 12) • duplicates (n = 4) • could not contact authors for data (n = 1) Trials included (n = 5) • 3 randomised, controlled trials • 2 cluster-randomised, controlled trials Figure 1. Flow of studies through the review. study22 was independently assessed and scored by two experi- posture while performing daily tasks,22 manual handling aids,21 and enced PEDro raters. The mean PEDro score was 6.2 (SD 1.3) with job rotation.21 One of the two trials investigating exercise programs23 blinding, concealed allocation, and adequate follow-up being the evaluated neck muscle stretching and endurance training. This was main items scored as high risk of bias. The PEDro scale responses delivered at work twice a day for each working dayand twice a week at for individual items and the total score for each included home over the 12-month study period. The second trial investigating randomised, controlled trial are available in Table 1. exercise24 evaluated a generalised aerobic program, including: body awareness and aerobic, strengthening, stabilising and stretching Participants exercises, supplemented by health information/stress management All of the included trials (3852 participants) examined a training, and a practical examination of the workplace. The exercise program was delivered in 1-hour sessions, three times per week for working-age population with the mean age around 40 years, about 9 months, and the health information/stress management compo- 42% of whom were female. Four trials investigated prevention nent was delivered in 1-hour sessions, once per week for 4 months. strategies in a population of office workers,20–23 while one trial investigated a sample of nursing personnel.24 Table 2 provides Outcome measures details about the characteristics of each trial. Raw data on the number of new events (eg, neck pain episodes) Intervention and number of participants were available for four20,22–24 of the The included trials investigated the effect of two neck pain five trials. For these four studies, ORs (95% CI) were calculated. For the remaining study,21 an OR with 95% CI and p-value was prevention strategies: ergonomic programs20–22 and exercise pro- provided, but raw data on the number of new events were not grams.23,24 The three trials20–22 assessing ergonomic programs used presented. No eligible trials were identified that reported outcome multiple prevention strategies: adjustment of workstation,20,22 data on the number of new episodes of neck pain leading to care ergonomic redesign or modification,20–22 evaluation of participant
162 de Campos et al: Interventions to prevent neck pain Table 1 PEDro scores of included trials. [D$5TF1]S_I tudy Eligibility Random Concealed Baseline Blind Blind Blind Adequate Intention- Between- Point Total score criteria allocation allocation comparability subjects therapists assessors follow-up to-treat group estimates (0 to 10) and source analysis comparisons and variability Pillastrini N Y N Y NN Y Y Y Y Y7 et al (2007)22_TD[]1F$I4 Y Y6 Y Y5 Conlon Y YN Y NN Y N Y Y Y5 et al (2008)20 Y Y8 Tveito N YY Y NN N N N et al (2009)24 Driessen Y YN Y NN N N Y et al (2011)21 Sihawong Y Y Y Y NN Y Y Y et al (2014)23 N = no, PEDro = Physiotherapy Evidence Database, Y = yes. Table 2 Characteristics of the included trials. Study ParticipantsaF]D$T_51[I Outcome definition Experimental group Control group Time and frequency of T_FI]D$61[ Follow-up interventions period Pillastrini et al n = 99 Neck pain episode: Ergonomic intervention: No intervention Exp: One ergonomic 5 months (2007)22 Mean age = 42 yrs Indicated the presence workstation evaluation and intervention session of 12 months Gender = 71% females of neck pain on a pain adjustments, and postural 30 mins for each operator, Administrative drawing. Outcome evaluation while with twice a month personnel of the city’s assessed at 5 months’ performing daily tasks, by a supervision and Town Hall follow-up. physiotherapist. consultation of 5 to 10 mins. Conlon et al n = 206 Neck pain episode: A Ergonomic program Minimal intervention: Exp/Con: Participants were (2008)20 Mean age = 43 yrs neck disorder (implementation of an workstation with a asked to use the work Gender = 28% females diagnosed on the adapted workstation): (i) conventional mouse station while on duty. Driessen et al Office workers physical examination if an alternative mouse; (ii) a (2011)21 (engineers) neck discomfort > 5/10 conventional Minimal intervention: Exp: Participants were 12 months reported at weekly mouse + forearm support educational movies asked to use the ergonomic 9 months Tveito et al (2009)24 n = 3047 assessment. board; (iii) an alternative about prevention of program while on duty Mean age = 42 yrs mouse + forearm support neck pain (first 3 months to Sihawong et al Gender = 41% females Neck pain episode: board – aimed to prevent implement the ergonomic (2014)23 Participants recruited Presence of neck pain at musculoskeletal disorders. No intervention measures) through four Dutch least 3 on a 4-point Ergonomic program: Con: 3 Â 45 s educational companies scale (DMQ). Outcome implementation of No intervention movies. assessed every Stay@Work participatory Exp: Aerobic program n = 40 3 months. ergonomic program 3 Â week for 1 hour for Mean age = N/S (evaluation and prioritise 9 months. A total of Gender = 100% females Neck pain episode: the risk factors and 15 hours of information Employees (nursing Severity was scored on ergonomic measures to (1 hour/week for 3 months) personnel) in a nursing a 4-point scale (0 = no prevent neck pain). on stress, coping, health home for older people complaint to 3 = severe Integrated Health Program: and lifestyle and a in Norway complaints); no cut-off physical exercise (based on workplace practical point. a standardised aerobic examination. n = 567 dancing program) to Mean age = 37 yrs Neck pain episode: An improve physical capacity, Exp: Neck muscles 12 months Gender = 50% females incident episode was strength and flexibility, stretching exercise twice Office workers with defined as having including: body awareness, daily for each working day, lower neck flexion pain > 30 mm on a 100- aerobic, strength, and muscle endurance range and muscle mm visual analogue stabilising and stretching training ten times, twice endurance, from scale and had no exercises. Supplemented by per wk, during the 12-mth 12 large-scale weakness or numbness health information/stress study period. enterprises in Bangkok in the upper limb. management training and a Outcome assessed practical examination of using a diary. the work place. Exercise program: stretching exercises and endurance exercises for the neck muscles. Included daily reminder messages for the first 3 months. Con = control group, DMQ = Dutch Musculoskeletal Questionnaire, Exp = experimental group, N/S = not stated. a Participants with no neck pain at baseline. seeking, activity limitation, or days lost from work. All trials Effect of ergonomic programs on preventing neck pain followed participants for 12 months (short-term follow-ups). Three randomised, controlled trials20–22 (3352 participants) The number of new events, sample size and ORs (95% CIs) for the included randomised, controlled trials are presented in Figure 2 on were included in the meta-analysis investigating the effect of the eAddenda. A summary of the findings and quality of evidence assessment (GRADE) are presented in Table 3. ergonomic programs compared to no or minimal intervention. One randomised, controlled trial20 had four intervention arms. The minimal intervention arm was used as the control group. Each of
Research 163 Table 3 Summary of findings and quality of evidence assessment. Prevention strategy Summary of findings Quality of evidence assessment (GRADE) Trials Participants OR Study limitation Inconsistency Imprecision Overall quality (n) (n) (95% CI) Low Ergonomic program 3 335220–221[D7TI$F_] 1.00 –1 –1 None Moderate Exercise 2 50023,24 (0.74 to 1.35) None None –1a 0.32 (0.12 to 0.86) GRADE = Grading of Recommendations Assessment, Development and Evaluation. The quality of evidence was downgraded one level (–1) if the study did not comply with each GRADE criteria. a Downgraded one level (–1) due to wide CI of pooled effect. three pairwise comparisons were separately included, with the hand searching, following the PRISMA recommendations, and the number of events and participants in the control group divided out evenly among the comparisons, as recommended in the Cochrane use of the GRADE system to appraise the overall quality of the Handbook for Systematic Reviews.18 The pooled results for ergonomic programs provided low-quality evidence of no protec- evidence. The risk of bias of included trials was assessed using the tive effect (OR 1.00, 95% CI 0.74 to 1.35) when compared to no or PEDro scale, which has acceptably high reliability and validity,15,16 minimal intervention in preventing new episodes of neck pain (Figure 3, Table 3). See Figure 2 in the eAddenda for a detailed and can be used as a continuous scale for measuring risk of bias in forest plot. randomised, controlled trials.17 Effect of exercise programs on preventing neck pain This systematic review and meta-analysis had some limitations. A Two randomised, controlled trials23,24$DF2[T]I_7 (500 participants) were small number of trials were included, despite the comprehensive included in the meta-analysis investigating the effect of exercise search strategy. The majority of the trials20–23 evaluated the programs compared to no intervention control. In one randomised, effectiveness of the intervention in office workers; thus, the controlled trial23 the intervention was restricted to exercise, while in generalisability of these findings to other populations is unclear. the other randomised, controlled trial,24 exercise was the primary Authors could not be contacted to gather information for one potentially eligible trial.25 Some included trials were not regis- intervention, supplemented by health information/stress manage- tered,20,22,24 and did not present a pre-specified published proto- col,20,22,24 leading to potential reporting bias. The two trials23,24 ment training, and a practical examination of the workplace. The evaluating exercise had different approaches to exercise: in one trial,23 the program was confined to neck exercises, whereas the other trial24 pooled results provided moderate-quality evidence of reduced risk evaluated a generalised whole body exercise program, supplemented of a future neck pain episode (OR 0.32, 95% CI 0.12 to 0.86) (Figure 4, by health information/stress management training, which means Table 3). See Figure 2 in the eAddenda for a detailed forest plot. there is uncertainty about which approach to recommend. Discussion It is believed that the current systematic review with meta- Five randomised, controlled trials investigating two interven- analysis is the first to have included only randomised, controlled trials evaluating prevention strategies for neck pain that have tion strategies to prevent neck pain were deemed eligible to be included asymptomatic participants at baseline (or at least included in this systematic review. The review found moderate- participants that did not meet all of the study’s criteria for an quality evidence that an exercise program substantially reduces episode of neck pain at baseline). Previous systematic reviews that the risk of a new episode of neck pain (OR 0.32, 95% CI 0.12 to 0.86). have investigated the effectiveness of interventions to prevent This evidence was derived from two trials that included neck pain have included trials with symptomatic participants at study entry.8–12$8I_]FDT2[ Some are also out of date,8–10 and some include 500 participants.23,24[D$F2T_7I] Pooled results from three trials20–22 with sub-optimal study designs (such as non-randomised trials or quasi-experimental studies).9,10 3352 participants produced low-quality evidence that ergonomic A recent review12 investigated the effectiveness of exercise for programs do not reduce the risk of a new episode of neck pain (OR preventing upper extremity musculoskeletal disorders, including 1.00, 95% CI 0.74 to 1.35). neck pain.12 That review found evidence of limited to strong quality The strengths of this systematic review included the use of a that exercise could prevent upper extremity symptoms; however, ps]IF$DT)3_erugi([G trrea-tsepgeyciufiseindgpmroutoltcipollereegleicstterorendic on PROSPERO, sensitive search it included studies with symptomatic participants at baseline (ie, databases with supplementary the studies evaluated treatment, not prevention), and also Study OR (95% CI) Random included study designs other than randomised, controlled trials. Furthermore, that review did not differentiate neck pain from o$DFT)4IG]_erugi[( ther body regions (eg, neck/shoulder) when assessing trials for Pillastrini22 OR (95% CI) Random Conlon20 Study Conlon20 Tveito24 Conlon20 Sihawong23 Driessen21 Pooled Pooled 0.01 0.1 1 10 100 0.01 0.1 1 10 100 Favours Exp Favours Con Favours Exp Favours Con Figure 4. Odds ratio for neck pain episode in trials of exercise, estimated by pooling data from two trials (n = 500). Note that one study24 administered exercise Figure 3. Odds ratio for neck pain episode in trials of ergonomic programs, supplemented by health information/stress management training, and a practical estimated by pooling data from three trials (n = 3352). examination of the workplace. 4[1_TD$IF]Exp = experimental group, Con = control group. Exp = experimental group, Con = control group.
164 de Campos et al: Interventions to prevent neck pain the effectiveness of exercise prevention strategies. As a result of the Conflict of interest: Nil. stricter inclusion criteria, the current review identified a substan- tially smaller number of randomised, controlled trials. Source(s) of support: Nil. A Cochrane review26 (with 13 randomised, controlled trials Acknowledgements: Mr Tarcisio F de Campos has a PhD involving 2397 workers) reported that most ergonomic interven- tions were not effective in preventing work-related upper limb and scholarship from Macquarie University (Macquarie University neck musculoskeletal disorders, which is in line with the results from the current review. However, one meta-analysis in the Research Excellence Scholarship (MQRES). Professor Chris G Maher Cochrane review, including two randomised, controlled trials,20,27 found moderate-quality evidence that the use of ergonomic holds a fellowship, Program grant and Centre for Research equipment may reduce the incidence of neck/shoulder pain. The difference in inclusion criteria, especially the inclusion of studies Excellence grant funded by Australia’s National Health and Medical that did not differentiate neck and shoulder pain, and studies of Research Council. participants with pain at study entry, may explain the somewhat different conclusions between the Cochrane review and the Provenance: Not invited. Peer reviewed. current systematic review and meta-analysis. Correspondence: Tarcisio F de Campos, Department of Health The results of the present systematic review on prevention of neck pain are similar to the results of a recently published Professions, 103D_$T]FMI[ acquarie University, Sydney, Australia. Email: systematic review on prevention of low back pain.28 Steffens and [email protected] colleagues also found that an exercise program alone (RR 0.65, 95% CI 0.50 to 0.86) or in combination with education (RR 0.55, References 95% CI 0.41 to 0.74) are effective for preventing low back pain. For a more direct comparison with the result of the Steffens review, 1. Vos T, Allen C, Arora M, Barber RM, Bhutta ZA, Brown A, et al. Global, regional, and the current meta-analysis for the exercise intervention was re- national incidence, prevalence, and years lived with disability for 310 diseases and calculated as RR (instead of OR as in Figures 2 and 4). Exercise injuries, 1990–2015: a systematic analysis for the Global Burden of Disease Study reduced the risk of a new episode of neck pain by 53% (RR 0.47, 2015. Lancet. 2016;388(10053):1545–1602. 95% CI 0.32 to 0.68). The calculation for the pooled RR result for the exercise intervention contrast is presented in Figure 5 on the 2. Hoy D, March L, Brooks P, Blyth F, Woolf A, Bain C, et al. The global burden of neck eAddenda. pain: estimates from the Global Burden of Disease 2010 Study. Ann Rheum Dis. 2014;73:1309–1315. Although the current systematic review found that exercise programs are likely to roughly halve the risk of a new episode of 3. Fejer R, Kyvik KO, Hartvigsen J. The prevalence of neck pain in the world popula- neck pain, the quality of the evidence is moderate and further high- tion: a systematic critical review of the literature. Eur Spine J. 2006;15:834–848. quality randomised, controlled trials are needed. One randomised, controlled trial24 evaluating exercise provided participants with 4. Leaver AM, Maher CG, McAuley JH, Jull G, Latimer J, Refshauge KM. People seeking health information/stress management training and a workplace treatment for a new episode of neck pain typically have rapid improvement in assessment as part of the intervention, which means there is symptoms: an observational study. J Physiother. 2013;59:31–37. uncertainty about the effectiveness of the exercise alone. The durations of the exercise programs were quite long – 9 months24 5. Carroll LJ, Hogg-Johnson S, van der Velde G, Haldeman S, Holm LW, Carragee EJ, et and 12 months23 – which needs to be borne in mind when al. Course and prognostic factors for neck pain in the general population: Results of considering this therapy. Additionally, there are no outcomes the Bone and Joint Decade 2000–2010 Task Force on Neck Pain and its Associated beyond 12 months, so the long-term effect is unknown. Further- Disorders. J Manipulative Physiol Ther. 2009;32:S87–S96. more, high-quality randomised, controlled trials are needed to investigate the potential benefit of interventions to prevent 6. Côté P, van der Velde G, Cassidy JD, Carroll LJ, Hogg-Johnson S, Holm LW, et al. The episodes of neck pain leading to care seeking, activity limitation, burden and determinants of neck pain in workers: Results of The Bone and Joint and days lost from work. Decade 2000–2010 Task Force on Neck Pain and Its Associated Disorders. J Manipulative Physiol Ther. 2009;32:S70–S86. In conclusion, the results of this review found moderate- quality evidence that an exercise program reduces the risk of a 7. Blanpied P, Gross A, Elliott JM, Devaney LL, Clewley D, Walton DM, et al. Neck Pain: new episode of neck pain. Ergonomic strategies do not appear to Revision 2017 – Clinical Practice Guidelines. J Orthop Sports Phys Ther. 2017;47:A1–A83. prevent neck pain. Additional trials with longer-term follow-up would more clearly establish the public health implications of 8. Linton SJ, van Tulder MW. Preventive interventions for back and neck pain this result. problems: what is the evidence? Spine. 2001;26:778–787. What was already known on this topic: Neck pain is 9. Brewer S, Eerd D, Amick III BC, Irvin E, Daum KM, Gerr F, et al. Workplace common, but clinical practice guidelines lack recommenda- interventions to prevent musculoskeletal and visual symptoms and disorders tions regarding prevention. Past systematic reviews of preven- among computer users: a systematic review. J Occup Rehabil. 2006;16:317–350. tive interventions for neck pain have had important flaws such as the inclusion of non-randomised studies. 10. Boocock MG, McNair PJ, Larmer PJ, Armstrong B, Collier J, Simmonds M, et al. What this study adds: Exercise programs substantially re- Interventions for the prevention and management of neck/upper extremity muscu- duce the risk of a new episode of neck pain. The evidence for loskeletal conditions: a systematic review. Occup Environ Med. 2007;64(5):291–303. this is of moderate quality and one of the included trials included some co-interventions with the exercise. Ergonomic 11. Kennedy CA, Amick BC, Dennerlein JT, Brewer S, Catli S, Williams R, et al. Systematic programs do not appear to significantly reduce the risk of a new review of the role of occupational health and safety interventions in the prevention episode of neck pain, but the evidence for this is of low quality. of upper extremity musculoskeletal symptoms, signs, disorders, injuries, claims and lost time. J Occup Rehabil. 2010;20:127–162. Footnote: a Comprehensive Meta-analysis, version 2.2.064, Biostat, Englewood, NJ, USA. 12. Van Eerd D, Munhall C, Irvin E, Rempel D, Brewer S, Van Der Beek AJ, et al. Effectiveness of workplace interventions in the prevention of upper extremity eAddenda: DITF]A2_$[9 ppendix 1, Figures 2 and 5 can be found online at musculoskeletal disorders and symptoms: an update of the evidence. Occup https://doi.org/10.1016/j.jphys.2018.05.003 Environ Med. 2016;73:62–70. Ethics approval: Not applicable as this is a systematic review. 13. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009;339:b2700. 14. Furlan DA, Malmivaara GA, Chou R, Maher CG, Deyo RA, Schoene M, et al. 2015 Updated method guideline for systematic reviews in the Cochrane Back and Neck Group. Spine. 2015;40:1660–1673. 15. Macedo LG, Elkins MR, Maher CG, Moseley AM, Herbert RD, Sherrington C. There was evidence of convergent and construct validity of Physiotherapy Evidence Database quality scale for physiotherapy trials. J Clin Epidemiol. 2010;63:920–925. 16. Maher CG, Sherrington C, Herbert RD, Moseley AM, Elkins M. Reliability of the PEDro scale for rating quality of randomized controlled trials. Phys Ther. 2003;83:713–721. 17. de Morton NA. The PEDro scale is a valid measure of the methodological quality of clinical trials: a demographic study. Aust J Physiother. 2009;55:129–133. 18. Higgins J, Green S, eds. Cochrane Handbook for Systematic Reviews of Interven- tions, Version 5.1.0. The Cochrane Collaboration; Updated March 2011. http:// handbook-5-1.cochrane.org/ [accessed 18 May 2018]. 19. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommen- dations. BMJ. 2008;336(7650):924–926. 20. Conlon CF, Krause N, Rempel DM. A randomised controlled trial evaluating an alternative mouse and forearm support on upper body discomfort and musculo- skeletal disorders among engineers. Occup Environ Med. 2008;65:311–318. 21. Driessen MT, Proper KI, Anema JR, Knol DL, Bongers PM, van der Beek AJ. The effectiveness of participatory ergonomics to prevent low-back and neck pain – results of a cluster randomized controlled trial. Scand J Work Environ Health. 2011;37:383–393. 22. Pillastrini P, Mugnai R, Farneti C, Bertozzi L, Bonfiglioli R, Curti S, et al. Evaluation of two preventive interventions for reducing musculoskeletal complaints in opera- tors of video display terminals. Phys Ther. 2007;87:536–544. 23. Sihawong R, Janwantanakul P, Jiamjarasrangsi W. Effects of an exercise programme on preventing neck pain among office workers: a 12-month cluster-randomised controlled trial. Occup Environ Med. 2014;71:63–70.
Research 165 24. Tveito TH, Eriksen HR. Integrated health programme: a workplace randomized body pain and incident musculoskeletal disorders among computer operators. controlled trial. J Adv Nurs. 2009;65:110–119. Occup Environ Med. 2006;63:300–306. 28. Steffens D, Maher CG, Pereira LS, Stevens ML, Oliveira VC, Chapple M, et al. 25. Robertson MM, Ciriello VM, Garabet AM. Office ergonomics training and a sit-stand Prevention of low back pain: a systematic review and meta-analysis. JAMA. workstation: Effects on musculoskeletal and visual symptoms and performance of 2016;176:199–208. office workers. Appl Ergon. 2013;44:73–85. Websites 26. Hoe VC, Urquhart DM, Kelsall HL, Sim MR. Ergonomic design and training for preventing work-related musculoskeletal disorders of the upper limb and neck in PEDro www.pedro.org.au adults. Cochrane Database Syst Rev. 2012;8:CD008570. 27. Rempel DM, Krause N, Goldberg R, Benner D, Hudes M, Goldner GU. A randomised controlled trial evaluating the effects of two workstation interventions on upper
Search
