Australian Journal Of Physiotherapy

tests and treatments. Use by physiotherapists Evidence source 152 Editorial ts overuse in society 69% use continuous passive motion for Systematic review of physiotherapists’ Yes patients following total knee treatment choices, including two replacement studies on total knee replacement3 No Yes 39% use whirlpools (indication was not Survey of 520 physical therapists in the Yes specified) USA8 13 to 79% use heat therapy for Systematic review of physiotherapists’ No musculoskeletal conditions treatment choices, including 94 studies on musculoskeletal conditions3 Yes No data Yes N/A Yes No data No N/A No 14 to 67% provide electrotherapy for LBP No Systematic review of physical 50% provide incentive spirometry to therapists’ treatment choices, patients following cardiac, thoracic or including 48 studies on LBP3 upper abdominal surgery Systematic review of physical therapists’ treatment choices, 79% use manual therapy and 14% use including six studies on cardiac or massage for frozen shoulder thoracic surgery9 Prospective cohort study of 125 25% order x-rays for non-specific LBP physiotherapists in the Netherlands treating 29 patients with frozen 27 to 70% of emergency department shoulder10 physicians do not use the Canadian C- Survey of 203 physiotherapists in Spine Rules, although there are no data Australia11 on use among physiotherapists 37% of emergency department physicians Survey of 1,150 emergency physicians do not use the Ottawa Ankle Rules, from the USA, Canada, UK and although there are no data on use among Australia12 physiotherapists Retrospective audit of 260 clinical records from academic emergency departments13

Table 1 (Continued) Recommendationa Test or treatment Targe Association Treatment Do not use passive interventions Brazilian Association of as a stand-alone approach for a Traumato-Orthopedic Physiotherapy long time in patients with low back pain Italian Association Do not use braces or belts to Treatment of Physiotherapists prevent or treat patients with Test chronic low back pain Test Do not suggest imaging exams Treatment for low back pain unless you Treatment suspect serious conditions such as trauma, cancer, infection, Treatment cauda equina syndrome and inflammatory conditions like Treatment ankylosing spondylitis and Treatment rheumatoid arthritis Do not use imaging exams to base your treatment of patients with chronic low back pain Do not recommend the use of insoles as prevention, as well as treatment, for patients with low back pain Do not use continuous passive mobilisation devices for postsurgical treatment of knee ligament reconstructions or uncomplicated hip or knee arthroplasties Do not use mechanical or manual traction as a single treatment or in combination with other treatments in patients with low back pain in the presence or absence of radicular pain Do not use specific exercises of selective strengthening of the vastus medialis obliquus in patellofemoral pain syndrome Do not use ultrasound therapy for rotator cuff tendinopathy, ankle sprains and low back pain Do not teach or have patients Treatment with acute or chronic respiratory diseases practise diaphragmatic breathing COPD = chronic obstructive pulmonary disease, LBP = low back pain, UK = United Kingdom, USA = United States of America. a All recommendations are ‘Don’t’ recommendations that do not provide an alternative.

ets overuse in society Use by physiotherapists Evidence source Editorial 153 Yes 5 to 67% of physiotherapists use passive Systematic review of physiotherapists’ interventions such as acupuncture, treatment choices, including 48 studies Yes traction, cold therapy, heat therapy, on LBP3 No ultrasound, electrical stimulation, transcutaneous electrical nerve Systematic review of physiotherapists’ No stimulation and inferential current for treatment choices, including 48 studies Yes LBP on LBP3 Yes 2 to 24% of physiotherapists use external Survey of 203 physiotherapists in support such as corsets, belts, braces or Australia11 Yes taping for subacute or chronic LBP 25% order x-rays for non-specific LBP Survey of 203 physiotherapists in Yes Australia11 Yes 25% order x-rays for non-specific LBP Yes Survey of 186 physiotherapists in . 24% give advice on footwear for LBP India14 69% use continuous passive motion for Systematic review of physiotherapists’ patients following total knee treatment choices, including two replacement studies on total knee replacement3 5 to 45% use traction for LBP Systematic review of physiotherapists’ treatment choices, including 48 studies No data on LBP3 26 to 44% use electrotherapy for shoulder N/A pain, and 14% use electrotherapy for acute ankle sprains Systematic review of physiotherapists’ treatment choices, including seven 100% prescribe breathing exercises for studies on shoulder pain and five patients with stable COPD and 72% studies on foot or ankle pain3 prescribe breathing exercises for patients Systematic review of physiotherapists’ with an acute exacerbation of COPD treatment choices, including four studies on COPD9

154 Editorial orthopaedic surgeons target low-value surgery; most recommenda- professionals (eg, imaging, surgery and medication) and more on tions target diagnostic tests. interventions routinely provided by physiotherapists. Another key issue is recommendations that are not measurable. A Ensure adherence to recommendations is measurable recent study evaluated the feasibility of translating Choosing Wisely recommendations into performance measures16 and found that six Improving the surveillance of treatment choices is a key issue in (13%) of the 45 assessed recommendations could be translated. Bar- physiotherapy. Without data on what treatments physiotherapists riers to translating recommendations included not being able to provide, it is impossible to measure adherence to Choosing Wisely capture relevant data in electronic health records (71% of recom- recommendations. Efficient ways to extract data on treatment choices mendations), use of non-specific terminology in recommendations from electronic medical records would be a major innovation within (18%), weak evidence to support recommendations (4%) and limited the physiotherapy profession. This could be achieved by using quality potential benefit on resource use and spending (16%).16 Similar bar- indicators to track concordance between treatment choices and riers apply to Choosing Wisely recommendations from physiotherapy guideline recommendations20 and linking electronic health records to associations. In many parts of the world, there is no routine system in a data dashboard where clinicians’ treatment choices can be tracked place for collecting data on treatments that physiotherapists provide. in real time.21 However, in the absence of such data, routine collec- Current evidence on physiotherapy practice relies on inefficient tion of physiotherapists’ treatment choices, through administrative means of collecting data, such as surveys and audits of clinical notes.9 databases and audits of clinical notes, could be performed to evaluate Furthermore, some Choosing Wisely recommendations from phys- the impact of the campaign. There is also a need for recommenda- iotherapy associations use qualified or vague language, which makes tions that can be translated into performance measures and that are it almost impossible to measure adherence, for example: ‘Avoid using detailed enough to measure. This could be achieved by increasing the electrotherapy modalities in the management of patients with low level of detail in current and future recommendations and avoiding back pain’; ‘Don’t routinely use incentive spirometry after upper vague language that gives physiotherapists latitude for not changing abdominal and cardiac surgery’; and ‘Don’t prescribe under-dosed their practice (eg, ‘Don’t routinely provide .’). strength training programs for older adults’. Optimise the language of the recommendations The challenge of measuring adherence to many Choosing Wisely recommendations could partially explain the minimal use of proper Marked variation in the language of Choosing Wisely recommen- implementation strategies to disseminate the recommendations. dations highlights the need to better understand how language could Although Choosing Wisely recommendations from physiotherapy maximise clinicians’ willingness to adopt them. For example, a content associations have been passively disseminated through websites and analysis of 1,293 Choosing Wisely recommendations found that only peer-reviewed journal articles,7 there is yet to be any active imple- 4% of recommendations provide alternatives to facilitate the de- mentation of the recommendations. Passive dissemination of evi- adoption of low-value care.15 This appears to be at odds with physio- dence rarely influences clinicians’ behaviour;17 hence, it is unlikely therapists’ preference for recommendations that offer encouragement the existing Choosing Wisely recommendations have had an impor- to provide high-value care.7 Recommendations that provide evidence- tant impact on physiotherapy practice. based alternatives to low-value care might increase physiotherapists’ willingness to adopt Choosing Wisely recommendations. The strength Solutions to maximise the impact of Choosing Wisely on of language used in Choosing Wisely recommendations also deserves physiotherapy practice attention. The above content analysis found that 42% of Choosing Wisely recommendations use qualified language (eg, ‘Don’t As the physiotherapy workforce expands globally (eg, in routinely.’, ‘Avoid.’). Some physiotherapists believe that qualified Australia18 and the United States19) and guidelines continue to pri- recommendations are more useful, as they are not too prescriptive and oritise non-pharmacological interventions for conditions commonly allow physiotherapists to use their clinical reasoning, while some managed by physiotherapists, it is vital that the profession takes the believe that unqualified recommendations (eg, ‘Don’t provide.’) are problem of healthcare overuse seriously. World Physiotherapy rep- more likely to change practice as they do not give physiotherapists the resents over 120 physiotherapy associations globally and could play a option to provide low-value care in ‘special’ cases.7 Before future rec- vital role in promoting Choosing Wisely and encouraging physio- ommendations are published and implemented, it is important to therapy associations to join the campaign. As new associations join explore how different aspects of language influence clinicians’ will- the campaign, we encourage them to learn from the missed oppor- ingness to adopt Choosing Wisely recommendations. tunities of existing member associations and consider the following suggestions for how to maximise the impact of Choosing Wisely on Active implementation strategies with proper evaluation future physiotherapy practice. The importance of active implementation strategies for changing Target recommendations appropriately clinician behaviour is well-established. A systematic review of nine trials (n = 1,815 physiotherapists)9 and an umbrella review (n = 19 The first step to maximising the impact of Choosing Wisely is systematic reviews including various health professionals)17 found perhaps the simplest: ensure that physiotherapy associations publish several active implementation strategies that were effective at recommendations that target inappropriate tests and treatments that changing clinicians’ behaviour, such as educational outreach visits, are frequently provided by physiotherapists. A 2019 systematic re- interactive educational meetings and multicomponent interventions view of 94 studies summarised the proportion of physiotherapists (eg, combination of peer assessment, performance monitoring, audit who provide various interventions for musculoskeletal conditions,3 and feedback). Developing and evaluating similar active strategies to and categorised interventions as appropriate or inappropriate. As increase adoption of Choosing Wisely recommendations (eg, the most comprehensive resource on physiotherapists’ treatment increasing physiotherapists’ access to education on the most practices, this review should be used to identify frequently provided currently recommended physiotherapy interventions) could be an low-value practices in physiotherapy such as acupuncture/dry important starting point to reduce overuse in physiotherapy. It would needling for low back pain (6 to 45% of physiotherapists) and elec- be a mistake for conversations about value in healthcare to only trotherapy for knee osteoarthritis (21 to 43%). This will also ensure consider research evidence. that Choosing Wisely lists from physiotherapy associations focus less on tests and treatments mostly provided by other health

Editorial 155 There are also societal and patient perspectives. Misaligned care is Source(s) of support: PK is a PhD student with a scholarship. JZ is an example of low-value care from patients’ perspective. It occurs supported by an NHMRC Investigator Grant (APP1194105). CM is when clinicians provide care that does not align with patients’ values supported by an NHMRC Principal Research Fellowship (APP1103022) and preferences.22 Explaining the evidence supporting various op- and NHMRC Program Grant (ID APP1113532). tions (including evidence against low-value interventions) could help align patients’ values and preferences with evidence-based care and Acknowledgements: Nil. improve patient satisfaction. It could also help foster a positive Provenance: Not invited. Peer-reviewed. patient-clinician relationship, which is an independent driver of the Correspondence: Priti Kharel, Institute for Musculoskeletal quality of care.23 However, in cases where patients’ preferences Health, Sydney Local Health District, Sydney, Australia. Email: remain discordant with the evidence, clinicians face the challenge of [email protected] balancing patient satisfaction and therapeutic alliance with the need to provide evidence-based care.24 References Conclusion 1. Choosing Wisely. An initiative of the ABIM Foundation. http://www. choosingwisely.org. Accessed 20 February, 2020. Choosing Wisely is rapidly expanding across various fields of medicine and health worldwide. The problem of overuse is as rele- 2. Zadro JR, et al. J Orthop Sports Phys Ther. 2020;50:113–115. vant in physiotherapy as it is in medicine, yet little is being done to 3. Zadro J, et al. BMJ Open. 2019;9:10. combat overuse in physiotherapy. Poor engagement with Choosing 4. Barry C, et al. Am J Clin Pathol. 2020;153:94–98. Wisely from physiotherapy associations and poorly targeted recom- 5. Hicks CW, et al. Am J Surg. 2017;214:571–576. mendations from those with Choosing Wisely lists is limiting the 6. Himelfarb J, et al. BMJ Open Qual. 2019;8:e000470. potential impact of the campaign. To ensure the rapidly expanding 7. Zadro J, et al. BMJ Open. 2019;9:9. global physiotherapy workforce is an opponent of overuse, more 8. Greco JL, et al. Phys Ther Rev. 2018;23:116–123. physiotherapy associations need to partner with Choosing Wisely and 9. Zadro JR, et al. Phys Ther. 2020;100:1516–1541. target recommendations at key examples of overuse in physio- 10. Karel Y, et al. Physiother. 2017;103:369–378. therapy. Recommendations need to be measurable, use language that 11. Keating JL, et al. Man Ther. 2016;22:145–152. maximises physiotherapists’ willingness to follow them and be 12. Eagles D, et al. Acad Emerg Med. 2008;15:1256–1261. implemented using active strategies (beyond passive dissemination). 13. Beutel BG, et al. West J Emerg Med. 2012;13:366–372. Ongoing evaluation of implementation strategies will ensure that the 14. Fidvi N, et al. Physiother Res Int. 2010;15:150–159. most effective and efficient strategies are used to translate Choosing 15. Zadro JR, et al. BMC Health Serv Res. 2019;19:707. Wisely recommendations into practice and reduce overuse in 16. Shetty KD, et al. Healthcare. 2015;3:24–37. physiotherapy. 17. Bero L, et al. BMJ. 1998;317:465–468. 18. Australian Health Practitioner Regulation Agency. Annual Reports 2010/11 to 2017/ Ethics approval: Nil. Competing interests: Nil. 18. https://www.ahpra.gov.au/Publications/Annual-reports/Annual-report-archive. aspx. Accessed 3 November, 2020. 19. American Physical Therapy Association. A model to project the supply and demand of physical therapists 2010–2025. https://www.apta.org/contentassets/a442158 0b60141b2a4c80ff9a05fc0a3/projecting-the-supply-and-demand-of-physical- therapists.pdf. Accessed 21 October, 2020. 20. Edwards JJ, et al. Rheumatology. 2015;54:844–853. 21. Machado GC, et al. Brit J Sports Med. 2020. bjsports-2019-101622. 22. Michaleff ZA, et al. J Orthop Sports Phys Ther. 2020;50:588–591. 23. Street RL, et al. Patient Educ Couns. 2008;74:295–301. 24. Saini V, et al. Lancet. 2017;390:178–190.

Journal of Physiotherapy 67 (2021) 163–176 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Invited Topical Review Physiotherapy management of Parkinson’s disease Marco YC Pang Department of Rehabilitation Sciences, Faculty of Health and Social Sciences, The Hong Kong Polytechnic University, Hong Kong KEY WORDS [Pang MYC (2021) Physiotherapy management of Parkinson’s disease. Journal of Physiotherapy 67:163– 176] Parkinson’s disease © 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under Exercise Review the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Rehabilitation Physical therapy Introduction caused by dysfunctions in the non-dopaminergic system, as it is often resistant to dopamine.7 The cardinal motor symptoms may give rise Parkinson’s disease (PD) is the second most common neurode- to a variety of secondary impairments such as muscle weakness (ie, generative disorder after Alzheimer’s disease.1 Physiotherapists play a reduced capacity of the muscles to generate force),8,9 altered gait (eg, very important role in the rehabilitation of people with PD, particu- reduced walking speed, stride length, cadence and level of indepen- larly in relation to the management of motor symptoms, promotion of dence in walking),9,10 reduced aerobic capacity (ie, reduced peak regular physical exercise and prevention of secondary impairments oxygen consumption rate and endurance)11–14 and falls.15,16 These and complications. This review summarises the: motor and non- secondary impairments, together with worsening of the motor motor impairments and secondary complications experienced by symptoms as the disease progresses, may trigger a vicious cycle of people with PD; research evidence about different interventions further decline in physical activity level, activity and participation. commonly used in the physiotherapy management of PD; and im- plications for research and practice. Apart from motor impairments, PD is also characterised by non- motor symptoms, including fatigue, depression, anxiety, sleep Epidemiology of Parkinson’s disease disturbance, cognitive impairments, behavioural issues and bladder/ bowel dysfunction.17 While the basal ganglia are also involved in Parkinson’s disease affects over 6 million people globally,2 with a regulating the non-motor functions such as the behavioural, cognitive prevalence of 51 to 439 per 100,000 people and an incidence of 2 to and emotional functions, increasing evidence suggests that the pe- 28 per 100,000 people, based on door-to-door surveys.1 Men are ripheral autonomic nervous system may be where the disease begins slightly more affected than women.3 Both the prevalence and inci- before the pathology spreads to the lower brainstem and eventually dence of PD increase with age, peaking in the seventh and eighth affects the substantia nigra.18 This may explain why some of the decades.1 Available data also suggest higher prevalence and incidence impairments in non-motor functions precede those of the motor of PD in western countries than eastern countries.1,4 Because of the functions by years or even decades.19 As the motor and non-motor rapidly ageing population, the number of people affected by PD is impairments continue to worsen, there may be severe disabilities in projected to double to over 12 million by 2040,2 which will inevitably different aspects of function, resulting in seriously compromised exacerbate the burden on healthcare systems and society. quality of life.20 Clinical features of Parkinson’s disease Physiotherapy management of Parkinson’s disease The cardinal motor symptoms of PD are bradykinesia, rigidity, This section summarises the evidence for a range of physiotherapy tremor and postural instability.5 The onset of the cardinal motor interventions that have been investigated for their effect on PD. symptoms arises from the loss of dopaminergic neurons of the sub- Where possible, the evidence from multiple similar trials has been stantia nigra pars compacta, leading to depletion of dopamine in the meta-analysed. The resulting evidence has been summarised in striatum. Therefore, the inhibitory influence from the basal ganglia to Figure 1. other brain regions that are involved in the control and execution of voluntary movements (eg, thalamus, brainstem and supplementary Meta-analytic approach motor area) becomes exaggerated, which may account for bradyki- nesia and rigidity.6 On the other hand, balance impairment could be In all meta-analyses presented in this review, only those trials with a comparison group that allowed the direct estimation of the https://doi.org/10.1016/j.jphys.2021.06.004 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

164 Pang: Physiotherapy management of Parkinson’s disease Benefit from the intervention Uncertain effects Aerobic training Resistance training (extremities) Yoga Nordic walking Treadmill training Dance Balance training Respiratory muscle training Action observational training Robot-assisted gait training tDCS (anodal stimulation of M1) Systematic review of Cueing and movement strategy training Aquatic exercise tDCS as an adjunct to walking training multiple RCTs Virtual reality ª Telerehabilitation Dual-task training Tai Chi rTMS (high-frequency stimulation of M1) Multiple high-quality RCTs Combined exercise training Single high-quality RCT or Ai Chi Qigong Virtual reality b several low-quality RCTs Figure 1. Interventions for management of Parkinson’s disease with associated levels of evidence. RCT = randomised controlled trials, M1 = primary motor cortex, rTMS = repetitive transcranial magnetic stimulation, tDCS = transcranial direct current stimulation. a centre-based, supervised. b home-based, minimally supervised. effects of the experimental intervention were included (ie, experi- wide confidence interval (Figure 2C). Aerobic exercise had little or no mental intervention versus no intervention/sham intervention/ effect on quality of life, as measured by the Parkinson’s Disease attentional control, or experimental intervention plus other in- Questionnaire-39 (PDQ-39) (Figure 2D). For detailed forest plots, see tervention(s) versus the same other intervention(s) only). The mean Figure 3 on the eAddenda. In all the above meta-analyses, the difference (MD), rather than the standardised mean difference (SMD), methodological quality of the majority of the studies was good was used in the analyses whenever possible because absolute mea- (PEDro scores  6). In addition to the effects on aerobic capacity and sures of effect may be more easily interpreted by clinicians than motor outcomes, a recent systematic review by Schootemeijer et al relative effects. The random-effects model was used if the I2 value found that aerobic exercise improved bone health and decreased was . 50% (ie, substantial heterogeneity); otherwise, the fixed-effect incidence of cardiovascular disease and mortality in people with model was used. PD;34 however, its effects on non-motor impairments were conflicting. As PD is chronic and progressive in nature, physiotherapy inter- vention involves management of motor impairments, promotion of The few aerobic exercise trials with a follow-up period had mixed regular physical exercise and prevention of secondary impairments results. Qutubuddin et al found no clear between-group difference in and complications.21,22 Physiotherapy may also have an important balance, motor symptoms and quality of life immediately after 8 role in delaying disease progression.22 Exercise intervention is a very weeks of training (intensity 61 to 80% of maximum heart rate, fre- important element in physiotherapy for people with PD.21,22 quency two sessions per week) and also at a 4-month follow-up.32 However, the methodological quality was only fair (PEDro score = Aerobic exercise 5) and the sample size was small (23 participants). In a recent small- scale randomised trial by Arfa-Fatollahkhani et al with 20 partici- Aerobic capacity, as measured by the maximal oxygen consump- pants,35 10 weeks of aerobic exercise at moderate intensity (intensity tion rate, is impaired in people with PD, particularly those with 60% heart rate reserve, frequency two sessions per week) resulted in moderate to severe PD.12,14 Compromised aerobic capacity among improvements in the Timed Up and Go (TUG) test and 6MWT that people with PD is also reflected in reduced walking endurance on were sustained for 2 months after the training had ended. More tests such as the 6-minute walk test (6MWT)11 and increased phys- research is required to explore the long-term effects of aerobic iologic cost when performing walking and other activities of daily exercise. living.23,24 Aerobic exercise may be a viable option to address these issues. Different modes of aerobic exercise have been studied in the Graded resistance exercise training PD population, with the most common ones being cycling on a sta- tionary bicycle, followed by walking on a treadmill.25 The final target People with PD have lower muscle strength than their peers exercise intensity was moderate to high, mostly at 50 to 80% heart without PD.8,36 Severe sarcopenia affects 20% of people with PD.37 rate reserve. The duration of each exercise session was 30 to 50 mi- Loss of muscle mass and strength is a key factor in the develop- nutes for most studies, while the frequency and duration of the ment of secondary osteoporosis,38 reduced functional performance36 program was typically three to five sessions per week for 8 to 24 and falls in this population.39 Graded resistance exercise training has weeks.25 been used to tackle the problem of muscle weakness and associated functional limitations in PD.40,41 A 2014 systematic review by Shu et al examined the effects of aerobic exercise in PD.25 In the current review, updated meta- Updated meta-analyses were generated by adding the results of analyses were performed by first extracting the relevant trials from randomised trials published after 201342–50 to the relevant studies Shu et al’s meta-analyses, and then adding the data obtained from extracted from the systematic reviews by Saltychev et al40 and Chung aerobic exercise trials published after 2013.26–33 The updated meta- et al.41 Most of the training programs consisted of 30 to 40 minutes of analyses estimated that aerobic exercise improved the peak oxygen resistance exercises in each session on two to three non-consecutive consumption rate by 2.9 ml/kg/min (95% CI 1.6 to 4.3) (Figure 2A) and days each week for 2 to 3 months, although a few studies had longer attenuated motor symptoms (ie, Unified Parkinson’s Disease Rating training durations up to 24 months. The initial resistance was typi- Scale Part III (UPDRS-III) or Movement Disorder Society-Sponsored cally set at 8 to 12 repetitions maximum (RM), which was then Revision of UPDRS Part III (MDS UPDRS-III)) by a SMD of 20.3 (95% gradually increased by 2 to 10% if the participants were able to CI 20.5 to 20.1) (Figure 2B). The effect of aerobic exercise on walking perform two to three sets of the same exercise at 8 to 12 RM with endurance (indicated by the 6MWT) was inconclusive due to the relative ease. The final target resistance at the end of the training

Invited Topical Review 165 A WMD (95% CI) period was typically 70 to 80% of 1 RM. The updated meta-analyses, Fixed mostly consisting of good-quality studies (PEDro scores  6), Study showed that progressive resistance exercise improved leg extensor Schenkman 2018 strength (SMD 0.71, 95% CI 0.21 to 1.21), TUG performance (MD 21.7 van der Kolk 2018 seconds, 95% CI 23.1 to 20.3), comfortable walking speed (SMD 0.39, Sacheli 2019 95% CI 0.12 to 0.66), fast walking speed (SMD 0.85, 95% CI 0.27 to Harvey 2019 1.43), motor symptoms (SMD 20.46, 95% CI 20.71 to 20.20), and van der Kolk 2019 PDQ-39 scores (MD 26, 95% CI 210 to 23), as shown in Figure 4. For Total detailed forest plots, see Figure 5 on the eAddenda. –10 –5 0 5 10 Research on the long-term effects of resistance training is scarce. Favours con (ml/kg/min) Favours exp Among the studies included in the meta-analyses here, Santos et al47 demonstrated that adding 8 weeks of progressive resistance training B SMD (95% CI) (two 60 to 70-minute sessions per week) to routine physiotherapy Fixed resulted in greater reduction in centre of pressure displacement Study during standing and greater improvements in fast walking speed and Fisher 2008 quality of life (PDQ-39) than routine physiotherapy alone, although Sage 2009 only the effect on fast walking speed was sustained at 1-month Qutubuddin 2013 follow-up. More studies are required to examine the effects of Cugusi 2015 resistance exercise training on other important outcomes such as van der Kolk 2018 balance and walking endurance, as well as on the long-term effects. Schenkman 2018 van der Kolk 2019 One particular aspect of resistance exercise training in PD is the Sacheli 2019 strengthening of respiratory muscles. This area has garnered atten- tion in research because people with PD often have respiratory dys- Total functions related to abnormal ventilator drive, restrictive changes, respiratory muscle weakness and upper airway obstruction.51 These –2 –1 0 1 2 may lead to impairments in swallowing and phonation.51 A 2020 systematic review only identified three randomised controlled trials C Favours exp Favours con of respiratory muscle strength training in PD. Although some positive effects on respiratory function were reported, two of these had PEDro Study WMD (95% CI) scores , 6.52 Therefore, the wide application of respiratory muscle Random training in PD cannot be supported until further high-quality research establishes its value.52 Cugusi 2015 Virtual reality and exergames van der Kolk 2019 Exercise that incorporates virtual reality (VR) is also gaining Harvey 2019 popularity in the rehabilitation of people with PD. The intervention often involves the use of computer-based games in a virtual reality Tollar 2019 environment. Some examples include the Nintendo Wii or Xbox Kinect, which are commercially available, and other customised VR Total tools specifically designed to address PD impairments. Incorporating VR into exercise may have potential advantages over conventional –200 –100 0 100 200 exercise by providing an interesting and interactive environment, which may increase patient motivation and engagement.53 Favours con (m) Favours exp Most of the VR intervention trials lasted 5 to 8 weeks (45 to 60 D WMD (95% CI) minutes per session, two to three sessions per week). A number of Fixed systematic reviews have addressed the use of VR in PD in recent Study years.53–58 A 2020 review by Canning et al57 found that VR rehabili- tation improved gait and balance when compared with inactive Qutubuddin 2013 controls in three facility-based trials, but such beneficial effects were not apparent in a home-based trial (PEDro scores 6 to 8); this may van der Kolk 2018 have been related to exercise underdosing due to inadequate super- vision in the home setting. A differential effect according to disease Harvey 2019 severity was also identified, with more beneficial effects for those with lower disease severity and potentially negative effects for those Tollar 2019 with higher disease severity. van der Kolk 2019 Although previous systematic reviews have attempted to estimate the effects of VR rehabilitation compared with other non-VR in- Total terventions,53–56,58 the meta-analyses are very difficult to interpret because the type and dose of the non-VR interventions used in –20 –10 0 10 20 different studies varied. Canning et al57 compared the effects of VR and non-VR rehabilitation of a similar type and dose based on the Favours exp Favours con findings of 11 trials and found no consistent evidence of VR rehabil- itation being more effective in improving gait or balance in PD. Figure 2. Effects of aerobic exercise compared with controls on (A) peak oxygen consumption rate, (B) Unified Parkinson’s Disease Rating Scale Part III (UPDRS-III), (C) Among the trials that compared VR rehabilitation and inactive 6-minute walk test and (D) PDQ-39, modified from Shu et al.25 Only the high-intensity control, only one incorporated a follow-up assessment after termi- group in Shenkman et al and only the cycling group in Tollar et al were used for meta- nation of training.59 A high-quality study (V-TIME trial) demonstrated analyses. Only randomised controlled trials that clearly specified the target training that participants in the VR treadmill training group had a lower intensity in the experimental group, and that the training intensity was distinct be- tween the experimental and comparison groups were included. See the main text section ‘Meta-analytic approach’ for further details.

166 Pang: Physiotherapy management of Parkinson’s disease A SMD (95% CI) E SMD (95% CI) Random Fixed Study Study Schilling 2010 Li 2012 Li 2012 Ni 2016 Hirsch 2003 Silva-Batista 2016 Paul 2014 Santos 2017 Ni 2016 Ferreira 2018 Silva-Batista 2018 Total Total –4 –2 0 2 4 –2 –1 0 1 2 Favours exp Favours con B Favours con Favours exp F WMD (95% CI) Fixed Study WMD (95% CI) Study Random Silva-Batista 2016 Schilling 2010 Santos 2017 Li 2012 Ferreira 2018 Paul 2014 de Lima 2019 Ni 2016 Total Silva-Batista 2018 de Lima 2019 –20 –10 0 10 20 Leal 2019 Favours exp Favours con Total Figure 4. (continued). –10 –5 0 5 10 incident rate of falls (8.1%, 95% CI 5.6 to 11.7) than those who received the same dose of treadmill training without VR (16.5%, 95% CI 9.9 to Favours exp (s) Favours con 27.3) in the PD subgroup during the 6-month follow-up period.59 C SMD (95% CI) A number of trials that compared VR and other non-VR in- Fixed terventions incorporated a follow-up period, mostly between 1 and 3 Study months.57,58 Only Shen et al had a follow-up period . 6 months Li 2012 beyond the training.60 They found that that their 3-month VR balance Paul 2014 and gait training reduced the fall rate and improved gait and balance Ni 2016 relative to conventional strengthening exercises; these beneficial ef- Santos 2017 fects were largely maintained at 3 months and 12 months after termination of training.60 Total In summary, the evidence currently supporting the use of VR in PD –2 –1 0 1 2 rehabilitation is largely limited to fully supervised conditions in facility-based settings only. Whether the VR interventions can induce Favours con Favours exp the same beneficial effects in less supervised or unsupervised con- ditions requires further investigation.57 D SMD (95% CI) Random Treadmill training Study Paul 2014 Common gait impairments in people with PD are: reduction in Ni 2016 gait speed, step length and cadence; shuffling gait; difficulty with Santos 2017 step initiation; and freezing of gait (FOG). Improving gait is one of the Leal 2019 primary goals of physiotherapy in PD. Some of the better-studied de Lima 2019 training approaches include treadmill training, sensory cueing with movement strategy training, Nordic walking and robot-assisted gait Total training. Studies on dual-task mobility training are also emerging. –2 –1 0 1 2 In gait training using a treadmill, a harness was used in some studies to ensure safety and prevent falling.61 Partial body weight Favours con Favours exp support was also used in some studies, whereas additional load was used in others.61 Many of the studies adopted a ‘speed-dependent Figure 4. Effects of resistance exercise compared with control on (A) leg extensor treadmill approach’, in which the belt speed was set as the highest strength, (B) Timed Up and Go Test, (C) comfortable walking speed, (D) fast walking speed at which the participant could walk safely without losing speed, (E) motor symptoms and (F) quality of life, modified from the systematic re- balance (ie, maximum achieved belt speed).61 Other studies used a views by Saltychev et al40 and Chung et al.41 The resistance training group without constant walking speed during the same training session.61 The added instability in Silva-Batista et al was used for analyses.48 The SMD, rather than treadmill walking program typically lasted for 4 to 8 weeks, with MD, was used in the analysis of leg extensor strength, comfortable walking speed, and three sessions (30 to 45 minutes) per week. The effects of treadmill fast walking speed because different outcome assessment methods involving different training in PD were examined in a number of systematic reviews,61,62 units of measurement were used. The SMD was also used in the analysis of motor including a 2015 Cochrane review of 18 randomised controlled symptoms because some studies used UPDRS III, while others used MDS UPDRS III as an outcome. See the main text section ‘Meta-analytic approach’ for further details.

Invited Topical Review 167 trials.61 The Cochrane review concluded that treadmill training A WMD (95% CI) improved gait speed and stride length, with moderate and low Fixed quality of evidence, respectively.61 However, the cadence and dis- Study tance did not improve with treadmill training.61 In the current re- Protas 2005 view, the evidence was further updated by extracting relevant data Fisher 2008 from the Cochrane review and adding the relevant randomised trials Piceli 2016 published after September 2014.62–66 The results seemed to suggest that treadmill training had a more pronounced effect on fast walking Total speed than comfortable walking speed, as shown in Figure 6. For detailed forest plots, see Figure 7 on the eAddenda. Specifically, the –0.50 –0.25 0 0.25 0.50 fast walking speed was improved by treadmill training by an average of 0.16 m/s (95% CI 0.04 to 0.28) (Figure 6A), whereas the comfortable Favours con Favours exp walking speed was improved by 0.11 m/s but the confidence interval (20.05 to 0.27) revealed a small chance of a better outcome in the B WMD (95% CI) control group (Figure 6B). However, the interpretation of this result Random should recognise that the quality of most trials was only fair (PEDro Study score 5). In a systematic review by Ni et al, it was recommended that Kurtais 2008 people with PD engage in treadmill training of 30 to 60 minutes per Fisher 2008 day, 2 to 3 days per week for more than 4 weeks.67 A self-selected Canning 2012 speed can be used initially, with gradual increases by 0.2 km/hour Ganesan 2015 as tolerated. Total Of the studies included in the meta-analyses, only Canning et al63 assessed the long-term effect by incorporating a follow-up assess- –0.2 –0.1 0 0.1 0.2 ment. It was found that their moderate-intensity, home-based treadmill walking exercise program did not improve walking capacity Favours con Favours exp but led to greater improvement in quality of life (PDQ-39) measured 6 weeks after termination of training. Further research is required to Figure 6. Effects of treadmill gait training on (A) fast walking speed and (B) comfort- assess the long-term effects of treadmill training for people with PD. able walking speed, compared with controls. See the main text section ‘Meta-analytic approach’ for further details. Sensory cueing and movement strategy training months.72 A recent systematic review by Radder et al showed that Gait hypokinesia in PD is thought to be attributable to the defi- movement strategy training including cueing induced a moderate cient formation of internal cues and reduced ability to activate the treatment effect on gait speed (six studies; SMD 0.45, 95% CI 0.13 to cortical motor set by the basal ganglia.68 External cues have long been 0.76); however, only three of these six studies had PEDro scores  6.73 used in PD rehabilitation to overcome the deficits in internal rhythm A moderately large effect on TUG performance was also found (six generation, and activation of the motor control system may be key to studies; SMD 0.53, 95% CI 0.23 to 0.82) in their meta-analysis of six facilitating a better gait pattern. Common types of cues that are used studies (four with PEDro scores  6).73 in PD rehabilitation of gait are largely auditory, visual and somato- sensory in nature.68–73 Sensory cueing is sometimes combined with High-quality studies are needed to examine sensory cueing in PD movement strategy training, which involves teaching people with PD and determine the most effective type of stimulus for different stages to focus their attention on movement and respond to sensory cues to of the disease.21 With the development of wearable technology (eg, improve performance in activities such as walking, obstacle negoti- wearable sensors, laser, augmented reality via Google glasses and ation and turning.74,75 This includes learning to: plan in advance for Halolens), there is much room for future research on their effective- movements to be performed, mentally rehearse the movements prior ness on gait rehabilitation for people with PD.78,79 to their execution and consciously focus the attention on movements while they are being performed. Complex movement sequences are Nordic walking also broken down into several components to facilitate better learning. The goal is to form a mental picture of the desired move- Nordic walking involves the use of two specially designed walking ment (eg, optimal step length) before its actual execution.75,76 poles with rubber tips during walking. It requires the individual to A number of systematic reviews have examined the effects of perform arm swings using the poles as they move forward, similar to sensory cueing on gait and falls in PD.69–73 The exercise interventions the movements observed in cross-country skiing.80 The use of poles with sensory cueing typically consisted of 30 to 60 minutes of during walking may promote postural adjustment and dissociation of training per session, two to three sessions per week for a period of 4 to 12 weeks, although the studies by Martin et al76 and Thaut et al77 the shoulder and pelvic girdles, and lessen axial rigidity, which may involved a longer training period of 6 months. The methodological facilitate a better gait pattern.80 Based on the 2017 systematic review quality of the reviewed studies varied (PEDro scores 2 to 8). Overall, by Bombieri et al,81 only four randomised controlled trials were both visual and auditory cues were found to be effective in improving identified. Improvements in motor outcomes (eg, walking speed and the kinematic gait parameters and turning execution, while reducing endurance, leg muscle strength) and non-motor outcomes (eg, the FOG and falls. The meta-analyses by Spaulding et al72 revealed depression, fatigue) were reported after Nordic walking training. The some differences in results between visual and auditory cueing. While visual cueing improved stride length only, auditory cueing also same systematic review also suggested that people with mild improved cadence and speed. With auditory cueing, the effect was disability or gait impairment seemed to benefit more from Nordic more pronounced when using high-intensity and high-frequency walking training.81 Their meta-analysis of these four randomised stimuli (10% above the patient’s normal walking cadence). Proprio- trials showed that Nordic walking was effective in reducing the ceptive cues in the form of rhythmic vibrations to the plantar surface UPDRS-III score (SMD 20.64, 95% CI 20.98 to 20.30).81 However, only of the foot skin synchronised with the step were shown to improve two of these trials (PEDro scores of 4 and 6) could delineate the ef- step length, speed, cadence and anticipatory postural adjustment (an important factor in step initiation).71 A number of studies also fects of Nordic walking by incorporating a relatively inactive control showed that the training effects from sensory cueing interventions group. No definitive conclusion could be made, partly due to the low can be largely maintained for a follow-up period varying from 1 to 6 methodological quality of the studies and small sample sizes used. Only two studies examined the long-term effects of Nordic walking training.80,82 Van Eihjkeren et al80 demonstrated that the gain in gait velocity and functional mobility after 6 weeks of training (12 hours in total) was maintained for 5 months after the training had

168 Pang: Physiotherapy management of Parkinson’s disease ended. However, there was no control group and therefore any motor learning facilitated with cueing and movement strategies, and improvement can be explained by other extraneous factors. Ebers- computerised training program such as Kinect) are effective in bach et al82 found that the benefits gained in cued reaction time after reducing FOG.89 However, recent work showed no difference in 4 weeks (16 hours in total) of Nordic walking was sustained for 16 cognitive function between those who have FOG and those who do weeks after program completion. not, after adjusting for covariates (particularly disease severity).89 There was also no correlation between FOG severity and cognitive Robot-assisted gait training performance.89 Thus, effective reduction in FOG may involve more complex strategies than a ‘cognitive-motor’ approach, but this will Application of robot-assistive devices in gait training has gained require more study. increasingly wide application in the rehabilitation of people with PD and its effects were examined in systematic reviews by Alwardat and Dual-task training Etoom.83,84 The robot-assisted gait training programs usually lasted 4 weeks (30 to 45 minutes per session, three to five sessions per Ambulation in daily living requires the ability to perform a week).83 The methodological quality of the reviewed trials was cognitive or motor task while walking (eg, carrying a glass of water or mostly good (PEDro scores  6 for six of seven studies).83 There is attending to traffic signals while walking). Relative to their peers high-quality evidence that robot-assisted gait training induced without PD, dual-task conditions cause people with PD to have more greater improvements in UPDRS-III score than conventional exercise exaggerated decreases in walking speed, stride length and cadence, (MD 25, 95% CI 23 to 28) but not treadmill training (MD 0, 95% CI 24 along with an increase in stride variability.90,91 Dual-task exercise to 3).84 Robot-assisted gait training also improved stride length (MD training in PD has gained increasing attention in recent years. It 9.3 cm, 95% CI 7.2 to 11.4; high-quality evidence) and gait speed (MD typically involves performing various walking and balance activities 0.17 m/s, 95% CI 0.09 to 0.24; low-quality evidence) and Berg Balance concurrently with cognitive tasks or upper limb motor tasks (eg, Scale (MD 3.6, 95% CI 0.4 to 6.7; moderate-quality evidence) when carrying an object).92 compared with conventional interventions.83 The training effects on UPDRS-III and Berg Balance Scale were sustained at the 1-month The effects of dual-task training were examined in a recent sys- follow-up but not at the 3-month follow-up. No clear treatment ef- tematic review of 11 randomised trials by Li et al.92 The majority of fects on stride time, cadence, TUG or balance confidence were these trials had a training duration of 6 to 12 weeks (45 to 60-minute detected.83 The positive effect on FOG was supported by case studies sessions, two to three sessions per week). All but one study had a and uncontrolled trials only.84 PEDro score of  6. Their meta-analyses showed that dual-task training improved gait speed in single-task conditions (SMD 20.29, Interventions for reducing freezing of gait 95% CI 20.47 to 20.10) but not in dual-task conditions (SMD 20.13, 95% CI 20.38 to 0.12) when compared with the control group (usual Various approaches have been used in attempts to reduce FOG, care or single-task training). The improvement was attributable to including balance exercises, treadmill training, sensory cueing, increase in cadence rather than step length. Dual-task training also movement strategy straining and action observation training. A 2020 improved motor symptoms, as measured by the UPDRS-III (SMD 0.56, systematic review by Consentino et al examined the effects of 95% CI 0.18 to 0.94) and Mini-BESTest score (SMD 20.44, 95% CI 20.84 different physiotherapy interventions on FOG. Their primary analysis to 20.05). However, the data of the control group in their meta- showed that physiotherapy interventions (nine trials examining analyses were derived from a combination of studies that used a cueing training/balance exercises/home-based exercises, with PEDro no-intervention control group and those that used a conventional/ scores  6 in seven of nine trials) had an overall treatment effect on single-task training group. As the conventional/single-task training reducing FOG compared with no intervention (SMD 20.29, 95% may have a beneficial effect on the outcomes, as clearly shown by CI 20.45 to 20.12).85 In the sub-group analyses focusing on different Strouwen et al in their DUALITY trial,93 their meta-analyses may have categories of intervention, home-based exercise interventions of underestimated the treatment effect of dual-task training. prolonged duration (ie, around 4 months), which included balance and gait exercises and cueing, showed a pronounced effect on FOG Regarding the long-term effects of dual-task training, the DUALITY compared with no intervention (three studies; SMD 20.30; 95% trial demonstrated that both dual-task training and single-task CI 20.53 to 20.07) but the quality of evidence as determined by the training led to similar improvements in dual-task walking speed, Grades of Recommendation, Assessment, Development and Evalua- which were sustained at the 12-week follow-up.93 Their study was tion (GRADE) was low due to risk of bias. In contrast, there was very also the only high-quality trial that measured falls and found no low-quality evidence that a shorter exercise program of 12 to 14 between-group difference in fall rate during the 24-week follow-up weeks in duration (three studies) or cueing training (two studies; period after training.93 In summary, more high-quality research is mean duration six to nine sessions) generated no effect.85 There was needed to investigate whether dual-task training is beneficial for no evidence that the effects of the above interventions are sustained improving gait and balance under dual-task conditions, and fall after termination of training. incidence. In action observation training, the participant is asked to first Overall, the gait training methods reviewed above have beneficial observe single motor actions (eg, rise from a chair, turn, walk through effects on certain aspects of gait, particularly walking speed, stride a narrow space) performed by a physiotherapist and then repetitively length and endurance. Ni et al67 examined different types of exercise practise the same movements.86,87 Therefore, action observation training in their systematic review of 40 randomised trials and, based training is considered a combined motor-cognitive approach in PD on the calculated effect sizes, the treatment effect from a task-specific rehabilitation. Adding action observation training to conventional exercise (eg, walking exercise, treadmill training) was generally exercises also led to greater reduction in FOG than the same exercises greater than that from a general exercise program (eg, cycling). with sham or no action observation training (four studies; Therefore, gait-specific training, rather than a general exercise pro- SMD 20.40, 95% CI 20.76 to 20.05), and the effect could be main- gram, may be required if gait is the outcome of interest. tained for 4 weeks (four studies; SMD 20.56, 95% CI 20.91 to 20.21), with moderate-quality evidence.85 However, adding visual/auditory Balance training cueing to treadmill or step training did not result in a better effect on FOG than treadmill or step training alone (very low quality of evi- People with PD may show deficits in different aspects of balance dence).85 Aquatic exercise also had no effect on FOG (very-low quality function, as described by Horak et al,94 namely: biomechanical con- of evidence).85 These findings are consistent with another systematic straints and postural orientation (ie, impaired flexibility, muscle review by Delgado-Alvarado et al,88 which showed that all available weakness, stooped posture); limits of stability and verticality (ie, studies classified as cognitive training (ie, action observation training, camptocormia, lateral trunk flexion, inability to hold an inclined posture); anticipatory postural adjustments (ie, diminished and delayed adjustments); reactive postural responses (ie, excessive

Invited Topical Review 169 muscle co-activations, decreased stepping reactions); sensorimotor Barnish et al indicated that Tango dance induced an overall treatment integration (ie, over-dependence on visual cues, impaired processing effect on UPDRS-III (MD 210, 95% CI 217 to 23) and that PD-specific of proprioceptive input); and dynamic control of gait (ie, hypokinetic, dance improved PDQ-39 (MD 28, 95% CI 212 to 24) but not per- shuffling gait, deficits in dual-task mobility).21 Thus, multidimen- formance in TUG (MD 22.1 seconds, 95% CI 26.3 to 2.1).103 Finally, the sional balance/agility exercise training is the most common physio- systematic review by de Almeida et al specifically assessed the effect therapy intervention to improve balance performance for people with of dance on postural control in PD and revealed a large effect size in PD.31,95–99 Examples of balance exercises typically included flexibility/ their meta-analysis (SMD 0.82, 95% CI 0.52 to 1.12).105 In summary, joint mobility exercises (ie, targeting biomechanical constraints), sustained dance practice is beneficial in improving various aspects weight shifting exercises (ie, limits of stability), exercises of self- of motor function 2 including balance, mobility and walking destabilisation of the centre of mass (ie, anticipatory postural ad- endurance 2 and delays the progression of motor symptoms. justments), tasks that involve external destabilisation of the centre of mass (ie, reactive postural responses), balance exercises on unstable Several mechanisms underlying the observed improvement support surfaces (ie, sensory orientation), balance activities during following dance therapy have been proposed. First, the external cues walking (eg, obstacle courses) that require continuous feedback and involved in dancing (eg, music with strong beats, physical contact feedforward postural adjustments (ie, dynamic control of gait).95–99 with partner, seeing the movements of partner, hearing the dance Other types of balance training methods that have been studied steps) may enable the participants to bypass the impaired basal include aquatic therapy, movement strategy training, motor-cognitive ganglia and access the cortical circuitry, thereby facilitating a better dual-task training, technology-assisted balance training, trunk/core gait pattern and reducing FOG.106 Dance often involves frequent exercises, treadmill walking with added perturbations, biofeedback stepping in different directions, turning, sudden cessation of step- and strengthening exercises with added instability.100–102 ping, shifting of centre of mass, and movements of varying speeds coordinated with arm movements. These may explain why sustained The training programs were typically 4 to 12 weeks in duration dance practice may result in improvement in balance and mobility. If (40 to 60 minutes per session, two to three sessions per week), the dance movements are intensive enough, they also impose a although a few studies involved a longer training period of  6 considerable demand on aerobic capacity and induce a training effect months. An earlier systematic review by Allen et al100 found that on endurance. balance training was effective in improving balance-related activity performance in PD, and that programs that involved highly chal- Dance practice makes the exercise more enjoyable and also pro- lenging balance activities had a tendency to induce greater motes social interaction.107,108 A qualitative research study has pro- improvement in balance. A more recent meta-analysis, by Shen vided important insights into the design and implementation of et al,101 of trials with moderate-to-high methodological quality found dance therapy classes in PD.109 It is important to consider the disease that exercise training generated significant short-term and long-term severity of the participants. One possible strategy is to stratify the improvements (up to 12 months) in balance performance (short-term dance groups according to individual ability level. It is critical that the effect: Hedges’ g = 0.303; long-term effect: Hedges’ g = 0.419). On the therapist is able to tailor the activities to the ability of the partici- other hand, a more recent systematic review by Flynn et al showed pants. Another important factor is the choice of music. Music with that home-based standing balance and/or gait exercises improved clear and strong beats is preferable because it helps participants to balance (SMD 0.21, 95% CI 0.10 to 0.32) and gait speed (SMD 0.30, 95% focus on movement and enables them to execute the movement se- CI 0.12 to 0.49) but not quality of life, when compared with no or quences more smoothly.109 sham intervention. Home-based exercise practice also induced similar improvement in balance to centre-based exercises Tai Chi, Qigong and yoga (SMD 20.04, 95% CI 20.36 to 0.27).102 Recent studies by Capato et al demonstrated that both people who freeze and non-freezers benefit Chinese traditional mind-body exercises, such as Qigong and Tai equally from the training.97–99 The treatment effect was greater and Chi, have become increasingly popular in neurological rehabilitation, lasted longer when the multidimensional balance exercises were including for PD. There are two theories underpinning these tech- supported by rhythmic auditory stimuli.97–99 Multidimensional bal- niques. The first is related to the concept of ‘qi’: in traditional Chinese ance exercise training induced improvement in balance not only in medicine, qi is an energy that flows through the meridian system in those with mild-to-moderate disease (Hoehn and Yahr stage 1 to the body.110 The blockage of the flow of qi is thought to contribute to 3)96–98 but also those with advanced PD (Hoehn and Yahr stage 4).99 illness. According to traditional Chinese medicine philosophies, Qigong facilitates the movement of qi throughout the body so that Dance exercise therapy health can be promoted.111 The second theory is related to the concept of ‘yin and yang’: they are the two complementary and Dance exercise therapy in PD is an emerging area of opposing elements that comprise the universe that needs to be research.103,104 Different types of dance have been studied, including maintained in harmony. Through practising Qigong, the yin and yang Tango, Irish dance, waltz, ballet, Turo PD (a qigong dance hybrid), of the body become balanced, thereby improving health.112 Ballu Sardu (a Sardinian folk dance), ballroom dancing, dance therapy specifically designed to address the PD symptoms, and a mixed genre. Tai Chi is a specific branch of Qigong and involves more martial Most dance therapy programs lasted for 8 to 13 weeks. Each session arts features, and involves the execution of slow and controlled body was typically 1 to 1.5 hour in duration, with training frequency at one movements coordinated with deep diaphragmatic breathing.21 The to two sessions per week. Argentine Tango was the most studied most common form used is the Yang style. The movements involved among the various dance programs. A 2020 systematic review of 16 often require shifting of the centre of gravity to the individual’s sta- trials (five trials having low risk of bias in  6 of 10 categories bility limit, reaching beyond the base of support, shifting of the body described in the Cochrane Collaboration risk of bias assessment tool) weight from one leg to another, changing the base of support from by Carapelloti et al revealed that dance exercise had favourable bilateral to unilateral stance, and sustained squatting movements. treatment effect on MDS UPDRS-III (MD 22, 95% CI 24 to 21), bal- With these manoeuvres, Tai Chi has the potential to improve balance ance (SMD 0.50, 95% CI 0.21 to 0.79), 6MWT (MD 50 m, 95% CI 15 to and muscle strength, which may in turn improve related functions 85), TUG (MD 21.1 seconds, 95% CI 22.1 to 0.2) and depressive such as mobility and endurance.21 Three systematic reviews were symptoms, as measured by the Beck Depression Inventory II published to examine the effect of Tai Chi in PD between 2014 and (MD 25.1, 95% CI 27.7 to 22.4).104 There was a tendency for dance 2015.113–115 All three reviews concluded that Tai Chi was effective in exercise to improve the forward and backward gait speed and stride improving balance and motor function for people with PD. length, but the confidence intervals did not exclude the possibility that usual care was slightly superior. The effect of dance exercise on To provide a more comprehensive picture of the therapeutic value PDQ-39 was not apparent. Another 2020 systematic review by of Tai Chi in PD based on the latest evidence, the meta-analyses presented in Yang et al114 were updated by including relevant rand- omised controlled trials that were published after 2013.116–120 Most trials involved regular Tai Chi practice for 30 to 60 minutes per

170 Pang: Physiotherapy management of Parkinson’s disease session, three times a week for 8 to 24 weeks. The updated meta- considered to be low or very low, except for the Berg Balance Scale analysis (PEDro scores 5 to 7) showed that Tai Chi improved bal- (high quality) and PDQ-39 (moderate quality).131 ance function as measured by the Berg Balance Scale (MD 4, 95% CI 1 to 6) (Figure 8A) and functional reach test (MD 5 cm, 95% CI 3 to 7) Combined exercise training (Figure 8B). Tai Chi also improved the TUG performance by 1.1 sec- onds (95% CI 20.4 to 21.8) (Figure 8C), and the 6MWT by 41 m (95% As the motor impairments and activity limitations are multifaceted CI 3 to 78) (Figure 8D). It also attenuated the motor symptoms among people with PD, the exercise training in the overall physio- (SMD 20.7, 95% CI 21.2 to 20.2) (Figure 8E). However, its effect on therapy intervention may be multidimensional, with incorporation of walking speed was inconclusive (Figure 8F). For detailed forest plots, different types of exercises (eg, aerobic, resistance, balance see Figure 9 on the eAddenda. As for the long-term effects of Tai Chi, training).43,134–142 Most of the trials of combined exercise training are of Li et al reported that the effect from 6 months of Tai Chi practice led good methodological quality (PEDro scores  6).43,134,135,137,139–140,142 to sustained effects for 3 months.121 Tai Chi also reduced the inci- The exercise sessions were typically conducted at a frequency of dence of falls after 12 weeks117 and 24 weeks121 of practice during a 3 two to three times per week for a duration of 4 to 16 weeks, although a to 6-month follow-up period (RR 0.33 to 0.44). few trials studied an exercise protocol of much longer duration (6 to 24 months).137–139 Most trials reported beneficial effects on strength, Baduanjin is a popular form of Qigong that involves fewer exer- gait, balance and endurance. Combined exercise training has also cises and is easier to master than Tai Chi.122 A few studies have been shown to have benefits on slowing the progression of motor examined the effect of Qigong for people with PD and identified impairments, as reflected by the change in MDS UPDRS-III scores.136 A beneficial effects on balance,122,123 mobility122,123 and motor func- few trials also reported positive impacts on the non-motor impair- tion124 after 8 to 24 weeks of Qigong practice (45 to 60 minutes per ments of fatigue,136 sleep quality,134 bone density136 and quality of session, two to four sessions per week); however, there was one high- life.43 quality trial122 and the others either have poor methodological quality (PEDro score 3)123 or a small sample size (, 40 Prevention of falls participants).124 Balance impairments are associated with falls in PD.100 Falls are A peaceful mind and focused attention are required during Qigong very common among people with PD, with an incidence rate of 343 and Tai Chi practice, leading researchers to postulate that these ex- per 100,000 people, representing more than twice the risk of falls ercise approaches may have a positive impact on non-motor out- than the reference population without PD.143 A systematic review of comes. The effects of these techniques on depression,119,125–127 22 PD studies showed that an average of 60.5% of participants cognition,119,126,127 sleep quality122,128 and quality of life (PDQ- experienced at least one fall, with 39% reporting recurrent falls.95 One 39)120,124,126 have been examined in a number of studies. The esti- of the detrimental consequences of falls is fragility fractures. After mated effects on these outcomes were mostly positive. adjusting for potential confounders, the incidence rate of fractures remained much greater among people with PD than their non-PD The application of yoga in PD has also garnered attention. Yoga counterparts (adjusted incidence rate ratio: 1.73), regardless of sex consists of prolonged muscle-stretching postures, diaphragmatic and age groups.143 The prevention of falls is thus a very important breathing and meditation.129 As a mind-body exercise intervention, goal of physiotherapy for PD. yoga also encourages body awareness and kinesthetic sense. A recent systematic review by Jin et al did not find strong enough evidence to A number of intervention studies used fall rate or number of support the use of yoga to reduce the risk of falls for people with fallers as an outcome measure.60,101,117,137,142,144–151 The intervention PD.130 The benefits of yoga on motor impairments, gait, muscle methods included balance/gait exercises,60,117,146,147,150 strengthening strength and quality of life were supported by pilot studies or low- exercises,149 a combination of balance/gait and strengthening exer- quality randomised trials. The study of yoga in PD is only emerging cises,137,142,144 Tai Chi,151 movement strategy training or exercise and more research is required to examine its health benefits. program combined with fall prevention education.137,146,148,149 The program duration showed great diversity, varying from 8 weeks to 6 Aquatic exercise months. The duration and frequency of the supervised sessions was usually 1 to 2 hours and one to two sessions per week, respectively. The use of aquatic exercise in PD has been examined in several The supervised sessions were often supplemented with a home ex- studies.131 The duration of the exercise programs varied between 4 ercise program. In some studies, the participants were required to and 10 weeks, with each session typically lasting 45 to 60 minutes at perform the exercises unsupervised following the intervention a frequency of 2 to 5 days per week. The exercises involved were period;60,137,145,148 in other studies, home-based exercise programs usually a mix of mobility, balance, strengthening and endurance ex- were supplemented with home visits or telephone calls.137,149 The ercises. One specific type of aquatic exercise is Ai Chi,125,132,133 which meta-analysis by Shen et al101 showed that exercise training reduced is a combination of Tai Chi and Qigong performed in the water rather the rate of falls (rate ratio 0.485 in the short term and 0.413 in the than on land. The Ai Chi trials were of good methodological quality long term) but not the number of fallers in the short or long term. (PEDro scores 6 to 7), although the sample sizes were relatively small Influence of disease severity seems to have an important association (29 to 40 participants in each study).125,132,133 Improvements in bal- with the outcome after exercise training. Those who have less severe ance (Berg Balance Scale, Single-leg-standing, Tinetti scores), TUG disease (UPDRS-III scores of 23 to 28 in Chivers-Seymour et al137 or  and motor function (UPDRS-III) were reported after 5 to 11 weeks (45 26 in Canning et al146, or Hoehn and Yahr scale 2 to 3 in Ashburn to 60 minutes per session, two to five sessions weekly) of Ai Chi et al148) experienced a reduction of fall rate with exercise interven- practice.125,132,133 tion, while those with more severe disease had a higher fall rate after training when compared with the control group. A 2019 systematic review by Cugusi et al examined the effects of aquatic exercises in PD.131 Most of the trials compared the effects of Effect of physical exercise on non-motor impairments aquatic exercise versus land-based exercise and were of good quality (all but one trial having PEDro scores  6).131 Their meta-analysis While the effects of specific types of exercise on non-motor im- revealed that aquatic exercise was superior to land-based exercise pairments were discussed in the respective sections above, a number in improving the Berg Balance Scale score (MD 2.7, 95% CI 1.6 to 3.9), of systematic reviews have attempted to synthesise the evidence Fall Efficacy Scale (MD 24.0, 95% CI 26.1 to 21.8) and PDQ-39 related to the effect of physical exercise training as a whole on non- (MD 26.0, 95% CI 211.3 to 20.6). There was a trend for aquatic ex- motor impairments in PD. Cusso et al showed that the effects of ex- ercise to be more beneficial than land-based exercise in reducing the ercise therapy on global non-motor impairments (indicated by UPDRS-III scores but the size of any additional benefit would be small UPDRS-I) and depressive symptoms were conflicting.152 The evidence (MD 21, 95% CI 22 to 0). The effects of aquatic exercise on the TUG and Activities-specific Balance Confidence Scale were similar to that of land-based exercise. However, the quality of evidence was

Invited Topical Review 171 A WMD (95% CI) E SMD (95% CI) Fixed Random Study Study Hackney 2008 Hackney 2008 Gao 2014 Li 2012 Amano 2013 Total Choi 2013 Gao 2014 –8 –4 0 4 8 Vergara-Diaz 2018 Favours con Favours exp Total B WMD (95% CI) Fixed Study Li 2012 –2 –1 0 1 2 Choi 2016 Favours exp Favours con Total F –10 –5 0 5 10 WMD (95% CI) Favours con Favours exp Study Random Li 2012 C Amano 2013 Study WMD (95% CI) Vergara-Diaz 2018 Hackney 2008 Fixed Li 2012 Total Gao 2014 Choi 2016 –0.2 –0.1 0 0.1 0.2 Vergara-Diaz 2018 Favours con Favours exp Total Figure 8. (continued). D –4 –2 0 2 4 stimulation is transcranial magnetic stimulation (TMS), which uses a Favours exp Favours con magnetic stimulator that contains an insulated coil wire with a Study connection to a large electrical capacitance. When a large pulse of Hackney 2008 WMD (95% CI) current runs through the coil, the rapidly changing magnetic field Choi 2013 Fixed generates an electrical current that has the ability to modulate Choi 2016 cortical excitability.155 TMS is conducted in patterns such as single pulse, repetitive pulse and theta burst. Low-frequency ( 1.0 Hz) Total repetitive TMS (rTMS) can reduce cortical excitability, while a high- frequency protocol can increase it.156,157 On the other hand, theta –100 –50 0 50 100 burst stimulation can be delivered intermittently or continuously. The Favours con Favours exp basic element of theta burst stimulation consists of three-pulse 50 Hz bursts applied with an interburst interval of 200 ms.158 In intermit- Figure 8. Effects of Tai Chi compared with controls on (A) Berg Balance Scale score, (B) tent TBS, which can increase cortical excitability, a train of bursts is Functional Reach, (C) Timed Up and Go, (D) 6-minute walk test, (E) Unified Parkinson’s delivered intermittently. Continuous theta stimulation, which can Disease Rating Scale Part III (UPDRS-III) and (F) walking speed. The SMD was used in reduce cortical excitability, is the uninterrupted delivery of the bursts the analysis of motor symptoms because some studies used UPDRS III, while others for a short period of time (eg, 20 seconds).158 Apart from its impact on used MDS UPDRS III as an outcome. See the main text section ‘Meta-analytic approach’ cortical excitability, rTMS can also affect the excitability of brain re- for further details. gions that are related to the site of stimulation, probably via the cortico-striato-thalamocortical circuitry, thus pointing to potential related to daytime sleepiness, fatigue, apathy and sleep quality was utility of rTMS in improving motor performance in PD.159,160 sparse and thus inconclusive.152 Another systematic review by da Silva et al153 revealed that adapted Tango, exergamesa, and treadmill A 2019 systematic review by Yang et al (23 studies, most with good methodological quality) showed that high-frequency rTMS, but training maintained or improved cognitive function among people at not low-frequency rTMS, was effective in improving motor perfor- mance (SMD 0.48, 95% CI 0.32 to 0.64).161 The frequency used in the a mild or moderate stage of PD. One common feature of these in- high-frequency rTMS protocol was typically between 5 and 10 Hz, and the intensity was 80 to 110% of the resting motor threshold. The terventions is their substantial demand on aerobic capacity. This is frequency of treatment sessions varied between two sessions per day to once a week. Multi-sessions of rTMS (typically three to ten ses- consistent with findings that moderate-intensity aerobic exercise led sions) were found to have stronger effects than a single session of to a more pronounced effect on improving global cognition among treatment. The stimulation site had an impact on the outcome. High- people with mild cognitive impairment or dementia.154 frequency rTMS only yielded a therapeutic effect on motor perfor- mance when the stimulation was applied to the primary motor cortex Transcranial magnetic stimulation (M1) but not the other brain regions. In addition, bilateral M1 stim- ulation yielded a greater treatment effect than unilateral M1 stimu- Increasing research has explored the use of brain stimulation for lation. The outcome was also affected by the stimulation dosage. The people with PD. One of the more investigated methods of brain effect of high-frequency rTMS applied to M1 was stronger when 18,000 to 20,000 pulses were delivered, compared with other

172 Pang: Physiotherapy management of Parkinson’s disease dosages. However, the medication state did not affect the magnitude number of treatment sessions, and tDCS intensity and duration. To of improvement.161 enhance the clinical utility of tDCS, more research is required to identify the optimal protocols for improving different aspects of Another systematic review by Chung et al (22 trials) examined not motor performance and also the neurophysiological mechanisms only motor impairments in general (eg, UPDRS-III) but also specific underlying the change in cortical excitability after tDCS aspects of motor activities (eg, upper limb function, walking).162 The intervention.169 methodological quality of the studies varied, but more than half of the criteria described in the Cochrane risk-of-bias checklist were Telerehabilitation fulfilled in 10 of the 22 trials reviewed. Their results revealed that rTMS induced improvements in: upper limb motor performance in Telehealth involves the use of electronic information and the short term (Hedges’ g 0.40); walking performance in the short telecommunication technologies to deliver healthcare services. It term (Hedges’ g 0.61) and long term (Hedges’ g 0.89) and UPDRS-III is particularly relevant to patients who have limited access to scores in the short term (Hedges’ g 0.31) and long term (Hedges’ g conventional face-to-face rehabilitation services for a variety of 0.54). Consistent with the meta-analysis by Yang et al,161 a stronger reasons, including financial burden, lack of required services effect was observed with stimulation at M1. Long-term improvement locally, and difficulties with travel due to long distance, impaired in UPDRS-III score was also associated with a greater number of total mobility or pandemic restrictions.176 A 2020 systematic review stimulation pulses.162 showed that telehealth interventions significantly lowered the degree of motor impairment (UPDRS-III).177 The interventions In a 2020 systematic review of 14 trials (13 with PEDro scores  7), were either phone-based or computer-based and involved a wide Xie et al examined the effect of TMS specifically on gait and found variety of arrangements: motor monitoring using wireless sensor improvement in walking speed after rTMS treatment, but only when technology, exercise DVD with weekly phone calls, gait training/ the walking speed was assessed in the ‘on’ state.163 In addition, the feedback via a smartphone application, virtual visits/care from a improvement in walking speed was not retained at 1 month follow- remote specialist, use of a Parkinson’s tracker app that enabled up. FOG and TUG were not improved with rTMS. The meta-analysis tracking of different self-monitoring measures (eg, cognition, by Goodwill et al revealed no effect of rTMS on cognition (five movement, mood) and medications, and delivery of a web-based studies, four with PEDro scores  7).164 Overall, rTMS does show course.177 some promise in improving motor performance, but the optimal protocol for people with different levels of disability requires further Two trials directly involved the participation of physiothera- research. pists.178,179 Ginis et al178 found that their Smartphone-delivered gait training system led to similar improvement in gait as standard Transcranial direct current stimulation physiotherapy after 6 weeks of training (30 minutes, three sessions per week) and the effects were retained after another 4 weeks. Their Another form of brain stimulation studied in PD is transcranial system also induced more gain in balance ability (mini-BESTest) than direct current stimulation (tDCS), which involves the delivery of a standard physiotherapy, but the effect was not maintained at follow- constant low-amplitude electric current (generally between 1 and 2 up. In a pilot study, Khalil et al179 studied the effects of an 8-week mA) via scalp electrodes. It can modulate excitability in both cortical program that involved the home use of an exercise DVD (three and subcortical brain regions.165,166 Anodal tDCS can increase times a week, 24 sessions) and walking program (once weekly, eight neuronal excitability, while cathodal tDCS can decrease it.165 tDCS sessions). Of the 32 sessions, eight sessions within the first 4 weeks may have relevance to the treatment of PD symptoms because anodal were supervised by a physiotherapist. The participants received a tDCS can enhance extracellular dopamine levels in the striatum167 weekly phone call from the physiotherapist to check the adherence to and inhibit GABAergic neurons.168,169 exercises and also to address areas of concern raised by the patients. The intervention was found to cause a greater reduction in MDS A systematic review by Broeder et al170 investigated the effects of UPDRS-III score than usual care. The retention rate and adherence to tDCS on different aspects of motor activities and cognition in PD (10 the above interventions were quite high, at 87 to 91% and 70 to 77%, studies). Two studies (PEDro scores 7 and 8) found that anodal tDCS respectively.178,179 However, there is a lack of economic analysis of over M1 reduced UPDRS-III scores compared to sham stimulation. these telerehabilitation interventions compared with conventional The other studies did not find significant results on UPDRS, but the physiotherapy. stimulation sites were different (dorsolateral prefrontal cortex, or a combination of premotor cortex, prefrontal cortex and M1). Of the More recently, Gandolfi et al addressed this limitation by seven studies (the majority with high quality, PEDro scores  7) that comparing both the therapeutic effects and cost of a home-based examined gait parameters, five reported beneficial effects. The results telerehabilitation virtual reality balance training program with an on upper limb activity were conflicting. Of the four included studies in-clinic sensory integration balance training regimen.180 After (the majority with high quality, PEDro scores  7), two reported 7 weeks of training (50 minutes per session, 3 days per week) that positive effects of anodal tDCS. A more recent systematic review by was remotely supervised by a physiotherapist, their post-hoc Goodwill et al164 revealed an overall improvement of motor activity analysis showed similar improvement in balance (Berg Balance with tDCS based on the results of nine studies (eight with PEDro Scale, Dynamic Gait Index), balance confidence (Activities-specific scores  7) and the degree of improvement was similar to that Balance Confidence Scale) and quality of life (PDQ-8) in both induced by rTMS. groups immediately after training and at 1-month follow-up. The total cost per patient was V384 for the telerehabilitation, which The effects of tDCS on cognitive function was mixed. Some studies was considerably lower than the in-clinic intervention (V602). An showed that anodal tDCS applied over the dorsolateral prefrontal important point to note is that the caregiver was always present cortex improved working memory171,172 and phononemic verbal during the training sessions to ensure safety. This may be a key fluency,171 but this was not observed in other studies.173,174 The sys- limitation of some telerehabilitation interventions, especially those tematic review by Goodwill et al164 revealed no effect of tDCS on that require a higher degree of monitoring of performance and cognition but the meta-analysis was based on the findings of three safety. studies. Overall, while home-based telerehabilitation may be a viable In clinical practice, tDCS is often combined with other physio- alternative to traditional face-to-face physiotherapy service delivery, therapy interventions. Beretta et al175 showed in their 2020 system- there is scarcity of research in this area. The search for innovative atic review of randomised trials that tDCS combined with cognitive methods of physiotherapy service delivery has never been so relevant and/or motor training may induce more pronounced treatment ef- because of the current COVID-19 pandemic. More research on PD fects on motor performance (gait, posture, upper limb activity) and telerehabilitation is urgently needed. cognition, suggesting a possible synergistic effect. However, the methodological quality of the studies was not evaluated. Similar to the rTMS trials, the stimulation protocol varied greatly in terms of the

Invited Topical Review 173 Future directions for clinical practice and research 17. Poewe W. Non-motor symptoms in Parkinson’s disease. Eur J Neurol. 2008;15(Suppl 1):14–20. A clinical regimen should be selected for each patient after consideration of the therapist time required for equipment set up and 18. Katzenschlager R, Head J, Schrag A, Ben-Shlomo Y, Evans A, Lees AJ, et al. Four- the cost-effectiveness and sustainability of the program. As PD is teen-year final report of the randomized PDRG-UK trial comparing three initial chronic and progressive in nature, patients should be equipped with treatments in PD. Neurology. 2008;71:474–480. the necessary skills to perform the exercises in home-based or community-based settings in the long run so that the therapeutic 19. Postuma RB, Aarsland D, Barone P, Burn DJ, Hawkes CH, Oertel W, et al. Identi- effects can be sustained. Therefore, delivery of physiotherapy service fying prodromal Parkinson’s disease: pre-motor disorders in Parkinson’s disease. through telecommunication technologies should be further explored. Mov Disord. 2012;27:617–626. Considering the promising preliminary results, more efforts should be directed towards developing innovative telerehabilitation services for 20. Song W, Guo X, Chen K, Chen X, Cao B, Wei Q, et al. The impact of non-motor people with PD delivered by physiotherapists and evaluating their symptoms on the health-related quality of life of Parkinson’s disease patients efficacy and cost-effectiveness. The use of Qigong, yoga and respira- from Southwest China. Parkinsonism Relat Disord. 2014;20:149–152. tory muscle training requires further research due to the limited evidence. Future research is also warranted to identify the optimal 21. Mak MKY, Wong-Yu ISK. Exercise for Parkinson’s disease. Int Rev Neurobiol. exercise protocols for improving more complex walking tasks (eg, 2019;147:1–44. dual tasking) and FOG. The use of rTMS and tDCS shows some promising results, but should undergo more research to identify the 22. Mak MK, Wong-Yu IS, Shen X, Chung CL. Long-term effects of exercise and optimal protocols for different outcomes and PD subgroups. More physical therapy in people with Parkinson disease. Nat Rev Neurol. 2017;13: work is also needed to investigate the effects of different types of 689–703. exercise training on non-motor outcomes in PD. 23. Katzel LI, Ivey FM, Sorkin JD, Macko RF, Smith B, Shulman LM. Impaired economy Footnotes: a Wii Fit, Nintendo, Kyoto, Japan. of gait and decreased six-minute walk distance in Parkinson’s disease. Parkinsons eAddenda: Figures 3, 5, 7 and 9 can be found online at https://doi. Dis. 2012;2012:241754. org/10.1016/j.jphys.2021.06.004 Ethics approval: Nil. 24. Christiansen CL, Schenkman ML, McFann K, Wolfe P, Kohrt WM. Walking econ- Competing interests: Nil. omy in people with Parkinson’s disease. Mov Disord. 2009;24:1481–1487. Source(s) of support: Nil. Acknowledgements: The author would like to thank Miss Ouyang 25. Shu HF, Yang T, Yu SX, Huang HD, Jiang LL, Gu JW, et al. Aerobic exercise for Huixi for assistance in data curation, and Dr Raymond Chung for Parkinson’s disease: a systematic review and meta-analysis of randomized assistance in statistical analysis. controlled trials. PLoS One. 2014;9:e100503. Provenance: Invited. Peer reviewed. Correspondence: Marco YC Pang, Department of Rehabilitation 26. Cugusi L, Solla P, Serpe R, Carzedda T, Piras L, Oggianu M, et al. Effects of a Nordic Sciences, Faculty of Health and Social Sciences, The Hong Kong Walking program on motor and non-motor symptoms, functional performance Polytechnic University, Hong Kong. Email: [email protected] and body composition in patients with Parkinson’s disease. NeuroRehabilitation. 2015;37:245–254. References 27. Harvey M, Weston KL, Gray WK, O’Callaghan A, Oates LL, Davidson R, et al. High- 1. Abbas MM, Xu Z, Tan LCS. Epidemiology of Parkinson’s Disease-East Versus West. intensity interval training in people with Parkinson’s disease: a randomized, Mov Disord Clin Pract. 2017;5:14–28. controlled feasibility trial. Clin Rehabil. 2019;33:428–438. 2. Dorsey ER, Sherer T, Okun MS, Bloem BR. The emerging evidence of the Parkinson 28. Schenkman M, Moore CG, Kohrt WM, Hall DA, Delitto A, Comella CL, et al. Effect pandemic. J Parkinsons Dis. 2018;8:S3–S8. of high-intensity treadmill exercise on motor symptoms in patients with de novo Parkinson disease: a phase 2 randomized clinical trial. JAMA Neurol. 3. Pringsheim T, Jette N, Frolkis A, Steeves TD. The prevalence of Parkinson’s 2018;75:219–226. disease: a systematic review and meta-analysis. Mov Disord. 2014;29:1583– 1590. 29. van der Kolk NM, de Vries NM, Penko AL, van der Vlugt M, Mulder AA, Post B, et al. A remotely supervised home-based aerobic exercise programme is feasible 4. Muangpaisan W, Hori H, Brayne C. Systematic review of the prevalence and for patients with Parkinson’s disease: results of a small randomised feasibility incidence of Parkinson’s disease in Asia. J Epidemiol. 2009;19:281–293. trial. J Neurol Neurosurg Psychiatry. 2018;89:1003–1005. 5. Kalia LV, Lang AE. Parkinson’s disease. Lancet. 2015;386:896–912. 30. van der Kolk NM, de Vries NM, Kessels RPC, Joosten H, Zwinderman AH, Post B, 6. Albin RL. The pathophysiology of chorea/ballism and Parkinsonism. Parkinsonism et al. Effectiveness of home-based and remotely supervised aerobic exercise in Parkinson’s disease: a double-blind, randomised controlled trial. Lancet Neurol. Relat Disord. 1995;1:3–11. 2019;18:998–1008. 7. Horak FB, Frank J, Nutt J. Effects of dopamine on postural control in parkinsonian 31. Tollár J, Nagy F, Hortobágyi T. Vastly different exercise programs similarly subjects: scaling, set, and tone. J Neurophysiol. 1996;75:2380–2396. improve parkinsonian symptoms: a randomized clinical trial. Gerontology. 8. Pang MY, Mak MK. Influence of contraction type, speed, and joint angle on ankle 2019;65:120–127. muscle weakness in Parkinson’s disease: implications for rehabilitation. Arch 32. Qutubuddin A, Reis T, Alramadhani R, Cifu DX, Towne A, Carne W. Parkinson’s Phys Med Rehabil. 2012;93:2352–2359. disease and forced exercise: a preliminary study. Rehabil Res Pract. 9. Mak MK, Pang MY, Mok V. Gait difficulty, postural instability, and muscle 2013;2013:375267. weakness are associated with fear of falling in people with Parkinson’s disease. Parkinsons Dis. 2012;2012:901721. 33. Sacheli MA, Neva JL, Lakhani B, Murray DK, Vafai N, Shahinfard E, et al. Exercise 10. Mak MK, Pang MY. Balance confidence and functional mobility are indepen- increases caudate dopamine release and ventral striatal activation in Parkinson’s dently associated with falls in people with Parkinson’s disease. J Neurol. disease. Mov Disord. 2019;34:1891–1900. 2009;256:742–749. 11. Mak MK, Pang MY. Balance self-efficacy determines walking capacity in people 34. Schootemeijer S, van der Kolk NM, Bloem BR, de Vries NM. Current perspectives with Parkinson’s disease. Mov Disord. 2008;23:1936–1939. on aerobic exercise in people with Parkinson’s disease. Neurotherapeutics. 12. Saltin B, Landin S. Work capacity, muscle strength and SDH activity in both legs 2020;17:1418–1433. of hemiparetic patients and patients with Parkinson’s disease. Scand J Clin Lab Invest. 1975;35:531–538. 35. Arfa-Fatollahkhani P, Safar Cherati A, Habibi SAH, Shahidi GA, Sohrabi A, 13. Protas EJ, Stanley RK, Jankovic J, MacNeill B. Cardiovascular and metabolic re- Zamani B. Effects of treadmill training on the balance, functional capacity and sponses to upper- and lower-extremity exercise in men with idiopathic Par- quality of life in Parkinson’s disease: a randomized clinical trial. J Complement kinson’s disease. Phys Ther. 1996;76:34–40. Integr Med. 2019;17:20180245. 14. Mavrommati F, Collett J, Franssen M, Meaney A, Sexton C, Dennis-West A, et al. Exercise response in Parkinson’s disease: insights from a cross-sectional com- 36. Inkster LM, Eng JJ, MacIntyre DL, Stoessl AJ. Leg muscle strength is reduced in parison with sedentary controls and a per-protocol analysis of a randomised Parkinson’s disease and relates to the ability to rise from a chair. Mov Disord. controlled trial. BMJ Open. 2017;7:e017194. 2003;18:157–162. 15. Mak MK, Pang MY. Parkinsonian single fallers versus recurrent fallers: different fall characteristics and clinical features. J Neurol. 2010;257:1543–1551. 37. Vetrano DL, Pisciotta MS, Laudisio A, Lo Monaco MR, Onder G, Brandi V, et al. 16. Chou KL, Elm JJ, Wielinski CL, Simon DK, Aminoff MJ, Christine CW, et al. Factors associated with falling in early, treated Parkinson’s disease: the NET-PD LS1 Sarcopenia in Parkinson disease: comparison of different criteria and association cohort. J Neurol Sci. 2017;377:137–143. with disease severity. J Am Med Dir Assoc. 2018;19:523–527. 38. Pang MY, Mak MK. Muscle strength is significantly associated with hip bone mineral density in women with Parkinson’s disease: a cross-sectional study. J Rehabil Med. 2009;41:223–230. 39. Pelicioni PHS, Menant JC, Latt MD, Lord SR. Falls in Parkinson’s disease subtypes: risk factors, locations and circumstances. Int J Environ Res Public Health. 2019;16:2216. 40. Saltychev M, Bärlund E, Paltamaa J, Katajapuu N, Laimi K. Progressive resistance training in Parkinson’s disease: a systematic review and meta-analysis. BMJ Open. 2016;6:e008756. 41. Chung CL, Thilarajah S, Tan D. Effectiveness of resistance training on muscle strength and physical function in people with Parkinson’s disease: a systematic review and meta-analysis. Clin Rehabil. 2016;30:11–23. 42. Ni M, Signorile JF, Mooney K, Balachandran A, Potiaumpai M, Luca C, et al. Comparative effect of power training and high-speed yoga on motor function in older patients with Parkinson Disease. Arch Phys Med Rehabil. 2016;97:345–354. 43. Ni M, Signorile JF, Balachandran A, Potiaumpai M. Power training induced change in bradykinesia and muscle power in Parkinson’s disease. Parkinsonism Relat Disord. 2016;23:37–44. 44. Paul SS, Canning CG, Song J. Leg muscle power is enhanced by training in people with Parkinson’s disease: a randomized controlled trial. Clin Rehabil. 2014;28: 275–288. 45. de Lima TA, Ferreira-Moraes R, Alves WMGDC, Alves TGG, Pimentel CP, Sousa EC, et al. Resistance training reduces depressive symptoms in elderly people with Parkinson disease: a controlled randomized study. Scand J Med Sci Sports. 2019;29:1957–1967.

174 Pang: Physiotherapy management of Parkinson’s disease 46. Leal LC, Abrahin O, Rodrigues RP, da Silva MC, Araújo AP, de Sousa EC, et al. Low- 75. Morris ME, Iansek R, Kirkwood B. A randomized controlled trial of movement volume resistance training improves the functional capacity of older individuals strategies compared with exercise for people with Parkinson’s disease. Mov with Parkinson’s disease. Geriatr Gerontol Int. 2019;19:635–640. Disord. 2009;24:64–71. 47. Santos L, Fernandez-Rio J, Winge K, Barragán-Pérez B, González-Gómez L, 76. Martin T, Weatherall M, Anderson TJ, MacAskill MR. A randomized controlled Rodríguez-Pérez V, et al. Effects of progressive resistance exercise in akinetic- feasibility trial of a specific cueing program for falls management in persons with rigid Parkinson’s disease patients: a randomized controlled trial. Eur J Phys Parkinson disease and freezing of gait. J Neurol Phys Ther. 2015;39:179–184. Rehabil Med. 2017;53:651–663. 77. Thaut MH, Rice RR, Braun Janzen T, Hurt-Thaut CP, McIntosh GC. Rhythmic 48. Silva-Batista C, Corcos DM, Roschel H, Kanegusuku H, Gobbi LT, Piemonte ME, auditory stimulation for reduction of falls in Parkinson’s disease: a randomized et al. Resistance training with instability for patients with Parkinson’s disease. controlled study. Clin Rehabil. 2019;33:34–43. Med Sci Sports Exerc. 2016;48:1678–1687. 78. Carpinella I, Cattaneo D, Bonora G, Bowman T, Martina L, Montesano A, et al. 49. Silva-Batista C, Corcos DM, Kanegusuku H, Piemonte MEP, Gobbi LTB, de Lima- Wearable sensor-based biofeedback training for balance and gait in Parkinson Pardini AC, et al. Balance and fear of falling in subjects with Parkinson’s disease is disease: a pilot randomized controlled trial. Arch Phys Med Rehabil. 2017;98:622– improved after exercises with motor complexity. Gait Posture. 2018;61:90–97. 630. 50. Ferreira RM, Alves WMGDC, de Lima TA, Alves TGG, Alves Filho PAM, 79. Ginis P, Nackaerts E, Nieuwboer A, Heremans E. Cueing for people with Parkin- Pimentel CP, et al. The effect of resistance training on the anxiety symptoms and son’s disease with freezing of gait: a narrative review of the state-of-the-art and quality of life in elderly people with Parkinson’s disease: a randomized novel perspectives. Ann Phys Rehabil Med. 2018;61:407–413. controlled trial. Arq Neuropsiquiatr. 2018;76:499–506. 80. van Eijkeren FJ, Reijmers RS, Kleinveld MJ, Minten A, Bruggen JP, Bloem BR. 51. Torsney KM, Forsyth D. Respiratory dysfunction in Parkinson’s disease. J R Coll Nordic walking improves mobility in Parkinson’s disease. Mov Disord. Physicians Edinb. 2017;47:35–39. 2008;23:2239–2243. 52. Rodríguez MÁ, Crespo I, Del Valle M, Olmedillas H. Should respiratory muscle 81. Bombieri F, Schena F, Pellegrini B, Barone P, Tinazzi M, Erro R. Walking on four training be part of the treatment of Parkinson’s disease? A systematic review of limbs: a systematic review of Nordic Walking in Parkinson disease. Parkinsonism randomized controlled trials. Clin Rehabil. 2020;34:429–437. Relat Disord. 2017;38:8–12. 53. Dockx K, Bekkers EM, Van den Bergh V, Ginis P, Rochester L, Hausdorff JM, et al. 82. Ebersbach G, Ebersbach A, Gandor F, Wegner B, Wissel J, Kupsch A. Impact of Virtual reality for rehabilitation in Parkinson’s disease. Cochrane Database Syst physical exercise on reaction time in patients with Parkinson’s disease-data from Rev. 2016;12:CD010760. the Berlin BIG Study. Arch Phys Med Rehabil. 2014;95:996–999. 54. Lei C, Sunzi K, Dai F, Liu X, Wang Y, Zhang B, et al. Effects of virtual reality 83. Alwardat M, Etoom M, Al Dajah S, Schirinzi T, Di Lazzaro G, et al. Effectiveness of rehabilitation training on gait and balance in patients with Parkinson’s disease: a robot-assisted gait training on motor impairments in people with Parkinson’s systematic review. PLoS One. 2019;14:e0224819. disease: a systematic review and meta-analysis. Int J Rehabil Res. 2018;41:287– 296. 55. Chen Y, Gao Q, He CQ, Bian R. Effect of virtual reality on balance in individuals with Parkinson disease: a systematic review and meta-analysis of randomized 84. Alwardat M, Etoom M. Effectiveness of robot-assisted gait training on freezing of controlled trials. Phys Ther. 2020;100:933–945. gait in people with Parkinson disease: evidence from a literature review. J Exerc Rehabil. 2019;15:187–192. 56. Triegaardt J, Han TS, Sada C, Sharma S, Sharma P. The role of virtual reality on outcomes in rehabilitation of Parkinson’s disease: meta-analysis and systematic 85. Cosentino C, Baccini M, Putzolu M, Ristori D, Avanzino L, Pelosin E. Effectiveness review in 1031 participants. Neurol Sci. 2020;41:529–536. of physiotherapy on freezing of gait in Parkinson’s disease: a systematic review and meta-analyses. Mov Disord. 2020;35:523–536. 57. Canning CG, Allen NE, Nackaerts E, Paul SS, Nieuwboer A, Gilat M. Virtual reality in research and rehabilitation of gait and balance in Parkinson disease. Nat Rev 86. Pelosin E, Barella R, Bet C, Magioncalda E, Putzolu M, Di Biasio F, et al. Effect of Neurol. 2020;16:409–425. group-based rehabilitation combining action observation with physiotherapy on freezing of gait in Parkinson’s disease. Neural Plast. 2018;2018:4897276. 58. Wang B, Shen M, Wang YX, He ZW, Chi SQ, Yang ZH. Effect of virtual reality on balance and gait ability in patients with Parkinson’s disease: a systematic review 87. Di Iorio W, Ciarimboli A, Ferriero G, Feleppa M, Baratto L, Matarazzo G, et al. and meta-analysis. Clin Rehabil. 2019;33:1130–1138. Action observation in people with Parkinson’s disease. A motor-cognitive combined approach for motor rehabilitation. A preliminary report. Diseases. 59. Mirelman A, Rochester L, Maidan I, Del Din S, Alcock L, Nieuwhof F, et al. Addition 2018;6:58. of a non-immersive virtual reality component to treadmill training to reduce fall risk in older adults (V-TIME): a randomised controlled trial. Lancet. 88. Delgado-Alvarado M, Marano M, Santurtún A, Urtiaga-Gallano A, Tordesillas- 2016;388:1170–1182. Gutierrez D, Infante J. Nonpharmacological, nonsurgical treatments for freezing of gait in Parkinson’s disease: a systematic review. Mov Disord. 60. Shen X, Mak MKY. Technology-assisted balance and gait training reduces falls in 2020;35:204–214. patients with Parkinson’s disease: a randomised controlled trial with 12 month follow-up. Neurorehabil Neural Repair. 2015;29:103–111. 89. Morris R, Smulders K, Peterson DS, Mancini M, Carlson-Kuhta P, Nutt JG, et al. Cognitive function in people with and without freezing of gait in Parkinson’s 61. Mehrholz J, Kugler J, Storch A, Pohl M, Elsner B, Hirsch K. Treadmill training disease. NPJ Parkinsons Dis. 2020;6:9. for patients with Parkinson’s disease. Cochrane Database Syst Rev. 2015;8:CD007830. 90. Wild LB, de Lima DB, Balardin JB, Rizzi L, Giacobbo BL, Oliveira HB, et al. Char- acterization of cognitive and motor performance during dual-tasking in healthy 62. Robinson AG, Dennett AM, Snowdon DA. Treadmill training may be an effective older adults and patients with Parkinson’s disease. J Neurol. 2013;260:580–589. form of task-specific training for improving mobility in people with Parkinson’s disease and multiple sclerosis: a systematic review and meta-analysis. Physio- 91. Kleiner AFR, Pagnussat AS, Prisco GD, Vagnini A, Stocchi F, De Pandis MF, et al. therapy. 2019;105:174–186. Analyzing gait variability and dual task interference in patients with Parkinson’s disease and freezing by means of the wordcolor Stroop test. Aging Clin Exp Res. 63. Canning CG, Allen NE, Dean CM, Goh L, Fung VS. Home-based treadmill training 2018;30:1137–1142. for individuals with Parkinson’s disease: a randomized controlled pilot trial. Clin Rehabil. 2012;26:817–826. 92. Li Z, Wang T, Liu H, Jiang Y, Wang Z, Zhuang J. Dual-task training on gait, motor symptoms, and balance in patients with Parkinson’s disease: a systematic review 64. Protas EJ, Mitchell K, Williams A, Qureshy H, Caroline K, Lai EC. Gait and step and meta-analysis. Clin Rehabil. 2020;34:1355–1367. training to reduce falls in Parkinson’s disease. NeuroRehabilitation. 2005;20:183– 190. 93. Strouwen C, Molenaar EALM, Münks L, Keus SHJ, Zijlmans JCM, Vandenberghe W, et al. Training dual tasks together or apart in Parkinson’s disease: results from the 65. Kurtais Y, Kutlay S, Tur BS, Gok H, Akbostanci C. Does treadmill training improve DUALITY trial. Mov Disord. 2017;32:1201–1210. lower-extremity tasks in Parkinson disease? A randomized controlled trial. Clin J Sport Med. 2008;18:289–291. 94. Horak FB, Wrisley DM, Frank J. The Balance Evaluation Systems Test (BESTest) to differentiate balance deficits. Phys Ther. 2009;89:484–498. 66. Picelli A, Varalta V, Melotti C, Zatezalo V, Fonte C, Amato S, et al. Effects of treadmill training on cognitive and motor features of patients with mild to 95. Allen NE, Schwarzel AK, Canning CG. Recurrent falls in Parkinson’s disease: a moderate Parkinson’s disease: a pilot, single-blind, randomized controlled trial. systematic review. Parkinsons Dis. 2013;2013:906274. Funct Neurol. 2016;31:25–31. 96. Wong-Yu IS, Mak MK. Multi-dimensional balance training programme improves 67. Ni M, Hazzard JB, Signorile JF, Luca C. Exercise guidelines for gait function in balance and gait performance in people with Parkinson’s disease: a pragmatic Parkinson’s disease: a systematic review and meta-analysis. Neurorehabil Neural randomized controlled trial with 12-month follow-up. Parkinsonism Relat Disord. Repair. 2018;32:872–886. 2015;21:615–621. 68. Morris ME, Iansek R, Matyas TA, Summers JJ. The pathogenesis of gait hypo- 97. Capato TTC, de Vries NM, IntHout J, Ramjith J, Barbosa ER, Nonnekes J, et al. kinesia in Parkinson’s disease. Brain. 1994;117:1169–1181. Multimodal balance training supported by rhythmic auditory stimuli in Par- kinson disease: effects in freezers and nonfreezers. Phys Ther. 2020;100:2023– 69. Muñoz-Hellín E, Cano-de-la-Cuerda R, Miangolarra-Page JC. Guías visuales como 2034. herramienta terapéutica en la enfermedad de Parkinson. Una revisión sistemá- tica [Visual cues as a therapeutic tool in Parkinson’s disease. A systematic re- 98. Capato TTC, de Vries NM, IntHout J, Barbosa ER, Nonnekes J, Bloem BR. Multi- view]. Rev Esp Geriatr Gerontol. 2013;48:190–197. modal balance training supported by rhythmical auditory stimuli in Parkinson’s disease: a randomized clinical trial. J Parkinsons Dis. 2020;10:333–346. 70. Gómez-González J, Martín-Casas P, Cano-de-la-Cuerda R. Effects of auditory cues on gait initiation and turning in patients with Parkinson’s disease. Neurologia. 99. Capato TTC, Nonnekes J, de Vries NM, IntHout J, Barbosa ER, Bloem BR. Effects of 2019;34:396–407. multimodal balance training supported by rhythmical auditory stimuli in people with advanced stages of Parkinson’s disease: a pilot randomized clinical trial. 71. Rocha PA, Porfírio GM, Ferraz HB, Trevisani VF. Effects of external cues on gait J Neurol Sci. 2020;418:117086. parameters of Parkinson’s disease patients: a systematic review. Clin Neurol Neurosurg. 2014;124:127–134. 100. Allen NE, Sherrington C, Paul SS, Canning CG. Balance and falls in Parkinson’s disease: a meta-analysis of the effect of exercise and motor training. Mov Disord. 72. Spaulding SJ, Barber B, Colby M, Cormack B, Mick T, Jenkins ME. Cueing and gait 2011;26:1605–1615. improvement among people with Parkinson’s disease: a meta-analysis. Arch Phys Med Rehabil. 2013;94:562–570. 101. Shen X, Wong-Yu IS, Mak MK. Effects of exercise on falls, balance, and gait ability in Parkinson’s disease: a meta-analysis. Neurorehabil Neural Repair. 73. Radder DLM, Lígia Silva de Lima A, Domingos J, Keus SHJ, van Nimwegen M, 2016;30:512–527. Bloem BR, et al. Physiotherapy in Parkinson’s disease: a meta-analysis of present treatment modalities. Neurorehabil Neural Repair. 2020;34:871–880. 102. Flynn A, Allen NE, Dennis S, Canning CG, Preston E. Home-based prescribed exercise improves balance-related activities in people with Parkinson’s disease 74. Morris ME, Iansek R, Matyas TA, Summers JJ. Stride length regulation in Par- and has benefits similar to centre-based exercise: a systematic review. kinson’s disease. Normalization strategies and underlying mechanisms. Brain. J Physiother. 2019;65:189–199. 1996;119:551–568.

Invited Topical Review 175 103. Barnish MS, Barran SM. A systematic review of active group-based dance, 134. Amara AW, Wood KH, Joop A, Memon RA, Pilkington J, Tuggle SC, et al. Ran- singing, music therapy and theatrical interventions for quality of life, functional domized, controlled trial of exercise on objective and subjective sleep in Par- communication, speech, motor function and cognitive status in people with kinson’s disease. Mov Disord. 2020;35:947–958. Parkinson’s disease. BMC Neurol. 2020;20:371. 135. Serrao M, Pierelli F, Sinibaldi E, Chini G, Castiglia SF, Priori M, et al. Progressive 104. Carapellotti AM, Stevenson R, Doumas M. The efficacy of dance for improving modular rebalancing system and visual cueing for gait rehabilitation in Parkin- motor impairments, non-motor symptoms, and quality of life in Parkinson’s son’s disease: a pilot, randomized, controlled trial with crossover. Front Neurol. disease: a systematic review and meta-analysis. PLoS One. 2020;15:e0236820. 2019;10:902. 105. de Almeida HS, Porto F, Porretti M, Lopes G, Fiorot D, Bunn PDS, et al. Effect of 136. Landers MR, Navalta JW, Murtishaw AS, Kinney JW, Pirio Richardson S. A high- dance on postural control in people with Parkinson’s disease: a meta-analysis intensity exercise boot camp for persons with Parkinson disease: a phase II, review. J Aging Phys Act. 2020;31:1–11. pragmatic, randomized clinical trial of feasibility, safety, signal of efficacy, and disease mechanisms. J Neurol Phys Ther. 2019;43:12–25. 106. Hackney ME, Earhart GM. Effects of dance on movement control in Parkinson’s disease: a comparison of Argentine tango and American ballroom. J Rehabil Med. 137. Chivers Seymour K, Pickering R, Rochester L, Roberts HC, Ballinger C, Hulbert S, 2009;41:475–481. et al. Multicentre, randomised controlled trial of PDSAFE, a physiotherapist- delivered fall prevention programme for people with Parkinson’s. J Neurol Neu- 107. Hackney ME, Earhart GM. Health-related quality of life and alternative forms of rosurg Psychiatry. 2019;90:774–782. exercise in Parkinson disease. Parkinsonism Relat Disord. 2009;15:644–648. 138. Coe S, Franssen M, Collett J, Boyle D, Meaney A, Chantry R, et al. Physical activity, 108. McGill A, Houston S, Lee RY. Dance for Parkinson’s: a new framework for fatigue, and sleep in people with Parkinson’s disease: a secondary per protocol research on its physical, mental, emotional, and social benefits. Complement Ther analysis from an intervention trial. Parkinsons Dis. 2018;2018:1517807. Med. 2014;22:426–432. 139. Rafferty MR, Prodoehl J, Robichaud JA, David FJ, Poon C, Goelz LC, et al. Effects of 2 109. Rocha PA, Slade SC, McClelland J, Morris ME. Dance is more than therapy: years of exercise on gait impairment in people with Parkinson disease: the PRET- qualitative analysis on therapeutic dancing classes for Parkinson’s. Complement PD randomized trial. J Neurol Phys Ther. 2017;41:21–30. Ther Med. 2017;34:1–9. 140. Winward C, Sackley C, Meek C, Izadi H, Barker K, Wade D, et al. Weekly exercise 110. Tsang HW, Cheung L, Lak DC. Qigong as a psychosocial intervention for depressed does not improve fatigue levels in Parkinson’s disease. Mov Disord. 2012;27:143– elderly with chronic physical illnesses. Int J Geriatr Psychiatry. 2002;17:1146– 146. 1154. 141. DiFrancisco-Donoghue J, Lamberg EM, Rabin E, Elokda A, Fazzini E, Werner WG. 111. Kemp CA. Qigong as a therapeutic intervention with older adults. J Holist Nurs. Effects of exercise and B vitamins on homocysteine and glutathione in Parkin- 2004;22:351–373. son’s disease: a randomized trial. Neurodegener Dis. 2012;10:127–134. 112. Seki K, Chisaka M, Eriguchi M, Yanagie H, Hisa T, Osada I, et al. An attempt to 142. Ashburn A, Fazakarley L, Ballinger C, Pickering R, McLellan LD, Fitton C. integrate Western and Chinese medicine: rationale for applying Chinese medicine A randomised controlled trial of a home based exercise programme to reduce the as chronotherapy against cancer. Biomed Pharmacother. 2005;59(S1):S132–S140. risk of falling among people with Parkinson’s disease. J Neurol Neurosurg Psy- chiatry. 2007;78:678–684. 113. Ni X, Liu S, Lu F, Shi X, Guo X. Efficacy and safety of Tai Chi for Parkinson’s dis- ease: a systematic review and meta-analysis of randomized controlled trials. 143. Kalilani L, Asgharnejad M, Palokangas T, Durgin T. Comparing the incidence of PLoS One. 2014;9:e99377. falls/fractures in Parkinson’s disease patients in the US population. PLoS One. 2016;11:e0161689. 114. Yang Y, Li XY, Gong L, Zhu YL, Hao YL. Tai Chi for improvement of motor function, balance and gait in Parkinson’s disease: a systematic review and meta-analysis. 144. Goodwin VA, Richards SH, Henley W, Ewings P, Taylor AH, Campbell JL. An PLoS One. 2014;9:e102942. exercise intervention to prevent falls in people with Parkinson’s disease: a pragmatic randomised controlled trial. J Neurol Neurosurg Psychiatry. 115. Zhou J, Yin T, Gao Q, Yang XC. A meta-analysis on the efficacy of tai chi in patients 2011;82:1232–1238. with Parkinson’s disease between 2008 and 2014. Evid Based Complement Alternat Med. 2015;2015:593263. 145. Wong-Yu ISK, Mak MKY. Multisystem balance training reduces injurious fall risk in Parkinson disease: a randomized trial. Am J Phys Med Rehabil. 2019;98:239– 116. Choi HJ. Effects of therapeutic Tai chi on functional fitness and activities of daily 244. living in patients with Parkinson disease. J Exerc Rehabil. 2016;12:499–503. 146. Canning CG, Sherrington C, Lord SR, Close JC, Heritier S, Heller GZ, et al. Exercise 117. Gao Q, Leung A, Yang Y, Wei Q, Guan M, Jia C, et al. Effects of Tai Chi on balance for falls prevention in Parkinson disease: a randomized controlled trial. and fall prevention in Parkinson’s disease: a randomized controlled trial. Clin Neurology. 2015;84:304–312. Rehabil. 2014;28:748–753. 147. Smania N, Corato E, Tinazzi M, Stanzani C, Fiaschi A, Girardi P, et al. Effect of 118. Amano S, Nocera JR, Vallabhajosula S, Juncos JL, Gregor RJ, Waddell DE, et al. The balance training on postural instability in patients with idiopathic Parkinson’s effect of Tai Chi exercise on gait initiation and gait performance in persons with disease. Neurorehabil Neural Repair. 2010;24:826–834. Parkinson’s disease. Parkinsonism Relat Disord. 2013;19:955–960. 148. Ashburn A, Pickering R, McIntosh E, Hulbert S, Rochester L, Roberts HC, et al. 119. Choi HJ, Garber CE, Jun TW, Jin YS, Chung SJ, Kang HJ. Therapeutic effects of tai chi Exercise- and strategy-based physiotherapy-delivered intervention for pre- in patients with Parkinson’s disease. ISRN Neurol. 2013;2013:548240. venting repeat falls in people with Parkinson’s: the PDSAFE RCT. Health Technol Assess. 2019;23:1–150. 120. Vergara-Diaz G, Osypiuk K, Hausdorff JM, Bonato P, Gow BJ, Miranda JG, et al. Tai Chi for reducing dual-task gait variability, a potential mediator of fall risk in 149. Morris ME, Menz HB, McGinley JL, Watts JJ, Huxham FE, Murphy AT, et al. Parkinson’s disease: a pilot randomized controlled trial. Glob Adv Health Med. A randomized controlled trial to reduce falls in people with Parkinson’s disease. 2018;7:2164956118775385. Neurorehabil Neural Repair. 2015;29:777–785. 121. Li F, Harmer P. Economic Evaluation of a Tai Ji Quan intervention to reduce falls in 150. Nieuwboer A, Kwakkel G, Rochester L, Jones D, van Wegen E, Willems AM, et al. people with Parkinson disease, Oregon, 2008-2011. Prev Chronic Dis. Cueing training in the home improves gait-related mobility in Parkinson’s 2015;12:E120. disease: the RESCUE trial. J Neurol Neurosurg Psychiatry. 2007;78:134–140. 122. Xiao CM, Zhuang YC. Effect of health Baduanjin Qigong for mild to moderate 151. Li F, Harmer P, Fitzgerald K, Eckstrom E, Stock R, Galver J, et al. Tai chi and Parkinson’s disease. Geriatr Gerontol Int. 2016;16:911–919. postural stability in patients with Parkinson’s disease. N Engl J Med. 2012;366:511–519. 123. Liu XL, Chen S, Wang Y. Effects of health qigong exercises on relieving symptoms of Parkinson’s disease. Evid Based Complement Alternat Med. 2016;2016:5935782. 152. Cusso ME, Donald KJ, Khoo TK. The impact of physical activity on non-motor symptoms in Parkinson’s disease: a systematic review. Front Med. 2016;3: 124. Lee HJ, Kim SY, Chae Y, Kim MY, Yin C, Jung WS, et al. Turo (Qi Dance) program 35. for Parkinson’s disease patients: randomized, assessor blind, waiting-list control, partial crossover study. Explore. 2018;14:216–223. 153. da Silva FC, Iop RDR, de Oliveira LC, Boll AM, de Alvarenga JGS, Gutierres Filho PJB, et al. Effects of physical exercise programs on cognitive function in Parkin- 125. Pérez-de la Cruz S. A bicentric controlled study on the effects of aquatic Ai Chi in son’s disease patients: a systematic review of randomized controlled trials of the Parkinson disease. Complement Ther Med. 2018;36:147–153. last 10 years. PLoS One. 2018;13:e0193113. 126. Nocera JR, Amano S, Vallabhajosula S, Hass CJ. Tai Chi exercise to improve non- 154. Law CK, Lam FM, Chung RC, Pang MY. Physical exercise attenuates cognitive decline motor symptoms of Parkinson’s disease. J Yoga Phys Ther. 2013;3:10. and reduces behavioural problems in people with mild cognitive impairment and dementia: a systematic review. J Physiother. 2020;66:9–18. 127. Zhu M, Zhang Y, Pan J, Fu C, Wang Y. Effect of simplified Tai Chi exercise on relieving symptoms of patients with mild to moderate Parkinson’s disease. 155. Rothwell J. Transcranial brain stimulation: past and future. Brain Neurosci Adv. J Sports Med Phys Fitness. 2020;60:282–288. 2018;2:2398212818818070. 128. Moon S, Schmidt M, Smirnova IV, Colgrove Y, Liu W. Qigong exercise may reduce 156. Maeda F, Keenan JP, Tormos JM, Topka H, Pascual-Leone A. Modulation of cor- serum TNF-a levels and improve sleep in people with Parkinson’s disease: a pilot ticospinal excitability by repetitive transcranial magnetic stimulation. Clin Neu- study. Medicines. 2017;4:23. rophysiol. 2000;111:800–805. 129. Van Puymbroeck M, Walter AA, Hawkins BL, Sharp JL, Woschkolup K, Urrea- 157. Peinemann A, Reimer B, Löer C, Quartarone A, Münchau A, Conrad B, et al. Long- Mendoza E, et al. Functional improvements in Parkinson’s disease following a lasting increase in corticospinal excitability after 1800 pulses of subthreshold 5 randomized trial of yoga. Evid Based Complement Alternat Med. Hz repetitive TMS to the primary motor cortex. Clin Neurophysiol. 2018;2018:8516351. 2004;115:1519–1526. 130. Jin X, Wang L, Liu S, Zhu L, Loprinzi PD, Fan X. The impact of mind-body exercises 158. Huang YZ, Edwards MJ, Rounis E, Bhatia KP, Rothwell JC. Theta burst stimulation on motor function, depressive symptoms, and quality of life in Parkinson’s dis- of the human motor cortex. Neuron. 2005;45:201–206. ease: a systematic review and meta-analysis. Int J Environ Res Public Health. 2019;17:31. 159. González-García N, Armony JL, Soto J, Trejo D, Alegría MA, Drucker-Colín R. Ef- fects of rTMS on Parkinson’s disease: a longitudinal fMRI study. J Neurol. 131. Cugusi L, Manca A, Bergamin M, Di Blasio A, Monticone M, Deriu F, et al. Aquatic 2011;258:1268–1280. exercise improves motor impairments in people with Parkinson’s disease, with similar or greater benefits than land-based exercise: a systematic review. 160. Strafella AP, Paus T, Barrett J, Dagher A. Repetitive transcranial magnetic stim- J Physiother. 2019;65:65–74. ulation of the human prefrontal cortex induces dopamine release in the caudate nucleus. J Neurosci. 2001;21:RC157. 132. Pérez de la Cruz S. Effectiveness of aquatic therapy for the control of pain and increased functionality in people with Parkinson’s disease: a randomized clinical 161. Yang C, Guo Z, Peng H, Xing G, Chen H, McClure MA, et al. Repetitive transcranial trial. Eur J Phys Rehabil Med. 2017;53:825–832. magnetic stimulation therapy for motor recovery in Parkinson’s disease: a meta- analysis. Brain Behav. 2018;8:e01132. 133. Kurt EE, Büyükturan B, Büyükturan Ö, Erdem HR, Tuncay F. Effects of Ai Chi on balance, quality of life, functional mobility, and motor impairment in patients with Parkinson’s disease. Disabil Rehabil. 2018;40:791–797.

176 Pang: Physiotherapy management of Parkinson’s disease 162. Chung CL, Mak MK. Effect of repetitive transcranial magnetic stimulation on 172. Boggio PS, Ferrucci R, Rigonatti SP, Covre P, Nitsche M, Pascual-Leone A, et al. physical function and motor signs in Parkinson’s disease: a systematic review Effects of transcranial direct current stimulation on working memory in patients and meta-analysis. Brain Stimul. 2016;9:475–487. with Parkinson’s disease. J Neurol Sci. 2006;249:313–318. 163. Xie YJ, Gao Q, He CQ, Bian R. Effect of repetitive transcranial magnetic stimulation 173. Doruk D, Gray Z, Bravo GL, Pascual-Leone A, Fregni F. Effects of tDCS on executive on gait and freezing of gait in Parkinson disease: a systematic review and meta- function in Parkinson’s disease. Neurosci Lett. 2014;582:27–31. analysis. Arch Phys Med Rehabil. 2020;101:130–140. 174. Ferrucci R, Cortese F, Bianchi M, Pittera D, Turrone R, Bocci T, et al. Cerebellar and 164. Goodwill AM, Lum JAG, Hendy AM, Muthalib M, Johnson L, Albein-Urios N, motor cortical transcranial stimulation decrease levodopa-induced dyskinesias et al. Using non-invasive transcranial stimulation to improve motor and in Parkinson’s disease. Cerebellum. 2016;15:43–47. cognitive function in Parkinson’s disease: a systematic review and meta- analysis. Sci Rep. 2017;7:14840. 175. Beretta VS, Conceição NR, Nóbrega-Sousa P, Orcioli-Silva D, Dantas LKBF, Gobbi LTB, et al. Transcranial direct current stimulation combined with 165. Nitsche MA, Paulus W. Excitability changes induced in the human motor cortex physical or cognitive training in people with Parkinson’s disease: a sys- by weak transcranial direct current stimulation. J Physiol. 2000;527:633–639. tematic review. J Neuroeng Rehabil. 2020;17:74. 166. Bolzoni F, Pettersson LG, Jankowska E. Evidence for long-lasting subcortical 176. Samii A, Ryan-Dykes P, Tsukuda RA, Zink C, Franks R, Nichol WP. Telemedicine facilitation by transcranial direct current stimulation in the cat. J Physiol. for delivery of health care in Parkinson’s disease. J Telemed Telecare. 2013;591:3381–3399. 2006;12:16–18. 167. Tanaka T, Takano Y, Tanaka S, Hironaka N, Kobayashi K, Hanakawa T, et al. 177. Chen YY, Guan BS, Li ZK, Yang QH, Xu TJ, Li HB, et al. Application of telehealth Transcranial direct-current stimulation increases extracellular dopamine levels intervention in Parkinson’s disease: a systematic review and meta-analysis. in the rat striatum. Front Syst Neurosci. 2013;7:6. J Telemed Telecare. 2020;26:3–13. 168. Stagg CJ, O’Shea J, Kincses ZT, Woolrich M, Matthews PM, Johansen-Berg H. 178. Ginis P, Nieuwboer A, Dorfman M, Ferrari A, Gazit E, Canning CG, et al. Feasibility Modulation of movement-associated cortical activation by transcranial direct and effects of home-based smartphone-delivered automated feedback training current stimulation. Eur J Neurosci. 2009;30:1412–1423. for gait in people with Parkinson’s disease: a pilot randomized controlled trial. Parkinsonism Relat Disord. 2016;22:28–34. 169. Filmer HL, Dux PE, Mattingley JB. Applications of transcranial direct current stimulation for understanding brain function. Trends Neurosci. 2014;37:742–753. 179. Khalil H, Busse M, Quinn L, Nazzal M, Batyha W, Alkhazaleh S, et al. A pilot study of a minimally supervised home exercise and walking program for people with 170. Broeder S, Nackaerts E, Heremans E, Vervoort G, Meesen R, Verheyden G, et al. Parkinson’s disease in Jordan. Neurodegener Dis Manag. 2017;7:73–84. Transcranial direct current stimulation in Parkinson’s disease: neurophysiological mechanisms and behavioral effects. Neurosci Biobehav Rev. 2015;57:105–117. 180. Gandolfi M, Geroin C, Dimitrova E, Boldrini P, Waldner A, Bonadiman S, et al. Virtual reality telerehabilitation for postural instability in Parkinson’s disease: a 171. Pereira JB, Junqué C, Bartrés-Faz D, Martí MJ, Sala-Llonch R, Compta Y, et al. multicenter, single-blind, randomized, controlled trial. Biomed Res Int. Modulation of verbal fluency networks by transcranial direct current stimulation 2017;2017:7962826. (tDCS) in Parkinson’s disease. Brain Stimul. 2013;6:16–24.

Journal of Physiotherapy 67 (2021) 158–159 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Editorial Prevention of spinal pain Tarcisio F de Campos a, Mark R Elkins b,c a Department of Health Professions, Faculty of Medicine, Health and Human Sciences, Macquarie University, Sydney, Australia; b Editor, Journal of Physiotherapy; c Faculty of Medicine and Health, University of Sydney, Sydney, Australia Journal of Physiotherapy’s latest article collection relates to the recurrence and the need for effective strategies to prevent re- prevention of spinal pain. This collection: facilitates access to recent currences of spinal pain. important findings in the field; highlights trends in the study designs, methodology, populations and interventions addressed by the Despite the clear evidence that spinal pain is a long-term problem research; and provides a scoping overview of avenues for further characterised by recurrent episodes,4,9–13 there has been minimal research. An article collection on low back pain was curated in 2018.1 attention given to investigating effective strategies for the prevention Despite its recency, a remarkable number of papers have addressed of spinal pain. Over the last two decades, the number of randomised prevention of low back pain and neck pain in the interim. Further- controlled trials investigating interventions for spinal pain has grown more, prevention of spinal pain has been a particular focus among the rapidly; however, the vast majority have investigated interventions more recent papers. looking at spinal pain treatment and very few have investigated spinal pain prevention strategies.14,15 Therefore, greater understand- Most people experience low back pain and neck pain at some ing regarding effective strategies to prevent spinal pain represents an point in their lives. Spinal pain, including low back pain and neck important research priority.16–19 pain, are among the leading causes of disability, affecting over half a billion people around the world.2,3 It is estimated that the global 1- This led to further investigation into ways to prevent episodes of year mean prevalence of activity-limiting low back pain is around spinal pain. Two recently published high-quality systematic reviews 40%, and neck pain is approximately 11.5%.4,5 Despite much research with meta-analysis have investigated interventions aimed at pre- over the past decades dedicated to understanding spinal pain, the venting a new episode of low back pain and a new episode of neck burden associated with this condition has failed to reduce. To reduce pain.14,20 The review of prevention of low back pain found moderate- the global burden associated with spinal pain, effective strategies to quality evidence that an exercise program in combination with edu- prevent spinal pain are important. Furthermore, given the recurrent cation reduces the risk of a new episode of low back pain by 45% (RR nature of spinal pain,4,6 interventions that can reduce the risk of 0.55, 95% CI 0.41 to 0.74), and low-quality evidence that an exercise recurrence in those who have previously experienced an episode are program alone may reduce the risk by 35% (RR 0.65, 95% CI 0.50 to also particularly important. 0.86). Most other intervention strategies (eg, education alone, use of back belts, use of shoe insoles, and ergonomic programs) either Although most cases of acute low back pain or neck pain improve lacked evidence or appeared to be ineffective.14 considerably in the first few weeks, rates of recurrence within 12 months are high.4,6 A recent systematic review of the literature The second study in this article collection is a systematic review investigating the risk of recurrence of low back pain included eight and meta-analysis by de Campos et al,20 which summarises the re- studies.7 This review reported that only one study8 was considered to sults of five trials investigating the evidence for interventions aiming have an appropriate estimate for rate of a recurrence of low back pain to reduce the risk of a new episode of neck pain.20 It appears that within one year as the authors used a short inception period. This exercise programs are also likely to prevent neck pain episodes. In study conducted by Stanton et al8 reported an estimated recurrence this review, only two trials were pooled in the meta-analysis for the rate of 33%. The authors of that review, however, suggested that it exercise intervention contrast. The included trials in this literature was not yet possible to obtain reliable estimates of recurrence pro- review investigated different approaches to the exercise. In one trial, portions as most included studies have small sample sizes, and low Sihawong et al21 investigated an exercise program involving methodological quality. stretching and endurance exercises restricted to the muscles within the neck region, while the trial by Tveito et al22 evaluated an inte- To overcome this gap in the literature, Da Silva et al4 conducted a grated health program including a generalised whole-body exercise high-quality prospective inception cohort study in Australia including program. Consequently, the most effective exercise program to 250 participants who had recovered from an episode of low back pain reduce future neck pain episodes remains unclear. within the previous month.4 This first study in the article collection investigated how commonly low back pain recurrences occur within The current evidence on prevention of low back pain demon- 1 year of recovering from a previous episode of low back pain, using strates that exercise alone and exercise in combination with educa- three different definitions of low back pain recurrence. The study tion are effective in reducing the risk of low back pain episodes (35 found that by 1 year, 69% (95% CI 62 to 74) of participants experienced and 45% risk reduction, respectively, at one year); however, the ma- a recurrence of any episode of low back pain, 40% (95% CI 33 to 46) of jority of the exercise programs in these trials are relatively costly, participants had a recurrence of an episode of low back pain leading inflexible and time-consuming, potentially making uptake of such to at least moderate activity limitation, and 41% (95% CI 34 to 46) of programs difficult.14 To overcome these barriers, de Campos et al,23 in participants had a recurrence of low back pain for which healthcare the third study in this article collection, investigated the estimated was sought. Results from this study confirmed the high rates of effect of a McKenzie-based self-management exercise and educa- tional approach compared to a minimal intervention control group, in https://doi.org/10.1016/j.jphys.2021.06.012 1836-9553/© 2021 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

Editorial 159 preventing recurrences of low back pain and future care seeking in Some of the focus of prevention studies can be on preventing future people recently recovered from an episode of non-specific low back episodes of spinal pain,14,20,23 preventing flare-ups in those with pain. The authors also aimed to investigate if the approach reduces current mild pain or reducing future impact of spinal pain and asso- the impact of low back pain over 1 year. The authors in this trial found ciated disability.25 It is unclear if prevention interventions are equally that the best estimate is that a McKenzie-based self-management effective for these different populations and outcomes. Future studies exercise and education program does not produce a substantial that include mixed populations such as those with and without a reduction in the risk of an activity-limiting episode of low back pain previous history of spinal pain, or those with and without current mild (HR 1.11, 95% CI 0.80 to 1.54) but may produce a substantial reduction pain, could investigate whether these factors are moderators for the in recurrence of an episode of low back pain leading to care-seeking effectiveness of interventions aiming to prevent spinal pain. (HR 0.69, 95% CI 0.46 to 1.04). The authors also found clear evidence that any effect on the impact of low back pain over 1 year is In summary, the work presented in this article collection includes negligible. important developments in research into the prevention of spinal pain. The study designs address the probability and prognostic factors The remaining article in this collection investigates the prefer- for a recurrence of low back pain over 1-year period,4 systematic ences of people presenting with a recent history of low back pain, for reviews on the evidence of prevention strategies to reduce the risk of exercise programs to reduce the risk of future episodes of low back low back pain14 and neck pain,20 the effectiveness of a prevention pain, using a discrete choice experiment design. This study by Ferreira strategy to reduce the risk of low back pain and its impact over a 12- et al24 explored the influence of program features and participant- month period,23 and finally the preference of people presenting with level characteristics on their preferences for low back pain exercise a recent history of low back pain, for exercise programs to reduce the prevention programs and determined their willingness to pay for the risk of future episodes of low back pain, using a discrete choice programs. This study reported data on over 640 participants and experiment.24 Importantly, each paper has clear implications for showed that people generally preferred brief, low-cost and home- clinical physiotherapists, which are identifiable in the paper’s ‘What based exercise programs. The results of this study diverged from this study adds’ summary box and discussion section. the current evidence from randomised controlled trials, which in- dicates that exercise programs that reduce the risk of a new episode Competing interest: Nil. of low back pain are typically of long duration and high frequency and Source of support: Nil. have been conducted in settings such as outpatient clinics or the Acknowledgement: Nil. workplace. Provenance: Invited. Not peer reviewed. Correspondence: Mark R Elkins, Centre for Education & Workforce This article collection on prevention of spinal pain identifies Development, Sydney Local Health District, Sydney, Australia. Email: important implications for future research and helps identify priorities [email protected] for future studies investigating the prevention of spinal pain. The first research implication is that the best exercise approach and dosage to References prevent spinal pain is still unclear. Whilst the majority of existing literature suggests exercise interventions can help reduce spinal pain, 1. Hush J, et al. J Physiother. 2018;64:208–209. some approaches such as that investigated in de Campos et al23 trial 2. Vos T, et al. Lancet. 2016;388:1545–1602. are not effective, so future trials investigating head-to-head compar- 3. Vos T, et al. Lancet. 2015;386:743–800. isons of different exercise interventions to prevent spinal pain are 4. Da Silva T, et al. J Physiother. 2019;65:159–165. required. Moreover, investigation of mediators within these trials may 5. Hogg-Johnson S, et al. Spine. 2008;33:S39–S51. also advance our understanding of the causal mechanisms involved in 6. Carroll LJ, et al. J Manipulative Physiol Ther. 2009;32:S87–S96. effective spinal pain prevention programs. Another, research impli- 7. Da Silva T, et al. J Orthop Sports Phys Ther. 2017;47:305–313. cation is that, despite the promising results in de Campos et al23 trial 8. Stanton TR, et al. Spine. 2008;33:2923–2928. in terms of preventing future healthcare seeking, the trial was not 9. Kongsted A, et al. BMC Musculoskelet Disord. 2016;17:2–11. adequately powered to address that question. This is an important 10. Dunn KM, et al. Am J Epidemiol. 2006;163:754–761. outcome and future prevention studies should investigate whether a 11. Dunn KM, et al. BMJ Open. 2013;3, e003838. prevention program can reduce healthcare seeking related to spinal 12. Kongsted A, et al. Spine J. 2015;15:885–894. pain. Furthermore, there is a lack of high-quality studies investigating 13. Ailliet L, et al. Eur J Pain. 2018;22:103–113. the impact of prevention strategies for spinal pain on other important 14. Steffens D, et al. JAMA Internal Medicine. 2016;176:199–208. outcomes such as quality of life, work ability and days lost from work. 15. Linton SJ, et al. Spine. 2001;26:778–787. Finally, an important research implication is to better define what is 16. Foster NE, et al. Lancet. 2018;391:2368–2383. meant by prevention of spinal pain and to determine what types of 17. Vassilaki M, et al. Hawaii J Med Public Health. 2014;73:122–126. prevention studies are most important. Prevention can include par- 18. Hoy D, et al. Arthritis Rheumatol. 2012;64:2028–2037. ticipants who have never experienced spinal pain, those who have 19. Hoy D, et al. Ann Rheum Dis. 2014;73:1309–1315. had previous episodes, and those with current low levels of pain. 20. de Campos TF, et al. J Physiother. 2018;64:159–165. 21. Sihawong R, et al. Occup Environ Med. 2014;71:63–70. 22. Tveito TH, et al. J Adv Nurs. 2009;65:110–119. 23. de Campos TF, et al. J Physiother. 2020;66:166–173. 24. Ferreira GE, et al. J Physiother. 2020;66:249–255. 25. de Campos TF, et al. Br J Sports Med. 2021;55:468.

Journal of Physiotherapy 67 (2021) 160 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Readers’ Choice Award goes to an extraordinary paper The Editorial Board is pleased to announce the annual Readers’ Choice Award, which recognises the paper published in Journal of Physio- therapy that generates the most interest by readers of the journal. The winning paper is chosen based on the number of times that each paper published in a given year is downloaded in the 6 months after its day of publication. The Readers’ Choice Award for 2020 is therefore announced in mid-2021. The winning paper from among those published in 2020 is the Invited Topical Review ‘Physiotherapy management for COVID-19 in the acute hospital setting: clinical practice recommendations’ by Peter Thomas and colleagues from Australia, Belgium, Canada and the United Kingdom.1 Consistent with previous years, this and the other Invited Topical Reviews from 2020 all received a high number of downloads.1–3 However, by compiling recommendations for clinical physiotherapy practice in relation to COVID-19 at the beginning of the pandemic, the winning paper harnessed a prodigious amount of interest. It was downloaded 8,000 times on the day it was published. Since then, it has been downloaded over 180,000 times and cited over 300 times. It has been translated into 26 languages, and endorsed by a range of organisations including World Physiotherapy, the Australian Physiotherapy Association, the Canadian Physiotherapy Association, the Association of Chartered Society of Physiotherapists in Respiratory Care UK, Associazione Riabilitatori dell’ Insufficienza Respiratoria, Koninklijk Nederlands Genootschap voor Fysiotherapie, the International Confederation of Cardiorespiratory Physical Therapists, AXXON Physical Therapy in Belgium, and Société de Kinésithérapie de Réanimation. The Altmetric score is approaching 1,000. The win is particularly impressive for one author, Dr Ianthe Boden, who also won the Paper of the Year award in the same year,4 for her cost- effectiveness analysis of preoperative physiotherapy.5 The Editorial Board of Journal of Physiotherapy congratulates Mr Thomas and his co-authors on their success. References 1. Thomas P, Baldwin C, Bissett B, Boden I, Gosselink R, Granger CL, Hodgson C, Jones AY, Kho ME, Moses R, Ntoumenopoulos G. Physiotherapy management for COVID-19 in the acute hospital setting: clinical practice recommendations. J Physiother. 2020;66:73–82. 2. Ostelo RW. Physiotherapy management of sciatica. J Physiother. 2020;66:83–88. 3. Bø K. Physiotherapy management of urinary incontinence in females. J Physiother. 2020;66:147–154. 4. Paper of the Year 2020. J Physiother. 2021;67:4. 5. Boden I, Robertson IK, Neil A, Reeve J, Palmer AJ, Skinner EH, Browning L, Anderson L, Hill C, Story D, Denehy L. Preoperative physiotherapy is cost-effective for preventing pulmonary complications after major abdominal surgery: a health economic analysis of a multicentre randomised trial. J Physiother. 2020;66:180–187. https://doi.org/10.1016/j.jphys.2021.06.001 1836-9553/

Journal of Physiotherapy 67 (2021) 224–227 Appraisal j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research Note: Individual participant data (IPD) meta-analysis A systematic review is a robust method with which to search for, requirements (based on published information). Any reasons for identify, extract, and synthesise evidence from individual studies to study exclusion should be clearly documented.4 answer a specific research question.1 Meta-analysis is a statistical analysis approach used in some systematic reviews to combine Accessing individual participant data quantitative information across studies, in order to produce overall summaries of the evidence (eg, of a treatment’s effect). Meta-analyses IPD can be collected through direct contact with primary study are most often conducted using data that has been extracted from authors with the use of data-sharing agreements or through data- peer-reviewed publications included in systematic reviews; such data sharing platforms (eg, disease-specific data repositories, journal are often called aggregate data, since they represent information websites). It is preferable to deal directly with study authors, as the combined across all participants in a particular study. The extracted IPD in a data-sharing platform may be incomplete (eg, not all out- data typically include a small number of data pieces from each study, comes provided) and unavailable to analyse and store locally. IPD such as the change in pain (mean, standard deviation) between projects should clearly describe how data were requested, collected treated and untreated study groups, which would allow a treatment and managed (eg, list and define all study-level and participant-level effect estimate and its confidence interval to be calculated. An data that were sought, including baseline and follow-up information). aggregate data meta-analysis is a useful approach with which to summarise the average overall effect of a treatment. However, having Data verification, manipulation and replication only aggregated group data limits the analyses that are possible, and in particular makes it problematic to examine relationships where Data from each primary study are evaluated, once received, to individual participant-level covariates are of interest. To address this, confirm comparability to the study publication(s) for descriptive another option to synthesise evidence across studies is to use the baseline data, range of included variables and missing observations. original, participant-level study data, using an approach called indi- Any discrepancies or missing information from those presented in the vidual participant data (IPD) meta-analysis. original publication(s) should be clarified with the primary study authors. Risk of bias classifications should also be updated at this This Research Note describes the steps involved in an IPD meta- stage, based on the IPD itself. analysis, explains when this research approach is most useful, and discusses key advantages, challenges and potential future directions. Study variable mapping Table 1 provides definitions of some key terms. Although this Research Note focuses on meta-analysis of randomised trials evalu- Once IPD are separately finalised for each study, they need to be ating treatment effectiveness, most points apply more broadly. harmonised as far as possible, so that included variables are consis- tently defined across studies. The homogeneity of the resulting What is an IPD meta-analysis project and what are the steps? master dataset depends on data availability of common measure- ments in the primary studies. Whenever possible, variables measured IPD meta-analysis projects are often considered to be the ‘gold continuously should be kept in their continuous data form. Any standard’ for synthesising evidence across studies.2 Table 2 highlights standardisation and translation process should be described in detail key differences in the steps of a systematic review with IPD meta- in a data management plan. analysis and a systematic review with a traditional aggregate data meta-analysis. With an IPD project, raw individual-level data about Analysis each participant (including baseline demographics, health conditions, prognostic factors and other relevant characteristics, as well as IPD meta-analyses are planned and described a priori and include outcome data) are obtained for each published study, then cleaned, descriptive, analytic and, possibly, exploratory analyses. There are harmonised and synthesised together. The steps involved in con- two main approaches to conduct IPD meta-analysis: one-stage or ducting a high-quality IPD project are described below, and additional two-stage. In a two-stage IPD meta-analysis, an estimate of the effect detail is provided in a forthcoming handbook.3 of interest (eg, treatment effect or interaction between treatment and participant-level characteristic) is first separately calculated for each Search and selection trial; these can be presented in forest plots and then (in the second stage) combined in a similar manner to a traditional meta-analysis to A well conducted IPD meta-analysis is part of a systematic review produce an overall summary. In one-stage IPD meta-analysis, a single with clear research questions, inclusion criteria, comprehensive multi-level model is estimated based on the IPD for all studies search, systematic selection, transparent study-level data extraction together, while accounting for the clustering of data within each trial. and risk of bias assessment. Selection criteria for systematic reviews The one-stage approach is appealing as it simultaneously estimates using IPD are similar to traditional reviews, typically including pop- multiple parameters in a single step and allows more flexibility in ulation, intervention, comparison and outcome (PICO) criteria, and modelling assumptions. The two-stage approach is appealing as it may include additional criteria such as sample size or risk of bias https://doi.org/10.1016/j.jphys.2021.04.001 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Appraisal Research Note 225 Table 1 Definitions of key terms related to individual participant data meta-analysis. Term Definition Meta-analysis Statistical technique for combining quantitative evidence across multiple studies Aggregate data Information averaged or summarised across a group of participants in a study, such as the overall treatment effect, mean age or the proportion of females Typically as reported in publications of the primary studies Traditionally extracted and used for meta-analysis Individual participant Raw participant-level data collected from the primary studies data (IPD) Made available for IPD meta-analysis, most commonly from the study researchers One-stage, two-stage Different statistical approaches to conducting the meta-analysis of IPD Two-stage: obtain estimates (aggregate data) for each study separately, then pool these across studies One-stage: single, usually hierarchical or multi-level, statistical model synthesises data from all IPD in a single step, whilst accounting for clustering of participants within trials Treatment effect When the treatment effectiveness size differs based on another factor (eg, participant characteristic) modification Table 2 Key differences between a traditional aggregate data meta-analysis and an individual participant data (IPD) meta-analysis for treatment effectiveness questions. Characteristic Traditional aggregate data meta-analysis Differences when using an IPD meta-analysis Study identification Systematic review with comprehensive search to identify None – usually IPD meta-analysis projects also use a systematic review with all studies that answer the research question comprehensive search to identify all studies that answer the research question. Study inclusion All available studies, including published and (where identifiable) In addition, due to the painstaking nature of retrieving IPD, the project may apply unpublished studies additional selection criteria to identify a subset of the studies (those low risk of bias, larger sample size) to prioritise for IPD collection. Data collected and Aggregate data extracted from each study published report, Original participant data are requested, obtained and cleaned from each study in close collaboration with the original study researchers. analysed or sometimes requested from study authors directly Analysis aims To summarise the average treatment effect across all studies, In addition, IPD meta-analyses usually aim to model patient-level relationships; for each outcome reported in particular, to obtain treatment effects adjusted for prognostic factors, and to To quantify any between-study heterogeneity in the treatment examine potential treatment effect modification by participant-level effect, and potentially assess subgroups defined by study or characteristics (treatment-covariate interactions). Furthermore, any outcomes population-level characteristics (eg, mean age) or time points available in the IPD can be considered, regardless of whether they were reported upon publication. enables familiar meta-analysis methods and the presentation of in- Primary study data may be missing for traditional meta-analysis if dividual study results, and so eases interpretation. When they use the a study was not published, or results (or outcomes) were partially or same estimation methods and assumptions, one-stage and two-stage selectively published. In the latter situation, a larger primary study approaches garner generally similar results.5 When most studies sample (eg, if only a study population subset was published) or are small (ie, patients and events are few), a one-stage approach is additional outcome measures, measured but not reported, may be more exact and therefore preferable. When examining potential available by collecting and analysing IPD. participant-level modifiers of treatment effect, this is performed by estimating treatment-covariate interactions that quantify how treat- Larger amounts of more homogeneous data may be available for ment effectiveness varies as participant-level characteristics change. syntheses using IPD if exposure, covariate or outcome measures Riley et al describe how to perform this in a one-stage or two-stage selected for presentation in publications differ across studies. If study approach,6 and stress the importance of separating within-trial results are reported across relevant primary studies using different from across-trial information to avoid aggregation bias. scales, follow-up time points or cut-off values, these cannot easily be synthesised using traditional meta-analysis and use of IPD can Reporting potentially allow for standardisation of these measures (eg, Levis 2020; Holden 2021).8,9 The PRISMA reporting guideline and checklist extension for IPD (PRISMA-IPD) should be followed for reporting.7 Additionally, analyses that require large sample sizes, such as development and testing of prediction models (eg, Hudda 201910) or When is an IPD meta-analysis the best approach to address a analyses of treatment effect modifiers (eg, Hayden 201911), can health research question? benefit from the use of IPD to increase power. These types of analyses investigate individual characteristics at a participant level in ways IPD studies are time and resource intensive; it often takes more beyond what is typically planned and feasible for the primary studies. than 2 years to obtain, clean and synthesise IPD. The best value for Very few randomised controlled trials are designed to detect treat- this considerable expense is when: the aggregate data needed to ment effect modifiers.12 An adequately powered primary study (ie, answer the research question are not reported/available in the pri- trial) to assess treatment effect modification or compare multiple mary study publications, and/or the IPD are needed to go beyond the treatment approaches would need to be extremely large (and more analyses or aims of the original primary studies. These situations expensive than IPD).13,14 particularly occur in health research questions related to diagnosis, prognosis or treatment effectiveness, as presented in Table 3, where What are the advantages of an IPD meta-analysis project? participant-level covariates and relationships are of interest for modelling. There are several advantages of systematic reviews using IPD over traditional meta-analyses based on aggregate data.15 Systematic re- views including IPD may improve data availability and quality to reduce publication bias. Re-analyses of the original raw participant

226 Appraisal Research Note Table 3 Examples of individual participant data (IPD) meta-analysis projects for different types of questions. Question type Research question and use of IPD Diagnostic test accuracy Comparison of two depression scales: assessment of diagnostic accuracy in a manner that was not analysed in the primary studies (heterogeneous tool measurements; multiple cut-offs rather than only what was published; no previous meta-analysis and only two primary studies were available to answer this question).8 Predictive modelling Development and validation of a prediction model for fat mass in children: IPD from four primary studies provided sufficient data to develop and validate a model using routinely available risk factors in existing datasets. The final model including height, weight, age, sex and ethnicity was reported to have high predictive accuracy.10 Overall prognosis and Prognostic factors for non-traumatic adolescent knee pain: IPD enabled selection and harmonising of prognostic factors and both short-term prognostic factors and long-term outcomes across studies that had not been reported in a homogeneous way.9 Intervention effectiveness Exercise treatment for chronic LBP: this IPD meta-analysis was able to identify potential treatment effect modifiers for exercise therapy that were not available in the primary studies. There was consistent evidence that heavy work demands and use of pain medications each modify the effectiveness of exercise therapy compared with other interventions.11 data can allow more consistent analyses or data presentation, insurmountable cost to prospectively collect adequate data for ef- participant subgroups, or definition of outcomes, and can reduce risk fect modification analyses, IPD analyses are very efficient. of bias concerns. Also, unpublished trials or unreported/incomplete data (and outcomes) can be included if IPD are available for these The future of IPD meta-analysis studies. Most systematic reviews summarise aggregate data.22 However, IPD meta-analysis can enable better investigation of subgroups availability of IPD is likely to increase in the future due to an increase and treatment effect modifiers, considering effects in participants in data repositories, data-sharing initiatives and expectations from with different characteristics (disentanglement of participant-level research funders. The International Committee of Medical Journal Edi- and study-level sources of heterogeneity in treatment effect). The tors (ICMJE) member journals required a data-sharing statement as of greater power for these types of analyses is difficult to achieve from July 2018, and a data-sharing plan for trials registered after January individual studies or from aggregate data. Traditional meta-analysis 2019. While they do not require data sharing, they have established can explore treatment-covariate interactions by meta-regression on minimal requirements intended to move the medical research field study-level characteristics. These analyses are limited to study/ toward the goal of universal data sharing. In addition to top-down population-level data (with limited variation at the aggregate mandates, there is increased recognition of the academic value of level – eg, mean age) and are prone to study-level confounding dataset development and sharing. Data sharing may increase the (causing aggregation bias).6 IPD gives more power, a larger number of feasibility and conduct of IPD meta-analysis through more readily individual-level covariates and a wider range of covariate values. IPD available datasets, and through a reduction in current barriers meta-analysis enables direct derivation of desired information at (ethical, regulatory).23 There is still a long way to go to achieve this; different time points, the correlation amongst multiple outcomes or even with a pandemic, the willingness to share IPD is low.24 time-points to be accounted for in the analysis (which can lead to efficiency gains),16 the modelling of continuous outcomes and vari- Future prospective coordination and collaboration for more ables on their continuous scale (thereby avoiding the use of arbitrary consistent data collection will make the efforts of IPD meta-analysis cut-off values), and the adjustment for prognostic factors and effect more beneficial to the clinical community.18 IPD meta-analysis is an modifiers to improve consistency for network meta-analysis.17 All of ethical, efficient research method that reuses data and enables in- these advantages have the potential to provide better data, which can vestigations of participant-level characteristics, treatment effect and enhance clinical decision-making. outcomes that would otherwise not be feasible with most current trial sizes. They generate additional knowledge from existing studies, The collaborative nature of IPD projects provides the advantage of thereby reducing the need for more primary studies and limiting a larger team, with input and engagement on the project from research waste. The clinical utility of IPD meta-analysis depends upon experienced trialists. Furthermore, this may enable more unpublished well-conducted and planned trials, the collection of a set of minimum data to be included, facilitate prospective planning of multisite and variables, and ethics and informed consent in primary trials that multi-country trials, support standardisation of prognostic factor allow for future data sharing. measurement, and foster sharing of data through accessible repositories.18 Competing interests: Nil. Source(s) of Support: Nil. What are the challenges of an IPD meta-analysis project? Acknowledgements: We thank Rachel Ogilvie and Heather Taylor for helpful comments and suggestions on this Research Note. A potential limitation of IPD studies is the inconsistent availability Provenance: Invited. Peer reviewed. and measurement of some individual variables. This can limit the Correspondence: Jill A Hayden, Community Health & Epidemi- ability to include prognostic factors or to assess all potential treat- ology, Dalhousie University, Canada. Email: [email protected] ment effect modifiers with the most valid and reliable measures. Jill A Haydena and Richard D Rileyb IPD meta-analysis projects may not be able to retrieve all study aCommunity Health & Epidemiology, Dalhousie University, Canada data; selection bias/availability bias may be a challenge if relevant bCentre for Prognosis Research, School of Medicine, Keele University, data are not available for all trials.19 To address this challenge, where possible, IPD meta-analysis should compare or combine United Kingdom data from the included IPD trials with aggregate data from other trials.20 IPD meta-analysis should report the proportion of eligible References included studies and reasons for unavailability. The reader should consider if bias was likely to be introduced by unavailable studies 1. Higgins JPT, et al, eds. Cochrane handbook for systematic reviews of interventions. and/or data. Cochrane; 2020. Version 6.1. An additional challenge of the IPD approach, acknowledged by all 2. Riley RD, et al. BMJ. 2010;340:221–228. researchers with experience in this study design, is the considerable 3. Riley RD, et al, eds. Individual participant data meta-analysis: A handbook for amount of time and effort that is involved in gathering, testing and compiling data from individual studies.21 Relative to the often- healthcare research. Chichester: Wiley; 2021. 4. Tierney JF, et al. PLoS Med. 2015;12:e1001855. 5. Burke DL, et al. Stat Med. 2017;36:855–875.

Appraisal Research Note 227 6. Riley RD, et al. Stat Med. 2020;39:2115–2137. 15. Tierney JF, et al. PLoS Med. 2020;17:e1003019. 7. Stewart LA, et al. JAMA. 2015;313:1657–1665. 16. Riley RD, et al. Res Synth Methods. 2015;6:157–174. 8. Levis B, et al. JAMA. 2020;323:2290–2300. 17. Riley RD, et al. BMJ. 2017;358:j3932. 9. Holden S, et al. Pain. 2021. https://doi.org/10.1097/j.pain.0000000000002184. 18. Seidler AL, et al. BMJ. 2019;367:l5342. 19. Ventresca M, et al. BMC Med Res Methodol. 2020;20:113–131. Online only. 20. Ahmed I, et al. BMJ. 2012:344:d7762. 10. Hudda MT, et al. BMJ. 2019;366:4293–4303. 21. Stewart LA, Clarke MJ. Stat Med. 1995;14:2057–2079. 11. Hayden JA, et al. Br J Sports Med. 2019;54:1277–1293. 22. Nevitt SJ, et al. BMJ. 2017;357:j1390. 12. Saragiotto BT, et al. J Clin Epidemiol. 2016;79:3–9. 23. Scutt P, et al. BMJ Open. 2020;10:e038765. 13. Gurung T, et al. Physiotherapy. 2015;101:243–251. 24. Li R, et al. Trials. 2021;22:153–157. 14. Mistry D, et al. Spine. 2014;39:618–629.

Journal of Physiotherapy 67 (2021) 190–196 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research Transcranial direct current stimulation provides no clinically important benefits over walking training for improving walking in Parkinson’s disease: a systematic review Lucas Rodrigues Nascimento a,b, Willian Assis do Carmo a, Gabriela Pinto de Oliveira a, Fernando Zanela da Silva Arêas a,c, Fernanda Moura Vargas Dias a a Center of Health Sciences, Discipline of Physiotherapy, Universidade Federal do Espírito Santo, Vitória, Brazil; b NeuroGroup, Discipline of Physiotherapy, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil; c Laboratory of Cognitive Sciences and Neuropsychopharmacology, Department of Physiological Sciences, Universidade Federal do Espírito Santo, Vitória, Brazil KEY WORDS ABSTRACT Parkinson’s disease Questions: Does walking training combined with transcranial direct current stimulation (tDCS) improve Rehabilitation walking (ie, speed, cadence and step length) and reduce falls and freezing, compared with no/sham inter- Gait vention, in people with Parkinson’s disease? Is walking training combined with tDCS superior to walking tDCS training alone? Are any benefits carried over to social participation and/or maintained beyond the inter- Exercise therapy vention period? Design: A systematic review with meta-analyses of randomised clinical trials. Participants: Ambulatory adults with a clinical diagnosis of Parkinson’s disease. Intervention: tDCS combined with walking training. Outcome measures: Primary outcomes were walking speed, cadence and step length. Secondary outcomes were number of falls, fear of falling, freezing of gait and social participation. Results: Five trials involving 117 participants were included. The mean PEDro score of the included trials was 8 out of 10. Participants undertook training for 30 to 60 minutes, two to three times per week, on average for 4 weeks. Moderate-quality evidence indicated that the addition of tDCS to walking training produced negligible additional benefit over the effect of walking training alone on walking speed (MD 20.01 m/s, 95% CI 20.05 to 0.04), step length (MD 1.2 cm, 95% CI 21.2 to 3.5) or cadence (MD 23 steps/minute, 95% CI 26 to 1). No evidence was identified with which to estimate the effect of the addition of tDCS to walking training on freezing of gait, falls and social participation. Conclusion: The addition of tDCS to walking training provided no clinically important benefits on walking in ambulatory people with Parkinson’s disease. Registration: PROSPERO CRD42020162908. [Nascimento LR, do Carmo WA, de Oliveira GP, Arêas FZdS, Dias FMV (2021) Transcranial direct current stimulation provides no clinically important benefits over walking training for improving walking in Parkinson’s disease: a systematic review. Journal of Physiotherapy 67:190–196] © 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction A Cochrane systematic review based on 18 randomised trials demonstrated benefits from treadmill walking training on walking Parkinson’s disease is the second most common degenerative speed (MD 0.09 m/s, 95% CI 0.03 to 0.14) and step length (MD 0.05 m, disease of the central nervous system and the most common move- 95% CI 0.01 to 0.09) but no clear effects on walking distance and ment disorder, resulting from the death of dopamine-producing cells cadence.6 Although that review highlighted the efficacy of walking in the substantia nigra.1 Over the past 30 years, the number of in- training, it may be a more clinically helpful intervention if the dividuals with Parkinson’s disease globally has more than doubled to magnitude of effect could be augmented. Therefore, a combination of over 6 million.2 Advances associated with medical treatment have interventions, such as exercise and pharmacological or non- resulted in improvements in motor symptoms such as bradykinesia pharmacological therapies, is becoming a popular practice to and rigidity.3 However, as the disease progresses over time, people enhance the effects of interventions, especially in the early stages of with Parkinson’s disease face impaired balance and walking limita- Parkinson’s disease.7 tions,4,5 which are associated with an increased risk of falls, social isolation and poorer quality of life.5 Rehabilitation aims to maintain Non-invasive brain stimulation by transcranial direct current community ambulation by improving walking parameters such as stimulation (tDCS), which modulates cortical excitability by applying a walking speed, step length and cadence.6 direct current to the skull, could be combined with walking training and has the potential to enhance its benefits.8 Previous studies have https://doi.org/10.1016/j.jphys.2021.06.003 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Research 191 suggested that tDCS modulates cortical excitability during stimulation Box 1. Inclusion criteria. by non-synaptic changes of the cells, and increasing evidence indicates that the after-effects of tDCS are driven by synaptic modification.9,10 It Design has been suggested that tDCS continuously applied for  5 minutes can  Randomised clinical trial provoke sustained changes in neuronal firing rates that last after the Participants current is switched off.11 Synaptic plasticity could therefore induce  People with Parkinson’s disease long-lasting excitability changes in the central nervous system and  Ambulatory effects may be carried over to improving activity limitations.11 Intervention  tDCS combined with walking training There have been four systematic reviews examining tDCS on Outcomes walking in people with Parkinson’s disease.12–15 Two reviews did not  Primary: walking speed, cadence and step length have the benefit of any meta-analysis.12,13 Two reviews reported  Secondary: number of falls, fear of falling, freezing of gait and short-term benefits of tDCS on overall walking activity, with a SMD ranging from 0.3 to 0.8.14,15 The reviews either included non- social participation randomised trials or excluded studies that examined exercise as an Comparisons intervention, which justifies the need for a systematic review to  tDCS combined with walking training versus no/sham examine the effects of the combination of tDCS with walking training on spatiotemporal outcomes related to walking. Given that the intervention methodological limitations of those reviews could bias their results  tDCS combined with walking training versus walking training and conclusions, a rigorous systematic review with meta-analysis of the current high-quality evidence is warranted. only The aim of this systematic review was to examine the efficacy of is an 11-item scale designed for rating the methodological quality tDCS combined with walking training for improving walking in (internal validity and statistical information) of randomised trials. people with Parkinson’s disease. In order to make recommendations Each item, except for Item 1, contributes one point to the total score based on a high level of evidence, this review was planned to include (range 0 to 10 points). Where a trial was not included on the data- only randomised controlled trials. base, it was scored by a reviewer who had completed the PEDro scale training tutorial. Therefore, the specific research questions for this systematic review were: Participants Trials involving adults diagnosed with Parkinson’s disease were 1. Does walking training combined with tDCS improve walking (ie, speed, cadence and step length) and reduce falls and freezing, included, while trials involving those with Parkinsonism or Parkin- compared with no/sham intervention, in people with Parkinson’s son’s plus disorders were excluded.16 The number of participants, age, disease? disease progression, baseline walking speed and on/off phase of medication were recorded to characterise and assess the similarity of 2. Is walking training combined with tDCS superior to walking the studies. training alone? Intervention 3. Are any benefits carried over to social participation and/or main- Trials were included if the experimental intervention was tDCS in tained beyond the intervention period? addition to task-specific walking training. To be included, tDCS had to Method be delivered over  5 minutes, and electrode positioning had to be in accordance with the International EEG 10/20 system.17 Stimulation This systematic review is reported according to the guidelines on could be applied before, during or after the walking training. Walking the Preferred Reporting Items for Systematic Reviews and Meta- training had to comprise planned, structured and repetitive exercises Analyses (PRISMA) statement website. delivered with the purpose of improving walking.18 Session duration, session frequency, program duration and tDCS parameters (eg, in- Identification and selection of trials tensity and position) were recorded to assess the similarity of the studies. The control intervention was defined according to the Searches were conducted on MEDLINE (1946 to June 2020), AMED research questions: to estimate the efficacy of tDCS plus walking (1985 to May 2019), EMBASE (1947 to May 2019), Cochrane (2005 to training, the control intervention could be nothing or sham-tDCS May 2019), PsycINFO (1806 to May 2019) and the Physiotherapy Ev- intervention; and to estimate the effect of tDCS plus walking idence Database (PEDro) (to June 2020) databases for relevant studies training relative to walking training alone, the control intervention without date or language restrictions. The search strategy was had to be walking training alone. registered at PubMed/Medline, so the authors received monthly no- tifications about potential papers related to this systematic review. Outcome measures Search terms included words related to Parkinson’s disease, trans- The primary outcomes were walking spatiotemporal parameters: cranial direct current stimulation and walking. The detailed search strategy is presented in Appendix 1 on the eAddenda. Title and ab- walking speed, cadence and step length, typically obtained using a stracts were displayed and screened to identify relevant studies. Full timed walk test or movement analysis systems. Secondary outcomes paper copies of peer-reviewed relevant papers were retrieved and were the number of falls, fear of falling, freezing of gait and social their reference lists were screened to identify further relevant studies. participation, typically obtained using diaries or questionnaires. The The methods sections of the retrieved papers were extracted and timing of the measurements and the procedure used to measure the independently reviewed by two reviewers (WAC and FMVD) using outcomes were recorded to assess the appropriateness of combining the predetermined criteria shown in Box 1. Both reviewers were studies in a meta-analysis. blinded to authors, journals and results. Disagreement or ambiguities were resolved through discussion with a third reviewer (LRN). Data extraction Assessment of characteristics of trials Information about the method (ie, design, participants, interven- tion and measures) and results (ie, number of participants and means Quality (SD) of outcomes) were extracted by two reviewers (GPO and WAC) The methodological quality of the included trials was assessed by and checked by a third reviewer (LRN). Where information was un- available in the published trials, details were requested from the extracting the PEDro scores from the PEDro website. The PEDro scale corresponding author.

192 Nascimento et al: tDCS and walking training in Parkinson’s disease Data analysis Titles and abstracts screened (n = 1,090) y MEDLINE, CINAHL, EMBASE, The change or post-intervention scores were used to obtain the pooled estimate of the effect of the intervention, using a random Cochrane, PsycINFO (n = 1,044) effects model. A visual inspection of the distribution of effect sizes in y PEDro (n = 46) the forest plots was performed and the I2 value was calculated to indicate the proportion of variance that was due to heterogeneity.19,20 Duplicate papers between Values of I2 . 50% were considered indicative of important hetero- databases (n = 12) geneity.20,21 The analyses were performed using Review Manager softwarea. The pooled data for each outcome were reported as the Papers excluded after screening weighted mean difference (MD) between the groups, with a 95% titles/abstracts (n = 1,068) confidence interval (CI). Where the data from a trial could not be included in a pooled analysis, the between-group result was reported. Potentially relevant papers retrieved for evaluation of full text (n = 22) The Grading of Recommendations Assessment, Development and y from electronic databases (n = 18) Evaluation (GRADE) system was used to summarise the overall cer- y from reference lists (n = 4) tainty of the evidence for each outcome. The GRADE system’s scores range from high to very low quality.22 We downgraded the score from Trials excluded after evaluation of full high quality by one level each time one of the following prespecified text (n = 17) a criteria was present: low methodological quality (if the majority of y study design is not a randomised studies in the meta-analysis had a PEDro score , 6); inconsistency of estimates among pooled studies (I2 . 50%), or when assessment was controlled trial (n = 10) not possible (no pooling); indirectness of participants (. 50% of the y experimental intervention is not studies did not describe the inclusion criteria); and imprecision (pooling , 400 participants for each outcome).23 Two reviewers (GPO walking training (n = 11) and WAC) assessed the quality of the evidence using the GRADE y experimental intervention did not system, with potential disagreements resolved by consensus with a third reviewer (LRN). include tDCS (n = 2) y outcome measure is not of interest Results (n = 4) Flow of trials through the review Papers included (n = 5) Comparisons included (n = 6) The electronic search strategy identified 1,090 papers. After screening titles, abstracts and reference lists, 22 potentially relevant Figure 1. Flow of studies through the review. full papers were retrieved. Seventeen papers failed to meet the in- a Papers may have been excluded for failing to meet more than one inclusion criterion. clusion criteria. These papers and the reasons for their exclusion are presented in Appendix 2 on the eAddenda. The remaining five papers Intervention were included in the review. One trial24 provided data for two In all trials, the experimental intervention was overground comparisons and, therefore, six comparisons were performed in the review, as shown in Figure 1. walking training combined with tDCS. The anode electrode was al- ways placed at the Cz position, which is consistent with the primary Characteristics of included trials motor cortex area. The duration of brain stimulation ranged from 13 to 30 minutes among trials, with an intensity of 2 mA. Stimulation The five studies involved 117 participants and investigated the was delivered before (n = 2) or during walking training (n = 2), but effects of walking training combined with tDCS for improving one trial25 did not mention timing of stimulation. Participants un- walking speed (n = 5), cadence (n = 4) and step length (n = 5) in dertook training for 30 to 60 minutes, two to three times per week, people with Parkinson’s disease, as shown in Table 1. There were no on average for 4 weeks (SD 1.5). The control group always received studies that examined the effects on falls, freezing or social partici- sham-tDCS24 or sham-tDCS plus overground walking training.24–28 pation. Additional information was requested from the author of one study,25 but no information was received. Outcome measures Measures of walking speed and cadence were obtained using a Quality The mean PEDro score of the included trials was 8 out of 10 (range movement analysis system (n = 2), a timed walk measure (n = 2) or foot sensors during a timed walk test (n = 1). Measures of step length 6 to 10), as presented in Table 2. All the trials had randomly allocated were obtained using a movement analysis system (n = 2), video participants, had similar groups at baseline, had less than 15% drop- analysis (n = 2) or foot sensors during a timed walk test (n = 1). outs, reported between-group differences, and reported point esti- Walking speed was converted to m/s, cadence to steps/minute and mate and variability. The majority of trials reported concealed step length to cm. allocation (80%), had blinded participants (60%) and blinded assessors (80%). On the other hand, the majority of trials did not report ther- Effect of walking training combined with tDCS compared with no/ apist blinding (60%) or whether an intention-to-treat analysis was sham intervention undertaken (80%). The effect of walking training combined with tDCS was examined Participants by one trial24 with a very small sample (n = 16) and a PEDro score of 6 The mean age of participants ranged from 61 to 67 years among out of 10. The quality of the evidence was considered low on GRADE. Results indicated that walking training combined with tDCS signifi- included trials, and the disease progression ranged from 6 to 8 years. cantly increased walking speed (MD 0.26 m/s, 95% CI 0.05 to 0.47) and The trials included participants at stages 1 to 3 of the Hoehn and Yahr step length (MD 18 cm, 95% CI 9 to 27) immediately after intervention scale, with a mean baseline walking speed ranging from 0.62 m/s to compared with sham-tDCS. Maintenance of benefits beyond the 1.4 m/s. One trial24 did not report the mean age of participants or intervention period was not examined. information regarding disease progression.

Table 1 Characteristics of the included trials (n = 5). Study Design Participantsa Costa-Ribeiro et al (2017) RCT n = 24 Frequency and duration Age (y) = 61 (13) Exp = active-tDCS 1 walking tr da Silva et al (2018) RCT Disease progression (y) = 6 (4) Hoehn and Yahr = 1 to 3 43 min 3 3/wk 3 4 wks Walking speed (m/s) = 1.38 (0.3) Con = sham-tDCS 1 walking tra Medication (on/off) = on 43 min 3 3/wk 3 4 wks n = 21 Age (y) = 66 (7) Exp = active-tDCS 1 walking tr Disease progression (y) = 6 (3) 60 min 3 3/wk 3 4 wks Hoehn and Yahr = 2 to 3 Walking speed (m/s) = 0.88 (0.12) Con = sham-tDCS 1 walking tra Medication (on/off) = on 60 min 3 3/wk 3 4 wks Kaski et al (2014) RCT n = 16 Exp = active-tDCS 1 walking tr Age (y) = NR 30 min 3 2/wk 3 6 wks Disease progression (y) = NR Hoehn and Yahr = NR Con1 = sham-tDCS Walking speed (m/s) = 0.62 (0.15) 30 min 3 2/wk 3 6 wks Medication (on/off) = on Con2 = sham-tDCS 1 walking t 30 min 3 2/wk 3 6 wks Schabrun et al (2016) RCT n = 16 Exp = active-tDCS 1 walking tr Age (y) = 67 (8) 60 min 3 3/wk 3 3 wks Disease progression (y) = 6 (4) Hoehn and Yahr = 2 to 3 Con = sham-tDCS 1 walking tra Walking speed (m/s) = 1.40 (0.17) 60 min 3 3/wk 3 3 wks Medication (on/off) = on Yotnuengnit et al (2018) RCT n = 40 Exp = active-tDCS 1 walking tr Age (y) = 65 (9) 60 min 3 3/wk 3 2 wks Disease progression (y) = 8 (4) Hoehn and Yahr = 2 to 3 Con = sham-tDCS 1 walking tra Walking speed (m/s) = 0.68 (0.17) 60 min 3 3/wk 3 2 wks Medication (on/off) = NR Con = control group, Exp = experimental group, NR = not reported, RCT = randomised clinical trial, 6MWT = 6-Minute Walk a Data with parentheses are mean (SD). b The intervention groups and outcome measures that are listed are those that were analysed in this systematic review; t

Interventionb Outcome measuresb n Parameters Speed = 10 m WT (m/s) Cadence = 10 m WT (steps/min) raining Setting: rehabilitation center Step length = video analysis (m) aining Anode electrode: Cz position Measurements = 0, 4 weeks Cathode electrode: supraorbital area Brain area: NR Speed = optoelectronic system (Qualisys) (m/s) Intensity: 2 mA Cadence = optoelectronic system (Qualisys) (steps/s) Duration: 13 min Step length = optoelectronic system (Qualisys) (m) Timing: before walking training Measurements = 0, 4 weeks raining Setting: rehabilitation center Speed = 6MWT (m/s) aining Anode electrode: Cz position Step length = video analysis (m) Cathode electrode: supraorbital area Measurements = 0, 6 weeks Brain area: supplementary motor area and primary motor cortex Speed = electronic walkway (GAITRite) (m/s) Cathode electrode: supraorbital area Cadence = electronic walkway (GAITRite) (steps/min) Intensity: 2 mA Step length = electronic walkway (GAITRite) (m) Duration: 15 min Measurements = 0, 3, 12 weeks Timing: NR Speed = optoelectronic system (Qualisys) (m/s) raining Setting: hospital Cadence = optoelectronic system (Qualisys) (steps/min) training Anode electrode: Cz position Step length = optoelectronic system (Qualisys) (m) Cathode electrode: NR Measurements = 0, 2, 4, 8 weeks Brain area: supplementary motor area and Research 193 primary motor cortex Intensity: 2 mA Duration: 15 min Timing: during walking training raining Setting: university aining Anode electrode: Cz position Cathode electrode: supraorbital area Brain area: primary motor cortex Intensity: 2 mA Duration: 20 min Timing: during walking training raining Setting: hospital aining Anode electrode: Cz position Cathode electrode: supraorbital area Brain area: lower limb motor cortex Intensity: 2 mA Duration: 30 min Timing: before walking training k Test, 10 m WT = 10-m Walk Test. there may have been other groups or measures in the paper.

194 Nascimento et al: tDCS and walking training in Parkinson’s disease Table 2 PEDro criteria and scores for the included trials (n = 5). Study Random Concealed Groups Participant Therapist Assessor , 15% Intention- Between-group Point Total allocation allocation similar blinding blinding blinding dropouts to-treat difference estimate (0 to 10) at baseline analysis reported and variability reported Costa-Ribeiro et al (2017) Y Y Y Y Y YY N Y Y9 da Silva et al (2018) YY Y Y NYY N Y Y8 Kaski et al (2014) YY Y N NNY N Y Y6 Schabrun et al (2016) Y Y Y Y Y YY Y Y Y 10 Yotnuengnit et al (2018) Y N Y N NYY N Y Y6 Y = yes, N = no. Effect of walking training combined with tDCS compared with Although this systematic review planned to answer three clinical walking training alone questions, only one randomised trial,24 with a very small sample (n = 16), provided information addressing the first research question (ie, Walking speed estimating the effects of walking training combined with tDCS The effect of walking training combined with tDCS, compared compared with no/sham intervention). The results suggested that walking training combined with tDCS may result in large improve- with walking training alone, on walking speed was examined by ments in walking speed (MD 0.26 m/s) and step length (MD 18 cm). pooling post-intervention data from five trials24–28 involving 111 However, the confidence intervals were wide and the quality of the participants (mean PEDro score 8 out of 10). Moderate-quality evi- evidence was graded as low, which indicates that future research is dence suggested that the addition of tDCS has no or negligible effect very likely to change the estimate of the effect on these outcomes.19 over walking training on walking speed immediately after the inter- vention (MD 20.01 m/s, 95% CI 20.05 to 0.04, I2 = 2%), as shown in Moreover, because previous results6 have already demonstrated Figure 2A. For a more detailed forest plot, see Figure 3A on the that walking training alone is effective for improving walking in eAddenda. Similarly, moderate-quality evidence suggested that the people with Parkinson’s disease, answering the second research addition of tDCS has no or negligible effect over walking training on question (ie, estimating the effect of adding tDCS to walking training) walking speed beyond the intervention period (MD 20.04 m/s, 95% seems more relevant. In other words: patients and clinicians should CI 20.11 to 0.03, I2 = 14%), as shown in Figure 4A. For a more detailed ask if spending time and money adding tDCS to walking training is forest plot, see Figure 5A on the eAddenda. worthwhile. Meta-analyses of five randomised clinical trials24–28 demonstrated that the addition of tDCS did not provide additional Step length benefits over walking training on walking speed, step length or The effect of walking training combined with tDCS compared with cadence. These results could be explained by the fact that non- invasive brain stimulation is superficial. While invasive brain stimu- walking training alone on step length was examined by pooling post- lation deeply modulates the subthalamic nucleus or internal globus intervention data from five trials24–28 involving 111 participants pallidus, tDCS activates cortical neurons located up to 2 to 3 cm (mean PEDro score 8 out of 10). Moderate-quality evidence suggested beneath the scalp,29 which may explain the lack of effects on walking that the addition of tDCS has no or negligible effect over walking ability. In all included trials, the anode electrode was placed at the Cz training on step length immediately after intervention (MD 1.2 cm, position, which is consistent with the primary motor cortex area. It 95% CI 21.2 to 3.5, I2 = 0%), as shown in Figure 2B. For a more detailed has been suggested that adding simultaneous stimulation to the forest plot, see Figure 3B on the eAddenda. Similarly, moderate- supplementary motor, dorsolateral prefrontal or premotor cortices quality evidence suggested that the addition of tDCS has no or may enhance the effects of tDCS, which should be examined in negligible effect over walking training on step length beyond the further randomised trials. Lack of significant results could also be intervention period (MD 20.3 cm, 95% CI 22.6 to 2.1, I2 = 0%), as explained by the characteristics of the participants. Three trials, for shown in Figure 4B. For a more detailed forest plot, see Figure 5B on instance, included participants with average walking speeds . 0.8 the eAddenda. m/s, who could be classified as unlimited community ambulators. The results suggest that when participants with few walking limitations Cadence are included, there is little room for improvement. Further rando- The effect of walking training combined with tDCS compared with mised clinical trials should examine whether effects are enhanced in moderately disabled individuals. As the quality of evidence found in walking training alone on cadence was examined by pooling post- this systematic review was moderate, further research may have an intervention data from four trials25–28 involving 95 participants important impact or change estimates of effect on walking ability.22 (mean PEDro score 8 out of 10). Moderate-quality evidence suggested that the addition of tDCS has no or negligible effect over walking This systematic review also aimed to estimate the effects of tDCS training on cadence immediately after intervention (MD 23 steps/ combined with walking training on freezing of gait, falls and social minute, 95% CI 26 to 1, I2 = 17%), as shown in Figure 2C. For a more participation (ie, the third research question), but no randomised detailed forest plot, see Figure 3C on the eAddenda. A small significant trials were found that measured these outcomes. Recent systematic decrease in cadence was observed beyond the intervention period reviews30,31 that examined nonpharmacological/nonsurgical treat- (MD 25 steps/minute, 95% CI 28 to 22, I2 = 0%), as shown in Figure 4C. ments in people with Parkinson’s disease suggested that either non- For a more detailed forest plot, see Figure 5C on the eAddenda. invasive brain stimulation or walking training with cues may improve freezing of gait and mobility. Currently, evidence is insufficient to Discussion support or refute the addition of tDCS to walking training for improving freezing of gait, falls and social participation. This systematic review aimed to examine the effects of the addi- tion of tDCS to walking training for improving walking, falls and social This systematic review had both strengths and weaknesses. The participation in people with Parkinson’s disease. One randomised mean PEDro score of 8 for the included trials represents high meth- trial demonstrated that walking training combined with tDCS odological quality. The more common sources of bias were lack of improved walking speed and step length, but meta-analyses of five blinding of therapists and describing whether an intention-to-treat randomised trials demonstrated that tDCS provided no additional analysis was undertaken. The number of participants per group var- benefits over walking training alone. ied across trials (mean 23, range 16 to 40), which could have opened the results to small trial bias. In addition, three trials included

Research 195 A WMD (95% CI) B WMD (95% CI) Random Random Study Study Costa-Ribeiro 2017 Costa-Ribeiro 2017 –20 –10 0 10 20 da Silva 2018 da Silva 2018 Favours con (cm) Favours exp Kaski 2014 Kaski 2014 Schabrun 2016 Schabrun 2016 Younuengnit 2017 Younuengnit 2017 Total Total –0.2 –0.1 0 0.1 0.2 Favours con (m/s) Favours exp C WMD (95% CI) Random Study Costa-Ribeiro 2017 da Silva 2018 Schabrun 2016 Younuengnit 2017 Total –20 –10 0 10 20 Favours con (steps/min) Favours exp Figure 2. Weighed mean difference (95% CI) in the effect of walking training combined with tDCS versus sham-tDCS 1 walking training immediately after intervention on (A) walking speed (m/s; n = 111), (B) step length (cm; n = 111) and (C) cadence (steps/min; n = 95). A WMD (95% CI) B WMD (95% CI) Random Random Study Study Schabrun 2016 Schabrun 2016 Younuengnit 2017 Younuengnit 2017 Total Total –0.2 –0.1 0 0.1 0.2 –20 –10 0 10 20 Favours con (m/s) Favours exp Favours con (cm) Favours exp C WMD (95% CI) Random Study Schabrun 2016 Younuengnit 2017 Total –20 –10 0 10 20 Favours con (steps/min) Favours exp Figure 4. Weighted mean difference (95% CI) in the effect of walking training combined with tDCS versus sham-tDCS 1 walking training beyond the intervention period on (A) walking speed (m/s; n = 56), (B) step length (cm; n = 56) and (C) cadence (steps/min; n = 56). participants with mean walking speeds . 0.8 m/s, which may give duration (range 30 to 60 minutes), session frequency (two to three/ little room for improvement. On the other hand, this review included week), and program duration (mean 4 weeks, SD 1.5) were also only randomised controlled trials, which were similar regarding their similar among the included trials. clinical characteristics. All trials delivered tDCS with an intensity of 2 mA at the Cz position and included mildly or moderately disabled In conclusion, moderate-quality evidence suggested that the participants at stages 1 to 3 on the Hoehn and Yahr scale. Session addition of tDCS, with an intensity of 2 mA at the Cz position, to walking training provided no clinically important benefits on walking

196 Nascimento et al: tDCS and walking training in Parkinson’s disease speed, step length or cadence in individuals with mild to moderate 9. Sánchez-Kuhn A, Pérez-Fernández C, Cánovas R, Flores P, Sánchez-Santed F. disability related to Parkinson’s disease. The provided estimates may Transcranial direct current stimulation as a motor neurorehabilitation tool: an change with larger high-quality studies. Evidence is missing to sup- empirical review. Biomed Eng Online. 2017;16:1–22. port or refute the addition of tDCS for improving freezing of gait and falls or informing whether benefits are carried over to social partic- 10. Stagg CJ, Nitsche MA. Physiological basis of transcranial direct current stimulation. ipation. Further randomised clinical trials may examine whether Neuroscientist. 2011;17:37–53. simultaneous stimulation of two or more brain areas would poten- tiate the effects of walking training. 11. Lang N, Siebner HR, Ward NS, Lee L, Nitsche MA, Paulus W, et al. How does transcranial DC stimulation of the primary motor cortex alter regional neuronal What was already known on this topic: People with Par- activity in the human brain? Eur J Neurosci. 2005;22:495–504. kinson’s disease typically have walking limitations, which are associated with falls, social isolation and poorer quality of life. 12. Beretta VS, Conceição NR, Nóbrega-Sousa P, Orcioli-Silva D, Dantas LK, Gobbi LT, Walking training on a treadmill improves walking speed and step et al. Transcranial direct current stimulation combined with physical or cognitive length, but the effects on other outcomes are unclear. Pre- training in people with Parkinson’s disease: a systematic review. J Neuroeng liminary evidence suggests that transcranial direct current stim- Rehabil. 2020;17:1–15. ulation might increase the benefits of walking training. What this study adds: In ambulatory people with Parkinson’s 13. Broeder S, Nackaerts E, Heremans E, Vervoort G, Meesen R, Verheyden G, disease, the addition of transcranial direct current stimulation to Nieuwboer A. Transcranial direct current stimulation in Parkinson’s disease: walking training provided no or negligible additional improvement neurophysiological mechanisms and behavioral effects. Neurosci Biobehav Rev. in walking speed, step length or the number of steps per minute. 2015;57:105–117. There is a lack of evidence about whether the addition of transcranial direct current stimulation to walking training im- 14. Lee HK, Ahn SJ, Shin YM, Kang N, Cauraugh JH. Does transcranial direct current proves freezing of gait, falls or social participation. stimulation improve functional locomotion in people with Parkinson’s disease? A systematic review and meta-analysis. J Neuroeng Rehabil. 2019;16:1–13. Footnotes: a Review Manager V.5.3, The Nordic Cochrane Centre, Copenhagen, Denmark. 15. Goodwill AM, Lum JAG, Hendy AM, Muthalib M, Johnson L, Albein-Urios N, et al. Using non-invasive transcranial stimulation to improve motor and cognitive eAddenda: Figures 3 and 5 and Appendices 1 and 2 can be found function in Parkinson’s disease: a systematic review and meta-analysis. Sci Rep. online at https://doi.org/10.1016/j.jphys.2021.06.003. 2017;7:4840. Ethics approval: Not applicable. 16. Flynn A, Allen NE, Dennis S, Canning CG, Preston E. Home-based prescribed ex- Competing interests: Nil. ercise improves balance-related activities in people with Parkinson’s disease and Source(s) of support: Brazilian Government Funding Agency – has benefits similar to centre-based exercise: a systematic review. J Physiother. Fundação de Amparo à Pesquisa e Inovação do Espírito Santo – FAPES 2019;65:189–199. (Universal - 021/2018), Brazil. Acknowledgements: Nil. 17. Klem GH, Lüders HO, Jasper HH, Elger C. The ten-twenty electrode system of the Provenance: Not invited. Peer reviewed. International Federation of Clinical Neurophysiology. Electroencephalogr Clin Neu- Correspondence: Lucas Rodrigues Nascimento, Center of Health rophysiol Suppl. 1999;52:3–6. Sciences, Discipline of Physiotherapy, Universidade Federal do Espír- ito Santo, Vitória, Brazil. Email: [email protected] 18. Chodzko-Zajko WJ, Proctor DN, Fiatarone Singh MA, Singh MA, Minson CT, Nigg CR, et al. Exercise and physical activity for older adults. Med Sci Sports Exerc. References 2009;41:1510–1530. 1. Tysnes OB, Storstein A. Epidemiology of Parkinson’s disease. J Neural Transm. 19. Bagg MK, Mclachlan AJ, Maher CG, Kamper SJ, Williams CM, Henschke N, et al. 2017;124:901–905. Paracetamol, NSAIDS and opioid analgesics for chronic low back pain: a network meta-analysis. Cochrane Database Syst Rev. 2018;6:CD013045. 2. Dorsey ER, Elbaz A, Nichols E, Abd-Allah F, Abdelalim A, Adsuar JC, et al. Global, regional, and national burden of Parkinson’s disease, 1990–2016: a systematic 20. Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 2018;17:939– Version 5.1. 0 [updated March 2011]. London: The Cochrane Collaboration; 2011. 953. 21. Balshem H, Helfand M, Schünemann HJ, Oxman AD, Kunz R, Brozek J, et al. GRADE 3. Morgan JC, Currie LJ, Harrison MB, Bennett JP, Trugman JM, Wooten GF. Mortality in guidelines: 3. Rating the quality of evidence. J Clin Epidemiol. 2011;64:401–406. levodopa-treated Parkinson’s disease. Parkinsons Dis. 2014;2014:1–8. 22. Henschke N, Ostelo RWJG, van Tulder MW, Vlaeyen JW, Morley S, Assendelft WJ, 4. Tomlinson CL, Patel S, Meek C, Herd CP, Clarke CE, Stowe R, et al. Physiotherapy et al. Behavioural treatment for chronic low-back pain. Cochrane Database Syst Rev. versus placebo or no intervention in Parkinson’s disease. Cochrane Database Syst 2010;7:CD002014. Rev. 2013;9:CD002817. 23. Nascimento LR, Boening A, Galli A, Polese JC, Ada L. Treadmill walking improves 5. Bayle N, Patel AS, Crisan D, Guo LJ, Hutin E, Weisz DJ, et al. Contribution of step walking speed and distance in ambulatory people after stroke and is not inferior to length to increase walking and turning speed as a marker of Parkinson’s disease overground walking: a systematic review. J Physiother. 2021;67:95–104. progression. PLoS One. 2016;11:1–13. 24. Kaski D, Dominguez RO, Allum JH, Islam AF, Bronstein AM. Combining physical 6. Mehrholz J, Kugler J, Storch A, Pohl M, Elsner B, Hirsch K. Treadmill training for training with transcranial direct current stimulation to improve gait in Parkinson’s patients with Parkinson’s disease. Cochrane Database Syst Rev. 2015;8:CD007830. disease: a pilot randomized controlled study. Clin Rehabil. 2014;28:1115–1124. 7. Bello O, Sanchez JA, Lopez-Alonso V, Márquez G, Morenilla L, Castro X, et al. The 25. da Silva DCL, Lemos T, de Sá Ferreira A, Horsczaruk CH, Pedron CA, de Carvalho effects of treadmill or overground walking training program on gait in Parkinson’s Rodrigues E, et al. Effects of acute transcranial direct current stimulation on gait disease. Gait Posture. 2013;38:590–595. kinematics of individuals with Parkinson disease. Top Geriatr Rehabil. 2018;34:262–268. 8. Elsner B, Kugler J, Pohl M, Mehrholz J. Transcranial direct current stimulation (tDCS) for idiopathic Parkinson’s disease. Cochrane Database Syst Rev. 26. Costa-Ribeiro A, Maux A, Bosford T, Aoki Y, Castro R, Baltar A, et al. Transcranial 2016;7:CD010916. direct current stimulation associated with gait training in Parkinson’s disease: a pilot randomized clinical trial. Dev Neurorehabil. 2017;20:121–128. 27. Schabrun SM, Lamont RM, Brauer SG. Transcranial direct current stimulation to enhance dual-task gait training in Parkinson’s disease: a pilot RCT. PLoS One. 2016;11:1–14. 28. Yotnuengnit P, Bhidayasiri R, Donkhan R, Chaluaysrimuang J, Piravej K. Effects of transcranial direct current stimulation plus physical therapy on gait in patients with Parkinson disease. Am J Phys Med Rehabil. 2018;97:7–15. 29. Chen KS, Chen R. Invasive and noninvasive brain stimulation in Parkinson’s Disease: clinical effects and future perspectives. Clin Pharmacol Ther. 2019;106: 763–775. 30. Delgado-Alvarado M, Marano M, Santurtún A, Urtiaga-Gallano A, Tordesillas- Gutierrez D, Infante J. Nonpharmacological, nonsurgical treatments for freezing of gait in Parkinson’s disease: a systematic review. Mov Disord. 2020;35:204–214. 31. Cosentino C, Baccini M, Putzolu M, Ristori D, Avanzino L, Pelosin E. Effectiveness of physiotherapy on freezing of gait in Parkinson’s Disease: A systematic review and meta-analyses. Mov Disord. 2020;35:523–536. Websites PEDro. www.pedro.org.au PRISMA. www.prisma-statement.org

Journal of Physiotherapy 67 (2021) 156–157 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Editorial Virtual hospitals: why we need them, how they work and what might come next Alla Melman, Chris G Maher, Gustavo C Machado Institute for Musculoskeletal Health, Sydney Local Health District and The University of Sydney, Sydney, Australia Globally, approximately one in three adults suffers from a chronic record system, shared care planning and remote monitoring tools, condition. People with chronic conditions are at higher risk of and are equipped with videoconferencing and telephones. Telephone requiring inpatient admission and increased length of hospital stay;1 or video consultations replace or reduce face-to-face visits. Clinical however, a significant portion of hospital admissions may be pre- ventable. The Australian Institute of Health and Welfare, for example, staff can remotely monitor multiple patients at once on their ‘virtual ward’. Video calls are used to collect clinical observations and pa- reported that in 2017 to 2018, around 748,000 admissions to public tients are advised to call the virtual hospital or an ambulance if they and private hospitals were classified as potentially preventable, deteriorate.9,10 In addition to these remote consultations, virtual accounting for 1 in 15 admissions or 2.9 million bed days, and costing over A$2.3 billion.2 In the US, 3.5 million potentially preventable hospitals may have a team of community nurses, allied health staff and general practitioners who deliver care in patients’ homes.9 adult inpatient stays accounted for US$33.7 billion in aggregate hospital costs in 2017, representing 12.9% of admissions.3 The COVID-19 pandemic has rapidly accelerated evolution of the New care models are needed to decrease the growing burden on HiTH models of care, resulting in the establishment of ‘virtual hos- hospitals, enabling patients to avoid or reduce ‘bricks and mortar’ pitals’, although evidence on the benefits and risks is only beginning hospital admission, whilst receiving high-value care.4 The ‘hospital in to emerge.11 It is imperative that virtual hospitals link in with existing the home’ (HiTH) model of care is now well established worldwide.5,6 In Australia, 3.7% of all admissions from 2011 to 2017 included HiTH community support services and general practitioners to reduce care, most frequently provided to patients requiring hospital treat- avoidable hospital admissions.11,12 Patients may be flagged by their community nurse, general practitioner or other provider as being at ment related to infections, venous thromboembolism or post-surgical care.7 A comparison of HiTH admissions (n = 80,000) with traditional risk of imminent hospital admission, and triaged to a virtual hospital admission, avoiding emergency department presentation.10 Alter- admissions (n = 2.1 million) demonstrated lower mortality (0.3 versus 1.4%) and decreased rates of readmission within 28 days (2.3 versus nately, patients may be transferred to a virtual hospital admission in 3.6%) with HiTH.7 an early supported discharge model, either directly from the emer- HiTH models may focus on admission avoidance or early sup- gency department or inpatient wards.10 The virtual hospital admis- ported discharge. A review of models that focus on early supported sion process requires strict patient suitability criteria to minimise risk hospital discharge (32 trials, n = 4,746, 12 countries) concluded that when admitting high acuity patients.11,13 Patients with low acuity can HiTH for a broad range of medical conditions decreased length of stay be admitted to a ‘virtual observation unit’ managed remotely with by 6.68 days (95% CI 10.19 to 3.17 days); the effect on health service daily telemedicine-supported symptom monitoring by nursing staff. costs was uncertain.5 A systematic review of HiTH models that focus Those stratified as higher acuity, requiring services such as supple- on admission avoidance (16 RCTS, n = 1,814) reported reduced mental oxygen or other medical treatments, may receive daily home healthcare costs; however, this excluded informal carer costs.6 There visits as well as remote vital sign monitoring on a ‘virtual acute care is also uncertainty regarding success of HiTH in reducing read- unit’.11 Patients at high risk of clinical deterioration would not be missions, due to mixed populations used in meta-analysis.5,6 An ev- considered suitable for virtual admission.13 idence synthesis of 10 systematic reviews of both HiTH models found Virtual hospital clinical data are collected from the patient via that for suitable patients, HiTH generally results in similar or improved clinical outcomes compared with inpatient treatment.8 The online survey tools and from remote monitoring devices. For umbrella review recommends prioritisation of admission-avoidance example, vital signs can be remotely monitored via wireless pulse models; however, further research is required to clarify cost savings and the burden on caregivers at home. oximeter (peripheral oxygen saturation and pulse rate), telemetry Virtual hospitals are differentiated from traditional HiTH models and a wearable continuous temperature-monitoring device applied to by their use of technologies for remote patient monitoring, with the axilla.9,14,15 A physical activity monitor in the form of a wearable device16 or adhesive skin patch13 can also be used to remotely ‘virtual wards’ managed remotely by hospital clinical staff. Virtual hospitals may include technology-enabled multidisciplinary team monitor physical activity levels and episodes of falls. These devices rooms, handover areas, tracking boards and ‘care pods’.9 Care pods are delivered to the patient’s home or hotel accommodation,15 or are physical spaces configured to facilitate privacy during conversa- applied in the emergency department, with technical assistance to tions, from which nurses remotely monitor patients. The clinician workstations in these pods provide access to the electronic medical download the monitoring application and be paired to a smartphone or tablet if needed. Data may be uploaded automatically to a web- based dashboard,9 or integrated directly into the hospital’s elec- tronic medical record system,14 enabling escalation of medical care if required. Alternately, patients may be asked to regularly enter their vital signs into an online application.10 https://doi.org/10.1016/j.jphys.2021.06.018 1836-9553/Crown Copyright © 2021 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Editorial 157 A recent implementation trial in the US evaluated acute care for and consultation. Virtual hospital admission provides a safety net in selected health conditions (cardiorespiratory, renal and gastrointes- that clinical care can be escalated without repeat attendance to the tinal), randomising 43 patients to virtual hospital care and 48 to usual emergency department. Virtual hospital services for those with low hospital care.13 The trial tested a hybrid model that included virtual back pain have the potential to reduce unnecessary inpatient ad- consultations (telephone and video) combined with daily clinical missions, risk of harm and functional decline, and healthcare home visits from medical and nursing staff, as well as physiothera- expenditure, while improving health outcomes. pists by referral. The trial’s primary outcome, the adjusted mean cost of the acute care episode, was 38% lower for virtual hospital patients. There are several barriers to implementing virtual hospitals, Virtual hospital patients also had fewer laboratory orders, imaging such as patient access to reliable internet services and an internet- studies and consultations. Of particular interest to physiotherapists is enabled device. Patients with low levels of expertise with tech- that virtual hospital patients were less sedentary (12 versus 23% of nology would require assistance with installing, accessing and the day) and spent less time lying down (18 versus 55% of the day). navigating an online application. Service access and delivery may There were no appreciable differences in safety and patient satisfac- be facilitated by providing basic internet-enabled electronic tab- tion compared with traditionally hospitalised patients; no patients lets.11 A patient ‘welcome pack’, including a video regarding what admitted under the virtual hospital required emergency medical to expect during a virtual hospital admission, with tips on setting services or were transferred back to the hospital during their episode up a camera for videoconferencing and remote monitoring devices, of care. is a useful strategy to optimise patient experience and physio- therapy time.19 In Australia, Sydney Local Health District launched the country’s first virtual hospital (rpavirtual) in February 2020. It has provided Virtual hospitals are a new option to meet the healthcare needs of care for over 16,000 patients across multiple cohorts. The initial co- patients in both cities and regional areas, which have traditionally horts included palliative care patients, adult patients with cystic been restricted by physical infrastructure. Further evaluation of novel fibrosis and patients at risk of recurring lower leg wounds. This has implementations is required to develop robust models of care, been expanded to patients receiving antenatal, paediatric, aged care, ensuring that they continue to be safe, clinically effective and cost- drug and alcohol, and mental health services. The team consists of effective. Development of suitability criteria for virtual care for over 55 clinicians: nurses, medical officers, psychologists, social specific conditions is needed. Care of patients admitted to virtual workers, midwives, physiotherapists, occupational therapists and hospitals under an ‘early supported discharge’ model needs to be speech pathologists. rpavirtual is well placed to respond to the evaluated for safety, patient acceptability, clinical effectiveness and emerging healthcare needs of quarantined COVID-19 patients.9 overall cost-effectiveness compared with traditional inpatient care. Similar virtual hospital models have now been adopted in rural and regional settings, with the launch of the New England Virtual Health Ethics approval: Not applicable. Network, integrating student training, research and healthcare.17 Source(s) of support: GCM and CGM are supported by fellowships from the National Health and Medical Research Council. Virtual physiotherapy management of patients across all sub- Competing interests: The authors declare that they have no disciplines has accelerated due to the COVID-19 pandemic.18 The competing interests. overall role of a physiotherapist in a virtual hospital is similar to usual Acknowledgements: We would like to acknowledge the contri- inpatient care: to provide triage and diagnosis; engage the patient bution of Andrew Di Lollo, a consumer representative of the rpavir- and the family or carer in treatment; and monitor patient outcomes tual. Andrew reviewed the editorial for readability and clarity, and and readiness for discharge. This includes community liaison and suggested further elaboration regarding barriers to technology- referral to transition services.12 Physiotherapists not yet accustomed enabled care for older persons. We would also like to acknowledge to virtual care may be concerned about completing screening and the assistance of Freya Raffan, Patient Experience and Service diagnosis remotely. Many resources are now available to rapidly Development Manager at rpavirtual. upskill physiotherapists in remote consultations. For example, a vir- Provenance: Not invited. Peer reviewed. tual spinal assessment template has been developed in Canada to Correspondence: Alla Melman, Institute for Musculoskeletal guide remote neurological assessment of spinal conditions19 and a Health, Sydney, Australia. Email: [email protected] system for remote monitoring of pulse oximetry during exercise has been developed and validated.20 Physiotherapy exercise prescription References can be facilitated via various online platforms, ideally integrated into the patient’s electronic medical record. This enables patients to log 1. Hajat C, et al. Prev Med Rep. 2018;12:284–293. their sessions and leave messages for their physiotherapists regarding 2. Australian Institute of Health and Welfare 2019. https://www.aihw.gov.au/reports/ symptoms and progress. Physiotherapists are able to modify exercise programs outside of direct consultation sessions. Home visits by primary-healthcare/potentially-preventable-hospitalisations physiotherapists may also be indicated in some clinical situations. 3. McDermott KW, et al. Characteristics and Costs of Potentially Preventable Inpatient A common reason for inpatient admission in Australia is low back Stays, 2017: Statistical Brief #259. In: Healthcare Cost and Utilization Project (HCUP) pain. In 2019 to 2020, low back pain was the sixth most common Statistical Briefs. Rockville (MD): Agency for Healthcare Research and Quality (US); reason for all emergency department visits in Australia; a third of 2006. these patients were subsequently admitted to hospital wards.21 The 4. Martin MC, et al. Stud Health Technol Inform. 2019;264:556–560. rate of hospitalisation for low back pain increased for the tenth 5. Gonçalves-Bradley DC, et al. Cochrane Database Syst Rev. 2017;6:CD000356. consecutive year to a record high in 2017 to 2018 at 664/100,000 6. Shepperd S, et al. Cochrane Database Syst Rev. 2016;9:CD007491. population. These admissions are more common among older pa- 7. Montalto M, et al. Med J Aust. 2020;213:22–27. tients, last an average of 9 days,22 and cost and average of $15,000 per 8. Leong MQ, et al. BMJ Open. 2021;11:e043285. episode of care.23 This generates a huge burden on healthcare sys- 9. Hutchings OR, et al. J Med Internet Res. 2021;23:e21064. tems, as admitted patients account for half of the total healthcare 10. Thornton J. BMJ. 2020;369:m2119. expenditure on low back pain in Australia.21 11. Sitammagari K, et al. Ann Intern Med. 2021;174:192–199. 12. Moore G, Du Toit A, et al. https://www.saxinstitute.org.au/publications/evidence- In order to address this issue, a virtual hospital model of care has check-library/the-effectiveness-of-virtual-hospital-models-of-care/ been proposed for patients with low back pain in Sydney Local Health 13. Levine DM, et al. Ann Intern Med. 2020;172:77–85. District. The new model will assist emergency department physicians 14. Baumgart DC. NPJ Digit Med. 2020;3:114. and rheumatologists in identifying suitable candidates for virtual 15. Fotheringham P, et al. BMC Public Health. 2021;21:225. admission. Patient eligibility criteria will include non-serious low 16. Mehta SJ, et al. JAMA Network Open. 2020;3:e2028328. back pain (with or without radiculopathy) requiring inpatient 17. University of New England. https://www.une.edu.au/about-une/faculty-of- admission but deemed safe to transfer home with remote monitoring medicine-and-health/nevihn 18. Eccleston C, et al. Pain. 2020;161:889–893. 19. ISAEC. https://www.isaec.org/uploads/1/3/1/2/13123559/final_doc_august_9_202 0_lb_rac_virtual_care_toolkit.pdf 20. Bonnevie T, et al. J Physiother. 2019;65:28–36. 21. Australian Institute of Health and Welfare 2020. https://www.aihw.gov.au/reports- data/myhospitals/sectors/emergency-department-care 22. Ferreira GE, et al. BMJ Qual Saf. 2019;28:826. 23. Coombs DM, et al. Lancet Reg Health – Western Pacific. 2021;7:100089.

Journal of Physiotherapy 67 (2021) 210–216 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research Women report sustained benefits from attending group-based education about pelvic floor muscles: a longitudinal qualitative study Ana Carolina Nociti Lopes Fernandes a, Domingo Palacios-Ceña b, Jean Hay-Smith c, Caroline Caetano Pena a, Mayra Feltrin Sidou a, Amanda Lima de Alencar a, Cristine Homsi Jorge Ferreira a a Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil; b Department of Physical Therapy, Occupational Therapy, Rehabilitation and Physical Medicine, Research Group of Humanities and Qualitative Research in Health Science, Universidad Rey Juan Carlos, Madrid, Spain; c Physiotherapist Rehabilitation Teaching and Research Unit, University of Otago, Wellington, New Zealand KEY WORDS ABSTRACT Pelvic floor Question: Among women who have participated in group-based education about the pelvic floor, what are Women’s health physiotherapy their perceptions of the program and the group format? Design: Exploratory longitudinal qualitative study. Health education Participants: Community-dwelling women aged  18 years who participated in three or four sessions of Physical therapy pelvic floor education in a group format at a university clinic. Data extraction and analysis: Semi-structured Qualitative group or individual interviews were conducted at three time points: 1 week, 3 months and  5 months after the education activity. Data were inductively content analysed and independently coded, with iterative theme development. Results: Women considered the content and delivery appropriate and useful. New knowledge was assimilated and shared with others, and many tried to adopt pelvic floor muscle training in daily life. The women felt that the education sessions might benefit other women, with and without pelvic floor dysfunction symptoms, and that such education would ideally be more widely available. A perception of the value of the education persisted over time, even though maintenance of some health-promoting be- haviours, such as pelvic floor muscle training, decreased. Conclusion: The pelvic floor group education sessions appeared to fulfil the purpose of increasing knowledge about pelvic floor (dys)function and applying this in daily life. Overall, the participants, who had completed three or four of the four sessions, found the program to be useful. A unique feature of this study was longitudinal data collection and it seemed that the perception of value persisted over time. [Fernandes ACNL, Palacios-Ceña D, Hay-Smith J, Pena CC, Sidou MF, de Alencar AL, Ferreira CHJ (2021) Women report sustained benefits from attending group- based education about pelvic floor muscles: a longitudinal qualitative study. Journal of Physiotherapy 67:210–216] © 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction rehabilitation program.7,12–14 A number of authors have specifically developed, described and tested the effects of education about PFM Pelvic floor muscles (PFM) support the pelvic organs, help main- and PFD on women’s knowledge,15–17 function and symptoms.12,15–19 tain urinary and anal continence, and contribute to sexual function.1,2 Most educational activities described in the literature aim to inform PFM function has a clearly established relationship with development women about at least one aspect of PFD.15–17,19 Some programs are of pelvic floor dysfunctions (PFDs), such as pelvic organ prolapse, underpinned by educational and/or health behaviour theory;12,18 urinary incontinence, constipation, anal incontinence and sexual dysfunction.2–5 however, there is no consensus on content or how best to deliver The first-line treatment for urinary incontinence, which is the an education program. most prevalent PFD, is pelvic floor muscle training (PFMT).1,6 PFMT efficacy requires sufficient exercise adherence,6–8 which appears Education delivery may be as important as content. For instance, more likely if patients are knowledgeable about their pelvic floor and education can be delivered one-to-one12 or in groups.15,16 A primary see an effect on their PFD symptoms.7 However, women have been found to have limited knowledge about the pelvic floor and PFDs.9,10 argument for group-based delivery is efficiency (eg, saving therapist time) but other advantages are also proposed, such as reducing Education is an intervention11 that may support patient capability, motivation and behavioural skill for PFMT at the start of the feelings of stigma or isolation and the behavioural support partici- pants offer each other.20 Simple, community-based educational interventions have demonstrated effectiveness on the prevention and management of https://doi.org/10.1016/j.jphys.2021.06.010 1836-9553/© 2021 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

Research 211 symptoms of PFD.19 There is, however, a notable lack of direct patient conducted by a single interviewer, observed by a physiotherapy stu- involvement in developing such educational interventions.18 There- dent and audio-recorded. After the interviews, participants were fore, it seems appropriate to conduct studies that seek the perspec- invited to communicate with the researchers (eg, letters, electronic tives of women who take part in such education programs. messages) if they wanted to supplement or modify the content dis- cussed in the interviews. Data collection continued until no new This study aimed to explore: the perspectives of women partici- codes or themes were evident in data analysis.21,22,25 pating in a group education program about the pelvic floor and PFD; what they found valuable and less valuable in the program content; All education sessions that the women had attended were its impact on their lives and habits; and the barriers and enablers of observed by an assistant researcher. The observed education content understanding and applying the information in real life (ie, per- was noted in a field diary to compare with the protocol of what forming home PFMT). should be included in the education session. Therefore, the research question for this longitudinal qualitative Data analysis study was: Demographic data and variations in the educational program Among women who have participated in group-based education content were descriptively analysed. Interviews were conducted in about the pelvic floor, what are their perceptions of the program Portuguese, transcribed verbatim and then inductively content ana- and group format? lysed26 by one researcher and independently checked by a second researcher in three steps: data familiarisation through transcription Methods and reading completed transcripts; inductive coding to systematically document the interview content; and grouping the codes from the Design previous step into higher-order categories representing potential themes. Next, the research team: reviewed and verified the themes; In this exploratory longitudinal qualitative study,21,22 women who agreed on the description and naming of the themes; and selected had participated in a pelvic floor education program were inter- illustrative quotes. Four researchers were involved in the analysis of viewed at three time points: 1 week after program completion, 3 the qualitative data and the process of data analysis was recorded in a months after completion and then  5 months after completion. The codebook, which was later verified by an external researcher. Quotes study was reported according to standards for qualitative were translated into English by one researcher and verified by research.23,24 another researcher. No qualitative software was used. The research team comprised five researchers (four women and Table 3 details the methods used in design and conduct of the one man), two physiotherapists, two undergraduate physiotherapy study to establish the trustworthiness of the data, including credi- students and one research nurse. Two had experience in qualitative bility, transferability, dependability and confirmability.21,22,25 study design and two had clinical and research experience with PFD. None had been involved in delivery of the education sessions that Results study participants had attended, and only one researcher had known the participants previously. Prior to the study, the positioning of the Flow of participants through the study researchers (presented in Table 1 on the eAddenda) was established through two briefing sessions addressing the theoretical framework, Fifty-three women were potentially eligible, of whom 29 agreed to their prior experience and their motivation for the research.21,22 participate and 19 attended the first interview (Figure 1). The 19 women had a mean age of 55 years and a mean parity of 2.6. Two- The study took place in the School Health Centre of the Ribeirão thirds were Caucasian, three-quarters were married, just over half Preto Medical School, University of São Paulo, Brazil. The educational had completed high school or further education, and nearly all program context is detailed in Appendix 1 on the eAddenda. In brief, identified with Catholic or Protestant faith groups (Tables 4 and 5). community dwelling women were offered four 1-hour education Nearly all the women had urinary incontinence and about half had sessions on the pelvic floor and PFD, delivered by undergraduate received some treatment. Sexual dysfunction was more common than physiotherapy students after training from a senior women’s health either pelvic organ prolapse or anal incontinence, and less than half physiotherapist. The sessions included instruction on PFMT. with these symptoms had received treatment for them. Participants, recruitment and sampling strategies A total of 32 interviews were conducted and 1,165 minutes of interviews were recorded: 646 minutes at the first set of interviews, To be eligible for the study, women had to be aged  18 years and 207 minutes at the second set and 312 minutes at the third set. After have attended three or four sessions of the educational program. Only the interviews, no participants contacted the researchers to supple- women who took part in the first interview (at program completion) ment or modify the content discussed in their interviews. Eight ed- were eligible to take part in the follow-up interviews (3 months,  5 ucation groups (four sessions of 1 hour each, total 32 sessions) were months). Recruitment occurred in the last session of the educational observed. Information coverage was generally very good (Table 6), program. A researcher then followed up each woman by phone to with between 15 and 18 of the 19 education items delivered at each of invite participation, explain the study, answer questions, and seek the eight groups. The PFMT protocol was delivered in every group. written informed consent. Of the 19 women who took part in the first set of interviews, 14 Data collection took part in the second set and nine in the third set (Figure 1). For the first set of interviews, there were 11 individual and four dyadic in- Data were collected over an 18-month period between February terviews. The second set of interviews were all conducted individu- 2017 and August 2018, at three time points (1 week, 3 months and  ally by phone. For the third set of interviews, three of the interviews 5 months after the educational group). Semi-structured interviews included two to four participants each. Data saturation in the first, based on a question guide developed by the researchers were used to second and third set of interviews was obtained after 18, 14 and three obtain information regarding specific topics of interest.25 The ques- interviews, respectively. tion guide is presented in Table 2 on the eAddenda. Themes The first and third interviews were conducted face-to-face in a reserved room at the School Health Centre, and the second by tele- The interview data were presented in two themes: ‘perceptions of phone or in person according to the women’s preference. Women the pelvic floor education’ and ‘using the new knowledge in daily life’. could choose an individual interview or to be interviewed with Each theme had contributing subthemes and illustrated using another woman from her education group. All interviews were

212 Fernandes et al: Group-based pelvic floor muscle education Table 3 Trustworthiness of the data. Criteria Techniques performed and application procedures Credibility  Investigator triangulation: each interview was analysed by two researchers and verified by another researcher. Thereafter, team meetings were performed during which the analyses were compared and categories were identified. Transferability Dependability  Participant triangulation: the study included participants belonging to a different diagnosis. Thus, multiple perspectives were obtained with a common Confirmability link (educational group about pelvic floor).  Triangulation of methods of data collection: semi-structured interviews were conducted at different timepoints and researcher field notes were kept.  Participant validation: participants were asked to confirm the data obtained during the stages of data collection and analysis. None of the participants made additional comments.  In-depth descriptions of the study performed, providing details of the characteristics of researchers, participants, contexts, sampling strategies, and the data collection and analysis procedures.  Audit by an external researcher: an external researcher assessed the study research protocol, focusing on aspects concerning the methods applied and study design. Also, an external researcher specifically checked the description of the coding tree, the major themes, patients’ quotations, quotations’ identification and the descriptions of themes.  Investigator triangulation, participant triangulation and data collection triangulation were used.  Researcher reflexivity was encouraged via the performance of reflexive reports and by describing the rationale behind the study. Assessed for eligibility (n = 53) detail is now presented in two subthemes: ‘delivery and content’ and ‘uptake’. Excluded (n = 24) • did not meet the inclusion criteria (n = 7) Delivery and content • declined to participate (n = 10) At the first interview, women were generally positive about the • telephone contact was not possible (n = 7) delivery of pelvic floor education content. They felt that the content Consented (n = 29) was presented naturally and the presenters used good analogies to explain anything that was potentially difficult to understand. The Excluded (n = 10) visual content (eg, posters) was praised, particularly as supporting • did not attend the first interview (n = 10) knowledge acquisition. Interview at 1 week (n = 19) I found it highly creative on their part to make the drawings and explain everything to us. Explain, show, talk. wow! (Interview 1, P7, Interview at 3 months (n = 14)a 65 years) Interview at ≥5 months (n = 9)b The educational content was valued because it addressed gaps in the women’s knowledge, and they were pleased to know more about Figure 1. Recruitment procedure, data collection process, and loss to followup. PFD, treatment options, PFMT and why it might help. The most useful a Three participants declined to attend the interview due to health issues and two information was that which the women could immediately apply in could not be contacted at the time of the interview, but all agreed to remain in the understanding or addressing their symptoms. Women with PFD study. complaints felt that the group helped them to value themselves and b Six participants declined to attend the interview and four could not be contacted. to speak more freely about their symptoms. representative quotes. The quotes were translated from Portuguese to It clarified a lot of the reasons why you lose urine and why you have English for reporting and presented in both languages in Table 7 on to perform the exercises. I used to be, ‘Why am I doing this?’ the eAddenda. (Interview 1, P12, 18 years) Perceptions of pelvic floor education Some participants were having treatment for PFD and even they Most women liked the way the pelvic floor education content was felt that the educational group helped them to better understand what was happening during their treatment sessions. Women also delivered and felt that they gained valuable knowledge. The most noted the value of education for those without PFD symptoms common criticism was that there was not enough time to practise because such information might help some women to prevent the PFMT. Generally, the women thought that the education would be onset of symptoms, and for those who did develop symptoms they useful for a much wider range of women (ie, not just those with might be more likely to seek treatment. For group participants who current symptoms of PFD) and they suggested the group be more had symptoms and had not previously sought help, the education led widely advertised. Women who were interviewed more than once to seeking more information and help from health professionals. continued to feel that the knowledge gained was worthwhile. More But that doesn’t mean that in a few years it can’t happen to me either. If I am already strengthening all this musculature, maybe I will be able to prevent it, won’t I?! (Interview 3, P8, 39 years) The main criticism was that there needed to be more time to practise PFMT. In addition, some participants found it difficult to have such a wide range of ages, existing levels of knowledge, or symptom experiences in the same group. For these women, a more homoge- neous participant profile was preferred, as they felt that the group interactions and discussion would have been even more valuable if the other participants were more like them. A few felt that the con- tent could be delivered in fewer than four sessions. Well, I thought the theory was great, but along with the theory there should be more practice because just explaining the exercises we

Research 213 Table 4 Table 5 Demographic characteristics of the participants. Clinical characteristics of the participants. Characteristics Participants Characteristic Participants (n = 19) (n = 19) Age (y), mean (SD) 55 (15) UI complaint, n (%) 17 (89) Self-reported ethnicity, n (%) yes 12 (63) 9 (53) white 0 (0) UI treatment, n (%)a 1 (6) black 7 (37) none 1 (6) other surgery 4 (24) Marital status, n (%) 3 (16) medication 1 (6) single 14 (74) physiotherapy 1 (6) married 2 (11) surgery and physiotherapy widowed medication and physiotherapy Formal education, n (%) 4 (21) elementary school 4 (21) POP complaint, n (%) 3 (16) middle school 7 (37) yes high school 4 (21) 1 (33) college POP treatment, n (%)a 1 (33) Religion, n (%) 8 (42) none 1 (33) Catholic 7 (37) surgery Evangelical 2 (11) Spiritist 2 (11) medication Jehovah’s Witness Parity, mean (SD) 2.6 (1.7) SD complaint, n (%) 8 (42) pregnancies 2.0 (1.2) yes births 1.05 (0.97) 7 (88) 0.94 (1.31) SD type, n (%)a 1 (13) caesarean birth 0.63 (0.95) dyspareunia vaginal birth anorgasmia 5 (63) abortions 2 (25) SD treatment, n (%)a 1 (13) none surgery medication Some percentages do not sum to 100 due to the effects of rounding. AI complaint, n (%) yes really can’t assimilate very well. I think there has to be a little more 3 (16) practice [of PFMT] for us to practise more. (Interview 1, P15, 54 AI treatment, n (%)a years) none 2 (67) 1 (33) It is important to point out that, during the other two sets of in- medication terviews, women confirmed their perception of the adequacy of the delivery and content of pelvic floor education. Constipation complaint, n (%) 7 (37) yes Uptake 1 (14) Most participants found out about the education program while Constipation treatment, n (%)a 1 (14) none 1 (14) on the waiting list for an appointment at a women’s health centre, medication 3 (43) and some were specifically invited to attend by a physiotherapist. alternative treatments 1 (14) Already having PFD symptoms or wanting to know more about PFD food adequacy were reasons to attend. enema To find out about myself. Because everything that begins has a so- Some percentages do not sum to 100 due to the effects of rounding. lution, but after everything is damaged is more difficult, isn’t it? I find myself like this. (Interview 1, P4, 63 years) AI = anal incontinence, POP = pelvic organ prolapse, SD = sexual dysfunction, UI = The women felt that the education could be more widely urinary incontinence. advertised as a broad range of women were likely to benefit. a Percentage of those with the complaint. Women who attended became information multipliers. Most passed on information gained through the group to others they thought Using the new knowledge in daily life might benefit. For women without PFD symptoms, participating alongside those with symptoms was an important stimulus to All women reported using their new knowledge, particularly performing PFMT. trying to perform PFMT at home. Women were able to identify the challenges to making changes in everyday life based on new [My aunt] said ‘Oh I’m losing pee’ then I told her about the exercises knowledge, but it was much more difficult to problem solve ways to you do. I told her about the exercise you have to do. I told her about address those barriers. Three subthemes captured application in daily my case, I explained, why I was coming to do the treatment here life: ‘useful knowledge’, ‘performing PFMT at home’ and ‘what makes every week and not. (Interview 1, P12, 18 years) change possible and worthwhile’. Older age, the day/time of the education sessions, travel to the Useful knowledge venue and lack of interest were mentioned as likely barriers to up- The women applied a range of knowledge in everyday life; what take. The credibility of the organisation offering the education, family support and willingness to learn were considered to be facilitators. was most useful was not the same for everyone. At the first interview, all reported trying to do PFMT at home, and a range of other skills – (.) But sometimes the person is interested [in the educational such as abdominal massage to reduce constipation, attention to diet, group] but she does not have the problem so she leaves for modification of voiding habits, better positioning on the toilet for tomorrow. (Interview 3, P8, 39 years) bowel evacuation – were also being practised. At the second inter- view, the women reported that similar skills were being practised, with the addition of further examples such as taking up general physical exercise and seeking help for PFD symptoms. By the third interview, the women were not introducing new lifestyle changes. Over time women were more likely to be sharing useful knowledge with others and feeling better about themselves based on the changes they had made. I perform it four, five times. If I’m sitting and I remember, I perform it, if I’m standing, I remember and perform it. Before getting up, I

214 Fernandes et al: Group-based pelvic floor muscle education Table 6 Information discussed in each of the eight educational groups. Information topics Educational group 12345678 Anatomy of female perineum Y YY Anatomy of female pelvic organs YYYYYY Y Difference between smooth and striated muscle YYYYYYY Pelvic floor muscle structure YY YYYYY Pelvic floor muscle function YYYYYYYY Physiology of micturition YYYYYYYY Stop test YYYYYYYY Healthy urinary habits YY YYY Female sexual response Y YYYYY Urinary incontinence YYYYYYYY Risk factors for urinary incontinence Y YYYYYY Conservative strategies for urinary incontinence YY YY Anal incontinence and risk factors YYYYYYYY Intestinal constipation and management strategies YYYYYYYY Pelvic organ prolapse YYYYYYYY Female sexual dysfunction YYYYYYYY Physiotherapy evaluation of pelvic floor muscle YYYYYYYY Physiotherapy treatment for pelvic floor dysfunction Y Y Y Y Y Y Y Y Pelvic floor muscle training protocol YYYYYYYY Y = yes. perform that on the bed. That’s it, six times slow and six times fast. All in PFD symptoms. Most often this was that doing regular PFMT ones that I remember that was oriented and that I understood well, improved symptoms, and stopping PFMT meant that the improvement I perform. (Interview 1, P7, 65 years) stopped or symptoms worsened. However, there were many challenges identified that meant it was difficult to maintain any behavioural So, I’m liking myself more. I stepped away like this. It was sad that changes in daily life. These included: keeping PFMT in mind (ie, obesity has something to do [with it] and this gave me a warning. I’m developing a mental habit); forgetting PFMT in the rush of daily life; walking, stopped eating fatty foods like frying. I will not say that I and the competing priorities for energy, time and attention. stopped 100%, but it was about 40%. (Interview 2 P5, 46 years) Then I had to stop [PFMT] due to personal issues too, and then I’m like Before the group, I had never spoken about [urinary incontinence] this: totally doing nothing, and then I’m going back to the problem [of with anyone. After the group and the information, I started to share urinary incontinence] from square one. (Interview 3, P3, 40 years) about this with my friends.’ (Interview 3, P3, 40 years) It was hard for women to see how these difficulties could be Performing PFMT at home successfully overcome. However, there were also facilitators of sus- Over time the effort to perform PFMT at home changed. All tained change. For instance, PFMT was more likely to be remembered if there was some association with a particular activity or time of day, women had tried home PFMT within a week of ending the program, and women were more likely feel that PFMT was worth it when they and at 3 months most had tried at least once in the past week. By the understood why they had to do it. third interview, fewer women were doing PFMT at home. For those who were doing PFMT at home, what they did varied a lot. ‘Because, [the trainee] said: ‘Ah, you have to perform the exercise at home’, but I was like ‘ah, why?.’ [murmuring with dissatisfaction Sometimes, when I’m washing in the kitchen or doing laundry, I with the trainee’s orientation] only when [others in the group] sometimes get [gestures of opening and closing the hand] con- explained that I understood why I have to perform it every day, did tracting. I like to keep contracting, I remember, there I get [gestures you get it? Then I started to perform it. because I understood the of opening and closing the hand]. When I remember I contract. importance. what occurs there. (Interview 1, P12, 18 years) (Interview 1, P2, 54 years) Discussion The number of contractions ranged from six to ten for fast con- tractions, while the number of sustained contractions (6 to 20-second This longitudinal qualitative study explored the perspectives of hold) was usually 10. Exercise frequency ranged from one to three women participating in an education program about the pelvic floor sets of exercises per day, with a target of three times per week. Body and PFD. Women felt that the four-session group education program positioning and incorporating pelvic floor muscle contractions in was delivered well and they gained useful knowledge from it that daily activities varied considerably, and women most commonly they applied in daily life and shared with others. They thought what commented that they just tried to fit it in whenever and however they learned could be useful to other women for prevention and they remembered. There was concern that some participants management of PFD. appeared to confuse a urine stop test, muscle stretching or dia- phragmatic breathing as measures to strengthen the pelvic floor No existing qualitative studies were found about women’s expe- muscles. riences of a group educational intervention for pelvic floor or PFD; this study may be the first to document this. A particular strength of Oh, once a day when I’m going to pee, not the first pee, regardless of this study was its longitudinal nature and, therefore, the women’s time and whatever pee I am, I hold it. I cut it, I hold it a little and then experience over time. Some findings were consistent with those from I let it go. I cut it about three times during a urination, that I go and previous authors: for instance, the women experienced similar stop. (Interview 1, P15, 54 years) problems to those previously documented6,7,27 when trying to sustain PFMT in daily life over several months. Whatever difficulties women What makes change possible and worthwhile face with maintaining behaviour change, this study adds the The most consistent influence on sustained change in daily life perception of persistent value of education for women. It was suffi- cient for them to feel that this was a public health intervention that was when women noticed a difference – improvement or worsening –

Research 215 should be made available at suitable times and places, which would What was already known on this topic: Pelvic floor muscle enable a wide range of women to attend. This finding may not training is well established as first-line therapy for urinary in- represent the perceptions of those who did not take part in the in- continence and is helpful for other pelvic floor disorders. The terviews; this was the primary limitation of the study. efficacy of pelvic floor muscle training depends on sufficient exercise adherence. Adherence appears more likely if patients are Health promotion and education about PFD may be a way to deal knowledgeable about their pelvic floor. Group-based education with myths and misinformation about the pelvic floor and about the pelvic floor has the potential benefits of efficiency and PFMT.15,18,28–31 The participants thought that the educational activity peer support. was valuable because it increased their knowledge related to the pelvic What this study adds: Women who participated in group- floor, reduced their uncertainty about treatment options and offered based education about the pelvic floor found the content and solutions to their concerns. This confidence in what they knew is format appropriate and useful. Many shared their new knowl- potentially important for treatment satisfaction. In a focus group study edge with others. Many felt that the intervention would benefit of women experiencing overactive bladder symptoms, Smith et al30 other women, with and without pelvic floor dysfunction, and found that those who felt confused about the condition, associated that group education would ideally be more widely available as a medical tests and care were more dissatisfied with their care. public health intervention. Their perception of the value of the education persisted over time, even though maintenance of some The benefit of group activities and education to empower and health-promoting behaviours, such as pelvic floor muscle support women and reduce the stigma and isolation associated with training, decreased. some health conditions is well-documented (eg, obesity, human im- munodeficiency virus and overactive bladder).30,32,33 This study eAddenda: Tables 1, 2 and 7 and Appendix 1 can be found online found that some women would prefer the groups to be more ho- at https://doi.org/10.1016/j.jphys.2021.06.010 mogeneous (eg, similar in age and symptoms) and some felt that the mix of those with and without symptoms was particularly useful for Ethics approval: The study was approved by the Research Ethics the symptom-free women, as it helped them understand the value of Committee of the School Health Centre of the Ribeirão Preto Medical the information and PFMT behaviour for preventing PFD. For women School, University of São Paulo. Participants gave informed consent with symptoms, the group gave them an opportunity to express what before commencing the study. they were feeling and be supported by others with similar experi- ences. It seems that group-based incontinence-specific educational Competing interests: Nil. interventions may be effective for treating and preventing urinary Source(s) of support: The authors Ana Carolina Nociti Lopes Fer- incontinence34 and this means that it is important to understand nandes and Caroline Caetano Pena received scholarship from Coor- what it is about group delivery that increases or decreases the effect. dination for the Improvement of Higher Education Personnel (CAPES) and Amanda Lima de Alencar received a scholarship from Unified The main criticism by participants in our educational program was Program of Scholarship to Support undergraduate students (PUB) of that there was not enough time to practise PFMT within each session. the University of São Paulo. Contact with health professionals for supervision of PFMT may be an Acknowledgements: Nil. important contributor to the effect of training;35 however, a lack of Provenance: Not invited. Peer reviewed. self-efficacy for correct PFM contraction and PFMT in general seems Correspondence: Cristine Homsi Jorge Ferreira, Ribeirão Preto common.7 While a PFM assessment was not offered to verify a correct Medical School, University of São Paulo, Ribeirão Preto, Brazil. Email: contraction, the four-session program offered an opportunity to teach [email protected] and follow-up with women, to increase their confidence and skill in performing PFMT, and most women did try PFMT at home. However, References there was concern that some women confused PFMT with other ac- tivities, such as diaphragmatic breathing, over time. Incorporating 1. Abrams P, Andersson KE, Birder L, Brubaker L, Cardozo L, Chapple C, et al. Fourth sufficient practice and feedback on practice of PFMT within an edu- international consultation on incontinence recommendations of the interna- cation program is essential to support uptake and maintenance of tional scientific committee: evaluation and treatment of urinary incontinence, PFMT, and consistent with theories of supporting self-efficacy36 and pelvic organ prolapse, and fecal incontinence. Neurourol Urodyn. 2010;29:213– behaviour change.11 240. If, as the women in this study suggested, such group PFD edu- 2. Bø K, Berghmans B, Morkved S, Van Kampen M. Evidence-based physical therapy cation is made more widely made available as a public health pelvic floor bridging science and clinical practice. In: Bo K, Sherburn M, eds. intervention, further study is needed to understand where the value Overview of physical therapy for pelvic floor dysfunction. Oxford: Elsevier; 2015:1. lies from the perspective of those taking part. One potential study is to investigate the effect of such a program on help-seeking. Some 3. Herbruckl F. Stress urinary incontinence: prevention, management, and provider beliefs about urinary incontinence act as a barrier to help-seeking, education. Urol Nurs. 2008;28:200–206. despite bothersome symptoms;37 the education sessions in the current study appeared to encourage further help-seeking behav- 4. Woodley SJ, Boyle R, Cody JD, Mørkeved S, Hay-Smith EJC. Pelvic floor muscle iour from those with symptoms. It is unclear, for instance, if training for prevention and treatment of urinary and faecal incontinence in participants’ help-seeking behaviour arose from gaining more antenatal and postnatal women. Cochrane Database Syst Rev. 2017;12:CD007471. confidence to talk about their symptoms, having a positive experi- ence with health professionals regarding their PFD symptoms, be- 5. Wu JM, Vaughan CP, Goode PS, Redden DT, Burgio KL, Richter HE, et al. Prevalence ing educated to dispel myths that nothing could be done, or other and trends of symptomatic pelvic floor disorders in U.S. women. Obstet Gynecol. reasons. This study suggested that women without PFD symptoms 2014;123:141–148. also found education useful and this might also change their help- seeking behaviour over time. 6. Dumoulin C, Hay-Smith J, Frawley H, McClurg D, Alewijnse D, Bo K, et al. 2014 consensus statement on improving pelvic floor muscle training adherence: Inter- In conclusion, women who participated in a specific pelvic national Continence Society 2011 State-of-the-Science Seminar. Neurourol Urodyn. floor education program considered the content and delivery 2015;34:600–605. appropriate and useful. The participants had assimilated new knowledge and shared it with others, and many had tried to 7. Hay-Smith J, Dean S, Burgio K, McClurg D, Frawley H, Dumoulin C. Pelvic-floor- adopt PFMT as part of daily life. They felt that the education muscle-training adherence ‘modifiers’: a review of primary qualitative studies- sessions might benefit other women, with and without PFD 2011 ICS State-of-the-Science Seminar research paper III of IV. Neurourol Urodyn. symptoms, and that pelvic floor education would ideally be 2015;34:622–631. more widely available. A perception of the value of the edu- cation persisted over time, even though maintenance of some 8. Kim H, Yoshida H, Suzuki T. The effects of multidimensional exercise treatment on health-promoting behaviours such as PFMT decreased. community-dwelling elderly Japanese women with stress, urge, and mixed urinary incontinence: a randomized controlled trial. Int J Nurs Stud. 2011;48:1165–1172. 9. de Freitas LM, Bø K, Fernandes ACNL, Uechi N, Duarte TB, Ferreira CHJ. Pelvic floor muscle knowledge and relationship with muscle strength in Brazilian women: a cross-sectional study. Int Urogynecol J. 2019;30:1903–1909. 10. Fante JF, Silva TD, Mateus-Vasconcelos ECL, Ferreira CHJ, Brito LGO. Do women have adequate knowledge about pelvic floor dysfunctions? A systematic review. Rev Bras Ginecol Obstet. 2019;41:508–519. 11. Michie S, Van Stralen MM, West R. The behaviour change wheel: a new method for characterising and designing behaviour change interventions. Implement Sci. 2011;6:42. 12. Alewijnse D, Metsemakers JF, Mesters IE, van den Borne B. Effectiveness of pelvic floor muscle exercise therapy supplemented with a health education program to

216 Fernandes et al: Group-based pelvic floor muscle education promote long-term adherence among women with urinary incontinence. Neuro- 24. O’Brien BC, Harris IB, Beckman TJ, Reed DA, Cook DA. Standards for reporting urol Urodyn. 2003;22:284–295. qualitative research: a synthesis of recommendations. Acad Med. 2014;89:1245– 13. Alewijnse D, Mesters I, Metsemakers J, van den Borne B. Predictors of long-term 1251. adherence to pelvic floor muscle exercise therapy among women with urinary incontinence. Health Educ Res. 2003;18:511–524. 25. Moser A, Korstjens I. Series: practical guidance to qualitative research. Part 3: 14. Doshani A, Pitchforth E, Mayne CJ, Tincello DG. Culturally sensitive continence sampling, data collection and analysis. Eur J Gen Pract. 2018;24:9–18. care: a qualitative study among South Asian Indian women in Leicester. Fam Pract. 2007;24:585–593. 26. Thomas DR. A general inductive approach for analyzing qualitative evaluation data. 15. de Andrade RL, Bø K, Antonio FI, Driusso P, Mateus-Vasconcelos ECL, Ramos S, et al. Am J Eval. 2006;27:237–246. An education program about pelvic floor muscles improved women’s knowledge but not pelvic floor muscle function, urinary incontinence or sexual function: a 27. Bø K, Kvarstein B, Nygaard I. Lower urinary tract symptoms and pelvic floor muscle randomised trial. J Physiother. 2018;64:91–96. exercise adherence after 15 years. Obstet Gynecol. 2005;105:999–1005. 16. Zhang N, He Y, Wang J, Zhang Y, Ding J, Hua KQ. Effects of a new community-based reproductive health intervention on knowledge of and attitudes and behaviors 28. Neels H, Wyndaele JJ, Tjalma WA, De Wachter S, Wyndaele M, Vermandel A. Knowl- toward stress urinary incontinence among young women in Shanghai: a cluster- edge of the pelvic floor in nulliparous women. J Phys Ther Sci. 2016;28:1524–1533. randomized controlled trial. Int Urogynecol J. 2016;27:545–553. 17. Berzuk K, Shay B. Effect of increasing awareness of pelvic floor muscle function on 29. Neels H, Tjalma WA, Wyndaele JJ, De Wachter S, Wyndaele M, Vermandel A. pelvic floor dysfunction: a randomized controlled trial. Int Urogynecol J. Knowledge of the pelvic floor in menopausal women and in peripartum women. 2015;26:837–844. J Phys Ther Sci. 2016;28:3020–3029. 18. Chiarelli P, Cockburn J. The development of a physiotherapy continence promotion program using a customer focus. Aust J Physiother. 1999;45:111– 30. Smith AL, Nissim HA, Le TX, Khan A, Maliski SL, Litwin MS, et al. Misconceptions 119. and miscommunication among aging women with overactive bladder symptoms. 19. Diokno AC, Newman DK, Low LK, Griebling TL, Maddens ME, Goode PS, Urology. 2011;77:55–59. et al. Effect of group-administered behavioral treatment on urinary in- continence in older women: a randomized clinical trial. JAMA Intern Med. 31. Pintos-Díaz MZ, Alonso-Blanco C, Parás-Bravo P, Fernández-de-Las-Peñas C, Paz- 2018;178:1333–1341. Zulueta M, Fradejas-Sastre V, et al. Living with urinary incontinence: potential 20. Griffiths F, Pepper J, Jørstad-Stein EC, Smith JF, Hill L, Lamb SS. Group versus in- risks of women’s health? A qualitative study on the perspectives of female patients dividual sessions delivered by a physiotherapist for female urinary incontinence: seeking care for the first time in a specialized center. Int J Environ Res Public Health. an interview study with women attending group sessions nested within a rand- 2019;16:3781. omised controlled trial. BMC Women’s Health. 2009;9:25. 21. Carpenter C, Suto M. Qualitative research for occupational and physical therapists: A 32. Miller-Kovach K, Hermann M, Winick M. The psychological ramifications of weight practical guide. Oxford: Blackwell Publishing; 2008. management. J Womens Health Gend Based Med. 1999;8:477–482. 22. Creswell JW, Poth CN. Qualitative Inquiry and research design choosing among five approaches. 4th ed. Thousand Oaks: SAGE Publications; 2018. 33. Crittenden KS, Kaponda CP, Jere DL, McCreary LL, Norr KF. Participation and 23. Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research diffusion effects of a peer-intervention for HIV prevention among adults in rural (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Malawi. Soc Sci Med. 2015;133:136–144. Care. 2007;19:349–357. 34. Demain S, Smith JF, Hiller L, Dziedzic K. Comparison of group and individual physiotherapy for female urinary incontinence in primary care. Physiotherapy. 2001;87:235–242. 35. Hay-Smith EJ, Herderschee R, Dumoulin C, Herbison GP. Comparisons of ap- proaches to pelvic floor muscle training for urinary incontinence in women. Cochrane Database Syst Rev. 2011;12:CD009508. 36. Bandura A. Self-efficacy: toward a unifying theory of behavioral change. Psychol Rev. 1977;84:191–215. 37. Koch LH. Help-seeking behaviors of women with urinary incontinence: an integrative literature review. J Midwifery Womens Health. 2006;51:39–44.