Journal of Physiotherapy 69 (2023) 42–46 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 A comprehensive exercise program improves foot alignment in people with flexible flat foot: a randomised trial Tanya Brijwasi, Pradeep Borkar Orthopaedic Physiotherapy Department, Dr. APJ Abdul Kalam College of Physiotherapy, Pravara Institute of Medical Sciences Loni, Ahmednagar, India KEY WORDS ABSTRACT Flexible flat foot Question: In people with flexible flat foot, what is the effect of a comprehensive exercise program on Longitudinal arch angle navicular drop height and medial longitudinal arch angle compared with a control regimen of brief active Navicular drop height range of motion exercises? Design: Randomised controlled trial with concealed allocation, blinding of as- Physical therapy sessors and intention-to-treat analysis. Participants: Fifty-two people with flexible flat foot. Intervention: Exercises The experimental group undertook 30-minute exercise sessions three times per week for 6 weeks. The exercises involved active dorsiflexion and plantarflexion, foot shortening exercises, gluteal muscle strengthening, and stretching. The control group performed active dorsiflexion and plantarflexion only for 6 weeks. Outcome measures: Navicular drop height and longitudinal arch angle. Results: Randomisation allocated 26 participants to each group. One participant from the experimental group and two from the control group did not complete the study. After 6 weeks, the participants in the experimental group improved their navicular drop height by 0.4 cm (95% CI 0.4 to 0.5) more than those in the control group. These participants also improved their longitudinal arch angle by 16 deg (95% CI 13 to 19) more than those in the control group. Conclusion: In people with flexible flat foot, a comprehensive 6-week exercise program improved the navicular drop height and longitudinal arch angle more than active dorsiflexion and plan- tarflexion alone. This improved the cosmetic appearance of the foot and reduced progression towards more severe flat foot, which typically becomes symptomatic. Trial registration: CTRI/2021/07/034599. [Brijwasi T, Borkar P (2023) A comprehensive exercise program improves foot alignment in people with flexible flat foot: a randomised trial. Journal of Physiotherapy 69:42–46] © 2022 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 foot. In flexible flat foot, the medial longitudinal arch is present in open kinematic chain conditions (non-weight-bearing) and lost in The human foot is a complex structure, which enables it to serve many diverse functions. During standing, it provides a base of sup- closed kinematic chain conditions (weight-bearing). Rigid flat foot port. During gait, the foot must be stable at foot strike and push off has a loss of the medial longitudinal arch height in both open and and be able to absorb shock.1 The bones of the foot are classified into closed kinematic chain conditions.5,6 three segments: the hindfoot (talus and calcaneus), the midfoot (navicular, cuboid and the three cuneiforms) and the forefoot The feet appear to be flat in infants due to presence of fat, (metatarsals and phalanges). These bones and the associated liga- although this normally begins to resolve after 2 years of age. The ments form three arches: the medial longitudinal arch, the lateral longitudinal arch and the transverse arch.2 The structure and arches start to appear when the child starts walking and the foot dynamicity of foot arches are essential for functions of the foot like shock absorption, body weight transmission, forward propulsion begins to regularly bear the body weight. Arches rapidly develop during locomotion and provision of a base of support. The height of the medial longitudinal arch is higher than the lateral longitudinal between 2 and 6 years of age and mature around 12 to 13 years of age. arch, and its curvature may change to a variable degree during weight-bearing.3,4 Prevalence of flat foot in children is high due to ligament laxity, which reduces with age.3 The prevalence of flexible flat foot in children aged Pes planus is a condition where the curvature of the medial lon- 2 to 6 years is 21 to 57%.7 During primary school, the prevalence gitudinal arch is flatter than normal and the entire sole of the foot declines to 13 to 28%.7 Prevalence among those aged 11 to 16 years is comes into near complete or complete contact with the ground.4 Flat foot can be classified into two subtypes: flexible flat foot and rigid flat 42%, among which 23% have bilateral flat foot and 19% have unilateral flat foot.8 Prevalence among young adults aged 18 to 21 years is 13.6%, with a slightly higher prevalence in females (14.4%) than males (12.8%).4 In the wider adult population, the prevalence has been re- ported between 5 and 14% by different researchers.4 The bony structures, ligamentous support and extrinsic and intrinsic foot muscles control excessive pronation and maintain the https://doi.org/10.1016/j.jphys.2022.11.011 1836-9553/© 2022 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 43 medial longitudinal arch during weight-bearing and locomotion ac- to the group to which each participant had been allocated. Data were tivities. Various ligaments have been shown to maintain this struc- analysed with an intention-to-treat approach. tural integrity, including the deltoid ligament, plantar fascia and spring ligament. Extrinsic muscles, including the anterior and pos- Participants terior tibialis and peroneus longus muscles, help to stabilise the midtarsal joint and provide dynamic support to the medial longitu- Participants were eligible for the study if they were aged between dinal arch during stance phase. Various intrinsic foot muscles, such as 18 and 21 years and had navicular drop height . 10 mm and medial abductor hallucis, flexor digitorum brevis and the interosseous longitudinal arch angle , 130 deg. Exclusion criteria were: any muscles, help to stabilise the foot arch during propulsion.9 congenital deformity of the lower limb, any recent injuries, any recent surgeries, any neurological deficit and pregnancy. No exclusion was There are various causes of flat foot, which relate to posterior tibial made due to gender or weight. tendon dysfunction, mid-foot laxity, external rotation of the hindfoot and trauma.10 The incidence of pain is increased around the lower Interventions extremity and the ankle/foot complex. There may be difficulty in walking and an altered gait pattern. Flat foot is also associated with Experimental group osteoarthritis, as the posture and motion of the foot and knee are A comprehensive exercise program was administered, which coupled within a closed kinematic chain during most weight-bearing activities. Closed chain coupling may link excessive planus foot included active dorsiflexion/plantarflexion, four short foot exercises morphology to excessive internal rotation.11 (Figure 2), gluteal strengthening exercises and calf stretches. Exer- cises were progressed after 2 weeks with alteration in the position The gluteus muscles stabilise the hip by counteracting gravity’s (ie, sitting, standing and single-leg stance) and/or an increase in hold hip adduction torque and maintaining proper leg alignment by time and number of repetitions (Table 1). The intervention was per- eccentrically controlling adduction and internal rotation of the thigh. formed at three sessions per week for 6 weeks. Gluteus muscle weakness internally rotates the hip joint, inducing foot pronation. Re-establishing correct muscle recruitment patterns Control group and enhancing strength and performance can be achieved by re- The control group intervention included active dorsiflexion/plan- activating the gluteal muscles.10 Alternative methods have also been used in the treatment of flexible flat foot (eg, passive supports such as tarflexion and calf stretching in long sitting (Table 2). The interven- foot orthoses, taping and motion control footwear have been sug- tion was performed at three sessions per week for 6 weeks. gested in the treatment of flexible flat foot). However, active exercise intervention has additional advantages over passive supports because Outcome measures of the improved foot arches due to strengthening of the intrinsic foot muscles; these are often prescribed by the therapist.11 The height of the medial longitudinal arch was assessed using the navicular drop height (ICC 0.83 to 0.95).12,13 The perpendicular dis- Therefore, the research question for this randomised trial was: tance between the floor and the navicular tuberosity was measured with a ruler first in non-weight-bearing then in weight-bearing. A In people with flexible flat foot, what is the effect of a compre- distance of . 10 mm was diagnostic of flat foot. To measure the hensive exercise program on navicular drop height and medial medial longitudinal arch angle, the centre of the goniometer was longitudinal arch angle compared with a control regimen of brief placed at the navicular tuberosity, and the ends of the goniometer active range of motion exercises? followed the landmarks on the centre of the medial malleolus and the head of the first metatarsal. The angle formed between the line from Method the medial malleolus to the navicular tuberosity and the line con- necting the head of the first metatarsal bone and the navicular tu- Design berosity was measured in degrees. A randomised control trial was conducted from July 2021 to Data analysis January 2022 in the Orthopaedic Physiotherapy Department of the Pravara Institute of Medical Sciences. Potential participants were Analyses were conducted using commercial softwarea. The assessed according to the eligibility criteria. Eligible people who were Shapiro-Wilk test was used to assess the normality of data distribu- willing to participate in the study were provided with verbal infor- tion. Quantitative variables were reported as mean (SD). The mean mation about the study and a written information sheet, and were between-group difference between the experimental and control required to give informed consent before undergoing baseline groups was calculated with unpaired data and reported with a 95% assessment and being allocated to a group. Randomisation was per- confidence interval. formed using simple random sampling into two groups: experi- mental and control. In order to conceal the upcoming random Results allocations, the randomised allocations were concealed in envelopes. Whenever researchers enrolled a new participant, they were required Deviations from the study protocol to contact a researcher who had no other involvement in the study with the new participant’s enrolment details before receiving the Apart from three participants who withdrew from the study due random allocation. Before the intervention period, demographic data to losing interest in participation, there were no deviations from the and baseline assessment of the study outcome measures were study protocol. The registered study question was addressed, all recorded. Participants in the experimental group were prescribed a 6- participants were prescribed their randomly allocated intervention, week comprehensive exercise program and those in the control and both registered outcomes were measured at the scheduled time group received brief, basic exercises for 6 weeks (Figure 1). In order to points. limit the impact of knowing whether they were in the experimental or control group, participants were advised that the study would Flow of participants through the study compare two exercise regimens and they received no information about the exercise intervention to which they were not allocated. The Among the 110 people who were screened for the study, 52 met same researchers reassessed the outcome measures after completion the eligibility criteria and were randomised into two groups: 26 in the of the 6-week intervention period. Outcome assessors were blinded experimental group and 26 in the control group (Figure 1). The
44 Brijwasi and Borkar: Corrective exercises for flexible flat foot Assessed for eligibility (n = 110) Excluded (n = 58) x did not meet eligibility criteria (n = 45) x declined to participate (n = 10) x other reasons (n = 3) Measured navicular drop height and longitudinal arch angle Week 0 Randomised (n = 52) (n = 26) (n = 26) Lost to follow-up (n = 1) Experimental group Control group Lost to follow-up (n = 2) x withdrew (n = 1) x withdrew (n = 2) x Active dorsiflexion and x Active dorsiflexion and plantarflexion plantarflexion x Foot shortening x Calf stretches exercises x 8 min x 3 sessions/wk x Gluteal muscle x 6 wk strengthening x Calf stretches x 30 min x 3 sessions/wk x 6 wk Measured navicular drop height and longitudinal arch angle Week 6 (n = 25) (n = 24) Figure 1. Design and flow of participants through the trial. groups were comparable at baseline, as presented in Table 3. Three height than in the control group. The mean between-group difference participants were lost to follow-up: one from the experimental group in navicular drop height of 0.4 cm in favour of the experimental group and two from the control group (Figure 1). At the 6-week assessment, was a very precise estimate (95% CI 0.4 to 0.5) (Table 4). For individual 25 participants in the experimental group and 24 in the control group participant data, see Table 5 on the eAddenda. were available for assessment and all were measured. Compliance with the prescribed regimen was not assessed. Although the medial longitudinal arch angle increased (ie, improved) in both groups, the mean between-group difference fav- Effect of intervention oured the experimental group by indicating 16 deg greater improvement than in the control group. The mean between-group Although the navicular drop height reduced (ie, improved) in both difference in navicular drop height of 16 deg in favour of the groups, the mean between-group difference favoured the experi- experimental group was a very precise estimate (95% CI 13 to 19) mental group by showing a 0.4 cm greater reduction in navicular drop (Table 4). For individual participant data, see Table 5 on the eAddenda. Figure 2. Short foot exercises: a) lesser toe extension b) toe spread c) doming and d) hallux extension.
Research 45 Table 1 Table 3 Content and progression of the experimental group intervention. Demographic characteristics of participants at baseline. Exercises Weeks Position Contraction Volume Frequency Characteristic Exp (n = 26) Con (n = 26) duration (s) (sessions/wk) Age (yr), mean (SD) 21.1 (0.4) 20.5 (0.9) 161 (5) 162 (6) Warm-up 1 to 6 - - 7 min 3 Height (cm), mean (SD) 60 (7) 62 (8) 23.0 (2.3) Active exercise 1 to 6 sitting 10 15 reps 3 Weight (kg), mean (SD) 23.5 (2.4) Body mass index (kg/m2), mean (SD) dorsiflexion 3 2 sets plantarflexion Exp = experimental group, Con = control group. Short foot exercises 1 to 2 sitting 5 15 reps 3 10 3 2 sets toe spread standing 15 hallux extension 3 to 4 the foot. They work at segmental levels to stabilise the medial longi- single leg tudinal arch and have an important neuromuscular role in fine-tuning doming 5 to 6 stance the position of the arch during weight-bearing functional activities. lesser toe Among the intrinsic foot muscles, abductor hallucis receives much extension attention owing to its various functions. It is believed that the re- covery of this muscle is important in intervention against pes planus. Gluteal muscle 1 to 2 side lying 0 8 to 10 reps 3 The abductor hallucis muscle plays a role in weight-bearing and pushing the body forward during push off in gait. The flexor hallucis strengthening prone lying 3 2 sets brevis muscle maintains the medial longitudinal arch during the terminal stance in gait to maintain foot stability. Hence, training hip abduction 3 to 4 side lying 5 10 to 12 reps abductor hallucis may contribute to reducing navicular drop height and improving the longitudinal arch angle.14 hip extension prone lying 3 2 sets Intrinsic foot muscles also play an important role in static balance, 5 to 6 side lying 10 10 to 12 reps such as standing on one leg, and in adjusting posture. Goldmann et al investigated the effects of toe flexion exercises in normal subjects and prone lying 3 2 sets found improvements in walking, running and jumping.15 Research con- ducted by Kelly et al showed that the activation of abductor hallucis was Cool down 1 to 6 - - 7 min 3 greater in standing on one leg rather than standing on both legs, with range of motion activation patterns being highly correlated with medial postural sway.16 exercises Lynn et al and Kim et al found that short foot exercises for 4 to 5 calf stretches weeks improved balance and navicular drop in patients with flexible flatfoot.17–19 Lee et al found that intrinsic muscle exercises given 5 The three session per week were prescribed to occur on Monday, Wednesday and days/week for 6 weeks improved plantar pressure distribution and Friday. dynamic balance in adults with pes planus.20 Min = minutes, reps = repetitions. Koh et al stated that weakness and dysfunction of the hip external All participants in each group who were followed up at the end of rotator can lead to hip adduction and medial rotation and dynamic the study reported no adverse events during the intervention period. knee valgus, which can affect foot pronation.21 The gluteal muscles (maximus, medius and minimus) stabilise the hip by counteracting Discussion gravity’s hip adduction torque and maintaining proper leg alignment by eccentrically controlling adduction and internal rotation of the This study aimed to estimate the effect of a comprehensive exer- thigh, and externally rotating the alignment of the lower extremity, cise program on navicular drop height and medial longitudinal arch reducing foot pronation. Gluteal muscle weakness internally rotates angle compared with a control regimen of brief active range of mo- the hip joint and induces foot pronation. Re-activating the gluteal tion exercises only. The study generated very precise estimates for muscles will re-establish correct muscle recruitment patterns and each of these outcomes, which both showed clearly greater benefit enhance strength and performance of gluteal muscles. Hence, gluteal from the comprehensive exercise program. The comprehensive ex- muscle strengthening indirectly strengthens the kinetic chain and ercise program was safe and there was little attrition among those helps in improving flat foot. using it over a 6-week period. Previous studies have demonstrated that neuromuscular changes It was difficult to nominate a smallest worthwhile effect for the in gluteus medius are associated with ankle hypermobility, ankle two outcome measures in this study because they were purely injury, iliotibial band friction syndrome and patellofemoral pain biomechanical rather than symptomatic or functional measures syndrome. Thus, strengthening gluteus medius is recommended to about which patients might have been able to give an opinion on the prevent and manage various lower extremity dysfunctions related to clinical worth of improvement of various magnitudes. Nevertheless it excessive pronation of the subtalar joint.21 is worth considering these biomechanical outcome measures because they reflect the severity of the condition and the cosmetic appearance Although the sample size of the study could be considered a of the foot. Perhaps more importantly, improvements in these limitation of the study, it was sufficient to produce very precise es- biomechanical outcomes presumably indicated delay in the pro- timates of the treatment effect. The study’s focus on unilateral flexible gression to more severe flat foot, which typically becomes symp- flat foot could also be seen as a limitation, although this did improve tomatic and may affect gait. The findings of the study will also be of the uniformity of the study participants. interest to those who consider excessive pronation to be a contrib- uting factor to lower extremity injuries.14 In conclusion, navicular drop height and longitudinal arch angle improved substantially after 6 weeks of a comprehensive exercise Short foot exercise training targets the plantar intrinsic muscles program in people with flexible flat foot. The comprehensive exercise (abductor hallucis, flexor digitorum brevis and quadratus plantae) of program was safe and incurred little attrition among those using the program. This evidence supports the belief that these exercises re- Table 2 cruit the intrinsic foot muscles and provides further insight into the Content of the control group intervention. role of the intrinsic foot muscles as local stabilisers of the foot. Im- provements in navicular drop height and longitudinal arch angle Exercises Weeks Position Contraction Volume Frequency presumably indicate delay in the progression to more severe flat foot, which typically becomes symptomatic and may affect gait. duration (s) (sessions/wk) Active exercise 1 to 6 long sitting 10 15 reps 3 dorsiflexion 3 2 sets plantarflexion Cool down 1 to 6 long sitting - 3 min 3 calf stretches The three session per week were prescribed to occur on Monday, Wednesday and Friday. Min = minutes, reps = repetitions.
46 Brijwasi and Borkar: Corrective exercises for flexible flat foot Table 4 Mean (SD) of groups, mean (SD) within-group difference and mean (95% CI) between-group difference. Outcome Groups Within-group difference Between-group difference Week 6 minus Week 0 Week 0 Week 6 Week 6 minus Week 0 Exp minus Con Exp (n = 25) Con (n = 24) Exp (n = 25) Con (n = 24) Exp Con –0.4 (–0.5 to –0.4) Navicular drop height (cm) 1.5 (0.3) 1.6 (0.2) 1.0 (0.2) 1.5 (0.3) –0.5 (0.1) –0.1 (0.1) 16 (13 to 19) Medial longitudinal arch angle (deg) 120 (6) 115 (11) 139 (7) 119 (9) 20 (4) 4 (5) Exp = experimental group, Con = control group. What is known on this topic: Flexible flat foot is a common 3. Patel M, Shah P, Ravaliya S, Patel M. Relationship of Anterior Knee Pain and Flat condition in which the medial longitudinal arch is present when foot: A Cross-Sectional Study. Int J Health Sci Res. 2021;11:86–92. not weight-bearing but lost when weight-bearing. Although mild flat foot may be asymptomatic, as the severity of the condition 4. Aenumulapalli A, Kulkarni MM, Gandotra AR. Prevalence of flexible flat foot in increases it typically becomes symptomatic and affects gait. adults: a cross-sectional study. J Clin Diagnost Res. 2017;11:AC17. What this study adds: In people with flexible flat foot, a comprehensive 6-week exercise program (involving active dor- 5. Bhoir T, Anap DB, Diwate A. Prevalence of flat foot among 18-25 years old phys- siflexion and plantarflexion, foot shortening exercises, gluteal iotherapy students: cross sectional study. Indian J Basic Appl Med Res. 2014;3:272– muscle strengthening, and stretching) improved the navicular 278. drop height and longitudinal arch angle more than a control regimen of active dorsiflexion and plantarflexion. 6. Sivachandiran S, Kumar G. Effect of corrective exercises programme among ath- letes with flat feet on foot alignment factors. Int J Phys Educ Sports Health. Footnotes: a SPSS software windows V.28.0.1.0 version (Statistical 2016;3:193–196. Package for Social Science), SPSS, Chicago, USA. 7. Vittore D, Patella V, Petrera M, Caizzi G, Ranieri M, Putignano P, et al. Extensor eAddenda: Table 5 can be found online at https://doi.org/10.1016/j. deficiency: first cause of childhood flexible flat foot. Orthopaedics. 2009;32:28. jphys.2022.11.011 8. Ibrahim S, Khan MS, Asif M, Hussain F. Prevalence of flat feet among school chil- Ethics approval: The Institutional Ethics Committee(s) approved dren. Indian J Physiother Occup Ther. 2019;13:207–211. this study (MPT/2021/01). All participants gave written informed consent before data collection began. 9. Jacob HA. Forces acting in the forefoot during normal gait–an estimate. Clin Bio- mech. 2001;16:783–792. Competing interests: Nil. Source(s) of support: Nil. 10. McCormack AP, Ching RP, Sangeorzan BJ. Biomechanics of procedures used in adult Acknowledgements: I wish to express my deep gratitude to my flatfoot deformity. Foot Ankle Clin. 2001;6:15–23. guide Dr. Pradeep Borkar and all the participants who participated in my study. 11. Gross KD, Felson DT, Niu J, Hunter DJ, Guermazi A, Roemer FW, et al. Association of Provenance: Not invited. Peer reviewed. flat feet with knee pain and cartilage damage in older adults. Arthritis Care Res. Correspondence: Tanya Brijwasi, Orthopaedic Physiotherapy 2011;63:937–944. Department, Dr. APJ Abdul Kalam College of Physiotherapy, Pravara Institute of Medical Sciences, Loni, Ahmednagar, India. Email: 12. Deng J, Joseph R, Wong CK. Reliability and validity of the sit-to-stand navicular [email protected] drop test: Do static measures of navicular height relate to the dynamic navicular motion during gait. J Stud Phys Ther Res. 2010;2:21–28. References 13. Brody DM. Techniques in the evaluation and treatment of the injured runner. 1. McKeon PO, Hertel J, Bramble D, Davis I. The foot core system: a new paradigm for Orthop Clin North Am. 1982;13:541–558. understanding intrinsic foot muscle function. Brit J Sports Med. 2015;49:290–290. 14. Mulligan EP, Cook PG. Effect of plantar intrinsic muscle training on medial 2. Lehmkuhl LD, Smith LK. Brunnstrom’s Clinical Kinesiology. Philadelphia: Davis; longitudinal arch morphology and dynamic function. Man Ther. 2013;18:425– 1984. 430. 15. Goldmann JP, Sanno M, Willwacher S, Heinrich K, Brüggemann GP. The potential of toe flexor muscles to enhance performance. J Sports Sci. 2013;31:424–433. 16. Kelly LA, Kuitunen S, Racinais S, Cresswell AG. Recruitment of the plantar intrinsic foot muscles with increasing postural demand. Clin Biomech. 2012;27:46–51. 17. Lynn SK, Padilla RA, Tsang KK. Differences in static-and dynamic-balance task performance after 4 weeks of intrinsic-foot-muscle training: the short-foot exer- cise versus the towel-curl exercise. J Sport Rehabil. 2012;21:327–333. 18. Kim EK, Kim JS. The effects of short foot exercises and arch support insoles on improvement in the medial longitudinal arch and dynamic balance of flexible flatfoot patients. J Phys Ther Sci. 2016;28:3136–3139. 19. Kim JS, Lee MY. The effect of short foot exercise using visual feedback on the balance and accuracy of knee joint movement in subjects with flexible flatfoot. Medicine. 2020;99:e19260. 20. Lee DR, Choi YE. Effects of a 6-week intrinsic foot muscle exercise program on the functions of intrinsic foot muscle and dynamic balance in patients with chronic ankle instability. J Exerc Rehabil. 2019;15:709. 21. Koh EK, Weon JH, Jung DY. Effects of activation of gluteus maximus and abdominal muscle using EMG Biofeedback on lumbosacral and tibiocalcaneal angles in standing position. J Korean Phys Ther. 2013;25:411–416.
Journal of Physiotherapy 69 (2022) 60 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Appraisal of Clinical Practice Guideline: Clinical Practice Guidelines and Principles for the care of people with dementia Date of latest update: February/March 2016. Date of next update: 2023. Patient group: Adults approach to adapting guidelines in one setting and for use in different cultural and organ- with dementia and symptoms of dementia, including people of all ages with a variety of forms of isational contexts https://g-i-n.net/wp-content/uploads/2021/05/ADAPTE-Resource-toolkit- dementia, which include Alzheimer’s disease, vascular dementia, dementia with Lewy bodies, V2.1-March-2010-updated-disclaimer.pdf (the ADAPTE process). Of the recommendations, frontotemporal dementia and dementia with Parkinson’s disease. Intended audience: 73 are practice points, seven are consensus-based and 29 are evidence-based. There are Healthcare professionals, including physiotherapists, in primary care, aged care and hospital seven evidence-based recommendations with moderate certainty of evidence, 16 with low settings, and carers of people with dementia. Additional versions: N/A. Expert working group: certainty of evidence and six with very low certainty of evidence. Of the seven recom- A panel of 21 experts from various healthcare fields formed a Guideline Adaption Committee and mendations with moderate certainty, six were of a pharmacological nature. There was a a further panel of four analysed the scientific evidence. The panel of experts consisted of practice point about promoting functional independence and one recommendation (low epidemiologists, geriatricians, psychologists, nurses, occupational therapists, palliative care, certainty of evidence) to encourage exercise, with assessment and advice from a physio- psychogeriatrics, rehabilitation physicians, psychiatry, anthropology and sociology, and carer therapist or exercise physiologist. There were 17 clinical questions addressed by systematic representatives. Funded by: National Health and Medical Research Council (NHMRC) reviews, and five questions defined as background addressed by non-systematic reviews. The Partnership Centre for Dealing with Cognitive and Related Function and funding partners: Guideline Adaption Committee have identified recommendations that should be prioritised HammondCare, Alzheimer’s Australia, Bightwater Care Group and Helping Hand Aged Care for research translation. Consultation with: Public consultation as well as various professional organisations. Approved by: NHMRC. Location: https://cdpc.sydney.edu.au/research/clinical-guidelines-for- Provenance: Invited. Not peer reviewed. dementia/ Description: Dementia is the second leading cause of death in Australia. This document Terry Chan includes guidelines and principles for the care of people with dementia. The main aim is to Department of Physiotherapy, The Alfred Hospital, Melbourne, Australia provide recommendations for the diagnosis and treatment of dementia for Australian healthcare professionals and administrators who work with people with dementia. This https://doi.org/10.1016/j.jphys.2022.10.002 guide is also useful for administrators who provide services for people with dementia and their carers, and for people with dementia and their families. It is based on a systematic 1836-9553/Crown Copyright © 2022 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Appraisal of Clinical Practice Guideline from the American Physical Therapy Association (APTA): Physical therapist management of Parkinson disease Date of latest update: September 2021. Date of next update: 2026. Patient group: Adult care); moderate (for behaviour change approach); and weak (for flexibility exercise and individuals with idiopathic, typical Parkinson’s disease (PD). Intended audience: Physio- telerehabilitation). In the absence of reliable evidence, two best-practice statements were therapists, physiotherapy assistants, neurologists, primary care clinicians, geriatricians, formulated: more research is needed on the effects of physiotherapy after deep brain rehabilitation medicine providers, nurse practitioners, physician assistants, occupational stimulation and service delivered by physiotherapists with expertise in PD may result in therapists and speech language pathologists. Additional versions: None. Expert working better outcomes. group: Twelve physiotherapist members from APTA, seven members of the American Par- kinson’s Disease Association, a neurologist from the American Academy of Neurology and an Commentary: The guideline is very well written and comprehensive, covering the main individual with PD. Funded by: American Physical Therapy Association. Consultation with: components of the AGREE II tool and addressing suggestions on quality improvement, APTA Board of Directors (Board), the APTA Scientific and Practice Affairs Committee, all implementation, dissemination and future research. The recommendations specifically relevant APTA sections and academies, stakeholder organisations and the physiotherapy apply to individuals with early and mid-stages of PD (Hoehn and Yahr stages 1 to 3), community at large. More than 47 public comments were received. Approved by: Not re- not to advanced stages of PD (4 and 5). The authors refrained from recommendations ported. Location: Journal article: https://academic.oup.com/ptj/article/102/4/pzab3 on risk factors and motor learning in the absence of evidence from randomised 02/6485202?login=false. controlled trials. It is unfortunate that the patient’s and carer’s involvement in the process of the development of the Clinical Practice Guidelines was minimal or Description: This Clinical Practice Guideline evaluates the effectiveness of approaches in the absent. physiotherapy management of patients with PD. A systematic literature review was con- ducted and articles were critically analysed by an international panel. The body of evidence Provenance: Invited. Not peer reviewed. for this Clinical Practice Guideline comprised 242 articles. The outcome categories included in each recommendation statement were organised according to the World Health Organi- Rik Gosselink zation’s International Classification of Functioning, Disability and Health Model domains: University of Leuven, Belgium impairment, activity and participation level. A detailed description of the benefits, risks, harms and costs is provided for each recommendation. The quality of the evidence for 9 (out https://doi.org/10.1016/j.jphys.2022.11.002 of 11) recommendations was classified as high. Subsequently, the strength of the recom- mendations was graded as: strong (for aerobic exercise, resistance training, balance training, external cueing, community-based exercise, gait training, task-specific training, integrated 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2022) 60 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Appraisal of Clinical Practice Guideline: Clinical Practice Guidelines and Principles for the care of people with dementia Date of latest update: February/March 2016. Date of next update: 2023. Patient group: Adults approach to adapting guidelines in one setting and for use in different cultural and organ- with dementia and symptoms of dementia, including people of all ages with a variety of forms of isational contexts https://g-i-n.net/wp-content/uploads/2021/05/ADAPTE-Resource-toolkit- dementia, which include Alzheimer’s disease, vascular dementia, dementia with Lewy bodies, V2.1-March-2010-updated-disclaimer.pdf (the ADAPTE process). Of the recommendations, frontotemporal dementia and dementia with Parkinson’s disease. Intended audience: 73 are practice points, seven are consensus-based and 29 are evidence-based. There are Healthcare professionals, including physiotherapists, in primary care, aged care and hospital seven evidence-based recommendations with moderate certainty of evidence, 16 with low settings, and carers of people with dementia. Additional versions: N/A. Expert working group: certainty of evidence and six with very low certainty of evidence. Of the seven recom- A panel of 21 experts from various healthcare fields formed a Guideline Adaption Committee and mendations with moderate certainty, six were of a pharmacological nature. There was a a further panel of four analysed the scientific evidence. The panel of experts consisted of practice point about promoting functional independence and one recommendation (low epidemiologists, geriatricians, psychologists, nurses, occupational therapists, palliative care, certainty of evidence) to encourage exercise, with assessment and advice from a physio- psychogeriatrics, rehabilitation physicians, psychiatry, anthropology and sociology, and carer therapist or exercise physiologist. There were 17 clinical questions addressed by systematic representatives. Funded by: National Health and Medical Research Council (NHMRC) reviews, and five questions defined as background addressed by non-systematic reviews. The Partnership Centre for Dealing with Cognitive and Related Function and funding partners: Guideline Adaption Committee have identified recommendations that should be prioritised HammondCare, Alzheimer’s Australia, Bightwater Care Group and Helping Hand Aged Care for research translation. Consultation with: Public consultation as well as various professional organisations. Approved by: NHMRC. Location: https://cdpc.sydney.edu.au/research/clinical-guidelines-for- Provenance: Invited. Not peer reviewed. dementia/ Description: Dementia is the second leading cause of death in Australia. This document Terry Chan includes guidelines and principles for the care of people with dementia. The main aim is to Department of Physiotherapy, The Alfred Hospital, Melbourne, Australia provide recommendations for the diagnosis and treatment of dementia for Australian healthcare professionals and administrators who work with people with dementia. This https://doi.org/10.1016/j.jphys.2022.10.002 guide is also useful for administrators who provide services for people with dementia and their carers, and for people with dementia and their families. It is based on a systematic 1836-9553/Crown Copyright © 2022 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Appraisal of Clinical Practice Guideline from the American Physical Therapy Association (APTA): Physical therapist management of Parkinson disease Date of latest update: September 2021. Date of next update: 2026. Patient group: Adult care); moderate (for behaviour change approach); and weak (for flexibility exercise and individuals with idiopathic, typical Parkinson’s disease (PD). Intended audience: Physio- telerehabilitation). In the absence of reliable evidence, two best-practice statements were therapists, physiotherapy assistants, neurologists, primary care clinicians, geriatricians, formulated: more research is needed on the effects of physiotherapy after deep brain rehabilitation medicine providers, nurse practitioners, physician assistants, occupational stimulation and service delivered by physiotherapists with expertise in PD may result in therapists and speech language pathologists. Additional versions: None. Expert working better outcomes. group: Twelve physiotherapist members from APTA, seven members of the American Par- kinson’s Disease Association, a neurologist from the American Academy of Neurology and an Commentary: The guideline is very well written and comprehensive, covering the main individual with PD. Funded by: American Physical Therapy Association. Consultation with: components of the AGREE II tool and addressing suggestions on quality improvement, APTA Board of Directors (Board), the APTA Scientific and Practice Affairs Committee, all implementation, dissemination and future research. The recommendations specifically relevant APTA sections and academies, stakeholder organisations and the physiotherapy apply to individuals with early and mid-stages of PD (Hoehn and Yahr stages 1 to 3), community at large. More than 47 public comments were received. Approved by: Not re- not to advanced stages of PD (4 and 5). The authors refrained from recommendations ported. Location: Journal article: https://academic.oup.com/ptj/article/102/4/pzab3 on risk factors and motor learning in the absence of evidence from randomised 02/6485202?login=false. controlled trials. It is unfortunate that the patient’s and carer’s involvement in the process of the development of the Clinical Practice Guidelines was minimal or Description: This Clinical Practice Guideline evaluates the effectiveness of approaches in the absent. physiotherapy management of patients with PD. A systematic literature review was con- ducted and articles were critically analysed by an international panel. The body of evidence Provenance: Invited. Not peer reviewed. for this Clinical Practice Guideline comprised 242 articles. The outcome categories included in each recommendation statement were organised according to the World Health Organi- Rik Gosselink zation’s International Classification of Functioning, Disability and Health Model domains: University of Leuven, Belgium impairment, activity and participation level. A detailed description of the benefits, risks, harms and costs is provided for each recommendation. The quality of the evidence for 9 (out https://doi.org/10.1016/j.jphys.2022.11.002 of 11) recommendations was classified as high. Subsequently, the strength of the recom- mendations was graded as: strong (for aerobic exercise, resistance training, balance training, external cueing, community-based exercise, gait training, task-specific training, integrated 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2023) 15–22 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 Bobath therapy is inferior to task-specific training and not superior to other interventions in improving arm activity and arm strength outcomes after stroke: a systematic review Simone Dorsch a,b, Cameron Carling c, Zheng Cao d, Emma Fanayan a, Petra L Graham e, Annie McCluskey b,f, Karl Schurr b, Katharine Scrivener b,g, Sarah Tyson h a Faculty of Health Sciences, Australian Catholic University, Sydney, Australia; b The StrokeEd Collaboration, Sydney, Australia; c Concentric Rehabilitation Services, Perth, Australia; d Concentric Rehabilitation Services, Sydney, Australia; e School of Mathematical and Physical Sciences, Macquarie University, Sydney, Australia; f Faculty of Medicine and Health, The University of Sydney, Sydney, Australia; g Department of Health Sciences, Macquarie University, Sydney, Australia; h School of Health Sciences, University of Manchester, Manchester, UK KEYWORDS ABSTRACT Stroke Question: What is the effect of Bobath therapy on arm activity and arm strength compared with a dose- Physical therapy matched comparison intervention or no intervention after stroke? Design: Systematic review of rando- Occupational therapy mised trials with meta-analysis. Participants: Adults after stroke. Intervention: Bobath therapy compared Upper extremity with no intervention or other interventions delivered at the same dose as the Bobath therapy. Outcome Rehabilitation measures: Arm activity outcomes and arm strength outcomes. Trial quality was assessed with the PEDro scale. Results: Thirteen trials were included; all compared Bobath with another intervention, which were categorised as: task-specific training (five trials), arm movements (five trials), robotics (two trials) and mental practice (one trial). The PEDro scale scores ranged from 5 to 8. Pooled data from five trials indicated that Bobath therapy was less effective than task-specific training for improving arm activities (SMD –1.07, 95% CI –1.59 to –0.55). Pooled data from five trials indicated that Bobath therapy was similar to or less effective than arm movements for improving arm activities (SMD –0.18, 95% CI –0.44 to 0.09). One trial indicated that Bobath therapy was less effective than robotics for improving arm activities and one trial indicated similar effects of Bobath therapy and mental practice on arm activities. For strength outcomes, pooled data from two trials indicated a large benefit of task-specific training over Bobath therapy (SMD –1.08); however, this estimate had substantial uncertainty (95% CI –3.17 to 1.01). The pooled data of three trials indicated that Bobath therapy was less effective than task-specific training for improving Fugl-Meyer scores (MD –7.84, 95% CI –12.99 to –2.69). The effects of Bobath therapy relative to other interventions on strength outcomes remained uncertain. Conclusions: After stroke, Bobath therapy is less effective than task-specific training and robotics in improving arm activity and less effective than task-specific training on the Fugl-Meyer score. Registration: PROSPERO CRD42021251630. [Dorsch S, Carling C, Cao Z, Fanayan E, Graham PL, McCluskey A, Schurr K, Scrivener K, Tyson S (2023) Bobath therapy is inferior to task- specific training and not superior to other interventions in improving arm activity and arm strength outcomes after stroke: a systematic review. Journal of Physiotherapy 69:15–22] © 2022 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 Despite its extensive clinical use, the efficacy of Bobath therapy to improve outcomes for the affected arm after stroke has never Bobath therapy, developed in the 1950s, has been a predominant been established. Five systematic reviews have compared Bobath approach to stroke rehabilitation globally for many years. Bobath therapy with other interventions targeting arm outcomes after therapy is based on the principles of facilitating automatic and voli- stroke. Two of these reviews contain meta-analyses that pooled tional movement through specific handling techniques that are results from two trials, using different trials in each review.3,4 Both thought to optimise recovery.1 Although the approach has evolved reviews concluded that Bobath therapy is less effective than other over time, there is still an emphasis on the role of sensory input, interventions for improving arm outcomes after stroke. However, which is manipulated via therapists’ facilitation of movement, one concluded that Bobath is more effective than no intervention, focusing on postural and trunk control as the main contributors to based on a meta-analysis of four trials.4 Of the three reviews activity.2 without meta-analysis, two concluded that there is no clear https://doi.org/10.1016/j.jphys.2022.11.008 1836-9553/© 2022 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/).
16 Dorsch et al: Bobath therapy for arm outcomes after stroke Box 1. Inclusion criteria Assessment of characteristics of the trials Design Quality Randomised controlled trial The PEDro database was searched to identify the PEDro scale Participants score. Where a trial was not listed on the PEDro database, two au- Adults after stroke thors independently rated the trial using the PEDro scale. Intervention Participants One group received therapy based on the Bobath concept, Participants in the included trials were adults at any stage after targeting the affected arm stroke. The number of participants, their age, time since stroke and Outcome measures inclusion criteria were recorded to describe the sample. Arm activity Strength Intervention Trials of Bobath therapy that targeted the affected arm were Comparisons Bobath therapy versus a different intervention targeting the included. To determine whether Bobath therapy was used, trials had affected arm, delivered at the same dosage as the Bobath to meet one of the following criteria: the authors explicitly stated that therapy the intervention was based on Bobath or neurodevelopmental Bobath therapy versus no intervention training; or the authors referenced a Bobath textbook or publication when describing the intervention; or the intervention description indication of superiority of any approach,5,6 while the most recent7 suggested that it was based on Bobath therapy (ie, aimed to found moderate evidence that other interventions were more normalise movement, normalise tone, facilitate normal movement or effective than Bobath therapy for improving motor control and inhibit reflex activity). If it was unclear whether the intervention was dexterity of the arm. In summary, the reviews to date have based on Bobath therapy, the trial’s authors were contacted. If the concluded that Bobath therapy is less effective than other in- intervention was mixed types of therapy, the publication needed to terventions for arm outcomes after stroke but is more effective than clearly state that at least half of the intervention targeting the no intervention. affected arm was Bobath therapy. If the Bobath therapy targeted multiple activities including lower limb activities, the amount of None of the reviews discussed above contains a meta-analysis of therapy targeting the affected arm needed to be clearly stated. all the available trials that investigate the use of Bobath therapy specifically targeting the affected arm after stroke. Therefore, this Comparison systematic review primarily aimed to evaluate the effect of Bobath The comparisons of interests were ‘another intervention’ or ‘no therapy on arm outcomes after stroke by comparing it with no intervention or a dose-matched intervention, using all the available intervention’. Where another intervention targeting the affected arm trials. was used, it needed to be delivered at the same dose (ie, the same amount of time) as the Bobath therapy. Two authors reviewed all Therefore, the research questions for this systematic review of comparison interventions (SD, EF) and grouped them into broad randomised trials with meta-analysis were: categories. 1. What is the effect of Bobath therapy on arm activity outcomes Outcome measures compared with a dose-matched comparison intervention or no The outcomes of interest were arm activity and arm strength. intervention after stroke? Activity outcomes were defined as assessments of the ability to 2. What is the effect of Bobath therapy on arm strength outcomes perform a task. This included standardised outcome measures such as compared with a dose-matched comparison intervention or no the Wolf Motor Function Test or Box and Block test, or customised intervention after stroke? assessments such as a timed reaching task.8,9 Arm strength outcomes were defined as assessments of force production. The Fugl-Meyer Method motor assessment was included as a measure of arm strength, as this composite impairment outcome measure predominantly as- Identification and selection of trials sesses strength. Fifty-six of the 66 points are allocated to volitional movements of the arm, scored as 0 for no movement, 1 for partial The systematic review was prospectively registered with PROS- movement or 2 for full movement. PERO and reported according to the PRISMA statement. An electronic search for relevant trials was conducted in September 2020 and Data extraction and analysis updated in December 2021. The following databases were searched: Ovid MEDLINE, Embase, CINAHL and PEDro. The search included Two authors (EF and SD) independently examined the full-text stroke-related terms, randomised controlled trial-related terms and version of the included trials to extract data. A variety of arm activity therapy-related terms. See Appendix 1 on the eAddenda for the full and arm strength outcome measures were reported. Where multiple details of the search strategy. outcome measures were reported, a hierarchy of choice was applied to decide which outcome measure to include in the pooled analysis. For Titles and abstracts were screened independently by two authors the activity outcome measures, the measure used was (listed in order of (SD and EF) to identify relevant trials. Full-text copies of relevant preference): ARAT, Wolf Motor Function Test, Frenchay Arm Test, Upper papers were retrieved and reviewed independently by two of the Extremity function test, Box and Block test and customised outcome following authors (KS, ST, CC, EF, AM, KS, SD and ZC) using pre- measures that involved reaching tasks.8,9 For the arm strength outcome determined criteria to determine eligibility (Box 1). If the two re- measure, the strength measures used were (again listed in order of viewers disagreed about the eligibility of a trial, a discussion was held preference): grip strength, shoulder flexion strength, elbow extension with the author group until a consensus was reached. Where ab- strength, Motricity Index and the Fugl-Meyer upper limb motor score. stracts or full-text trials were only available in another language, Mean post-intervention and change scores and standard deviations Google Translate was used to translate the methods section into were retrieved where possible. Where standard deviations for change English. Where abstracts were only available in Chinese, one author scores were not provided, they were imputed as suggested in the (ZC) reviewed these trials. See Appendix 2 on the eAddenda for de- Cochrane Handbook,10 whereby the available pre and post standard tails of the full-text screening. deviations were combined with an estimated correlation calculated
Research 17 from other studies. Data were extracted from graphs for two trials.9,11 participants late after stroke (. 6 months). The average age of par- Seven authors were contacted to request additional data and one ticipants ranged from 49 to 73 years. author responded.12 Intervention Where a variety of outcome measures for activity and strength outcomes were reported, SMD based on change scores was used. Bobath therapy was described as ‘Bobath’ therapy or therapy Where only a single score was used to measure the outcome, WMD following ‘Bobath principles’ by authors in six trials.8,11,16–19 Bobath based on change scores was used. Pooled estimates of intervention therapy was described as ‘neurodevelopmental’ therapy in seven effect were calculated via DerSimonian and Laird random-effects trials.9,12,20–24 See Appendix 3 on the eAddenda for details of the meta-analyses.13 All pooled results were reported as SMD or WMD Bobath interventions. (Bobath therapy – comparator therapy) with 95% CI. For the few outcome measures in which lower scores indicated a better Comparison interventions were allocated to one of four cate- outcome, negative signs were added to the mean scores so that the gories; see Appendix 4 on the eAddenda for operational definitions of outcomes were all in the same direction. Where post-intervention each category. Interventions that involved general arm movements results were reported as medians and interquartile ranges, the without reaching, grasping or manipulation of everyday objects were methods described by Hozo and colleagues were used to convert called arm movements. The comparison interventions were cat- results into estimated means and standard deviations.14 Heteroge- egorised as arm movements in five trials.8,12,17,18,22 Interventions that neity between trials was assessed using Cochrane’s Q. Sensitivity involved reach, grasp and manipulation of everyday objects such as analyses were undertaken whereby studies with imputed SD were cups, combs and light switches were termed task-specific training; excluded from the analysis. R statistical software with the meta constraint-induced movement therapy shaping was included in this package was used for all analyses.15 category. The comparison interventions were categorised as task- specific training in five trials.11,19–21,24 Interventions that involved arm Results movements using a device that could assist movement and provided gaming interaction were termed robotics. The comparison in- Flow of trials through the review terventions were categorised as robotics in two trials.9,16 In- terventions that involved motor imagery of the affected arm whilst The electronic search strategy identified 1,684 papers. After performing reaching, grasping and manipulation activities were screening of titles and abstracts, 188 full-text publications were termed mental practice. The comparison intervention was categorised retrieved and screened for eligibility. After screening the reference as mental practice in one trial.23 See Appendix 5 on the eAddenda for lists of other systematic reviews, five additional papers were identi- details of the comparison interventions. With the exclusion of one fied. Titles and abstracts of these five papers were screened, and one trial that delivered 6 hours of intervention to both groups,20 the full-text publication was retrieved and screened for eligibility. A total average dosage of Bobath therapy and the comparison therapies was of 188 full-text publications were evaluated. To determine eligibility, 51 minutes per session (range 20 to 120 minutes) over 5 weeks (range 11 authors were contacted to confirm whether at least 50% of the 2 to 8 weeks). intervention was Bobath therapy in one group of their trial: three authors did not respond; one author responded that one group had Outcome measures received intervention based on Bobath therapy;12 and the remaining authors responded that their trial did not include an intervention Of the included trials, all provided data that could be included in where at least 50% was Bobath therapy. After screening full texts, 172 both the activity and arm strength analyses, except for one trial that papers failed to meet the inclusion criteria. Three trials had resulted had no activity level outcome measure16 and two trials that had no in two published papers each (one with the full data set and one with arm strength outcome measure.18,22 a subgroup of the data); the two papers that reported the full data set for those two trials were included. Therefore, a total of 13 trials were Effect of Bobath therapy versus other interventions on arm included in this systematic review. See Figure 1 for a summary of the activity outcomes flow of trials through the review. The effect of Bobath therapy compared with task-specific arm Characteristics of the included trials training was examined by pooling outcomes from five trials involving 247 participants (Figure 2). The pooled SMD was large Of the 13 included trials, all were published in English. No trials (SMD –1.07) in favour of task-specific training, with the confidence compared Bobath therapy with no intervention. Although trials of interval indicating that the effect was moderate to large (95% CI Bobath therapy compared with no intervention were mentioned in –1.59 to –0.55, I2 = 68%). For a detailed forest plot, see Figure 3 on the Introduction, none of these trials met our inclusion criteria, as the the eAddenda. Bobath therapy targeted multiple activities rather than specifically targeting the affected arm. Together the trials included 636 unique The effect of Bobath therapy compared with arm movements was participants. See Table 1 for details of the included trials. examined by pooling outcomes from five trials involving 262 par- ticipants (Figure 4). The pooled SMD indicated that arm movements Quality have a similar or better effect on arm activities than Bobath therapy (SMD –0.18, 95% CI –0.44 to 0.09, I2 = 11%). There was substantial The PEDro scores ranged from 5 to 8 out of 10. No trials used heterogeneity in the first pooled analysis but there were no pre- blinding of participants or therapists to group allocation. One trial did dictors to explore that could potentially explain this heterogeneity. not use blinded assessors, nine did not use concealed allocation, five For a detailed forest plot, see Figure 5 on the eAddenda. did not use intention-to-treat analysis and two did not have adequate follow-up (ie, 85% of the sample). See Table 2 for further details. The effect of Bobath therapy compared with robotics was exam- ined in one trial of 19 participants. The MD in forward reach distance Participants showed 6.6 cm more change from robotics than Bobath therapy, with the confidence interval indicating between 1.1 and 12.1 cm greater The participants were between 14 days and 4.5 years after their change in reach for robotics than Bobath therapy.9 stroke, with five trials including participants in the acute/sub-acute stages after stroke (, 6 months) and eight trials including The effect of Bobath therapy compared with mental practice was examined in one trial of 39 participants. The median and interquartile range of the change scores was converted to mean and SD.25,26 The mean difference in Wolf motor function test (time) scores indicated similar effects of Bobath therapy and mental practice; the test was completed a mean of 0.65 seconds faster in the mental practice group.
18 Dorsch et al: Bobath therapy for arm outcomes after stroke Titles and abstracts screened (n = 1,684) • from PEDro (n = 741) • from CINAHL (n = 101) • from Embase (n = 112) • from Medline (n = 725) • from reference lists (n = 5) Papers excluded after screening titles/abstracts (n = 1,496) Potentially relevant papers retrieved for evaluation of full text (n = 188) • from electronic databases (n = 187) • from reference lists (n = 1) Papers excluded after evaluation of full text (n = 172) a • no Bobath intervention (n = 149) • intervention dosages not equivalent (n = 96) • upper limb outcome not measured (n = 25) • design not a randomised trial (n = 17) • participants not adults aged > 18 years (n = 4) Papers included in review (n = 16) Trials included in review (n = 13) Trials included in meta-analysis (n = 12) Figure 1. Flow of trials through the review. a Papers may have met more than one exclusion criterion. However, this estimate had substantial uncertainty (MD –0.65 s, 95% estimate (95% CI –3.17 to 1.01, I2 = 95%). For a detailed forest plot, see CI –2.12 to 0.82).23 Figure 11 on the eAddenda. Sensitivity analyses The effect of Bobath therapy compared with arm movements was Sensitivity analyses of the effect of Bobath therapy compared with examined by pooling outcomes from three trials involving 179 par- ticipants (Figure 12). The pooled MD point estimate of Fugl-Meyer other interventions on arm activity using no imputed data were motor score indicated slightly more benefit from arm movements calculated. These meta-analyses had very similar results to the ana- (MD –2.46), although the confidence interval showed substantial lyses with imputed data. The SMD of Bobath therapy compared with uncertainty in this estimate (95% CI –7.09 to 2.16, I2= 76%). For a task-specific training (two trials involving 36 participants) was –1.16, detailed forest plot, see Figure 13 on the eAddenda. 95% CI –1.89 to –0.44. The SMD of Bobath therapy compared with arm movements (three trials involving 186 participants) was –0.13, 95% CI The effect of Bobath therapy compared with robotics was exam- –0.42 to 0.16. See Figures 6 and 7 on the eAddenda. ined by pooling strength outcomes from two trials involving 81 participants (Figure 14). The pooled SMD indicated slightly better Effect of Bobath therapy versus other interventions on arm benefit from robotics (SMD –0.25), although the confidence interval strength outcomes showed substantial uncertainty in this estimate (95% CI –0.69 to 0.19, I2 = 0%). There was substantial heterogeneity in some of the pooled The effect of Bobath therapy compared with task-specific training analyses but there were no predictors to explore that could poten- was examined by pooling outcomes from three trials involving 138 tially explain this heterogeneity. For a detailed forest plot, see participants (Figure 8). The pooled WMD of Fugl-Meyer motor Figure 15 on the eAddenda. outcome scores was in favour of task-specific training (MD –7.84 points), although the true size of the effect may be substantially The effect of Bobath therapy compared with mental practice was smaller or larger than this estimate (95% CI –2.69 to –12.99 points, I2 = examined in one trial of 39 participants. The median and interquartile 40%). For a detailed forest plot, see Figure 9 on the eAddenda. range of the change scores was converted to mean and SD.25,26 The MD of grip strength indicated similar effects of Bobath therapy and The effect of Bobath therapy compared with task-specific training mental practice (0.43 kg in favour of the mental practice group), was also examined by pooling strength outcomes from two trials although the estimate is unclear (95% CI –2.83 to 3.69).23 involving 109 participants (Figure 10). The SMD point estimate was a large effect in favour of task-specific training (SMD –1.08), although Discussion the confidence interval showed substantial uncertainty in this Bobath therapy was less effective than task-specific training and robotics in improving arm activity after stroke. Arm movements were
Research 19 Table 1 Intervention Characteristics of included trials. Bobath (n = 52): training based on the Trial Participants Bobath neurodevelopmental technique Outcome measures Arya 201219 Comparison (n = 51): task-specific N = 103 training (motor learning and shaping) Activity Impairment Age (yr) Dosage: 60 min, 4 to 5/wk, 4 wk ARATa FMA WMFT Bobath = 50 (SD 8) Bobath (n = 16): therapy based on Bobath Comparison = 52 (SD 8) therapeutic exercises MAL Chronicity (wk) Comparison (n = 13): arm movements Bobath = 12 (SD 7) (skill acquisition with EMG biofeedback) Timing: 0 wk, 4 wk Comparison = 12 (SD 6) Dosage: 45 min, 3/wk, 5 wk Basmajian 198718 N = 29 Bobath (n = 20): training based on Activity Impairment El-Bahrawy 201211 Age (yr) Bobath principles UEFT Finger oscillation test Lum 20029 Comparison (n = 20): task-specific Timing: 0 wk, 5 wk Moon 201824 All = 62 (39 to 79) training (drinking task involving reach, Impairment Piron 20108 Chronicity (wk) grasp and release) Activity Grip strength Platz 200517 Both: electrical stimulation at the wrist PPT Modified Ashworth scale All = 16 (4 to 44) for 30 min, 3/wk, 6 wk Dosage: 45 min, 3/wk, 6 wk Timing: 0 wk, 6 wk N = 40 Age (yr) Bobath (n = 14): training based on Activity Impairment neurodevelopmental therapy Reach distance FMAa Bobath = 49 (SD 3) Comparison (n = 13): robotics (reaching Shoulder muscle strengtha Comparison = 51 (SD 3) to target with upper limb in a robotic Timing: 0 wk, 8 wk Chronicity (mo) device) Elbow muscle strength Bobath = 8 (SD 1) Both: tone normalisation and limb Comparison = 9 (SD 2) positioning for 10 min, 24 in 8 wk Activity Impairment Dosage: 50 min, 24 in 8 wk MAL FMA N = 27 Age (yr) Bobath (n = 9): neurodevelopmental Timing: 0 wk, 4 wk therapy-based manual exercise Bobath = 66 (SD 2) Comparison (n = 9): task-specific training Activity Impairment Comparison = 63 (SD 4) (upper limb circuit training using putty, Timed reach task FMA Chronicity (mo) skate, incline board, stacking cones, range Bobath = 29 (SD 6) of motion arc and ring) Timing: 0 wk, 4 wk Comparison = 30 (SD 6) Dosage: 20 min, 5 to 6/wk, 4 wk Activity Impairment N = 18 Bobath (n = 23): specific exercises based ARAT FMA Age (yr) on Bobath principles Ashworth Scale Comparison (n = 27): arm movements Timing: 0 wk, 4 wk Bobath = 63 (SD 12) (reach and grasp movements with Comparison = 71 (SD 9) affected arm using motion-tracking Chronicity (d) equipment) Bobath = 21 (SD 5) Dosage: 60 min, 5/wk, 4 wk Comparison = 18 (SD 5) Bobath (n = 20): therapy following a N = 50 Bobath manual, supervised by a senior Age (yr) Bobath instructor. Comparison (n = 20): arm movements Bobath = 62 (SD 10) (repetitive training of arm movements Comparison = 59 (SD 8) through available range of motion) Chronicity (mo) Both: usual standard rehabilitation Bobath = 15 (SD 12) therapy (activities of daily living, arm Comparison = 15 (SD 13) activities, stance, gait, speech and cognition) N = 40 Dosage: 45 min, 20 in 4 wk Age (yr) Bobath (n = 32): therapy based on Bobath = 61 (SD 11) neurodevelopmental training principles Comparison = 63 (SD 13) Comparison (n = 22): arm movements Chronicity (wk) (virtual reality-based training of upper Bobath = 7 (SD 4) limb movements) Comparison = 6 (SD 4) Dosage: 45 min, 16 in 4 wk Schuster-Amft 201822 N = 54 Bobath (n = 36): activities based on the Activity Impairment Suputtitada 200420 Age (yr) neurodevelopmental training method BBTa CMSA Tariah 201021 Comparison (n = 33): task-specific CAHAI-13 Line bisection test Bobath = 61 (SD 11) training (CIMT) Comparison = 61 (SD 13) Dosage: 6hr, 5/wk, 2 wk Timing: 0 wk, 4 wk Chronicity (yr) Bobath = 3.6 (SD 3.7) Bobath (n = 8): facilitation of arm Activity Impairment Comparison = 2.4 (SD 2.4) movement based on ARAT Grip strengtha neurodevelopmental principles N = 69 Comparison (n = 10): task-specific Timing: 0 wk, 2 wk Pinch strength Age (yr) training (reach, grasp and manipulation tasks, restraint applied to intact arm) Activity Impairment Bobath = 59 (SD 4) Dosage: 120 min, 7/wk, 8 wk WMFTa FMA Comparison = 60 (SD 5) MAL Chronicity (yr) . 80% of participants Timing: 0 wk, 8 wk (range) = 1 to 3 N = 18 Age (yr) Bobath = 61 (SD 5) Comparison = 55 (SD 11) Chronicity (mo) Bobath = 9 (SD 6) Comparison = 10 (SD 4)
20 Dorsch et al: Bobath therapy for arm outcomes after stroke Table 1 (Continued) Trial Participants Intervention Outcome measures Taveggia 201616 N = 54 Bobath (n = 27): passive and active- Activity Impairment Timmermans 201323 Age (yr) assisted mobilisation based on the MIa Bobath concept Modified Ashworth Scale Whitall 201112 Bobath = 68 (SD 13) Comparison (n = 27): robotics (upper Hand pain Comparison = 73 (SD 10) limb movements in ‘Armeo Spring’ Chronicity (mo) robotic device) Timing: 0 wk, 6 wk All (range) = 0.5 to 12 Both: conventional treatment 30 min, 5/wk, 6 wk Activity Impairment N = 42 Dosage: 30 min, 5/wk, 6 wk FAI FMAa Age (yr) WMFTa Grip strengtha Bobath (n = 21): exercise based on Arm accelerometry Bobath = 59 (SD 10) neurodevelopmental training principles Impairment Comparison = 60 (SD 7) Comparison (n = 21): mental practice Timing: 0 wk, 6 wk FMAa Chronicity (d) (mental practice of functional arm Shoulder muscle strength Bobath = 32 (SD 18) movements) Activity Elbow muscle strength Comparison = 36 (SD 27) Both: usual therapy WMFT Wrist muscle strength Dosage: 30 min, 7/wk, 6 wk Shoulder ROM N = 92 Elbow ROM Age (yr) Bobath (n = 50): exercises based on Wrist ROM neurodevelopmental training principles Thumb ROM Bobath = 58 (SD 13) Comparison (n = 42): arm movements Comparison = 60 (SD 10) (reaching task within a training Chronicity (yr) apparatus of handlebars moved along Bobath = 4.1 (SD 5.2) linear tracks) Comparison = 4.5 (SD 4.1) Dosage: 60 min, 3/wk, 6 wk Timing: 0 wk, 6 wk ARAT = Action Research Arm Test, WMFT = Wolf Motor Function Test, MAL = Motor Activity Log, FMA = Fugl-Meyer Assessment, UEFT = Upper Extremity Function Test, PPT = Purdue Pegboard Test, MI = Motricity Index, FAI = Frenchay Activities Index, BBT = Box and Block Test, CAHAI-13 = Chedoke McMaster Arm and Hand Activity Inventory, CMAS = Chedoke- McMaster Stroke Assessment. a Indicates outcome measure used for analysis. Table 2 PEDro scores of included studies. Study Random Concealed Groups Participant Therapist Assessor , 15% Intention-to-treat Between-group Point estimate Total allocation allocation similar blinding blinding blinding dropouts analysis difference and (0 to 10) at baseline reported variability reported Ayra 201219 Y Y Y N NYY Y Y Y8 N Y Y5 Basmajian 198718 Y N Y N N Y N Y Y Y8 N Y Y6 El-Bahrawy 201211 Y Y Y N NYY N Y Y5 Y Y Y8 Lum 20029 Y N Y N NYY Y Y Y8 Y Y Y8 Moon 201824 Y N Y N NNY Y Y Y7 N Y Y6 Piron 20108 Y Y Y N NYY Y Y Y7 Y Y Y7 Platz 200517 Y Y Y N NYY Y Y Y6 Schuster-Amft 201822 Y Y Y N NYY Supittitada 200420 Y N Y N NYY Tariah 201021 Y N Y N NYY Taveggia 201616 Y N Y N NYY Timmermans 201323 Y N Y N NYY Whitall 201112 Y N Y N NYN Study SMD (95% CI) Study SMD (95% CI) Arya 2012 Random Basmajian 1987 Random El-Bahrawy 2012 Piron 2010 Moon 2018 Platz 2005 Supittitada 2004 Schuster-Amft 2018 Tariah 2010 Whitall 2011 Pooled Pooled –3 –2 –1 0 12 3 –1.5 –1.0 –0.5 0 0.5 1.0 1.5 Favours task- Favours Favours arm Favours specific training Bobath movements Bobath Figure 2. Standardised mean difference (95% CI) in the effect of Bobath therapy versus Figure 4. Standardised mean difference (95% CI) in the effect of Bobath therapy versus task-specific training on activity outcomes. arm movements on activity outcomes.
Research 21 Study WMD (95% CI) between Bobath therapy and task-specific training in improving arm Arya 2012 Random strength measured with specific strength measures remains very Moon 2018 unclear. Robotics and arm movements were similar to or better than Tariah 2010 Bobath therapy for improving arm strength after stroke. The relative efficacy of Bobath therapy and mental practice for improving arm Pooled strength after stroke remains unclear. –15 –10 –5 0 5 10 15 There is a large body of evidence to guide rehabilitation in- terventions following stroke. Clinical guidelines for stroke rehabili- Favours task- Favours tation universally recommend intensive task-specific training, with specific training Bobath no clinical guidelines recommending Bobath therapy. Previous sys- tematic reviews of rehabilitation for the arm after stroke generally Figure 8. Weighted mean difference (95% CI) in the effect of Bobath therapy versus conclude that other interventions are more effective than Bobath task-specific training on Fugl-Meyer Assessment score. therapy, or that there is insufficient evidence to make conclusions about the relative effectiveness of different interventions and treat- Study SMD (95% CI) ment approaches.3–7 However, these previous reviews have included El-Bahrawy 2012 Random relatively low numbers of trials with small sample sizes. The current Supittitada 2004 review was able to include 13 trials involving close to 600 partici- pants. Findings from this comprehensive review, combined with Pooled conclusions from previous reviews, confirm that task-specific training and robotics result in improved arm outcomes for stroke survivors –3 –2 –1 0 12 3 when compared with Bobath therapy. Favours task- Favours The results of this systematic review show that task-specific specific training Bobath training is superior to Bobath therapy for arm activity outcomes, with the 95% CI for the activity outcomes showing a moderate to large Figure 10. Standardised mean difference (95% CI) in the effect of Bobath therapy versus effect size (between 0.55 and 1.59) in favour of task-specific training task-specific training on strength outcomes. for activity. Additionally, task-specific training is superior or equiva- lent to Bobath therapy for arm strength outcomes. Task-specific Study WMD (95% CI) training involves intensive practice of the tasks a person is trying to Piron 2010 Random improve. When motor impairments do not allow individuals to Platz 2005 practise tasks in their entirety, there is evidence to support training of Whitall 2011 components of tasks without being passively moved or assisted. For example, ‘shaping’ strategies used in constraint-induced movement Pooled therapy involve the use of part-practice to target a person’s specific activity limitations. This differs fundamentally from Bobath therapy, –15 –10 –5 0 5 10 15 where therapists believe that facilitation of movement by passive (or active-assisted) guidance of movements in a specific way, focusing on Favours arm Favours postural control, will lead to improved outcomes for stroke movements Bobath survivors.2 Bobath therapy is reliant on the therapist assisting the stroke survivor and hence requires more staff time and is more Figure 12. Weighted mean difference (95% CI) in the effect of Bobath therapy versus resource intensive. The use of therapies such as task-specific training, arm movements on Fugl-Meyer Assessment score. arm movements, robotics and mental practice provides stroke survivors with increased opportunity to independently and Study SMD (95% CI) intensively practise the arm movements and tasks that they are Taveggia 2016 Random aiming to improve. Even a finding of equivalent outcomes would Lum 2002 indicate that therapists should prioritise these other therapies over Bobath therapy. Pooled As found in a previous review of lower limb outcomes,27 limita- –1.5 –1.0 –0.5 0 0.5 1.0 1.5 tions included the lack of clarity in definitions provided about Bobath therapy. Bobath therapy is an approach rather than one discrete Favours robotics Favours Bobath intervention; consequently, it is difficult to standardise the in- terventions between therapists or trial interventions. However, there Figure 14. Standardised mean difference (95% CI) in the effect of Bobath therapy versus is an underpinning belief of Bobath therapy that therapists need to arm robotics on strength outcomes. facilitate movements that focus on postural control and the trunk, and this appears to be common to the interventions described in similar to or better than Bobath at improving arm activity after these trials. The publication dates and quality scores of the included stroke. Bobath therapy and mental practice had similar efficacy for trials could be considered as limitations, as the trials were published improving arm activity after stroke. Bobath therapy was less effective between 1987 and 2018 and there was variety in the methodological than task-specific training in improving arm strength after stroke quality. However, eight of the 13 studies scored 7 or 8 on the PEDro measured with the Fugl-Meyer motor score but the difference scale and the lowest score was 5. A strength of this review was that the search was comprehensive, and it included a greater number of trials than previous reviews on this topic. Importantly, it only included trials where the dose of Bobath therapy targeting the affected arm and the comparison intervention were matched. Thus, differences in the dose of treatments cannot be considered a possible reason for the differences in measured outcomes. In conclusion, Bobath therapy is less effective than task-specific training and robotics for improving arm activities after stroke. Bobath therapy is less effective than task-specific training for
22 Dorsch et al: Bobath therapy for arm outcomes after stroke improving arm strength after stroke, as reflected in the Fugl-Meyer 8. Piron L, Turolla A, Agostini M, Zucconi CS, Ventura L, Tonin P, et al. Motor learning Assessment score. Use of Bobath therapy in preference to other in- principles for rehabilitation: a pilot randomized controlled study in poststroke terventions is not supported. patients. Neurorehabil Neural Repair. 2010;24:501–508. What is already known on this topic: Bobath therapy is 9. Lum PS, Burgar CG, Shor PC, Majmundar M, Van der Loos M. Robot-assisted widely used in stroke rehabilitation, despite a growing body of movement training compared with conventional therapy techniques for the evidence challenging its efficacy and underlying beliefs. rehabilitation of upper-limb motor function after stroke. Arch Phys Med Rehabil. What this study adds: This review shows that task-specific 2002;83:952–959. training and robotic training is more effective than Bobath ther- apy for improving upper limb activity outcomes after stroke. Task- 10. Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (Eds). specific training is also more effective than Bobath therapy for Cochrane Handbook for Systematic Reviews of Interventions version 6.3 (updated improving Fugl-Meyer Assessment score after stroke. It challenges February 2022). Cochrane, 2022. Available from www.training.cochrane.org/ the prioritisation of Bobath therapy in stroke rehabilitation. handbook. Accessed November 12, 2022. eAddenda: Appendices 1 to 5 and Figures 3, 5, 6, 7, 9, 11, 13 and 15 11. El-Bahrawy MNEA. Efficacy of motor relearning approach on hand function in can be found online at https://doi.org/10.1016/j.jphys.2022.11.008 chronic stroke patients. A controlled randomized study. Ital J Physiother. 2012;2:4. Ethics approval: Nil. 12. Whitall J, Waller SM, Sorkin JD, Forrester LW, Macko RF, Hanley DF, et al. Bilateral Competing interests: Authors SD, KS, AM, KS are directors of and Unilateral Arm Training Improve Motor Function Through Differing Neuro- StrokeEd and earn some income from workshops on evidence-based plastic Mechanisms: A Single-Blinded Randomized Controlled Trial. Neurorehabil therapy in stroke rehabilitation. Neural Repair. 2011;25:118–129. Source(s) of support: Nil. Acknowledgements: Izumi Ibayashi for reviewing non-English 13. DerSimonian R, Laird N. Meta-analysis in clinical trials revisited. Contemp Clin manuscripts. Trials. 2015;45:139–145. Providence: Not invited. Peer reviewed. Correspondence: Simone Dorsch, School of Allied Health, 14. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, Australian Catholic University, Sydney, Australia. Email: range, and the size of a sample. BMC Med Res Methodol. 2005;5:13. [email protected] 15. Schwarzer G. meta: An R Package for Meta-Analysis. R news. 2007;7:40–45. References 16. Taveggia G, Borboni A, Salvi L, Mulé C, Fogliaresi S, Villafañe JH, et al. Efficacy of robot- 1. Michielsen M, Vaughan-Graham JA, Holland A, Magri A, Suzuki M. The Bobath assisted rehabilitation for the functional recovery of the upper limb in post-stroke concept - a model to illustrate clinical practice: responding to comments on patients: a randomized controlled study. Eur J Phys Rehabil Med. 2016;52:767–773. Michielsen et al. Disabil Rehabil. 2019;41:2109–2110. 17. Platz T, Eickhof C, van Kaick S, Engel U, Pinkowski C, Kalok S, et al. Impairment- oriented training or Bobath therapy for severe arm paresis after stroke: a single- 2. Vaughan-Graham J, Cheryl C, Holland A, Michielsen M, Magri A, Suzuki M, et al. blind, multicentre randomized controlled trial. Clin Rehabil. 2005;19:714–724. Developing a revised definition of the Bobath concept: Phase three. Physiother Res 18. Basmajian JV, Gowland CA, Finlayson MA, Hall AL, Swanson LR, Stratford PW, et al. Int. 2020;25:e1832. Stroke treatment: comparison of integrated behavioral-physical therapy vs tradi- tional physical therapy programs. Arch Phys Med Rehabil. 1987;68:267–272. 3. Hiraoka K. Rehabilitation effort to improve upper extremity function in post-stroke 19. Arya KN, Verma R, Garg RK, Sharma VP, Agarwal M, Aggarwal GG. Meaningful task- patients: a meta-analysis. J Phys Ther Sci. 2001;13:5–9. specific training (MTST) for stroke rehabilitation: a randomized controlled trial. Top Stroke Rehabil. 2012;19:193–211. 4. Wattchow KA, McDonnell MN, Hillier SL. Rehabilitation interventions for upper 20. Suputtitada A, Suwanwela NC, Tumvitee S. Effectiveness of constraint-induced limb function in the first four weeks following stroke: a systematic review and movement therapy in chronic stroke patients. J Med Assoc Thai. 2004;87:1482–1490. meta-analysis of the evidence. Arch Phys Med Rehabil. 2018;99:367–382. 21. Tariah HA, Almalty A-M, Sbeih Z, Al-Oraibi S. Constraint induced movement therapy for stroke survivors in Jordon: a home-based model. Int J Ther Rehabil. 5. Kollen BJ, Lennon S, Lyons B, Wheatley-Smith L, Scheper M, Buurke JH, et al. The 2010;17:638–646. effectiveness of the Bobath concept in stroke rehabilitation what is the evidence? 22. Schuster-Amft C, Eng K, Suica Z, Thaler I, Signer S, Lehmann I, et al. Effect of a four- Stroke. 2009;40:e89–e97. week virtual reality-based training versus conventional therapy on upper limb motor function after stroke: A multicenter parallel group randomized trial. PloS 6. Luke C, Dodd KJ, Brock K. Outcomes of the Bobath concept on upper limb recovery one. 2018;13:e0204455. following stroke. Clin Rehabil. 2004;18:888–898. 23. Timmermans AAAP, Verbunt JAMDP, van Woerden RM, Moennekens MMD, Pernot DHMD, Seelen HAMP. Effect of mental practice on the improvement of 7. Díaz-Arribas MJ, Martín-Casas P, Cano-de-la-Cuerda R, Plaza-Manzano G. Effec- function and daily activity performance of the upper extremity in patients with tiveness of the Bobath concept in the treatment of stroke: a systematic review. subacute stroke: a randomized clinical trial. J Am Med Dir Assoc. 2013;14:204–212. Disabil Rehabil. 2020;42:1636–1649. 24. Moon JH, Park KY, Kim HJ, Na CH. The effects of task-oriented circuit training using rehabilitation tools on the upper-extremity functions and daily activities of pa- tients with acute stroke: a randomized controlled pilot trial. Osong Public Health Res Perspect. 2018;9:225–230. 25. Luo D, Wan X, Liu J, Tong T. Optimally estimating the sample mean from the sample size, median, mid-range and/or mid-quartile range. Stat Methods Med Res. 2018;27:1785–1805. 26. Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol. 2014;14:135. 27. Scrivener K, Dorsch S, McCluskey A, Schurr K, Graham PL, Cao Z, et al. Bobath therapy is inferior to task-specific training and not superior to other interventions in improving lower limb activities after stroke: a systematic review. J Physiother. 2020;66:225–235.
Journal of Physiotherapy 69 (2023) 4–5 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 Cardiac conditions Ana Paula Coelho Figueira Freire a, Mark R Elkins b,c a Department of Health Sciences, Central Washington University, Ellensburg, USA; b Editor, Journal of Physiotherapy; c Faculty of Medicine and Health, University of Sydney, Sydney, Australia This Editorial introduces another of Journal of Physiotherapy’s execution of aerobic and resistance exercise, which are commonly article collections.1,2 These are collections of papers on a particular performed on treadmills, cycle ergometers and weight machine topic, published in the Journal of Physiotherapy within the past decade equipment. Therefore, water-based exercises can be offered as an and compiled to: facilitate access to important recent findings on the alternative, where the effects of buoyancy reduce the weight-bearing topic; highlight trends in research designs, methods, populations and load on the lower limbs and spine.9 interventions; and suggest avenues for further research. The studies in this article collection relate to physiotherapy management of car- Scheer et al10 conducted a randomised controlled trial, published diac conditions. in 2021, investigating the effects of 12 weeks of water-based circuit training exercise on aerobic capacity, strength and body composition In 2011, there was a call for physiotherapists to become more of patients with stable coronary heart disease. The results were engaged in cardiology research.3 The Editorial highlighted issues compared with those of a group that performed aerobic and including the lack of research in this area being led by physiothera- strengthening exercises at a cardiac rehabilitation outpatient gym. pists, despite their extensive training in prevention and management Water-based exercises presented very similar improvements in of risk factors for coronary disease. Much important research into VO2peak compared with the traditional program. The mean differ- physiotherapy for people with heart conditions has been published ence between groups was 0.2 ml/kg/min (95% CI –1.5 to 1.9). Addi- since then. This article collection focuses on aspects of cardiac reha- tionally, both exercise modalities also showed similar gains in muscle bilitation from both centre-based and home-based perspectives, as strength and comparable reductions in total body fat.10 well as reviewing progress in perioperative care, complementing the Editorial on perioperative physiotherapy published in 2022.4 The Home-based cardiac rehabilitation important advances that have subsequently been published in Journal of Physiotherapy are summarised below and collated as free full-text Despite the well-known benefits of exercise-based cardiac reha- articles in the online article collection. Importantly, each paper has bilitation, these programs are not currently being used to their full clear implications for clinical physiotherapists, which are identifiable potential. The rates of patients undergoing rehabilitation vary from in the paper’s ‘What this study adds’ summary box. ,20% for heart failure to 40% in those with coronary heart dis- ease.11,12 Therefore, strategies are needed to reduce barriers and Centre-based cardiac rehabilitation widen access to cardiac rehabilitation programs, such as home-based exercises.13 Clinical guidelines on the management of the most prevalent heart conditions consistently endorse cardiac rehabilitation as an One of the first randomised controlled trials to compare a home- efficient and secure strategy, with exercise identified as a central based video telerehabilitation program with a conventional centre- element.5,6 Exercise training has a direct benefit on reducing risk based program for patients with chronic heart failure was carried factors for atherosclerotic disease and also improving heart and cor- out in 2017 by Hwang et al.7 After 12 weeks of intervention, no sig- onary vasculature.6 Therefore, it is crucial to advance physiothera- nificant between-group differences were observed in 6MWD (MD 15 pists’ knowledge on the effects of different modalities of exercise. The m, 95% CI –28 to 59). In addition, the home-based telerehabilitation first and second articles in this collection address investigations group had significantly higher attendance rates when compared with around the effects of different modalities of exercise-based programs the centre-based group, with a between-group mean difference of six for patients with coronary heart disease and heart failure. sessions (95% CI 2 to 9).7 In 2010, Hwang and colleagues7 performed a systematic review to Chien et al also reported improvements in 6MWD (MD 21 m, 95% evaluate whether resistance training, alone or as an adjunct to aer- CI 7 to 36) and quality of life by providing home-based exercises obic training, could improve cardiac function, exercise capacity and associated with a daily activity log and telephone monitoring every 1 quality of life in patients with chronic heart failure. Results from eight to 2 weeks when compared with a control group that did not trials involving 241 participants revealed that resistance training exercise.14 increased the 6-minute walk distance (6MWD) by an average of 52 m (95% CI 19 to 85) in comparison with no training. However, when Another study investigating the effects of home-based training resistance training was provided either alone or as an addition to was performed by Jones et al in 2010.15 In this randomised trial, pa- aerobic training, no other benefits were seen on cardiac function, tients with stage I or II hypertension performed slow deep breathing peak oxygen consumption or quality of life. at home, either unloaded or breathing against a load of 20 cmH2O provided by a threshold-loaded breathing device for 8 weeks. Both A common concern when addressing exercise therapy for patients training groups showed significant reductions in systolic and diastolic with cardiovascular diseases is the high prevalence of morbidities blood pressure. Additionally, adding an inspiratory load enhanced the such as overweight, obesity and arthritis.8 These factors can limit the decrease in systolic blood pressure, with a between-group difference of –5.3 mmHg (95% CI –1.0 to –9.6). https://doi.org/10.1016/j.jphys.2022.11.010 1836-9553/© 2022 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/).
Editorial 5 In the studies discussed above, the 6MWD was one of the most improvements in physical function as measured by the Short Physical investigated outcomes and is considered an independent predictor of Performance Battery (MD 0.4 points on a 0-to-12 scale, 95% CI –0.9 to mortality and cardiovascular hospitalisation in people with heart 1.6).20 Secondary outcomes including pain and kinesiophobia also did failure.16 Wegrzynowska-Teodorczyk performed a prospective not show important between-group differences. The study also observational study and showed that participants with a 6MWD showed that implementation of the modified sternal precautions did 468 m had a mortality hazard ratio of 3.22 (95% CI 1.17 to 8.86) at 1 not cause harm or adverse events. Consequently, the authors year and 2.18 (95% CI 1.18 to 4.03) at 3 years of follow-up.16 Therefore, concluded that until further research is performed, modified sternal it is important to highlight that the 6-minute walk test should be precautions might be an equally appropriate option.20 considered an important component of the complex evaluation of the cardiac patient due to its prognostic value. Future directions Cardiac surgery This article collection indicated several beneficial physiotherapy interventions at various phases of cardiac surgery and rehabilitation. Preoperative and postoperative cardiac rehabilitation have Although most studies presented a variety or combination of strate- become an essential part of care for patients undergoing cardiac gies that can be provided, exercise training can be identified as the surgery procedures. Such programs can reduce mortality, improve cornerstone of the interventions for heart conditions. While the role physical performance and increase quality of life.6,17 Physiotherapists of physiotherapy interventions for cardiac populations is well estab- have an important role in assessing and providing different elements lished, several studies have highlighted the need for further in- of these programs, including education, inspiratory muscle training vestigations in this area. These main aspects for further research and exercise training.4 include: addressing the effects of rehabilitation programs on sub- groups of patients with heart failure, especially females and more Preoperative care was investigated by Snowdon and colleagues in severe cases of the disease; determining the applicability of tele- a systematic review published in 2014.18 The review primarily aimed rehabilitation in rural and remote locations with variable internet to analyse whether preoperative intervention in patients undergoing service; and establishing the effectiveness of preoperative exercise cardiac surgery could reduce pulmonary complications and shorten training on cost-effectiveness, pulmonary complications, length of length of intensive care unit or hospital stay.18 Their review included hospitalisation and time to extubation after cardiac surgery. a meta-analysis of data from six trials involving 661 participants; it indicated a worthwhile reduction in the relative risk of developing Ethics approval: Not applicable. postoperative pulmonary complications with preoperative in- Competing interest: Nil. terventions (RR 0.39, 95% CI 0.23 to 0.66). Additionally, preoperative Source(s) of support: Nil. intervention reduced the time to extubation by a pooled mean dif- Acknowledgements: Nil. ference of 0.14 days (95% CI 0.01 to 0.26). The effect on length of Provenance: Invited. Not peer reviewed. intensive care unit stay was unclear (MD –0.15 days, 95% CI –0.37 to Correspondence: Mark Elkins, Centre for Education & Workforce 0.08), as was the effect on length of hospital stay (MD –0.55 days, 95% Development, Sydney Local Health District, Sydney, Australia. Email: CI –1.32 to 0.23), except among older participants aged . 63 years [email protected] (MD –1.32 days, 95% CI –2.36 to –0.28).18 Finally, when evidence for specific preoperative therapies was taken into account, inspiratory References muscle training promoted significant reductions in postoperative pulmonary complications (RR 0.42, 95% 0.21 to 0.82) and reduced 1. Dale MT, et al. J Physiother. 2021;67:84–86. hospital stay (MD –2.1 days, 95% CI –3.41 to –0.76). 2. Reubenson A, et al. J Physiother. 2022;68:153–155. 3. Redfern J, et al. J Physiother. 2011;57:209–211. Significant morbidity and mortality are linked to major surgical 4. Boujibar F, et al. J Physiother. 2022;68:218–219. procedures, including cardiac surgery. One common surgical pro- 5. Long L, et al. Cochrane Database Syst Rev. 2019;1:CD003331. cedure that physiotherapists often manage in postoperative care is 6. Dibben G, et al. Cochrane Database Syst Rev. 2021;11:CD001800. median sternotomy.19 A common concern among physiotherapists 7. Hwang R, et al. J Physiother. 2017;63:101–107. and other healthcare providers is whether or not to restrict upper 8. Sanderson BK, et al. J Cardiopulm Rehabil. 2003;23:281–289. limb and trunk movements after median sternotomy, due to the 9. Becker BE, et al. PM&R. 2009;1:859–872. conflicting evidence that is available.20 10. Scheer A, et al. J Physiother. 2021;67:284–290. 11. Golwala H, et al. Am Coll Cardiol. 2015;66:917–926. The SMART trial explored the effects of an intervention that 12. Kotseva K, et al. Eur J Prev Cardiol. 2018;25:1242–1251. included less restrictive use of the upper limbs and trunk for patients 13. Neubeck L, et al. Eur J Prev Cardiol. 2012;19:494–503. who had undergone cardiac surgery via median sternotomy. The 14. Chien CL, et al. J Physiother. 2011;57:157–163. experimental group was encouraged to use their arms during trans- 15. Jones CU, et al. J Physiother. 2010;56:179–186. fers and other tasks, in addition to perform upper limb exercise three 16. Wegrzynowska-Teodorczyk K, et al. J Physiother. 2013;59:177–187. times daily within the limits of pain and discomfort. The control 17. Hulzebos EHJ, et al. Cochrane Database Syst Rev. 2012;11:CD010118. group was instructed with usual precautions, including to restrict use 18. Snowdon D, et al. J Physiother. 2014;60:66–77. of the upper limbs for 4 to 6 weeks after surgery.20 After 12 weeks of 19. Mekontso Dessap A, et al. Clin Microbiol Infect. 2011;17:292–299. follow-up, there were negligible between-group differences in the 20. Katijjahbe MA, et al. J Physiother. 2018;64:97–106.
Journal of Physiotherapy 69 (2023) 57 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Clinimetrics: The Lawton-Brody Instrumental Activities of Daily Living Scale Summary The Lawton-Brody Instrumental Activities of Daily Living Scale The Lawton IADL may be scored in a few ways, such as dichoto- (Lawton IADL) was invented as a measurement tool to evaluate the mous (0 to 8 scale), trichotomous (0 to 16 scale)8 and polytomous summation (8 to 31 scale);4 however, the most widely used is the level of functional independence in the general elderly population. The Lawton IADL is widely used to monitor disease progression, 0 to 8 scale. The Lawton IADL (0 to 8 scale) has high inter-rater evaluate treatment effectiveness and plan care services in clinical and reliability among health professionals, with intraclass correlation community settings.1 coefficients (ICCs) of 0.85,1 a moderate test-retest probability with ICCs of 0.755 and good internal consistency with Cronbach’s The Lawton IADL scale comprises eight items: using a telephone, alpha of 0.9.4 shopping, cooking, housework, washing clothes, accessing trans- portation, self-managing medication and finances.1 The Lawton IADL The minimal detectable change for the Lawton IADL (0 to 8 scale) is a self-report or informant-based questionnaire requiring the assessors to interview the subjects or the primary carers.1–3 is 1.5 points, meaning the smallest observable difference attributive The most common scoring system in the Lawton IADL is the 0 to to a ‘real’ change in performance beyond a measurement error is 8 scale,4 where 0 is fully dependent and 8 is independent in these 2 points.5 The evidence of responsiveness in the English version of activities. For each item, a score of 1 suggests a higher level of the Lawton IADL of 0 to 8 scoring is scarce. A validation study on the functional dependence, whereas a score of 0 indicates a more severe form of functional impairment.1,4,5 Spanish version reports a moderate-to-large value of responsiveness, The Lawton IADL can be completed by clinicians and researchers with a standardised response mean of 0.38 to 0.84 in elderly patients with adequate knowledge and training, which often involves theory- who sustained hip or wrist fractures from falls.9 based learning on standardised administration and the scoring Two studies5,7 in different population groups suggest a ceiling method and several practice runs on real patients with a highly skilled occupational therapist.6 The test takes approximately 10 to 15 effect of the Lawton IADL, where a significant proportion of people minutes to complete.5 The scale is readily available online free of (approximately 28% and 31%) achieved the highest score. charge. The Lawton IADL scale (https://www.alz.org/careplanning/ downloads/lawton-iadl.pdf) has been translated and validated in In terms of validity, Lawton and Brody did not explain the selec- different languages, including: Japanese, Chinese, Spanish, Persian, tion method and rationale behind each item in the Lawton IADL scale Sinhala, Iranian and Turkish. in the original paper published in 1969; hence, content validity is Reliability and validity of the Lawton IADL have been established difficult to assess.3 For construct validity, the Lawton IADL demon- in different patient populations ranging from community-dwelling strates a moderate positive correlation with the Physical elderly,1,4 memory clinic clients,2,7 psychiatric patients5 and rural elder communities. self-maintenance scale and between the Lawton IADL and the Mental Status Questionnaire (p , 0.01, Physical self-maintenance scale Pearson’s r = 0.61, Mental Status Questionnaire Pearson’s r = 0.48).1 As there is no ‘gold standard’ measurement tool for IADL, criterion validity is inapplicable to the Lawton IADL.3 Commentary Lingjia Weia and Carol Hodgsona,b aAustralian and New Zealand Intensive Care Research Centre, School of The Lawton IADL is a widely used measurement tool for assessing functional ability in various populations and clinical settings, Public Health & Preventive Medicine, Monash University including inpatient, outpatient and community care.1,4–6 bAlfred Intensive Care Unit, Victoria, Australia The Lawton IADL is a valid, reliable1,5 and efficient5 tool, and References minimal training is required.5 There is no consensus scoring system, although the most common method is the 0 to 8 scale.4 The clini- 1. Lawton MP, Brody EM. Gerontologist. 1969;9:179–186. metric properties of the method may vary depending on the scoring 2. Tabert MH, et al. Neurology. 2002;58:758–764. system used while administering the test. 3. Sikkes SA, et al. J Neurol Neurosurg Psychiatry. 2009;80:7–12. 4. Vittengl JR, et al. Ageing & Mental Health. 2006;10:40–47. Several studies have reported a high ceiling effect,5,6 which in- 5. Huang SL, et al. Am J Occup Ther. 2018;72:1–7. dicates that the Lawton IADL has limited utility in tracking function in 6. Edwards MJ, et al. Occup Ther J Res. 1995;15:103–110. more independent patients. Future studies are needed to explore how 7. Hancock P, Larner AJ. Dement Geriatr Cogn Discord. 2007;23:133–139. to discriminate functional ability among patients with the highest 8. Ng TP, et al. J Gerontol A Biol Sci Med Sci. 2006;61:726–735. score on the Lawton IADL scale. 9. Vergara I, et al. Health & Qual Life Outcomes. 2012;10:130. Provenance: Invited. Not peer reviewed. https://doi.org/10.1016/j.jphys.2022.06.007 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2023) 58–59 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Clinimetrics: The quality adjusted life year The quality adjusted life year (QALY) is an internationally recognised to a preference-based QOL questionnaire such as the EQ-5D and the standard metric that integrates quantity of life and quality of life (QOL) into a single index.1 Developed in the 1970s, QALYs are commonly used Health Utility Index. The preference-based QOL questionnaires differ as outcomes for health policy analysis and cost-effectiveness analysis (measured as cost/QALY gained), as they facilitate comparisons across in terms of the number and types of questions asked, and the disparate diseases and populations, allowing for decisions about allo- cation of limited healthcare resources.2,3 A lower cost per QALY gained amalgamation of responses by the individual subjectively on QOL provides evidence of value for money, providing a rationale for gives rise to the health state.4 The questionnaire responses are implementing an intervention. expressed as a utility score using a value set – a set of weights allo- The QALY is derived from multiplication of the utility score with cated for the health valuation from a relevant reference group.4–6 The an individual’s length of life, which can be determined by the area under the graph. Figure 1 graphically demonstrates QALYs generated QALY incorporates the assumption that an individual’s health status with and without an intervention, with the difference between the changes over time, with each status having a value.7 area under each curve representing the QALYs gained from the intervention. A utility score for health states is defined as a preference Whilst surveys such as the EQ-5D were found to provide practi- weight – where each preference is related to a particular health outcome – and health states with higher weights are considered cally significant results and previously validated for use, there is a more desirable. The utility is measured along an interval scale of 0.0 dearth of scientific literature investigating the measurement prop- (death) to 1.0 (perfect health), to which any difference in utility is the erties of the QALY.8,9 Concurrently, many critics have posed meth- same, regardless of where the change occurred on the scale.4 Nega- tive utility scores, signifying outcomes worse than death, are also odological and ethical concerns surrounding QALYs, including possible.4 The utility score can be generated from direct methods, such as the time trade-off or standard gamble, or indirect methods, discrepancies in validity and reliability, and reduction in freedom of which are more commonly used. Indirect methods involve responses choice stemming from QALYs restricting allocation of innovative treatments.10–12 However, the urgency to resolve these issues is nascent, given the extensive applicability and interest of QALYs as a key outcome to assist in healthcare resource allocation.13 It is pertinent to note that despite the imperfections, QALYs are practically useful in the bigger picture of guiding decision-makers in decisions about the value for money of a particular healthcare intervention, aligning with patients’ QOL.14 Therefore, to enhance this construct for use in future economic Figure 1. Quality adjusted life years (QALYs) are the area under the graph. Patient A (without the intervention) lives for 1 year with 0.9 utility (1 year x 0.9 utility score = 0.9 QALYs) followed by 0.4 utility for another year (1 year x 0.4 utility score = 0.4 QALYs), generating 1.3 (undiscounted) QALYs (0.9 QALYs 1 0.4 QALYs). Patient B, who receives the inter- vention, lives longer with a higher QOL. They first live with 0.9 utility for 2 years (2 years x 0.9 utility score = 1.8 QALYs), followed by 2 years with 0.7 utility (2 years x 0.7 utility = 1.4 QALYs). Therefore, 3.2 (undiscounted) QALYs are generated for an individual receiving intervention (1.8 QALYs 1 1.4 QALYs). Hence, the QALYs gained with intervention is 1.9 QALYs (3.2 QALYs – 1.3 QALYs). https://doi.org/10.1016/j.jphys.2022.06.008 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/ 4.0/).
Appraisal 59 evaluations it is suggested that limitations surrounding QALYs should Sheraya De Silva and Alisa M Higgins be incorporated in current research.15 Australian and New Zealand Intensive Care Research Centre, School of Public Health and Preventive Medicine, Faculty of Medicine, Nursing and The QALY framework aids physiotherapists in effectively imple- menting interventions that improve outcomes for patients, by incorpo- Health Sciences, Monash University, Australia rating the patient’s quantity and quality of life. What uniquely sets QALYs aside from other measures of outcome is the capability to compare a References variety of interventions across different diseases and conditions.8 1. Prieto L, Sacristán JA. Health Qual Life Outcomes. 2003;1:1–8. It is pertinent to acknowledge the limitations and concerns sur- 2. Fanshel S, Bush JW. Oper Res. 1970;18:1021–1066. rounding QALYs as well as the lack of publications exploring its 3. McDowell I. Measuring health: a guide to rating scales and questionnaires. USA: measurement properties. There is developing urgency for addressing these issues, as the QALY is currently the key measure of health Oxford University Press; 2006. outcome in economic evaluation widely used worldwide. There is 4. Whitehead SJ, Ali S. Br Med Bull. 2010;96:5–21. also increasing scope to formulate new preference-based value-sets 5. Tolley K. London: Hayward Medical Communications; 2009. specifically for the paediatric population, as a QALY gained may not be 6. EuroQoL. https://euroqol.org/publications/key-euroqol-references/value-sets/. valued equally across children, adolescents and adults.16 It is crucial that QALYs are incorporated in future research, to understand the Accessed May 2, 2022. limitations, identify the measurement properties and propose 7. Araújo CDM, et al. Acta Ortop Bras. 2014;22:102–105. constructive solutions. Therefore, QALYs will be amplified for use in 8. Oliveira JS, Hayes A. J Physiother. 2020;66:133. allocating effective physiotherapy interventions in the best interests 9. Wilson R, et al. Health Qual Life Outcomes. 2014;12:1–7. of the patient (based on their QOL).15 10. McGregor M. Cmaj. 2003;168:433–434. 11. Beresniak A, et al. Pharmacoeconomics. 2015;33:61–69. Provenance: Invited. Not peer reviewed. 12. Pettitt D, et al. J Stem Cell Res Ther. 2016;6. 13. Sanghera S, Coast J. Value Health. 2020;23:343–350. 14. Neumann PJ, Cohen JT. JAMA. 2018;319:2473–2474. 15. Lipscomb J, et al. Value Health. 2009;12:S18–S26. 16. Petrou S. Cost Eff Resour Alloc. 2022;20:1–5.
Journal of Physiotherapy 69 (2023) 55 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: An educational program increases habitual physical activity and self-efficacy for physical activity in people with coronary artery disease Synopsis Summary of: Pitta NC, Furuya RK, Freitas ND, Dessotte CAM, Dantas RAS, Cio outcome was the Baecke Habitual Physical Activity Questionnaire, which is MA, et al. Effect of an educational program on physical activity in individuals scored from 3 (sedentary) to 15 (very active). Secondary outcome measures undergoing their first percutaneous coronary intervention: a randomized were the Self-efficacy Scale for Physical Activity (0 low to 10 high self-efficacy) clinical trial. Braz J Phys Ther. 2022;26:100443. and the Hospital Anxiety and Depression Scale (0 low to 21 high anxiety/ depression). Results: A total of 109 participants completed the study. At 6 Question: What are the effects of an educational program emphasising physical months after hospital discharge, the change in the Baecke questionnaire was activity with telephone follow-up compared with routine care in people un- 0.89 points higher in the experimental group (95% CI 0.32 to 1.46). Self-efficacy dergoing their first percutaneous coronary intervention? Design: Randomised improved by 2.30 points more in the experimental group than the control controlled trial with concealed allocation and blinded assessment of some group (95% CI 1.12 to 3.49). In the experimental group, anxiety and depression outcomes. Setting: A public teaching hospital in Brazil. Participants: Adults were similar to or slightly better than in the control group: MD –0.71 (95% CI undergoing their first percutaneous coronary intervention. Inability to exercise, –1.89 to 0.47) for anxiety and MD –0.78 (95% CI –2.14 to 0.60) for depression. use of supplemental oxygen and participation in cardiac rehabilitation were Conclusion: The educational program increased habitual physical activity and exclusion criteria. Randomisation of 125 participants allocated 63 to the self-efficacy for physical activity in people with coronary artery disease. experimental group and 62 to a control group. Interventions: Both groups received verbal information about the percutaneous coronary procedure, Provenance: Invited. Not peer reviewed. including hospitalisation routine, recovery after the procedure, prescribed medication and date for outpatient return. In addition, the experimental group Mark Elkins received an educational cardiac rehabilitation program based on social cogni- Centre for Education & Workforce Development, tive theory. The content included: patient assessment, nutritional guidelines, management of risk factors, smoking cessation, psychosocial support and the Sydney Local Health District, Australia importance of physical activity. The educational content was provided in oral and written forms. Three monthly telephone calls reinforced the information https://doi.org/10.1016/j.jphys.2022.12.004 and sought to persuade the participant to comply with the recommended physical activity and lifestyle changes. Outcome measures: The primary Commentary self-efficacy mean that clinicians should consider administering an educational program to this clinical population. Despite the well-established benefits and cost-effectiveness of cardiac rehabilitation, health inequities in low- and middle-income countries can Further investigations with larger samples using more reliable measure- hamper attendance. Therefore, studies investigating approaches to facilitate ments of physical activity levels are still needed, but educational programs access to cardiac rehabilitation and promote increases in physical activity levels using accessible, low-technology interventions are probably effective in are essential. In Brazil, where the study was conducted, 500,000 more places in increasing physical activity levels among patients after percutaneous coronary cardiac rehabilitation programs are still needed nationally.1 There can also be intervention. variable internet service and poor access to smartphones; therefore, low technology tools such as landline telephones and messaging via mobile Provenance: Invited. Not peer reviewed. phones can be an attractive alternative to provide educational programs for patients transitioning from hospital to home care. Ana Paula Coelho Figueira Freire Health Sciences Department, Central Washington University, The authors have shown that e-health can assist health professionals aiming to make cardiac rehabilitation more accessible and to encourage in- Ellensburg, USA creases in physical activity levels. Nevertheless, it is important to consider whether the effects observed in this study were clinically worthwhile. The https://doi.org/10.1016/j.jphys.2022.12.005 authors nominated a between-group difference of 0.5 points on the Baecke questionnaire as a worthwhile effect in their sample size calculation. The mean Reference between-group difference exceeded this threshold, but the 95% CI spanned it. The mean effect on self-efficacy (2.30) and the 95% CI (1.12 to 3.49) both 1. Britto RR, et al. Braz J Phys Ther. 2020;24:167–176. seemed worthwhile, given that the scale was 10 points wide. The effects on anxiety and depression were small and uncertain. Overall, the potentially worthwhile effect on physical activity combined with the worthwhile effect on 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2023) 55 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: An educational program increases habitual physical activity and self-efficacy for physical activity in people with coronary artery disease Synopsis Summary of: Pitta NC, Furuya RK, Freitas ND, Dessotte CAM, Dantas RAS, Cio outcome was the Baecke Habitual Physical Activity Questionnaire, which is MA, et al. Effect of an educational program on physical activity in individuals scored from 3 (sedentary) to 15 (very active). Secondary outcome measures undergoing their first percutaneous coronary intervention: a randomized were the Self-efficacy Scale for Physical Activity (0 low to 10 high self-efficacy) clinical trial. Braz J Phys Ther. 2022;26:100443. and the Hospital Anxiety and Depression Scale (0 low to 21 high anxiety/ depression). Results: A total of 109 participants completed the study. At 6 Question: What are the effects of an educational program emphasising physical months after hospital discharge, the change in the Baecke questionnaire was activity with telephone follow-up compared with routine care in people un- 0.89 points higher in the experimental group (95% CI 0.32 to 1.46). Self-efficacy dergoing their first percutaneous coronary intervention? Design: Randomised improved by 2.30 points more in the experimental group than the control controlled trial with concealed allocation and blinded assessment of some group (95% CI 1.12 to 3.49). In the experimental group, anxiety and depression outcomes. Setting: A public teaching hospital in Brazil. Participants: Adults were similar to or slightly better than in the control group: MD –0.71 (95% CI undergoing their first percutaneous coronary intervention. Inability to exercise, –1.89 to 0.47) for anxiety and MD –0.78 (95% CI –2.14 to 0.60) for depression. use of supplemental oxygen and participation in cardiac rehabilitation were Conclusion: The educational program increased habitual physical activity and exclusion criteria. Randomisation of 125 participants allocated 63 to the self-efficacy for physical activity in people with coronary artery disease. experimental group and 62 to a control group. Interventions: Both groups received verbal information about the percutaneous coronary procedure, Provenance: Invited. Not peer reviewed. including hospitalisation routine, recovery after the procedure, prescribed medication and date for outpatient return. In addition, the experimental group Mark Elkins received an educational cardiac rehabilitation program based on social cogni- Centre for Education & Workforce Development, tive theory. The content included: patient assessment, nutritional guidelines, management of risk factors, smoking cessation, psychosocial support and the Sydney Local Health District, Australia importance of physical activity. The educational content was provided in oral and written forms. Three monthly telephone calls reinforced the information https://doi.org/10.1016/j.jphys.2022.12.004 and sought to persuade the participant to comply with the recommended physical activity and lifestyle changes. Outcome measures: The primary Commentary self-efficacy mean that clinicians should consider administering an educational program to this clinical population. Despite the well-established benefits and cost-effectiveness of cardiac rehabilitation, health inequities in low- and middle-income countries can Further investigations with larger samples using more reliable measure- hamper attendance. Therefore, studies investigating approaches to facilitate ments of physical activity levels are still needed, but educational programs access to cardiac rehabilitation and promote increases in physical activity levels using accessible, low-technology interventions are probably effective in are essential. In Brazil, where the study was conducted, 500,000 more places in increasing physical activity levels among patients after percutaneous coronary cardiac rehabilitation programs are still needed nationally.1 There can also be intervention. variable internet service and poor access to smartphones; therefore, low technology tools such as landline telephones and messaging via mobile Provenance: Invited. Not peer reviewed. phones can be an attractive alternative to provide educational programs for patients transitioning from hospital to home care. Ana Paula Coelho Figueira Freire Health Sciences Department, Central Washington University, The authors have shown that e-health can assist health professionals aiming to make cardiac rehabilitation more accessible and to encourage in- Ellensburg, USA creases in physical activity levels. Nevertheless, it is important to consider whether the effects observed in this study were clinically worthwhile. The https://doi.org/10.1016/j.jphys.2022.12.005 authors nominated a between-group difference of 0.5 points on the Baecke questionnaire as a worthwhile effect in their sample size calculation. The mean Reference between-group difference exceeded this threshold, but the 95% CI spanned it. The mean effect on self-efficacy (2.30) and the 95% CI (1.12 to 3.49) both 1. Britto RR, et al. Braz J Phys Ther. 2020;24:167–176. seemed worthwhile, given that the scale was 10 points wide. The effects on anxiety and depression were small and uncertain. Overall, the potentially worthwhile effect on physical activity combined with the worthwhile effect on 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2023) 56 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: In adults receiving intensive care, increased early active mobilisation did not improve clinical outcomes but increased adverse events Synopsis Summary of: TEAM Study Investigators and the ANZICS Clinical Trials and out of hospital at day 180. Adverse events were recorded. Group, Hodgson CL, Bailey M, Bellomo R, Brickell K, Broadley T, Buhr Results: A total of 733 participants completed the study (369 in the H, et al. Early Active Mobilization during Mechanical Ventilation in experimental group and 364 in the control group). Negligible the ICU. N Engl J Med. 2022; Oct 26. https://doi.org/10.1056/NEJMoa22 between-group difference in days alive and out of hospital at day 09083 180 was observed (absolute difference 22 days, 95% CI 210 to 6). Patients with at least one adverse event potentially due to mobi- Question: In adults who are mechanically ventilated in the intensive lisation were more common in the early mobilisation group (OR care unit (ICU), does increased early active mobilisation compared 2.55, 95% CI 1.33 to 4.89). Cardiac arrhythmia and desaturation with usual care improve clinical outcomes? Design: Randomised were especially more common in the early mobilisation group. controlled trial with concealed allocation and blinded assessors. Conclusion: In adults who are mechanically ventilated in the ICU, Setting: Forty-nine ICUs in six countries. Participants: Adults increased early active mobilisation does not affect days alive and out intubated and expected to remain invasively mechanically venti- of hospital, and is associated with more adverse events than usual lated for at least 1 day, who were sufficiently stable to make mobilisation. mobilisation potentially possible. Randomisation of 750 participants allocated 372 to the experimental group and 378 to the control Provenance: Invited. Not peer reviewed. group. Interventions: The control group received usual mobilisation in the ICU. The experimental group received one or more individ- Vinicius Cavalheri ually tailored sessions of daily physiotherapy. The sessions were Curtin School of Allied Health, Curtin University, Australia conducted at the highest possible level of mobilisation deemed safe, for the longest time possible (dropping to lower levels if fatigued). https://doi.org/10.1016/j.jphys.2022.12.007 The intervention window was 28 days from randomisation in both groups. Outcome measures: The primary outcome was days alive Commentary Patients surviving an intensive care unit (ICU) stay are at risk of (maximum 28 days). The trial raises concern for more adverse events developing post-ICU syndrome.1 In 2009, a 104-patient randomised with early mobilisation; however, these events were rare, as docu- trial of occupational and physiotherapy activities that started within 1.5 mented in previous literature.4 Adverse events must be interpreted days after intubation (usual care started at 7.4 days) highlighted the within the context of the intervention, potential ascertainment bias importance of early rehabilitation to improve functional independence due to the unblinded intervention, information about exposure to at hospital discharge.2 mobility activities (number of days, types of activities) and consequences of these events. The TEAM trial stimulates the field to The TEAM trial is another important milestone in this field and is continue to rigorously evaluate different timing, frequency, intensity, the largest trial to date (750 participants). It investigated a duration and types of interventions, with consistent measurement physiotherapist-led 7 day/week intervention. There was blinding of and reporting of intervention and usual care, protocol fidelity, and primary outcome assessment (but not participants or personnel); adverse events in both experimental and control groups. participants were randomised within w2.5 days of ICU admission; Providence: Invited. Not peer reviewed. physiotherapist assessments in the experimental and control groups occurred on 94% and 81% of all days, respectively; and the duration of Michelle E Kho active mobilisation was 21 6 15 and 9 6 9 minutes/day, respectively School of Rehabilitation Science, McMaster University, Hamilton, Canada (12-minute difference). https://doi.org/10.1016/j.jphys.2022.12.006 TEAM tested a high-intensity approach to mobilisation from the start of each session; the highest level of mobility for the longest References duration possible until fatigue was targeted. Mobilisation was defined as activities including sitting at the edge of the bed or higher. In 1. Rousseau A-F, et al. Crit Care. 2021;25:108. contrast, previous studies have reported mobilisation as approaches 2. Schweickert WD, et al. Lancet. 2009;373:1874–1882. typically starting with passive movements, progressing to more 3. Reid JC, et al. J Intensive Care. 2018;6:80. intensive activities.3 No between-group difference in the primary 4. Nydahl P, et al. Ann Am Thorac Soc. 2017;14:766–777. outcome (days alive and out of hospital at day 180) was demonstrated. However, the intervention occurred only during the ICU stay 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/ 4.0/).
Journal of Physiotherapy 69 (2023) 56 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: In adults receiving intensive care, increased early active mobilisation did not improve clinical outcomes but increased adverse events Synopsis Summary of: TEAM Study Investigators and the ANZICS Clinical Trials and out of hospital at day 180. Adverse events were recorded. Group, Hodgson CL, Bailey M, Bellomo R, Brickell K, Broadley T, Buhr Results: A total of 733 participants completed the study (369 in the H, et al. Early Active Mobilization during Mechanical Ventilation in experimental group and 364 in the control group). Negligible the ICU. N Engl J Med. 2022; Oct 26. https://doi.org/10.1056/NEJMoa22 between-group difference in days alive and out of hospital at day 09083 180 was observed (absolute difference 22 days, 95% CI 210 to 6). Patients with at least one adverse event potentially due to mobi- Question: In adults who are mechanically ventilated in the intensive lisation were more common in the early mobilisation group (OR care unit (ICU), does increased early active mobilisation compared 2.55, 95% CI 1.33 to 4.89). Cardiac arrhythmia and desaturation with usual care improve clinical outcomes? Design: Randomised were especially more common in the early mobilisation group. controlled trial with concealed allocation and blinded assessors. Conclusion: In adults who are mechanically ventilated in the ICU, Setting: Forty-nine ICUs in six countries. Participants: Adults increased early active mobilisation does not affect days alive and out intubated and expected to remain invasively mechanically venti- of hospital, and is associated with more adverse events than usual lated for at least 1 day, who were sufficiently stable to make mobilisation. mobilisation potentially possible. Randomisation of 750 participants allocated 372 to the experimental group and 378 to the control Provenance: Invited. Not peer reviewed. group. Interventions: The control group received usual mobilisation in the ICU. The experimental group received one or more individ- Vinicius Cavalheri ually tailored sessions of daily physiotherapy. The sessions were Curtin School of Allied Health, Curtin University, Australia conducted at the highest possible level of mobilisation deemed safe, for the longest time possible (dropping to lower levels if fatigued). https://doi.org/10.1016/j.jphys.2022.12.007 The intervention window was 28 days from randomisation in both groups. Outcome measures: The primary outcome was days alive Commentary Patients surviving an intensive care unit (ICU) stay are at risk of (maximum 28 days). The trial raises concern for more adverse events developing post-ICU syndrome.1 In 2009, a 104-patient randomised with early mobilisation; however, these events were rare, as docu- trial of occupational and physiotherapy activities that started within 1.5 mented in previous literature.4 Adverse events must be interpreted days after intubation (usual care started at 7.4 days) highlighted the within the context of the intervention, potential ascertainment bias importance of early rehabilitation to improve functional independence due to the unblinded intervention, information about exposure to at hospital discharge.2 mobility activities (number of days, types of activities) and consequences of these events. The TEAM trial stimulates the field to The TEAM trial is another important milestone in this field and is continue to rigorously evaluate different timing, frequency, intensity, the largest trial to date (750 participants). It investigated a duration and types of interventions, with consistent measurement physiotherapist-led 7 day/week intervention. There was blinding of and reporting of intervention and usual care, protocol fidelity, and primary outcome assessment (but not participants or personnel); adverse events in both experimental and control groups. participants were randomised within w2.5 days of ICU admission; Providence: Invited. Not peer reviewed. physiotherapist assessments in the experimental and control groups occurred on 94% and 81% of all days, respectively; and the duration of Michelle E Kho active mobilisation was 21 6 15 and 9 6 9 minutes/day, respectively School of Rehabilitation Science, McMaster University, Hamilton, Canada (12-minute difference). https://doi.org/10.1016/j.jphys.2022.12.006 TEAM tested a high-intensity approach to mobilisation from the start of each session; the highest level of mobility for the longest References duration possible until fatigue was targeted. Mobilisation was defined as activities including sitting at the edge of the bed or higher. In 1. Rousseau A-F, et al. Crit Care. 2021;25:108. contrast, previous studies have reported mobilisation as approaches 2. Schweickert WD, et al. Lancet. 2009;373:1874–1882. typically starting with passive movements, progressing to more 3. Reid JC, et al. J Intensive Care. 2018;6:80. intensive activities.3 No between-group difference in the primary 4. Nydahl P, et al. Ann Am Thorac Soc. 2017;14:766–777. outcome (days alive and out of hospital at day 180) was demonstrated. However, the intervention occurred only during the ICU stay 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/ 4.0/).
Journal of Physiotherapy 69 (2023) 54 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Modified constraint-induced movement therapy is probably the most effective non-conventional upper limb intervention post-stroke Synopsis Summary of: Saikaley M, Pauli G, Sun H, Serra JR, Iruthayarajah J, these included modified constraint-induced movement therapy (MD Teasell R. Network meta-analysis of non-conventional therapies for 6.7 points, 95% CI 4.3 to 8.9), high-frequency repetitive transcranial improving upper limb motor impairment poststroke. Stroke. 2022; magnetic stimulation (MD 5.4 points, 95% CI 1.9 to 8.9), mental imagery Oct 14. https://doi.org/10.1161/STROKEAHA.122.040687. Online ahead (MD 5.4, 95% CI 1.8 to 8.9), bilateral arm training (MD 5.2 points, 95% CI of print. 2.3 to 8.1) and intermittent theta burst stimulation (MD 5.1 points, 95% CI 0.6 to 9.5), which were ranked as the top five most effective in- Objective: To compare the effectiveness of non-conventional in- terventions according to the surface under the cumulative ranking terventions with usual care for improving upper limb impairment after curve. Comparisons between non-conventional interventions found stroke. Data sources: PubMed, Scopus, Web of Science, Embase and that modified constraint-induced movement therapy was superior to CINAHL were searched from inception to 1st April 2021. The Evidence eight other interventions. Conclusion: Compared with usual care Based Review of Stroke Rehabilitation was searched to identify any (conventional therapy), modified constraint-induced movement missed articles. Study selection: Randomised controlled trials and therapy had the highest probability of being the most effective non- randomised cross-over studies primarily involving adults with stroke, conventional intervention for improving upper limb motor impair- in which a non-conventional upper limb intervention was compared ment after stroke. with conventional rehabilitation and used the upper limb Fugl-Meyer scale as the primary outcome. Studies without time-matched therapy Provenance: Invited. Not peer reviewed. dose control groups were excluded. Data extraction: Two independent reviewers extracted the data. Methodological quality was assessed Prudence Plummer using the PEDro scale. Data synthesis: Of 5,408 studies initially iden- Department of Physical Therapy, MGH Institute of Health Professions, tified by the search, 176 randomised controlled trials met the selection criteria and were included in the network meta-analysis. Methodo- USA logical quality ranged from 3 to 9 out of 10. Relative to conventional care, 11 of the 20 non-conventional interventions were more effective https://doi.org/10.1016/j.jphys.2022.12.002 for improving upper limb motor impairment on the Fugl-Meyer scale; Commentary induced movement therapy was highest ranked by the network meta-analysis approach. These findings can help the field narrow This network meta-analysis compared therapies that are not down candidate therapies that warrant further testing in appropri- traditionally delivered as part of usual care with dose-matched con- ately designed clinical trials.4 Ensuring that future trials follow best ventional therapies (usual care). Non-conventional therapies practice in systematic intervention development, address common included modified constraint-induced movement therapy, brain and populations and report experimental and control intervention electrical stimulation, and robotics. Conventional therapies included content in line with appropriate EQUATOR guidelines will enhance task-specific and Bobath training, and range of motion exercises. A confidence in the results of future network meta-analyses. robust approach was outlined to manage intervention heterogeneity. Inconsistent and inadequate reporting of usual care, and the het- Provenance: Invited. Not peer reviewed. erogeneity of ‘usual care’ interventions1 remain some of the biggest challenges that may impede the translation into clinical practice of Emily Dalton pooled analyses involving comparisons with usual care. Physiotherapy Department, University of Melbourne, Australia The inclusive eligibility criteria and use of a network meta- https://doi.org/10.1016/j.jphys.2022.12.003 analysis enabled a broad and novel comparison of intervention efficacy. Overall, the pooled participants had some upper limb References movement at baseline (ie, mild-moderate impairment).2 Therapies for people who have more severe impairment continue to elude the 1. Lohse K, et al. Arch Phys Med Rehabil. 2018;99:1424–1432. stroke recovery field. While time post-stroke and impairment 2. Bernhardt J, et al. Int J Stroke. 2017;12:444–450. severity were not important covariates in this review, there was 3. Dalton E, et al. Int J Stroke. 2022. limited variability across the included studies. The context that these 4. Bernhardt J, et al. Int J Stroke. 2019;14:792–802. knowledge units provide is important to guide the generalisability of trial findings in clinical practice.3 Eleven non-conventional therapies were found to perform significantly better than usual care (seven non-adjuvant behavioural and four adjuvant brain stimulation therapies). Modified constraint- 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2023) 54 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Modified constraint-induced movement therapy is probably the most effective non-conventional upper limb intervention post-stroke Synopsis Summary of: Saikaley M, Pauli G, Sun H, Serra JR, Iruthayarajah J, these included modified constraint-induced movement therapy (MD Teasell R. Network meta-analysis of non-conventional therapies for 6.7 points, 95% CI 4.3 to 8.9), high-frequency repetitive transcranial improving upper limb motor impairment poststroke. Stroke. 2022; magnetic stimulation (MD 5.4 points, 95% CI 1.9 to 8.9), mental imagery Oct 14. https://doi.org/10.1161/STROKEAHA.122.040687. Online ahead (MD 5.4, 95% CI 1.8 to 8.9), bilateral arm training (MD 5.2 points, 95% CI of print. 2.3 to 8.1) and intermittent theta burst stimulation (MD 5.1 points, 95% CI 0.6 to 9.5), which were ranked as the top five most effective in- Objective: To compare the effectiveness of non-conventional in- terventions according to the surface under the cumulative ranking terventions with usual care for improving upper limb impairment after curve. Comparisons between non-conventional interventions found stroke. Data sources: PubMed, Scopus, Web of Science, Embase and that modified constraint-induced movement therapy was superior to CINAHL were searched from inception to 1st April 2021. The Evidence eight other interventions. Conclusion: Compared with usual care Based Review of Stroke Rehabilitation was searched to identify any (conventional therapy), modified constraint-induced movement missed articles. Study selection: Randomised controlled trials and therapy had the highest probability of being the most effective non- randomised cross-over studies primarily involving adults with stroke, conventional intervention for improving upper limb motor impair- in which a non-conventional upper limb intervention was compared ment after stroke. with conventional rehabilitation and used the upper limb Fugl-Meyer scale as the primary outcome. Studies without time-matched therapy Provenance: Invited. Not peer reviewed. dose control groups were excluded. Data extraction: Two independent reviewers extracted the data. Methodological quality was assessed Prudence Plummer using the PEDro scale. Data synthesis: Of 5,408 studies initially iden- Department of Physical Therapy, MGH Institute of Health Professions, tified by the search, 176 randomised controlled trials met the selection criteria and were included in the network meta-analysis. Methodo- USA logical quality ranged from 3 to 9 out of 10. Relative to conventional care, 11 of the 20 non-conventional interventions were more effective https://doi.org/10.1016/j.jphys.2022.12.002 for improving upper limb motor impairment on the Fugl-Meyer scale; Commentary induced movement therapy was highest ranked by the network meta-analysis approach. These findings can help the field narrow This network meta-analysis compared therapies that are not down candidate therapies that warrant further testing in appropri- traditionally delivered as part of usual care with dose-matched con- ately designed clinical trials.4 Ensuring that future trials follow best ventional therapies (usual care). Non-conventional therapies practice in systematic intervention development, address common included modified constraint-induced movement therapy, brain and populations and report experimental and control intervention electrical stimulation, and robotics. Conventional therapies included content in line with appropriate EQUATOR guidelines will enhance task-specific and Bobath training, and range of motion exercises. A confidence in the results of future network meta-analyses. robust approach was outlined to manage intervention heterogeneity. Inconsistent and inadequate reporting of usual care, and the het- Provenance: Invited. Not peer reviewed. erogeneity of ‘usual care’ interventions1 remain some of the biggest challenges that may impede the translation into clinical practice of Emily Dalton pooled analyses involving comparisons with usual care. Physiotherapy Department, University of Melbourne, Australia The inclusive eligibility criteria and use of a network meta- https://doi.org/10.1016/j.jphys.2022.12.003 analysis enabled a broad and novel comparison of intervention efficacy. Overall, the pooled participants had some upper limb References movement at baseline (ie, mild-moderate impairment).2 Therapies for people who have more severe impairment continue to elude the 1. Lohse K, et al. Arch Phys Med Rehabil. 2018;99:1424–1432. stroke recovery field. While time post-stroke and impairment 2. Bernhardt J, et al. Int J Stroke. 2017;12:444–450. severity were not important covariates in this review, there was 3. Dalton E, et al. Int J Stroke. 2022. limited variability across the included studies. The context that these 4. Bernhardt J, et al. Int J Stroke. 2019;14:792–802. knowledge units provide is important to guide the generalisability of trial findings in clinical practice.3 Eleven non-conventional therapies were found to perform significantly better than usual care (seven non-adjuvant behavioural and four adjuvant brain stimulation therapies). Modified constraint- 1836-9553/© 2022 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2023) 53 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Respiratory waveform analysis guides settings for effective mechanical insufflation-exsufflation in patients with amyotrophic lateral sclerosis Synopsis Summary of: Sancho J, Ferrer S, Bures E, Fernanda-Presa L, Bañuls P, Cruz dysfunction was associated with upper airway collapse during exsufflation Gonzalez M, et al. Waveforms analysis in patients with amyotrophic lateral (OR 0.32, 95% CI 0.11 to 0.98). A Norris bulbar score . 27 was associated with sclerosis for enhanced efficacy of mechanically assisted coughing. Respir Care. no change in waveforms with adjustment of the MI-E settings (OR 0.87, 95% 2022;67:1226–1235. CI 0.79 to 0.96). From these results, an algorithm was developed: in those with Norris bulbar score , 27, change insufflation pressure and flow to Question: In patients with amyotrophic lateral sclerosis (ALS), how well do resolve obstruction and change exsufflation pressure to resolve upper airway waveform graphics generated by a mechanical insufflation-exsufflation (MI-E) collapse during expiration, followed by adjustment of insufflation and device aid in guiding settings to prevent obstruction during insufflation and exsufflation time. Applying this algorithm enabled 68 participants (99%) to collapse during exsufflation? Design: Prospective observational study. Setting: achieve a cough peak flow . 2.7 l/s. Conclusion: Analysis of waveforms A respiratory care unit in a Spanish tertiary hospital. Participants: Adult in- during MI-E in patients with ALS identified two phenotypes: obstruction patients with ALS. Home tracheostomy ventilation, dementia and contraindi- during insufflation (related to bulbar upper motor neuron dysfunction) and cations to MI-E were exclusion criteria. Consecutive recruitment for 2 years collapse during exsufflation (related to bulbar lower motor neuron enrolled 69 participants. Interventions: Participants underwent sessions of MI- dysfunction). An algorithm can guide pressure, flow and time settings for E with different insufflation-exsufflation pressures administered at each ses- effective MI-E in patients with ALS. sion: 6 20, 6 30, 6 40 and 6 50 cmH2O. Outcome measures: Flow-time and pressure-time waveforms were analysed to identify obstruction during insuf- Provenance: Invited. Not peer reviewed. flation and upper airway collapse during exsufflation, and predictors were sought. The MI-E pressures were then adjusted based on an algorithm linked to Alicia Spittle these predictors, followed by reassessment of whether MI-E was effective (ie, Department of Physiotherapy, University of Melbourne peak cough flow . 2.7 l/s). Results: All participants completed the study, among whom 51% demonstrated obstruction during insufflation and 26% https://doi.org/10.1016/j.jphys.2022.11.003 demonstrated upper airway collapse during exsufflation on waveform analysis. Bulbar upper motor neuron dysfunction was associated with obstruction dur- ing insufflation (OR 7.19, 95% CI 2.32 to 22.29). Bulbar lower motor neuron 1836-9553/© 2022 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/). Commentary In 2015, Australia had a burden of disease cost of w$1.94 billion for amyo- improve treatment efficacy. This work by Sancho et al is an important step in trophic lateral sclerosis (ALS).1 Respiratory compromise is the main contributor addressing current gaps in respiratory care. Future research should focus on the clinical effect of individualised MI-E settings in people with ALS. to morbidity and mortality in people with neuromuscular disease, emphasising the importance of optimising respiratory care.2 There is evidence that Provenance: Invited. Not peer reviewed. mechanical insufflation-exsufflation (MI-E) is an effective cough augmentation Elizabeth Lambrinos technique and its use is recommended in many guidelines as the treatment of Respiratory Support Service, Royal Prince Alfred Hospital, Sydney, choice for advanced neuromuscular weakness.3 In patients with ALS, particularly Australia when bulbar involvement is present, MI-E treatment failure due to upper airway collapse and/or laryngeal inspiratory closure has been demonstrated.4 Anderson https://doi.org/10.1016/j.jphys.2022.11.004 et al4 pioneered using transnasal fibreoptic laryngoscopy to visualise the larynx and titrate MI-E settings to mitigate these issues. However, this is an invasive References method and is not readily available to physiotherapists. 1. Deloitte Access Economics 2015. 2. Ambrosino N, et al. Eur Respir J. 2009;34:444–451. The paper by Sancho et al shows that a non-invasive method of assessment 3. Chatwin M, et al. Respir Med. 2018;136:98–110. 4. Andersen T, et al. Respir Care. 2021;66:1196–1213. (ie, analysis of flow and pressure waveforms) can be used to individualise MI-E 5. Nilsestuen JO, Holland VA. Respir Ther. 2020;15:32–40. settings in patients with ALS and thereby optimise treatment. The described 6. Lacombe M, et al. Arch Phys Med Rehabil. 2019;100:2346–2353. waveform characteristics that represent obstruction during insufflation and upper airway collapse during exsufflation correlate with previous research.5 The work by Sancho et al also supports that of Lacombe et al6 in highlighting that the use of cough peak flow assessment alone, without waveform analysis, may provide erroneously high measurements and incorrect assessment of cough efficacy. Waveforms can be readily obtained and interpreted in clinical practice to provide information about causes of MI-E failure, thereby assisting titration to 1836-9553/Crown Copyright © 2022 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2023) 53 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Critically appraised paper: Respiratory waveform analysis guides settings for effective mechanical insufflation-exsufflation in patients with amyotrophic lateral sclerosis Synopsis Summary of: Sancho J, Ferrer S, Bures E, Fernanda-Presa L, Bañuls P, Cruz dysfunction was associated with upper airway collapse during exsufflation Gonzalez M, et al. Waveforms analysis in patients with amyotrophic lateral (OR 0.32, 95% CI 0.11 to 0.98). A Norris bulbar score . 27 was associated with sclerosis for enhanced efficacy of mechanically assisted coughing. Respir Care. no change in waveforms with adjustment of the MI-E settings (OR 0.87, 95% 2022;67:1226–1235. CI 0.79 to 0.96). From these results, an algorithm was developed: in those with Norris bulbar score , 27, change insufflation pressure and flow to Question: In patients with amyotrophic lateral sclerosis (ALS), how well do resolve obstruction and change exsufflation pressure to resolve upper airway waveform graphics generated by a mechanical insufflation-exsufflation (MI-E) collapse during expiration, followed by adjustment of insufflation and device aid in guiding settings to prevent obstruction during insufflation and exsufflation time. Applying this algorithm enabled 68 participants (99%) to collapse during exsufflation? Design: Prospective observational study. Setting: achieve a cough peak flow . 2.7 l/s. Conclusion: Analysis of waveforms A respiratory care unit in a Spanish tertiary hospital. Participants: Adult in- during MI-E in patients with ALS identified two phenotypes: obstruction patients with ALS. Home tracheostomy ventilation, dementia and contraindi- during insufflation (related to bulbar upper motor neuron dysfunction) and cations to MI-E were exclusion criteria. Consecutive recruitment for 2 years collapse during exsufflation (related to bulbar lower motor neuron enrolled 69 participants. Interventions: Participants underwent sessions of MI- dysfunction). An algorithm can guide pressure, flow and time settings for E with different insufflation-exsufflation pressures administered at each ses- effective MI-E in patients with ALS. sion: 6 20, 6 30, 6 40 and 6 50 cmH2O. Outcome measures: Flow-time and pressure-time waveforms were analysed to identify obstruction during insuf- Provenance: Invited. Not peer reviewed. flation and upper airway collapse during exsufflation, and predictors were sought. The MI-E pressures were then adjusted based on an algorithm linked to Alicia Spittle these predictors, followed by reassessment of whether MI-E was effective (ie, Department of Physiotherapy, University of Melbourne peak cough flow . 2.7 l/s). Results: All participants completed the study, among whom 51% demonstrated obstruction during insufflation and 26% https://doi.org/10.1016/j.jphys.2022.11.003 demonstrated upper airway collapse during exsufflation on waveform analysis. Bulbar upper motor neuron dysfunction was associated with obstruction dur- ing insufflation (OR 7.19, 95% CI 2.32 to 22.29). Bulbar lower motor neuron 1836-9553/© 2022 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/). Commentary In 2015, Australia had a burden of disease cost of w$1.94 billion for amyo- improve treatment efficacy. This work by Sancho et al is an important step in trophic lateral sclerosis (ALS).1 Respiratory compromise is the main contributor addressing current gaps in respiratory care. Future research should focus on the clinical effect of individualised MI-E settings in people with ALS. to morbidity and mortality in people with neuromuscular disease, emphasising the importance of optimising respiratory care.2 There is evidence that Provenance: Invited. Not peer reviewed. mechanical insufflation-exsufflation (MI-E) is an effective cough augmentation Elizabeth Lambrinos technique and its use is recommended in many guidelines as the treatment of Respiratory Support Service, Royal Prince Alfred Hospital, Sydney, choice for advanced neuromuscular weakness.3 In patients with ALS, particularly Australia when bulbar involvement is present, MI-E treatment failure due to upper airway collapse and/or laryngeal inspiratory closure has been demonstrated.4 Anderson https://doi.org/10.1016/j.jphys.2022.11.004 et al4 pioneered using transnasal fibreoptic laryngoscopy to visualise the larynx and titrate MI-E settings to mitigate these issues. However, this is an invasive References method and is not readily available to physiotherapists. 1. Deloitte Access Economics 2015. 2. Ambrosino N, et al. Eur Respir J. 2009;34:444–451. The paper by Sancho et al shows that a non-invasive method of assessment 3. Chatwin M, et al. Respir Med. 2018;136:98–110. 4. Andersen T, et al. Respir Care. 2021;66:1196–1213. (ie, analysis of flow and pressure waveforms) can be used to individualise MI-E 5. Nilsestuen JO, Holland VA. Respir Ther. 2020;15:32–40. settings in patients with ALS and thereby optimise treatment. The described 6. Lacombe M, et al. Arch Phys Med Rehabil. 2019;100:2346–2353. waveform characteristics that represent obstruction during insufflation and upper airway collapse during exsufflation correlate with previous research.5 The work by Sancho et al also supports that of Lacombe et al6 in highlighting that the use of cough peak flow assessment alone, without waveform analysis, may provide erroneously high measurements and incorrect assessment of cough efficacy. Waveforms can be readily obtained and interpreted in clinical practice to provide information about causes of MI-E failure, thereby assisting titration to 1836-9553/Crown Copyright © 2022 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Journal of Physiotherapy 69 (2023) 35–41 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 Education plus exercise for persistent gluteal tendinopathy improves quality of life and is cost-effective compared with corticosteroid injection and wait and see: economic evaluation of a randomised trial Ross Wilson a, J Haxby Abbott a, Rebecca Mellor b, Alison Grimaldi b,c, Kim Bennell d, Bill Vicenzino b a Centre for Musculoskeletal Outcomes Research, Department of Surgical Sciences, University of Otago, Dunedin, New Zealand; b University of Queensland School of Health and Rehabilitation Sciences, University of Queensland, Brisbane, Australia; c Physiotec, Brisbane, Australia; d Department of Physiotherapy, Melbourne School of Health Sciences, The University of Melbourne, Australia KEY WORDS ABSTRACT Physical therapy Question: For patients with gluteal tendinopathy, what is the cost utility from health system and societal Cost-effectiveness perspectives of three management approaches: education plus exercise, ultrasound-guided corticosteroid Gluteal tendinopathy injection or wait and see? Design: Economic evaluation alongside a three-group, parallel, randomised Education clinical efficacy trial. Participants: People aged between 35 and 70 years with image-confirmed gluteal Exercise tendinopathy were recruited via advertisements. Interventions: Education plus exercise, consisting of 14 visits to a physiotherapist, with detailed instruction on tendinopathy management, twice weekly supervised exercise sessions, daily home exercises, a handout and a CD; corticosteroid injection, consisting of one ultrasound-guided injection and a handout on general tendon care; and ‘wait and see’, consisting of one visit to a physiotherapist with assurance and advice on staying active whilst respecting pain. Outcome measures: Economic outcome measures were quality-adjusted life years (QALYs) calculated from EuroQol EQ-5D-3L using Australian population preference weights, and total economic costs obtained from participant- reported data collected over the 1-year follow-up period. Missing data (,12% per group) were imputed. Linear regression was used to estimate incremental QALYs and costs between interventions; uncertainty was assessed by calculating 90% confidence intervals, cost-effectiveness acceptability curves and confidence ellipses. Results: A total of 204 individuals (82% women) were enrolled. Incremental cost- effectiveness ratio favoured education plus exercise over corticosteroid injection (AU$12,719 and $5,592 on societal and health system perspectives, respectively) and over wait and see ($29,258 and $3,444 on societal and health system perspectives, respectively). Complete case analysis and varying the direct intervention costs did not change the (imputed analysis) results, with the exception that corticosteroid injection was less cost-effective. Conclusion: Education plus exercise for gluteal tendinopathy improves health-related quality of life and is cost-effective compared with corticosteroid injection and wait and see for treating gluteal tendinopathy. Registration: ACTRN12612001126808. [Wilson R, Abbott JH, Mellor R, Grimaldi A, Bennell K, Vicenzino B (2023) Education plus exercise for persistent gluteal tendinopathy improves quality of life and is cost-effective compared with corticosteroid injection and wait and see: economic evaluation of a randomised trial. Journal of Physiotherapy 69:35–41] © 2022 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 with those of people with severe osteoarthritis of the hip.1 Data on the economic burden of gluteal tendinopathy are unknown; given Gluteal tendinopathy (also known as greater trochanteric pain that reduced work participation and quality of life of gluteal ten- syndrome) is a painful condition of the hip that can impair physical dinopathy are comparable with that of hip osteoarthritis, it is plau- function. Common symptoms are pain at or around the lateral hip sible that the economic burden is somewhat comparable. A recent (especially with walking and stair climbing), sleep disturbance, and study of the economic burden of hip osteoarthritis in the Netherlands reduction in physical activity and social participation. This common reported an average of 159 sick leave calendar days and V12,482 in condition, estimated to affect 10 to 25% of the general population, is costs.2 Therefore, it is proposed that the high burden of gluteal ten- more frequent in women aged 40 to 60 years, and the reported dinopathy underpins the need for cost-effective interventions. reduced work participation levels, high levels of pain and dysfunction affecting physical activity, and reduced quality of life are comparable Three common approaches to the medical management of gluteal tendinopathy are corticosteroid injections, the physiotherapy https://doi.org/10.1016/j.jphys.2022.11.007 1836-9553/© 2022 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/).
36 Wilson et al: Cost-effectiveness of treatments for gluteal tendinopathy approach of exercise therapy plus patient education about the con- Table 1 dition, and to simply ‘wait and see’ whether the condition worsens Protocol for each management approach examined in the study. or spontaneously resolves. The results of a recent clinical trial comparing these three strategies found that both corticosteroid in- Approach Content jection and exercise plus patient education resulted in greater im- provements compared with a wait-and-see approach, with education Education plus 8-week program consisting of 14 visits to a physiotherapist. plus exercise achieving the better outcomes at 1-year follow-up.3,4 exercise Specific and detailed advice on tendon care; twice weekly supervised, progressed exercise sessions; daily home exercises; This study reports an economic evaluation performed alongside Corticosteroid handout and CD. that three-group, parallel, pragmatic, randomised clinical efficacy injection 1 ultrasound-guided injection of celestone chronodose plus trial. It performed a cost-utility analysis, from health system and lignocaine (local anaesthetic); basic advice on tendon care societal perspectives, of three alternatives – a program of education Wait and see (handout). plus exercise versus corticosteroid injection versus a wait-and-see 1 visit to a physiotherapist, providing assurance and general approach – for managing persistent gluteal tendinopathy. advice (verbal and handout) on staying active while respecting pain. Therefore, the research question for this economic evaluation of a randomised trial was: CD = compact disc For patients with gluteal tendinopathy, what is the cost utility 6-month follow-up periods. All costs are reported in 2015 Australian from health system and societal perspectives of three manage- dollars (1 AUD z 0.73 USD in 2015). Discounting of costs and QALYs ment approaches: education plus exercise, ultrasound-guided was unnecessary due to the 1-year follow-up. corticosteroid injection or wait and see? Statistical analysis Methods The statistical analysis plan was developed and published a priori Trial procedures (https://espace.library.uq.edu.au/view/UQ:409744). Statistical ana- lyses were conducted by an independent health economist (RW) not The protocol for the multicentre, parallel, three-group, pragmatic, involved in the study design and conduct. Data and analysis code are randomised clinical trial was prospectively registered and published.5 available on request. All statistical analyses were conducted on an Briefly, community dwelling people with lateral hip pain were intention-to-treat basis using R v4.2.8 recruited from Brisbane and Melbourne, Australia, via advertisements in print and social media. They were screened for eligibility and Multiple imputation was used to account for missing data. Im- randomly allocated to receive one of three strategies: education about putations were created using predictive mean matching for contin- load management plus exercise; one corticosteroid injection under uous variables and binary, ordinal or polytomous logistic regression ultrasound guidance; and a wait-and-see approach (Table 1). Partic- for categorical variables. The outcome analysis was conducted on ipants were followed for 1 year. each of 30 imputed datasets and the results combined using Rubin’s rules.9,10 The clinical outcomes have been reported elsewhere.3 In brief, education plus exercise outperformed corticosteroid injection and Estimates of the incremental QALYs and between-group costs wait and see on the global rating of change outcome at 8 and 52 were estimated by linear regression, adjusted for baseline covariates weeks, while corticosteroid injection and wait and see did not differ (baseline health utility, baseline costs, age, hip(s) affected, study site, at 52 weeks. On pain severity, the results paralleled global rating of symptom duration, mechanism of onset, menopausal status, occu- change at 8 weeks but not at 52 weeks, when education plus exercise pation, education level, country of birth, marital status, living was not different to corticosteroid injection (note: pain levels for all arrangement, height, weight, body mass index, waist circumference, groups at 52 weeks were lower than the inclusion criterion of 4/10 hip circumference, pelvic width and lateral width). Incremental cost- on a numerical rating scale where 10 is worst pain imaginable). effectiveness ratios (ICERs) and incremental net benefits (INBs) were calculated at the conventional policy-relevant willingness-to-pay Outcome measures (WTP) thresholds of one, two and three times gross domestic product (GDP) per capita per QALY gained. Uncertainty was assessed by The economic outcome measures were quality-adjusted life years calculating confidence intervals, cost-effectiveness acceptability (QALYs) and total economic costs accrued over the 1-year follow-up curves and confidence ellipses by applying the central limit theorem period; QALYs were calculated from the EuroQol EQ-5D-3L instru- to the estimates of incremental costs and QALYs.11 ment completed by participants at baseline and 4, 8, 12, 26 and 52 weeks. The EQ-5D responses were converted into health utility scores Sensitivity analyses using Australian population preference weights.6 To assess whether the results were sensitive to the imputation of Economic costs were obtained by patient self-report using a missing data for non-respondents, a complete case analysis (using modified version of the Osteoarthritis Costs and Consequences data only from participants responding to all follow-up surveys) was Questionnaire,7 administered at baseline and at 26 and 52 weeks. conducted. As patient-reported healthcare use (and costs) are highly Healthcare utilisation (other than medication), impact on working variable, a sensitivity analysis (from the health system perspective) hours or duties, and costs incurred by family or friends (eg, travel) considering only the direct intervention costs was also conducted. As were reported for the 6 months preceding each follow-up. Medica- healthcare costs can vary widely among providers and regions and tion use was reported for the week preceding each follow-up, to may differ from those paid to providers in the trial, sensitivity ana- minimise recall bias, and converted to a 6-month total. Intervention lyses varying the unit costs of the interventions were conducted. costs were calculated based on reimbursement rates paid to providers (AU$156.20 for ultrasound-guided corticosteroid injection, $840 for Patient and public involvement education plus exercise physiotherapy intervention and $60 for wait- and-see approach). All other costs were calculated by applying A patient representative at National Health and Medical Research reference unit costs for each item of care to participants’ reported Council program grant meetings was involved in the planning and healthcare utilisation (Appendix 1 on the eAddenda). Total economic development of the trial. Patients were not involved in the cost- costs over 1-year follow-up were calculated at both the societal level effectiveness analysis reported in this article. (primary; all reported costs) and health system level (secondary; healthcare costs only) by summing the costs incurred in each of the Results Participant recruitment and characteristics have been described previously.3 In brief, 204 participants aged between 35 and 70 years
Research 37 Table 2 Baseline demographic characteristics of trial participants, by treatment group. Characteristic All participants (n = 204) Education plus exercise (n = 69) Corticosteroid injection (n = 66) Wait and see (n = 69) Study site, n (%) 99 (49) 32 (46) 33 (50) 34 (49) Brisbane 105 (51) 37 (54) 33 (50) 35 (51) Melbourne 54.8 (8.8) 54.8 (8.1) 55.3 (9.4) 54.5 (9.1) 1.66 (0.08) 1.65 (0.09) 1.66 (0.07) 1.67 (0.08) Age (y), mean (SD) 75.9 (15.3) 75.9 (14.4) 74.4 (14.6) 77.3 (16.7) Height (m), mean (SD) 27.4 (5.1) 27.7 (4.8) 27.0 (5.1) 27.6 (5.5) Weight (kg), mean (SD) 88.6 (13.4) 87.7 (12.4) 88.9 (13.3) 89.2 (14.5) Body mass index, mean (SD) 104.6 (10.1) 104.2 (8.3) 104.0 (10.0) 105.5 (11.8) Waist girth (cm), mean (SD) Hip circumference (cm), mean (SD) 43 (21) 16 (23) 15 (23) 12 (17) Hormonal status, n (%) 24 (12) 7 (10) 5 (8) 12 (17) 93 (46) 31 (45) 26 (38) premenopausal 2 (3) 36 (55) perimenopausal 7 (3) 13 (19) 1 (2) 4 (6) postmenopausal 37 (18) 9 (14) 15 (22) unknown 37 (54) n/a 121 (59) 24 (35) 41 (62) 43 (62) Main occupation, n (%) 55 (27) 17 (26) 14 (20) manager/professional 16 (8) 6 (9) 7 (10) tradesperson/clerical worker 11 (5) 2 (3) 3 (5) transport, sales, service worker, labourer 0 (0) 4 (6) 5 (7) no paid job 1 (0) 1 (2) 0 (0) don’t know 0 (0) Education level, n (%) 2 (1) 21 (30) 1 (2) 1 (1) , 3 yrs high school 41 (20) 12 (17) 7 (11) 13 (19) 31 yrs high school 45 (22) 13 (19) 18 (27) 15 (22) some tertiary training 53 (26) 23 (33) 22 (33) 18 (26) graduated from uni/polytechnic 63 (31) 18 (27) 22 (32) any post-graduate study 50 (72) Marital status, n (%) 143 (70) 7 (10) 47 (71) 46 (67) married/civilly united 20 (10) 8 (12) 5 (8) 8 (12) living with significant other 27 (13) 1 (1) 8 (12) 11 (16) divorced/separated 3 (4) 0 (0) 2 (3) widowed 3 (1) 6 (9) 2 (3) single 11 (5) 7 (10) Living status, n (%) 31 (45) 9 (14) 6 (9) alone 22 (11) 31 (45) 31 (47) 33 (48) partner/spouse only 95 (47) 25 (38) 25 (36) partner and child(ren) 81 (40) 0 (0) children only 1 (2) 5 (7) 6 (3) with diagnosed gluteal tendinopathy were recruited between March corticosteroid injection group, 94% in the education plus exercise 2013 and September 2015. Most participants were women (82%). group and 88% in the wait-and-see group. Baseline characteristics Retention rates (completing follow-up at 52 weeks) were 95% in the were similar in the three groups (Tables 2 and 3). Table 3 Baseline clinical characteristics of trial participants, by treatment group. Characteristic All participants Education plus exercise Corticosteroid injection Wait-and-see (n = 204) (n = 69) (n = 66) (n = 69) Study hip, n (%) 105 (51) 36 (52) 33 (50) 36 (52) right 99 (49) 33 (48) 33 (50) 33 (48) left 157 (77) 55 (80) 46 (70) 56 (81) Symptoms, n (%) 47 (23) 14 (20) 20 (30) 13 (19) unilateral bilateral 24 (12) 5 (7) 11 (17) 8 (12) 51 (25) 21 (30) 16 (24) 14 (20) Symptom duration, n (%) 129 (63) 43 (62) 39 (59) 47 (68) 2 to 6 months 6 to 12 months 9 (4) 2 (3) 3 (5) 4 (6) .12 months 15 (7) 8 (12) 5 (8) 2 (3) 178 (87) 58 (84) 58 (88) 62 (90) Mechanism of onset, n (%) 2 (1) 1 (1) 0 (0) 1 (1) slip/fall 4.9 (1.0) 4.8 (1.0) 4.8 (1.0) 4.9 (1.2) change in activity 45.7 (16.2) 45.2 (17.3) 46.4 (15.1) 45.6 (16.4) insidious onset other 4.6 (1.9) 4.5 (2.1) 4.5 (2.0) 4.9 (1.6) Pain scale (0 to 10), mean (SD) NA (NA) NA (NA) 0.768 (0.096) 0.752 (0.131) Lateral hip pain questionnaire, activities of 13.6 (9.0) 13.4 (10.0) 13.7 (8.2) 13.6 (8.7) daily living (0 to 100), mean (SD) 4.7 (4.4) 4.6 (4.1) 5.0 (5.2) Patient specific functional scale average score 59.9 (12.3) 4.5 (4.0) 59.3 (10.9) 60.2 (13.1) (0 to 10), mean (SD) 47.7 (9.2) 60.2 (12.9) 47.9 (9.1) 47.8 (9.6) EQ-5D questionnaire (0 to 1), mean (SD) 462.5 (428.6) 47.3 (9.1) Pain catastrophising scale, (0 to 52), mean (SD) 434.4 (424.6) 368.8 (313.4) 580.2 (500.8) Patient health questionnaire (0 to 20), mean (SD) 0.81 (0.30) VISA-G (0 to 100), mean (SD) 10.4 (6.3) 0.77 (0.30) 0.82 (0.30) 0.84 (0.31) Pain self-efficacy questionnaire (0 to 60), mean (SD) 10.0 (6.5) 10.6 (6.3) 10.7 (6.1) Active Australia questionnaire, total time spent in overall activity in the past week, (minutes), mean (SD) Gluteal muscle torque (Nm/kg), mean (SD) Active lag (deg), mean (SD) EQ-5D = European quality of life-5; VISA-G = Victorian Institute for Sport Assessment – Gluteal; NA = not applicable
38 Wilson et al: Cost-effectiveness of treatments for gluteal tendinopathy Table 4 threshold of 2 3 GDP per capita) and the health system perspective Mean (SD) costs in AU$ (2015) and health outcomes (QALYs) through 1-year follow-up. (at 1 3 GDP per capita). Confidence intervals for the INB of cortico- steroid injection relative to wait and see spanned zero from both the Education plus Corticosteroid Wait and societal and health system perspectives, at all WTP thresholds. Con- exercise injection see fidence ellipses for the incremental QALYs and costs of the in- terventions are shown in Figure 1. Costs ($) 6,979 (9,077) 5,033 (6,372) 5,162 (6,471) intervention costs 840 (0) 156 (0) 60 (0) From the societal perspective, education plus exercise was esti- health system costs mated to have a probability of being cost-effective relative to wait and productivity costs 3,417 (4,927) 3,395 (3,533) 3,898 (4,510) see exceeding 85%, and relative to corticosteroid injection exceeding other costs 1,668 (5,884) 606 (2,629) 521 (2,897) 90%, at all considered WTP thresholds (Figure 2). From the health 1,053 (3,077) 876 (3,316) 683 (2,236) system perspective, these probabilities, relative to both wait and see QALYs 0.859 (0.108) 0.833 (0.116) 0.787 (0.145) and corticosteroid injection, exceeded 97% at all WTP levels. QALY = quality-adjusted life years Sensitivity analyses Resource use, economic costs and quality-adjusted life years Complete case analysis indicated that the results were robust to the imputation of missing data for non-respondents; the education Differences in healthcare use between the groups were generally plus exercise intervention compared with wait and see hardly small and variable, with no group showing consistently higher or differed, while corticosteroid injection was less cost-effective in lower use than others, although the education plus exercise group complete-case comparisons with education plus exercise physio- had lower use of pain medication, particularly non-steroidal anti- therapy or wait and see (Table 6). Overall conclusions were un- inflammatory drugs and supplements (Appendix 2). changed from the base case (imputed) analysis. Both mean costs (societal perspective) and mean QALYs over 1- Only the direct intervention costs reduced variability in the cost year follow-up were greatest in the education plus exercise group data, resulting in the INMB for education plus exercise relative to either (Table 4). Health system costs (other than the cost of the interven- wait and see or corticosteroid injection being strictly greater than zero tion) were greatest in the wait-and-see group, while productivity and at the 1x GDP per capita WTP threshold, with corticosteroid injection other costs were greatest in the education plus exercise group; not clearly better value than the wait-and-see approach. however, costs in all categories showed wide variation. Estimated QALY gains (after adjustment for baseline covariates) and their con- Varying the cost of delivering the education plus exercise phys- fidence intervals were greater than zero for the education plus ex- iotherapy intervention (to $100 per session) or the cost of delivering ercise group relative to both corticosteroid injection and wait and see corticosteroid injection either upward (to $279 per session)12 or (Table 5). No clear QALY difference was observed between cortico- downward (to $64, reflecting the lower cost of a non-US guided in- steroid injection and wait and see. There was little difference in costs jection) did not change any of the conclusions. between any of the groups from the health system perspective. Both education plus exercise and corticosteroid injection had higher costs Discussion than wait and see from the societal perspective, but the confidence intervals all spanned zero so the average between-group differences The results of this trial indicate that an education plus exercise in costs in the wider population of people with gluteal tendinopathy intervention is cost-effective relative to both wait and see and remain uncertain. corticosteroid injection for the management of gluteal tendinopathy, from both health system and societal perspectives over a 1-year time Cost-effectiveness horizon. This is due to its superior health-related quality of life gains, despite slightly higher treatment costs. Over 1-year follow-up, the education plus exercise intervention was highly cost-effective relative to both the wait-and-see approach Strengths of this economic evaluation include prospective regis- (ICER $29,258 from the societal perspective and $3,444 from the tration of the analysis plan, and conduct within a high-quality, health system perspective, both well below the GDP/capita threshold pragmatic, randomised controlled efficacy trial with low loss to for cost-effectiveness) and the corticosteroid injection intervention follow-up. The costs and effects estimates were based on compre- (ICERs $12,719 and $5,592, respectively) (Table 5). Confidence in- hensive cost and outcome data collection using validated in- tervals for the INBs of education plus exercise relative to both wait struments. The results were robust to several sensitivity analyses, and see and corticosteroid injection were strictly greater than zero, at giving a high degree of confidence in the primary finding that the the 90% confidence level, from the societal perspective (at the WTP Table 5 Incremental QALYs, costs, cost-effectiveness ratio, and net monetary benefits, through 1-year follow-up. Education plus exercise relative Corticosteroid injection relative Education plus exercise relative to to wait and see to wait and see corticosteroid injection Incremental QALYs 0.057 (0.025 to 0.090) 0.005 (20.027 to 0.038) 0.052 (0.019 to 0.085) Incremental costs $1,680 (2$860 to $4,220) $1,018 (2$1,440 to $3,475) $663 (2$1,876 to $3,201) societal perspective $198 (2$1,201 to $1,597) 2$94 (2$1,570 to $1,383) $291 (2$1,161 to $1,744) health system perspective Incremental cost-effectiveness ratio $29,258 $190,852 $12,719 societal perspective $3,444 2$17,551a $5,592 health system perspective Incremental net monetary benefit $2,225 (2$1,303 to $5,753) 2$655 (2$4,130 to $2,820) $2,880 (2$645 to $6,404) societal perspective $6,130 ($833 to $11,426) 2$292 (2$5,578 to $4,994) $6,422 ($1,085 to $11,759) 1x GDP per capita $9,964 ($2,573 to $17,356) 2x GDP per capita $10,035 ($2,736 to $17,333) $70 (2$7,254 to $7,394) 3x GDP per capita $3,251 ($438 to $6,064) health system perspective $3,707 ($936 to $6,478) $456 (2$2,395 to $3,307) $6,793 ($1,914 to $11,673) 1x GDP per capita $7,612 ($2,815 to $12,409) $819 (2$4,071 to $5,709) $10,336 ($3,280 to $17,392) 2x GDP per capita $11,517 ($4,592 to $18,442) $1,181 (2$5,856 to $8,219) 3x GDP per capita Cells report effect estimate (90% confidence interval). All effects are adjusted for baseline EQ-5D utility, healthcare costs, age, affected hip, study site, duration of symptoms, method of onset, sex, occupational category, educational level, country of birth, marital status, living status and body size measurements. GDP = gross domestic product in 2017 Australian dollars; QALYs = quality-adjusted life years. a Intervention dominates comparator (lower costs and higher QALYs).
Research 39 Figure 1. Cost-effectiveness plane. Points show estimated incremental costs (y-axis) and QALYs gained (x-axis) for the education plus exercise (green) and corticosteroid injection (orange) interventions relative to usual care and for the education plus exercise intervention relative to corticosteroid injection (purple), over 1-year follow-up. Shaded ellipses show uncertainty intervals at the 50%, 75% and 90% levels. The black line represents the 1x GDP per capita willingness-to-pay level per QALY gained; areas below and to the right of the line indicate cost-effectiveness of the intervention. All costs are in 2015 Australian dollars. GDP = Gross Domestic Product in 2015 Australian dollars ($68,000, zUS$49,700); QALY = quality-adjusted life years. education plus exercise strategy is cost-effective; there was no strong months, which may have resulted in underreporting of some evidence either way that corticosteroid injection results in either healthcare items (eg, medication).13 Participants in the trial had better or worse outcomes or higher or lower costs than the wait-and- gluteal tendinopathy confirmed by magnetic resonance imaging; the see approach. results may be moderated in clinical practice where patients may not be selected based on imaging. The results of this trial should be interpreted in light of its limi- tations. This was an efficacy trial recruiting participants from the The cost-effectiveness of interventions for gluteal tendinopathy community via advertisements, and thus the results may not be has seen very little attention. Two such studies reported on cortico- entirely generalisable to consecutive patients seeking care. The trial steroid injections in patients with greater trochanteric pain syn- was powered for the primary clinical outcome so, as is usual with drome14 and trochanteric bursitis12 – both terms have been trial-based economic evaluation, incremental cost estimates were interchangeably used for clinically determined gluteal tendinopathy. very uncertain. Nevertheless, the cost-effectiveness of education plus A secondary economic evaluation of a randomised clinical trial15 in exercise was estimated with a high degree of confidence and was patients with greater trochanteric pain syndrome compared corti- robust to sensitivity analyses. The validated questionnaire used to costeroid injections with usual care that consisted of analgesics as collect health service use data relied on retrospective recall over 6 needed. It reported no between-group differences in health-related Figure 2. Cost-effectiveness acceptability curves. Lines show the estimated probability of cost-effectiveness for the education plus exercise (green) and corticosteroid injection (orange) interventions relative to wait-and-see control and for the education plus exercise intervention relative to corticosteroid injection (purple), at each level of willingness-to-pay per QALY gained. The red vertical lines represent policy-relevant willingness-to-pay thresholds of 1x, 2x and 3x GDP per capita. All costs (and willingness-to-pay values) are in 2015 Australian dollars. GDP = Gross Domestic Product in 2015 Australian dollars ($68,000, zUS$49,700); QALY = quality-adjusted life years.
40 Wilson et al: Cost-effectiveness of treatments for gluteal tendinopathy Table 6 Incremental net monetary benefits through 1-year follow-up, sensitivity analyses. Education plus exercise Corticosteroid injection Education plus exercise relative relative to wait and see relative to wait and see to corticosteroid injection Base case $2,225 (2$1,303 to $5,753) 2$655 (2$4,130 to $2,820) $2,880 (2$645 to $6,404) societal perspective $3,707 ($936 to $6,478) $456 (2$2,395 to $3,307) $3,251 ($438 to $6,064) health system perspective $2,273 (2$900 to $5,446) 2$2,680 (2$5,875 to $515) $4,953 ($1,723 to $8,184) Complete case analysis $3,838 ($857 to $6,819) 2$1,055 (2$4,070 to $1,959) $4,894 ($1,848 to $7,939) societal perspective health system perspective $3,125 ($933 to $5,317) $266 (2$1,951 to $2,484) $2,859 ($613 to $5,104) Direct intervention costs only $1,793 (2$1,745 to $5,331) 2$592 (2$4,076 to $2,892) $2,385 (2$1,147 to $5,918) health system perspective $3,273 ($492 to $6,053) $516 (2$2,344 to $3,375) $2,757 (2$67 to $5,581) Increased physiotherapy cost $2,232 (2$1,296 to $5,760) 2$769 (2$4,244 to $2,705) $3,001 (2$523 to $6,526) societal perspective $3,714 ($943 to $6,485) $342 (2$2,509 to $3,193) $3,373 ($559 to $6,186) health system perspective $2,219 (2$1,309 to $5,748) 2$569 (2$4,044 to $2,905) $2,789 (2$736 to $6,313) Increased injection cost $3,702 ($931 to $6,473) $542 (2$2,309 to $3,393) $3,160 ($346 to $5,973) societal perspective health system perspective Reduced injection cost (unguided injection) societal perspective health system perspective Cells report incremental net monetary benefit (90% confidence interval), at willingness-to-pay threshold equivalent to Australian one GDP/capita. All effects are adjusted for baseline EQ-5D utility, healthcare costs, age, affected hip, study site, duration of symptoms, method of onset, sex, occupational category, educational level, country of birth, marital status, living status and body size measurements. GDP = gross domestic product in 2017 Australian dollars. quality of life or costs. The authors concluded that corticosteroid in- eAddenda: Appendices 1 and 2 can be found online at https://doi. jections have a low probability of cost-effectiveness.14 The other small org/10.1016/j.jphys.2022.11.007 trial of trochanteric bursitis assumed lower intervention costs and found that ultrasound-guided corticosteroid injection is less cost- Ethics approval: The trial was approved by the human research effective than injection using clinical landmarks.12 In the current ethics committees of the University of Queensland (#2012000930) study, corticosteroid injection would not have been cost-effective and the University of Melbourne (ID 1238598). relative to wait and see or education plus exercise, even at the lower cost of non-guided injection. We found no economic evalua- Competing interests: Nil. tions of education plus exercise for gluteal tendinopathy. Source(s) of support: The original trial was funded by a National Health and Medical Research Council (NHMRC) program grant The randomised clinical trial on which this analysis was based (#631717). KB was supported by an NHMRC principal research showed that education plus exercise delivered superior health- fellowship (#1058440). The funders of the study had no role in design related quality of life outcomes (measured by the EQ-5D) than and conduct of the study; collection, management, analysis and either corticosteroid injection or a wait-and-see approach, and that interpretation of the data; and preparation, review or approval of the there was negligible difference between the other two groups.3 This manuscript or the decision to submit for publication. result is consistent with the primary clinical outcome of global rating Acknowledgements: We acknowledge the group of authors from of change. The finding regarding corticosteroid injections is consis- the original LEAP trial from which the data were obtained. We thank tent with the only other comparable randomised controlled trial, the participants who volunteered to participate in the original trial which found benefits of injections at 12 weeks but not at 52 weeks.15 and the clinicians who carried out treatments and imaging. Success rates for the education plus exercise program (74% reporting Data sharing: Data are available in a public, open access re- improvement in hip condition at 12 and 26 weeks) were higher in the pository, with curated access. https://espace.library.uq.edu.au/view/ current study than in a previous clinical trial involving home-based UQ:409744 exercise for gluteal tendinopathy (41% at 4 months),16 suggesting Providence: Not invited. Peer reviewed. the importance of adequate supervision to enable both progression of Correspondence: Bill Vicenzino, School of Health and Rehabilita- the exercise program and appropriately deliver the education infor- tion Sciences, The University of Queensland, Brisbane, Australia. mation to the individual patient. The higher costs of the supervised Email: [email protected] physiotherapy program appear to be more than offset by the sus- tained improvements in patients’ health-related quality of life. References In conclusion, education plus exercise for gluteal tendinopathy is 1. Fearon AM, Cook JL, Scarvell JM, Neeman T, Cormick W, Smith PN. Greater likely to deliver meaningful improvements in patients’ health-related trochanteric pain syndrome negatively affects work, physical activity and quality of quality of life and provide good value to the health system when life: A case control study. J Arthroplasty. 2014;29:383–386. https://doi.org/10.1016/ compared with either a wait-and-see approach or corticosteroid in- j.arth.2012.10.016. jection. These results are consistent with clinical evidence supporting use of education plus exercise as an effective management approach 2. Hardenberg M, Speklé EM, Coenen P, Brus IM, Kuijer PPFM. The economic for gluteal tendinopathy. burden of knee and hip osteoarthritis: Absenteeism and costs in the Dutch workforce. BMC Musculoskelet Disord. 2022;23:364. https://doi.org/10.1186/ What was already known on this topic: Gluteal tendinop- s12891-022-05306-9. athy is a common and costly problem that is more frequent in women aged 40 to 60 years. It causes pain, reduces work ca- 3. Mellor R, Bennell K, Grimaldi A, Nicholson P, Kasza J, Hodges P, et al. Education plus pacity and limits physical activity, thereby reducing quality of life. exercise versus corticosteroid injection use versus a wait and see approach on What this study adds: Education plus exercise for gluteal global outcome and pain from gluteal tendinopathy: Prospective, single blinded, tendinopathy is likely to deliver meaningful improvements in randomised clinical trial. BMJ. 2018;360:k1662. patients’ health-related quality of life and provide good value to the health system when compared with either a wait-and-see 4. Mellor R, Bennell K, Grimaldi A, Nicholson P, Kasza J, Hodges P, et al. Education plus approach or corticosteroid injection. exercise versus corticosteroid injection use versus a wait and see approach on global outcome and pain from gluteal tendinopathy: Prospective, single blinded, randomised clinical trial. Br J Sports Med. 2018;52:1464–1472. 5. Mellor R, Grimaldi A, Wajswelner H, Hodges P, Abbott JH, Bennell K, et al. Exercise and load modification versus corticosteroid injection versus ‘wait and see’ for persistent gluteus medius/minimum tendinopathy (the LEAP trial): A protocol for a randomised controlled trial. BMC Musculoskelet Disord. 2016;17:196. https://doi. org/10.1186/s12891-016-1043-6. 6. Viney R, Norman R, King MT, Cronin P, Street DJ, Knox S, et al. Time trade-off derived EQ-5D weights for Australia. Value Health. 2011;14:928–936. https://doi. org/10.1016/j.jval.2011.04.009.
Research 41 7. Pinto D, Robertson MC, Hansen P, Abbott JH. Good agreement between question- 13. Evans C, Crawford B. Patient self-reports in pharmacoeconomic studies: Their use naire and administrative databases for health care use and costs in patients with and impact on study validity. PharmacoEconomics. 1999;15:241–256. https://doi. osteoarthritis. BMC Med Res Methodol. 2011;11:45. org/10.2165/00019053-199915030-00004. 8. R Core Team. R: A Language and Environment for Statistical Computing. R Foundation 14. Blok DJ, Brinks A, Rijn RM van, Verhaar JA, Koes BW, Bierma-Zeinstra SM, et al. for Statistical Computing; 2021. https://www.R-project.org/. Cost-effectiveness analysis of corticosteroid injections in patients with greater trochanteric pain syndrome. In: Brinks A, ed. Greater Trochanteric Pain 9. Carpenter JG, Kenward MG. Multiple Imputation and Its Application. Chichester: Syndrome in General Practice. PhD Thesis. Erasmus University Rotterdam; John Wiley & Sons; 2013. 2011:37–57. 10. Rubin DB. Multiple Imputation for Nonresponse in Surveys. Chichester: John Wiley & 15. Brinks A, Rijn RM van, Willemsen SP, Bohnen AM, Verhaar JA, Koes BW, et al. Sons; 1987. Corticosteroid injections for greater trochanteric pain syndrome: A randomized controlled trial in primary care. Ann Fam Med. 2011;9:226–234. https://doi.org/10. 11. Nixon RM, Wonderling D, Grieve RD. Non-parametric methods for cost- 1370/afm.1232. effectiveness analysis: The central limit theorem and the bootstrap compared. Health Economics. 2010;19:316–333. https://doi.org/10.1002/hec.1477. 16. Rompe JD, Segal NA, Cacchio A, Furia JP, Morral A, Maffulli N. Home training, local corti- costeroid injection, or radial shock wave therapy for greater trochanter pain syndrome. 12. Mitchell WG, Kettwich SC, Sibbitt WL, et al. Outcomes and cost-effectiveness of Am J Sports Med. 2009;37:1981–1990. https://doi.org/10.1177/0363546509334374. ultrasound-guided injection of the trochanteric bursa. Rheumatol Int. 2018;38:393–401. https://doi.org/10.1007/s00296-018-3938-z.
Journal of Physiotherapy 69 (2023) 1-3 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 How to facilitate mainstream media coverage of physiotherapy research Joshua R Zadro Institute for Musculoskeletal Health, Sydney School of Public Health, Faculty of Medicine and Health, The University of Sydney and Sydney Local Health District, Sydney, Australia Many physiotherapists seek to promote their clinical research for a approach. Some may even be convinced to run a story on research variety of reasons. They may wish to increase awareness of the that’s weeks or months old if they feel they have at least part of the findings of research among patients and/or their carers. They may story to themselves. Although you may sacrifice mass media coverage hope that promoting the research will encourage clinicians to use the with this approach, you may increase your chance of getting a high- new evidence in their patient management. They may want other profile media outlet to run a story on your research. researchers to incorporate the new evidence into current and future research endeavours. They may also wish to promote a current study Many news journalists have unpredictable schedules because they so that patients consider participating in it. Many physiotherapists must go where the news is happening. Make yourself available at use social media to promote their research, where they can largely short notice during a set period around the time the research is control how their research is summarised and presented. Failure to published so they can contact you when they have time. To assist communicate with mainstream media – such as television, radio, them with filming, audio recording or photographing appropriate newspaper and journalists’ podcasts – limits the exposure that the content to help tell the story of your research, think about how you research receives;1 however, it can be more difficult to control the could provide: an example of the morbidity caused by the clinical narrative in these settings. condition of interest, a demonstration of the treatment and/or an example of how the research has or will assist patients with the The purpose of this Editorial is threefold: to offer suggestions to condition. help physiotherapists involved in clinical research promote their research to mainstream media; to offer suggestions to help them You can assist journalists to feature your research in a news story engage with mainstream media; and to suggest ways in which they by writing a summary of your research and its findings in a format might help members of the media to portray the research accurately. that is as similar as possible to the format that they will use. For Overall, it encourages physiotherapists involved in clinical research to example, for a radio or podcast interview, suggest possible questions think about promoting their research on mainstream media. and what your responses would be. For a television news report, provide a summary of the research and its impact for patients, with Promoting your research in the media suggested images. Provide contact details for yourself and a patient who is willing to be interviewed about the research and what it Being proactive is essential to successfully promoting your means for them. research in the mainstream media. Planning should begin the day your paper is accepted for publication. Most good journalists will only run a Where possible, try to provide appealing images. If you cannot story on research the day it is published or a few days after at the supply live content for the journalist to film, provide high-resolution latest, so you need to be organised. Once your paper is accepted, video recording in landscape format. For television and print media, contact the journal editor to find out the exact publication date. you might also prepare an infographic, where you present the find- Journalists will ask for this so that they can time their news story with ings of your research with the correct interpretation of the data. the day your paper goes online. Next, you need to start preparing a press release. A good press release will provide a summary of the Before, during and after an interview findings, use plain language (no jargon) and be tailored to the target audience (eg, general audience, patient groups, clinicians, policy- The Sense About Science website provides helpful advice for re- makers). You should also consider finding a ‘hook’ that a general searchers being interviewed by journalists.3–5 audience could engage with, linking the paper to recent newsworthy events, preparing some quotes to convey key messages and identifying Before an interview any patient stories that could be used by journalists. If your institution has a media team, they may be able to help you with the above steps Journalists usually explain the subject of their article or what they and start pitching your idea to journalists. When you pitch to jour- would like to talk about to a researcher. Although some journalists nalists, you need to be confident and direct about your message so that may send through a few questions as prompts before the interview, they can understand how your story will benefit the audience.2 most will not and there are a few important reasons for this. First, journalists want the conversation to be unscripted and unpredictable. One proactive way to generate mass media coverage is a general The best quotes usually arise from unscripted conversation, and un- press release that allows you to pitch your research to journalists predictable conversations more often lead to a more compelling and from multiple outlets. This often results in many similar stories being nuanced narrative of new evidence. Second, a journalist’s job is to run. Another approach is to target your pitch to one high-profile uphold freedom of speech and press. If journalists provided a list of media outlet and offer an ‘exclusive’ interview about your research. questions to researchers in advance, answers may be manipulated to Journalists are largely driven by exclusivity so may prefer this serve the interests of universities, funders or the researchers themselves. https://doi.org/10.1016/j.jphys.2022.11.009 1836-9553/© 2022 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/).
2 Editorial Table 1 Tips for when you are contacted by a journalist (adapted from Primary Health Care Research and Information Service 20142). Ask Arrange Avoid what topics will be covered to respond to the inquiry in a timely fashion assuming that the journalist is an expert in the field for sample questions to schedule your interview in a quiet place being offended if you do not receive sample ques- who the audience is to look up the reporter and read their previous pieces what the deadline is to look up the journalist’s publications tions before the interview if the story is broadcast, news, feature or a policy preparing set responses to interview questions piece how long the interview will be So, what can you do before an interview? It is often helpful to Helping members of the media to portray the research accurately remember three to five points you want to get across in the interview and mention them if they do not come up. Having some examples, Journalists do not always present research appropriately.6–10 A analogies and stories can help reinforce these points. You should go systematic review of 44 studies assessing the quality of news reports into the interview with the attitude of building a lasting relationship about health interventions found that many reports do not mention with the journalist, so they have a trustworthy source for similar conflicts of interest (78%), alternative interventions (64%), harms stories in the future. Ensuring everything you talk about is covered in (60%) or costs (82%), and do not quantify effects (53%) or report ab- the article is less important. Ultimately, a story is meant to inform and solute effects (83%).6 Other studies have found similar results, high- entertain. Table 1 summarises some useful tips when contacted by a lighting that journalists generally focus on the benefits of an journalist. intervention while neglecting potential harms,7–9 use spin (eg, fail to acknowledge a non-significant effect on the primary outcome, focus At the interview on positive within-group effects, focus on subgroups, fail to acknowledge small sample sizes and imprecision),7–9 use sensation- Immediately before the interview begins, it may be important to alism (eg, buzz words, positive framing),8 and make inappropriate ask the journalist whether: they are recording the conversation, they extrapolations (eg, apply animal studies to humans, infer causation can share your quotes with you before publication, there will be a from uncontrolled studies).7–9 A popular segment from HBO’s Last follow-up conversation, and they know when the story will be pub- Week Tonight with John Oliver (aired on 8th May 2016) provided lished. You should also be in a quiet place with your phone and email several examples of how the media has taken scientific research out turned off. of context and misrepresented the findings to create a sound bite for TV.10 During an interview, keep in mind that journalists like back- ground information and details – these make for good stories. Do not It is important to avoid blaming journalists for this. Perhaps some be patronising if they ask simple questions, which may not imply instances are due to a desire to sensationalise a story. However, misunderstanding; journalists sometimes use simple questions to others could be due to a genuine misunderstanding of the statistical prompt you to provide analogies, quotes and stories. You should avoid results reported in published research,11 lack of training on how to jargon and technical details unless asked and ask for clarification if interpret and report research,11,12 and spin in the research abstract.9 you do not understand a question. Remember you are always being Some initiatives exist or are underway to address these issues (eg, recorded – unless you ask for something to be ‘off the record’ before tips and workshops from professional associations,13,14 a checklist for saying it – and anything you say can be quoted. appropriate reporting of research in press releases15). There are also some things you can do. Try to get on the front foot with journalists Sometimes the part of the story you find most interesting will not by converting your statistical results into lay terms wherever they be the focus of the article and sometimes you will not even be quoted might come across your research. On social media, include a lay after providing an interview. Do not take this to heart. The best you summary of what the results mean for patients. Try to publish in a can do is be informative, helpful and try build a good relationship journal that includes a lay summary of the main findings of the with the journalists. Journalists particularly like it if you can recom- research. mend someone else for them to speak to on this topic and if you have any useful images that could accompany the article. If new evidence A group of statisticians who volunteer their time to help jour- on this topic emerges in the future, let them know – they may want to nalists to get the right understanding of the statistical results in write another story. research operates through www.STATS.org. You could recommend this to journalists, especially if they ask for help to understand the After an interview statistics in your published research and/or express appreciation for the efforts you have taken to present the statistical results in an easily Do not expect to see the article before its published. Journalists are digestible format. often on extremely tight deadlines and do not have time to vet the opinions of all the people they have interviewed. Journalists may Effectively communicating physiotherapy research to the main- reach out to clarify a technical point or if they do not understand the stream media is essential to increase awareness of important new meaning of a quote. In cases where the subject matter is highly evidence among patients and the public, clinicians, researchers and technical and sensitive, they may show you the full article before policymakers. Researchers wanting to promote their work via main- publication. stream media should start planning their media strategy the day their paper is accepted for publication, seek support from their promotions If this has made you nervous about being interviewed, there are team where possible, and make their press release engaging and some things you can do. First, ask to see your quotes: a journalist can targeted. Researchers should aim to build positive relationships with show you the quotes they plan to use, but because the conversation journalists by: being flexible with interview times; helping them find was ‘on the record’ they are under no obligation to change them. images, videos and patient stories; writing summaries in a format Second, reach out to the journalist if you feel you made a factual that journalists use frequently; recommending other experts to speak mistake; these need to be corrected, including at the post-publication on a topic they are being interviewed for; and helping journalists stage. Finally, find out the rough publication date so you can quickly interpret and report research appropriately. respond to last minute requests (eg, double-check quotes and information). Ethics approval: Not applicable. Competing interests: Nil. Source(s) of support: Nil.
Editorial 3 Acknowledgements: Nil. 6. Oxman M, et al. F1000Research. 2022;10:433. Correspondence: Joshua R Zadro, Institute for Musculoskeletal Health, Sydney School of Public Health, Faculty of Medicine and 7. Woloshin S, Schwartz LM. Med J Aust. 2006;184:576–580. Health, The University of Sydney and Sydney Local Health District, Sydney, Australia. Email: [email protected] 8. Dempster G, et al. J Sci Comm. 2022;21:A06. References 9. Yavchitz A, et al. PLoS Med. 2012;9:9. 1. Seeman N. Electronic Healthcare. 2009;7:101–108. 10. Cordova M, et al. PNLA Quarterly. 2018;82. 2. Primary Health Care Research and Information Service (PHCRIS). Engaging with 11. Voss M. Am J Public Health. 2002;92:1158–1160. the media: Promoting your work to the media. February 2014. 3. Sense About Science USA. A Media Guide for Scientists: Before the Interview. 12. Kees B. Newsroom Training: where’s the investment? John S. and James L. Knight https://senseaboutscienceusa.org/wp-content/uploads/2016/02/SAS_BEFORE_02 Foundation; 2002. 0316.pdf. Accessed November 27, 2022. 4. Sense About Science USA. A Media Guide for Scientists: During the Interview. 13. Australasian Medical Writers Association. Professional development program https://senseaboutscienceusa.org/wp-content/uploads/2016/02/SAS_DURING_RGB_ 020116.pdf. Accessed November 27, 2022. AMWA; 2022. https://www.medicalwriters.org/about-us/professional- 5. Sense About Science USA. A Media Guide for Scientists: After the Interview. https://senseaboutscienceusa.org/wp-content/uploads/2016/02/SAS_AFTER_RGB_ development/. Accessed November 27, 2022. 020116-2.pdf. Accessed November 27, 2022. 14. Sweet M. Aust Prescr. 2000;23:70–71. 15. Equator Network. PR-Rx – Press Release Reporting Exemplar (registered 23 February 2021). https://www.equator-network.org/library/reporting-guidelines- under-development/reporting-guidelines-under-development-for-other-study- designs/. Accessed November 27, 2022. Websites Media guide for scientists senseaboutscienceusa.org/media-guide-for-scientists STATS check senseaboutscienceusa.ofg/stats-check/
Journal of Physiotherapy 69 (2023) 23–34 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 Mobile app use to support therapeutic exercise for musculoskeletal pain conditions may help improve pain intensity and self-reported physical function: a systematic review Debra Thompson a, Samuel Rattu a, Jared Tower a, Thorlene Egerton b, Jill Francis c,d, Mark Merolli b,e a Department of Physiotherapy, School of Health Sciences, Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, Australia; b Centre for Health, Exercise and Sports Medicine, Department of Physiotherapy, School of Health Sciences, Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, Australia; c School of Health Sciences, Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, Australia; d Centre for Implementation Research, Ottawa Hospital Research Institute – General Campus, Ottawa, Canada; e Centre for Digital Transformation of Health, Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, Australia KEY WORDS ABSTRACT Musculoskeletal pain Question: What is the effect of therapeutic exercise or tailored physical activity programs supported by a Exercise mobile app (compared with exercise or physical activity programs delivered using other modes) for people Mobile applications with musculoskeletal pain conditions? Design: Systematic review of published randomised controlled trials Physical therapy with meta-analysis. Participants: People of all ages with musculoskeletal pain conditions. Intervention: Meta-analysis Therapeutic exercise or tailored physical activity programs supported by a mobile app. Outcome measures: Pain intensity, pain interference, self-reported physical function, physical performance, adher- ence, psychosocial outcomes, health-related quality of life, work participation, physical activity, goal attainment and satisfaction. Results: Eleven studies were eligible for inclusion, with a total of 845 partici- pants. There was low certainty evidence that using mobile apps to deliver exercise programs helps to reduce pain intensity to a worthwhile extent (SMD –0.60, 95% CI –0.93 to –0.27). There was low certainty evidence that using mobile apps to deliver exercise programs helps to improve self-reported physical function to a worthwhile extent (SMD –0.92, 95% CI –1.57 to –0.27). Although the effect of using mobile apps to deliver exercise programs on pain interference was also estimated to be a worthwhile benefit (SMD –0.66), this estimate came with marked uncertainty (95% CI –1.52 to 0.19) so the effect remains unclear. The remainder of the outcomes were unclear due to sparse evidence. The most common behaviour change intervention functions in the mobile app interventions were: training, enablement and environmental restructuring. Conclusion: Mobile apps supporting therapeutic exercise or tailored physical activity programs for musculoskeletal pain conditions may help in reducing pain intensity and improving physical function. The mobile apps utilised a limited range of behaviour change intervention functions. Registration: CRD42021248046 [Thompson D, Rattu S, Tower J, Egerton T, Francis J, Merolli M (2023) Mobile app use to support therapeutic exercise for musculoskeletal pain conditions may help improve pain intensity and self-reported physical function: a systematic review. Journal of Physiotherapy 69:23–34] © 2022 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 can be improved by behaviour change interventions.6 Using innova- Musculoskeletal conditions affect approximately 1.7 billion people tive digital solutions to deliver these interventions may help support globally;1 they can cause pain, decrease physical function, reduce exercise and physical activity behaviours.7 psychological well-being, curtail social and economic participation, and are significant predictors of years lived with disability.2 Imaging, Digital health interventions, including mobile apps, use informa- surgery and opioids remain over-utilised in the care of musculo- tion and communication technologies to support healthcare.8 Three skeletal pain,3 despite safe, effective, widely available alternative management options. Therapeutic exercise and tailored physical ac- systematic reviews have recently addressed the effectiveness of dig- tivity are high-value, low-cost interventions for improving symptoms and functional outcomes in musculoskeletal pain conditions,2,4 pro- ital health interventions to support exercise for musculoskeletal pain vided that sufficient adherence is achieved.5 Adherence to exercise conditions;9–11 they found improved adherence in musculoskeletal conditions9 and a beneficial effect on pain, but mixed findings for function and health-related quality of life in people with knee oste- oarthritis.10,11 The reviews reported a limited range of outcomes9–11 and/or focused on a specific condition.10,11 Also, mobile apps were https://doi.org/10.1016/j.jphys.2022.11.012 1836-9553/© 2022 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/).
24 Thompson et al: Mobile apps supporting exercise for musculoskeletal pain Box 1. Nine behaviour change intervention functions and Identification and selection of studies their definitions (adapted from Michie et al, 201115). A comprehensive search of the bibliographic databases Medline Education: increasing knowledge or understanding (Ovid), EMBASE (Ovid), CINAHL complete (EBSCO), Cochrane Central Persuasion: using communication to induce positive or Register of Controlled Trials (CENTRAL), Web of Science, Physio- therapy Evidence Database (PEDro) and Scopus was conducted from negative feelings or stimulate action inception of indexing to 16 March 2021 and updated on 29 January Incentivisation: creating expectation of reward 2022. The search strategy consisted of key words (including exploded Coercion: creating expectation of punishment or cost terms) and Medical Subject Headings, relevant to three key concepts: Training: imparting skills musculoskeletal pain conditions, exercise and mobile apps (see Restriction: using rules to reduce the opportunity to engage Appendix 1 on the eAddenda for the detailed search strategy). Peer review of the search strategy was undertaken by an experienced in the target behaviour (or increase the target behaviour by librarian using the PRESS checklist.21 Clinical trial registries reducing the opportunity to engage in competing behaviours) (Clinicaltrials.org), ProQuest Dissertations and Theses, grey literature Environmental: restructuring the physical or social context databases (TROVE) and reference lists of relevant systematic reviews Modelling: providing an example for people to aspire to or and included clinical trials were also searched. imitate Enablement: increasing means/reducing barriers to increase Screening of each title and abstract for potential eligibility against capability or opportunity pre-published criteria was performed independently by any two of three reviewers (DT, SR and JT) using Covidence softwarea. Full-text grouped with a heterogeneous mix of digital interventions.9–11 articles were each independently screened against the eligibility However, mobile apps have different features to other digital health criteria by any two of three reviewers (DT, SR and JT). Reviewers were interventions that may enhance their ability to deliver instructional not blinded to the authors, journals or results of the studies. Conflicts content and behaviour change interventions.12,13 For example, mobile at each stage were resolved by discussion between the two reviewers apps are delivered via highly portable, accessible devices, making who screened the article, with the third and, where necessary, a support and social networking immediately available. Sophisticated fourth reviewer (MM) assisting in reaching consensus. Article authors monitoring is possible (eg, via in-built sensors and synchronisation were contacted via email if the full-text manuscript was unavailable with wearable devices) and prominent messaging (eg, via notifica- or if clarification was required on aspects of their study. tions and alerts) can effectively deliver information. Therefore, there is a need to investigate their effectiveness more specifically. The inclusion criteria are listed in Box 2. Studies with mixed populations were excluded unless a subgroup of patients meeting the Successful delivery of a therapeutic exercise or tailored physical eligibility criteria were analysed and reported separately. Studies activity program requires a response in the form of adherence to a set of were eligible if their experimental interventions delivered healthcare behaviours;5 therefore, it is worth investigating the ability of mobile professional-prescribed (either directly or via artificial intelligence), apps to deliver behaviour change content designed to foster these tailored, therapeutic exercise or physical activity via a patient-focused behaviours.14 The ‘Behaviour Change Wheel’ is a theory-informed app for mobile phone or tablet. Studies were excluded if: the exper- framework for developing behaviour change interventions and offers imental intervention involved exercise not delivered by an app (eg, a comprehensive model of behaviour change determinants.15,16 delivered via a website); exercise was not a target behaviour (eg, app Behaviour change intervention functions (Box 1) are broad categories for mindfulness); the exercise or physical activity intervention was of the mechanisms underlying behaviour change interventions and can assist with describing interventions in a standardised way.15 Box 2. Inclusion criteria. This review aimed to estimate the effects of using a mobile app to Design support therapeutic exercise or tailored physical activity programs in Randomised controlled trial people with musculoskeletal pain conditions, on pain intensity, pain Published in peer-reviewed journal interference, self-reported physical function, physical performance, Available in English full text adherence, psychosocial outcomes, health-related quality of life, work participation, physical activity levels, goal attainment and satisfac- Participants tion. Pain intensity is the predominant symptom for musculoskeletal Acute, subacute or persistent musculoskeletal pain conditions and its interference with daily activities is of key impor- condition tance to clinicians and patients.17 Since mobile apps may be theorised Any age to affect these outcomes through psychosocial changes,17,18 psycho- social outcomes are also considered to be important. Intervention Tailored therapeutic exercise or physical activity, prescribed This review also aimed to note which behaviour change inter- by healthcare professional vention functions9 are included in the investigated mobile apps. Delivered via mobile app Given that there is currently no consensus on the effectiveness of 6 in-app interventions such as wearable electronic devices within-app behaviour change interventions,19 this review aimed to and education describe interventions in the included studies using the Behaviour 6 additional interventions other than exercise/mobile app Change Wheel taxonomy of intervention functions.15 received by both groups Home or community settings Therefore, the research question for this systematic review was: Outcome measures What is the effect of therapeutic exercise or tailored physical ac- Pain severity tivity programs supported by a mobile app (compared with ex- Pain interference ercise or physical activity programs delivered using other modes) Self-reported physical function for people with musculoskeletal pain conditions? Physical performance Psychosocial outcomes Method Adherence to exercise program Physical activity levels This systematic review is reported according to the Preferred Health-related quality of life Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Work participation guidelines20 and was registered a priori with PROSPERO. Minor Goal attainment modifications to the protocol are described below. Comparator Equivalent therapeutic exercise or physical activity not delivered via mobile app
Research 25 generic and non-tailored (eg, recommending World Health Organi- possible. Group means with standard deviations for within-group zation physical activity guidelines); or if the app was designed for changes, between-group differences after the intervention and/or healthcare professional use (eg, for monitoring). Studies were eligible between-group differences in change scores, with 95% CI and/or if the app included features that were related to the exercise program standard errors, were extracted. Incomplete data were calculated (eg, education or wearable electronic devices) but not education or from the data provided if possible. A pooled estimate of effect was wearable devices alone. Studies were eligible only if any additional calculated using RevMan softwareb if three or more articles reported interventions other than exercise (eg, face-to-face physiotherapy) the same outcome domain. Where fewer than three articles reported were the same in both groups. the same outcome domain, findings were synthesised narratively only, rather than also being pooled quantitatively. Random effects Eligible comparator groups received similar exercise or physical models were used to determine weighted, standardised mean dif- activity interventions as the experimental group, but without mobile ferences (SMD) between groups immediately after the intervention app support, such that the mobile app was the primary point of and at the longest follow-up. Weighted mean differences (MD) were difference between the groups. Similarity of exercise or physical ac- calculated if the same measurement tool was used for the domain. tivity interventions between groups was judged in terms of type, Adherence was evaluated by comparing actual adherence to a target frequency, duration and intensity of exercise. Studies were excluded if level. This method was not planned a priori; however, given the ways they had an inactive comparator (eg, exercise and mobile app versus a that adherence data were reported in the studies, this method was no-exercise attention comparator), more intensive face-to-face considered most appropriate. Subgroup analyses according to age comparator conditions, or inadequately described comparator con- group, condition, risk of bias and the mobile app’s behaviour change ditions. This was changed from the registered protocol, which also intervention functions were planned, where possible. Data were included similar intensity ‘self-management’ interventions as eligible narratively synthesised by outcome, considering the magnitude and comparators. The change was made to increase directness of the re- precision of between-group differences in the context of risk of bias view by ensuring that exercise/physical activity components of the and clinical heterogeneity. interventions were similar. The smallest worthwhile effect of using an app to deliver an ex- Only studies measuring quantitative outcomes of interest (Box 2) ercise program is unclear. Apps are generally cheap and smartphones were included. A broad operational definition of physical perfor- are relatively common worldwide. Downloading an app is quick and mance was used, encompassing physical function (mobility, dexterity, using an app is arguably more convenient than using an exercise axial ability and ability to carry out instrumental activities of daily program prescribed on paper. Hence, in this context, any between- living)22 and physical fitness (cardiovascular endurance, muscular group difference might be seen as worthwhile. However, for people strength, muscular endurance, flexibility, body composition).22 Any who would need to invest in a smartphone and/or increase their measure of adherence to exercise was included (eg, self-reported digital literacy in order to use the app, an important difference might number of sessions, number of exercises, duration of exercise) and be larger. We nominated that a difference of 1.5 points on a 10-point physical activity included self-reported and objective measures. Work pain scale or an SMD of 0.2 would be considered worthwhile. participation included length of time to return to work and work limitations. Certainty of findings Assessment of characteristics of studies The certainty of the body of evidence for each outcome domain was assessed by any two of three reviewers (DT, JT or SR) using the Clinical characteristics GRADE approach.25 The risk of bias due to missing results in the The study characteristics extracted to enable consideration of synthesis was assessed by determining missing results from included studies, compared with published protocols, and the likelihood of clinical heterogeneity were: participant characteristics (age, sex/ publication bias was evaluated via qualitative signals, since graphical gender, condition, location and healthcare context); experimental and statistical methods were not appropriate.26 and comparator group exercise/physical activity program; details of the experimental app and other components in the interventions; Results relevant outcome measures; and measurement time points. Data describing study design, sources of funding and conflicts of interest Flow of studies through the review were also extracted. The search identified 7,978 unique titles and abstracts (Figure 1). Behaviour change components Eleven articles (10 studies) were included in the review. Reasons for The active behaviour change components of the mobile apps were exclusion during full-text screening are presented in Appendix 2 on the eAddenda. categorised independently according to the nine behaviour change intervention functions15 (Box 1) by any two of three reviewers (DT, SR or JT) for each included study. Differences were resolved through discussion and involvement of the third reviewer where needed. Risk of bias Characteristics of studies The risk of bias was independently assessed at the outcome level Study characteristics are summarised in Table 1. All studies were by any two of three reviewers (DT, SR or JT) for each included study published since 2016. The included studies randomised a total of 845 using the Cochrane Risk of Bias tool version 2 (ROB2).23 Differences participants. Sample sizes of included studies ranged from 2027 to were resolved through discussion, with involvement of a third 22028,29 participants. Mean ages ranged from 2727 to 6330 years. reviewer where necessary. Risk of bias was graphically represented Participants had persistent musculoskeletal conditions in eight of the using the ‘robvis’ tool.24 Where insufficient detail was published, an studies, and acute injuries in three studies. Conditions included knee attempt was made to contact study authors to obtain further osteoarthritis,31–33 low back pain,34,35 ankle sprains,28,29 frozen information. shoulder,36 neck pain27,35 and wrist, hand and finger injuries.37 Data extraction and analysis Exercise programs varied in their therapeutic focus, with muscular strengthening exercises being the most frequently prescribed.28–31,34 Outcome data from each included article were independently Other exercise programs included finger/hand mobilisation exercises extracted by any two of three reviewers (SR, DT or JT) for the time- for wrist/hand/finger injuries,37 passive range of motion exercises for points baseline, immediately after the intervention and at the longest frozen shoulder,36 ‘McKenzie’ neck exercises (range of motion and available follow-up. Data were cross-checked for differences, tabu- stretching) for neck pain,27 whole body physical activity,34 activities of lated by outcome domain and sub-grouped by condition where daily living,34 balance and proprioception,28,29 stretching30,35 and
26 Thompson et al: Mobile apps supporting exercise for musculoskeletal pain Figure 1. Flow of trials through the review.
Table 1 Participants Experimental Interve Study characteristics. Con N = Exp 20, Con 20 Exercises: Global postural Study Age (y) = Exp 41 (8), Con 40 (8) re-education Exercises: Sam Sex (%F) = Exp 50, Con 50 Technology: Smartphone, app. Abadiyan Chronic non-specific neck pain, . 3 mths ‘Seeb’ app 202132 duration, VAS 3 to 8 Exercises: Sam N = Exp 20, Con 20 Exercises: Progressive lower paper-based H Alasfour Age (y) = Exp 54 (4), Con 55 (5) limb strengthening, daily, 10 app 202031 Sex (%F) = Exp 100, Con 100 reps/exercise Unilateral or bilateral knee osteoarthritis, Technology: Tablet or Exercises: Stre Anan 6 mths duration, NPRS 7 smartphone, ‘MyDearKnee’ during break t 202135 app N = Exp 61, Con 60 Exercises: Chat-bot prompts Exercises: Sam Blanquero Age (y) = Exp 42 (9), Con 42 (8) for stretching and paper-based H 202037 Sex (%F) = Exp 19, Con 28 movement exercises daily app Neck/ shoulder pain/stiffness or low back Technology: Smartphone, Chhabra pain or both Secaide Ver.0.9TM app Exercises: Sam 201834 N = Exp 40, Con 34 Exercises: Clinician- paper-based H Age (y) = Exp 45 (11), Con 42 (11) prescribed exercises, app Chitkar Sex (%F) = Exp 32, Con 44 progressed by algorithm 202133 Bone and soft tissue injuries of wrist/ Technology: Tablet, ‘ReHand’ Exercises: Sam hand/finger that limit functional ability app face-to-face se Choi N = Exp 45, Con 48 Exercises: Physical activity 201936 Age (y) = Exp 41 (14), Con 41 (14) program, back exercises, Exercises: Sam Sex (%F) = NS activities of daily living exercise progr Lee Mechanical LBP 6 radicular symptoms, Technology: Smartphone, verbally 201727 . 12 wks duration, NPRS 5 ‘Snapcare’ app N = Exp 32, Con 32 Exercises: Exercises for knee Exercises: Nec Thiengwittayaporn Age (y) = Exp 59 (9), Con 58 (6) OA, unclear description brochure (incl 202130 Sex (%F) = Exp 100, Con 100 Technology: Smartphone program) Radiologically confirmed, symptomatic Other: 1 neck p knee OA Exercises: Passive range of session N = Exp 42, Con 42 motion stretches (forward Exercises: Sam Age (y) = Exp 54 (8), Con 55 (6) flexion, external rotation, paper-based H Sex (%F) = Exp 61, Con 73 cross-body adduction, app Frozen shoulder, 1 mth duration sleeper stretch). Technology: Smartphone N = Exp 11, Con 9 Exercises: McKenzie neck Age (y) = Exp 27 (5), Con 28 (5) exercise program, 2/wk. Sex (%F) = Exp 55, Con 45 Technology: Smartphone, Self-reported chronic neck pain, . 6 bespoke app mths duration, VAS 3 N = Exp 44, Con 45 Exercises: Lower limb Age (y) = Exp 62 (7), Con 63 (10) strengthening and Sex (%F) = Exp 86, Con 93 stretching, 3 exercises, 10 Unilateral or bilateral primary knee OA times each Technology Smartphone, ‘Love your knee’ app
entions Both Outcome(s) Time points Research 27 ntrol Neck pain brochure, PT 50 me as Exp, nil min/d, 4 d/wk for 8 wks VAS, NDI, PILE, SF-36 Baseline, 8 wks (post- intervention) me as Exp, Education, 2 supervised Arabic NPRS, WOMAC, 5 3 HEP in lieu of exercise sessions, resistance sit-stand test, adherence Baseline, 3 wks (mid- band (self-report logbook) intervention), 6 wks (post- etches 3 min/d intervention) time PT and OT 3 d/wk, 30 to Pain intensity (1 to 5), 60-min session difference in worst pain Baseline, 12 wks (post- me as Exp, score, absence of severe intervention) HEP in lieu of Written prescription from pain, subjective physician (medication, improvement Baseline, 2 wks (mid- me as Exp, physical activity, home VAS, QuickDASH, grip intervention), 4 wks (post- HEP in lieu of exercises) strength, pinch strength, 9- intervention) Education (Knee OA causes, hole peg test, Time taken to me as Exp, 2 3 risk factors, treatment, diet return to work Baseline, 12 wks (post- essions and exercises) NPRS, MODI intervention) me as Exp, Celecoxib NSAID WOMAC, SF-36 Baseline, 8 wks (post- ram delivered intervention) Weekly text message VAS, shoulder range of ck pain reminders motion (flexion, abduction, Baseline, 4 wks (mid- luding exercise external rotation at side, intervention), 8 wks (mid- pain education internal rotation at back) intervention), 12 wks (post- me as Exp, intervention) HEP in lieu of VAS, NDI, exercise adherence Baseline, 8 wks (post- intervention) ROM, KOOS, KSS Baseline, 4 wks (post- intervention)
28 Thompson et al: Mobile apps supporting exercise for musculoskeletal painTime pointsBaseline, weekly, 8 wks Con = control group, Exp = experimental group, FADI = Foot and Ankle Disability Index, HEP = home exercise program, KOOS = Knee Injury and Osteoarthritis Outcome Score, KSS = Knee Society Score, MODI = Modified Oswestry Disability Index, NDI = (post-intervention), monthly Neck Disability Index, NPRS = numerical pain rating scale, NS = not stated, NSAID = non-steroidal anti-inflammatory drugs, OA = osteoarthritis, OT = occupational therapy, PILE = progressive isoinertial lifting evaluation, PT = physiotherapy, QuickDASH = postural education32; or were not clearly described.33 The interventionuntil 12 mthsQuick Disabilities of the Shoulder, Arm and Hand questionnaire, SF-36 = Short Form 36 quality of life questionnaire, VAS = visual analogue scale, WOMAC = Western Ontario and McMaster Universities Arthritis Index, %F = percent female. durations ranged from 430,37 to 12 weeks34–36 (median 8 weeks). The dosage of exercise programs varied significantly between trials withOutcome(s)FADI, adherence frequency ranging from 2 to 3 times daily36 to ‘at least twice per week’.27 Additional components included education, face-to-faceBothBalance board physiotherapy and occupational therapy, prescription of medication, and text message reminders. Most studies delivered the mobile app viaInterventionsControlExercises: Same as Exp, mobile phones, one study delivered the mobile app via a tablet device37paper-based HEP in lieu of and one study allowed either device to be used.31 Comparator groupsapp received the same therapeutic exercise programs without the use of a mobile app in eight studies28–33,36,37 and a similar tailored therapeuticExperimentalExercises: Lower limb exercise or physical activity intervention in three studies.27,34,35strengthening, balance and proprioception Behaviour change intervention functions were classified for all(neuromuscular training). studies33 (Table 2) (see Appendix 3 on the eAddenda for further de-Technology: Smartphone, tails). All mobile app interventions incorporated ‘training’. ‘Enable-‘Strengthen your ankle’ app ment’ (six studies) and ‘environmental restructuring’ (five studies) were also commonly utilised. ‘Education’, ‘persuasion’ and ‘incentiv-ParticipantsN = Exp 110, Con 110 isation’ were each used in single studies and no study used ‘coercion’,Age (y) = Exp 38 (13), Con 38 (14) ‘restriction’ or ‘modelling’. All comparator interventions incorporatedSex (%F) = Exp 50, Con 50 training and two incorporated ‘enablement’.Self-reported ankle sprain in last 2 mths Table 1 (Continued) Risk of bias assessmentStudyVan Reijen 2016, 201728,29 All included studies were assessed as having a high risk of bias overall, resulting from high risk of bias or some concern in at least three of five ROB2 domains (Figure 2). Six studies28–31,33,35,37 had high risk of bias due to missing outcome data (Domain 3) because of high participant drop-out without appropriate analysis. Only two studies28,29,31 followed published a priori protocols, leading to some concerns for the majority of included studies in Domain 5 (reporting of results). Comparison of the results of included studies against available protocols was attempted to determine missing results (see Appendix 4 on the eAddenda). This was not possible for most studies as no protocol was published,27,33,34,37 the protocol was not published a priori32 or the protocol was unavailable.35,36 One study added additional outcomes after publication of the protocol.30 The remain- ing studies followed their published protocol.28,29,31 Outcome measures such as pain severity and pain interference were self-reported so the participants were the assessors and unable to be blinded to allocation (Domain 4), thus potential for bias in the domain related to measurement of outcomes was inevitable. In real- world implementation of mobile app interventions, patients will be aware of their purpose and therefore this source of bias will exist in practice as a contextual factor in mobile app intervention effective- ness. Risk of bias for each of the review’s outcomes are presented in more detail in Appendix 5 on the eAddenda. Effects of mobile app interventions Pain intensity and pain interference A meta-analysis of the effect of mobile apps to support therapeutic exercise or physical activity programs on pain intensity immediately after the intervention pooled the SMD of nine trials (Table 3, Figure 3) and demonstrated benefit in favour of mobile apps, with substantial heterogeneity (SMD –0.60, 95% CI –0.93 to –0.27, I2 = 70%). For a detailed forest plot, see Figure 4 on the eAddenda. No studies measured pain intensity beyond immediately after the intervention. A subgroup analysis of studies in knee osteoarthritis populations showed significant benefit, with substantial heterogeneity (SMD –0.82, 95% CI –1.45 to –0.18, I2 = 74%). The certainty of the body of evidence for pain intensity was low using the GRADE approach,25 after downgrading for risk of bias and inconsistency (see Appendix 6 on the eAddenda). The three studies measuring pain interference reported mixed results immediately after the intervention. Combining these studies in a meta-analysis again produced an estimate of substantial benefit (SMD –0.66); however, that estimate came with so much uncertainty (95% CI –1.52 to 0.19) that the true effect remained very uncertain (Figure 5); there was also considerable heterogeneity (I2 = 80%). For a detailed forest plot, see Figure 6 on the eAddenda. The GRADE
Research 29 Table 2 Comparison of intervention functions behaviour change between experimental (mobile app) and comparator interventions. Study Group Educationa Persuasion Incentivisation Coercion Training Restriction Environmental restructuring Modelling Enablement U Abadiyan Exp UU U 202132 Con U U Alasfour Exp U UU 202031 Con U U Anan Exp U U 202135 Con U U Blanquero Exp UU U 202037 Con U U Chhabra Exp UU U 201834 Con U Chitkar Exp U 202133 Con U Choi Exp UU 201936 Con U Lee Exp U U 201727 Con U Thiengwittayaporn Exp U U 202130 Con Van Reijen Exp U 2016, 201728,29 Con U Con = control group, Exp = experimental group. a Refers to increasing knowledge or understanding regarding the targeted exercise behaviour. certainty of the body of evidence for pain interference was very low Randomisation process eAddenda.) This estimate had considerable heterogeneity (I2 = 88%). after being downgraded for risk of bias, inconsistency and impreci- Deviations from intended interventions Subgroup analysis of studies in knee osteoarthritis populations fav- sion (see Appendix 6 on the eAddenda). Missing outcome data oured mobile apps (SMD –1.33). Although there was some uncer- Measurement of the outcome tainty in the estimate, all values contained with the confidence Self-reported physical function and physical performance Selection of the reported result interval indicated worthwhile benefits (95% CI –2.06 to –0.60). Again, The pooled effect of five studies reporting self-reported physical Overall there was heterogeneity among the studies (I2 = 78%). The GRADE certainty of the body of evidence for self-reported physical function function (Table 4, Figure 7) favoured mobile apps (SMD –0.92) and was low after being downgraded for risk of bias and inconsistency. this appeared to be a clinically worthwhile effect (95% CI –1.57 to Van Reijen et al29 also measured ankle function longer term (10 –0.27) (see Figure 7, or Figure 8 for a detailed forest plot on the months after the intervention) and found negligible between-group difference (MD 0.7, 95% CI –1.2 to 2.6) on the 0-to-100 FADI scale. Low risk of bias Some concerns Physical performance outcome measures were included in six High risk of bias studies27,30–32,36,37 (Table 5). Meta-analyses were not performed due to the heterogeneity of outcomes. Of the 12 outcome measures re- Abadiyan 202132 ported, only one showed a clear between-group difference.30 The Alasfour 202031 GRADE certainty of the body of evidence for physical performance was low after being downgraded for risk of bias and imprecision Anan 202135 (Appendix 6). Blanquero 202037 Adherence Chhabra 201834 Between-group differences in adherence to the prescribed exer- Chitkar 202133 Choi 201936 cise program were reported in two studies28,31 (Table 6). Delivering Lee 201727 exercises for knee osteoarthritis via a mobile app achieved better adherence than provision of printed materials.31 There was similar Thiengwittayaporn 202130 adherence between mobile app-delivered and booklet-delivered ex- Van Reijen 2016, 201728,29 ercises for ankle sprain.28 One study measured adherence in the experimental group only and found that both the number of exercise Figure 2. Risk of bias at overall study level. sessions and the exercise time exceeded their target value.27 The GRADE certainty of the body of evidence for adherence was very low after downgrading for study limitations, inconsistency and imprecision. Psychosocial outcomes One study found that the time taken to return to work was a mean of 18 days shorter (95% CI 3 to 33) in the experimental group (Table 6), which was argued by the original authors to be a worth- while effect.37 The GRADE certainty of evidence was very low after downgrading one level for study limitations and two levels for imprecision. One study reported fear avoidance beliefs27 and found no clear between-group difference (Table 6). The GRADE classification of this outcome was very low after downgrading once for study limitations and twice for imprecision.
30 Thompson et al: Mobile apps supporting exercise for musculoskeletal pain Table 3 Results for pain intensity and pain interference measures. Study Outcome measure Post-intervention mean (SD) Between-group difference mean (95% CI) Exp Con Pain intensity Pain VAS (0 to 10) 4.4 (1.7) 5.8 (1.1) –1.4 (–2.3 to –0.5) Abadiyan 202132 Pain NPRS (0 to 10) 3.6 (2.1) 5.2 (2.4) –1.6 (–3.1 to –0.1)a Alasfour 202031 Pain scale (1 to 5) 3.0 (1.1) 4.0 (0.8) –1.0 (–1.4 to –0.6) Anan 202135 Pain VAS (0 to 10) 2.7 (1.7) 3.6 (2.0) –0.9 (–2.0 to 0.2) Blanquero 202037 NPRS (0 to 10) 3.3 (1.7) 3.2 (2.7) Chhabra 201834 WOMAC – Pain (0 to 20) 11.8 (1.4)b 13.8 (1.4)b 0.1 (–0.8 to 1.0) Chitkar 202133 Pain VAS (0 to 10) 1.8 (2.5) 2.2 (1.7) –2.0 (–2.2 to –1.8) Choi 201936 Pain VAS (0 to 10) 2.7 (2.0) 3.7 (2.0) –0.4 (–1.3 to 0.5) Lee 201727 KOOS – Pain (0 to 100) 73.3 (7.2) 70.7 (5.9) –1.0 (–2.8 to 0.8) Thiengwittayaporn 202130 –2.6 (–5.4 to 0.2) Pain interference NDI (0 to 50) 19.3 (6.0) 28.5 (5.3) –9.2 (–12.8 to –5.6) Abadiyan 202132 MODI (0 to 50) 20.2 (17.8) 29.9 (20.1) –9.7 (–17.4 to –2.0) Chhabra 201834 NDI (0 to 50) 17.3 (8.3) 15.9 (8.7) Lee 201727 1.4 (–6.1 to 8.9) KOOS = Knee Injury and Osteoarthritis Outcome Score, MODI = Modified Oswestry Disability Index, NDI = Neck Disability Index, NPRS = Numerical Pain Rating Scale, VAS = Visual Analog Scale, WOMAC = Western Ontario and McMaster Universities Arthritis index. a Calculated from published means (SD). b SD calculated from published SE. Health-related quality of life people with musculoskeletal conditions, as well as avenues for Two studies reported health-related quality of life, using the SF- further investigation and development in this field. Meta-analyses found low certainty evidence that mobile app-supported exercise 3627,33 (Table 7). Physical functioning, bodily pain and vitality sub- interventions achieved greater improvements in pain intensity and scales were estimated to be better in the mobile app group in one self-reported physical function but no difference in pain interference study33 but there were no clear between-group differences in any compared with comparable exercise interventions without mobile subscale in another study.27 The certainty of this evidence using app support. For the other outcomes, the limited data from low- GRADE was very low after downgrading for study limitations and quality studies showed mixed results for adherence and health- imprecision. related quality of life, no benefit for fear avoidance beliefs, a reduc- tion in the time taken to return to work, and no data for physical Goal attainment, physical activity and satisfaction activity, goal attainment or satisfaction. None of the included studies measured goal attainment, physical Our meta-analyses suggest that using mobile apps to support activity levels or satisfaction in both experimental and comparator exercise may lead to a greater reduction in pain intensity, although groups. the low certainty of evidence precluded strong recommendations from being made. The SMD of –0.6 (95% CI –0.93 to –0.27) appears to Discussion be a worthwhile improvement. When converted to the Numerical Pain Rating Scale, this equates to an improvement of 1.4 points (95% This systematic review identified potential benefits of using mo- CI 0.6 to 2.2), which would be worthwhile for regular smartphone bile apps to support therapeutic exercise and physical activity for users. However, for those who would have to obtain a smartphone and/or improve their digital literacy, this result spans the smallest Subgroup SMD worthwhile effect that we nominated (1.5 points) or that others have Study (95% CI) nominated (2 points) for this outcome,38 indicating uncertainty about whether the effect would be large enough to justify the use of an app Other conditions to deliver the exercise intervention. The substantial statistical het- Abadiyan³² erogeneity may reflect differences in populations between studies Anan³⁵ and the content of the mobile apps, in addition to biases due to study Blanquero³⁷ design. The modest reduction in pain intensity with the use of mobile Chhabra³⁴ apps found in this review is consistent with previous systematic re- Choi³⁶ views investigating a wider range of digital health interventions in Lee²⁷ knee osteoarthritis populations.11,39 Subtotal The main estimate of the effect of mobile app-supported exercise Knee OA on pain interference was that it has worthwhile benefit (SMD –0.66) Alasfour³¹ but this estimate was so imprecise (95% CI –1.52 to 0.19) that the true Chitkar³³ Thiengwittayaporn³º Study SMD Subtotal (95% CI) Abadiyan³² Chhabra³⁴ Lee²⁷ Total –2 –1 0 1 2 –2 –1 0 1 2 Favours exp Favours con Favours exp Favours con Figure 3. Forest plot of the results of random effects meta-analysis on the effect of Figure 5. Forest plot of the results of random effects meta-analysis on the effect of intervention on pain intensity using post-intervention scores, demonstrated via SMD intervention on pain interference using post-intervention scores, demonstrated via (95% CI). OA = osteoarthritis. SMD (95% CI).
Research 31 Table 4 Results for self-reported physical function. Study Outcome measure Post-intervention mean (SD) Between-group difference mean (95% CI) Exp Con Alasfour 202031 WOMAC – PFa (0 to 28)c 3.0 (1.9) 5.2 (3.2) 2.20 (0.4 to 3.96)b Blanquero 202037 QuickDASH (0 to 100)c 26.6 (16.8)d 37.7 (23.4)d 11.1 (–0.9 to 23.1)d Chitkar 202133 WOMAC – PF (0 to 68)c 41.0 (2.1)e 45.4 (2.0)e Thiengwittayaporn 202130 KOOS – ADL (0 to 100)f 4.9 (3.9 to 5.9) Van Reijen 201628 (post-intervention) FADI (0 to 100)f 80.4 (9.8) 71.2 (7.0) 9.2 (5.5 to 12.9) Van Reijen 201729 (10 mths) FADI (0 to 100)f 1.3 (–0.6 to 3.2) 95.9 (6.4) 94.6 (6.9) 0.7 (–1.2 to 2.6) 97.2 (8.0) 97.9 (4.5) FADI = Foot and Ankle Disability Index, KOOS = Knee Injury and Osteoarthritis Outcome Score, PF = Physical Function subscale, QuickDASH = Shortened Disabilities of the Arm to Shoulder and Hand Questionnaire, WOMAC = Western Ontario and McMaster Universities Arthritis Index. Between-group differences adjusted so positive difference indicates greater benefit in experimental group. a Arabic version of the reduced WOMAC – Physical Function subscale. b Calculated from published means (SD). c Higher scores indicate greater disability. d Re-calculated from individual participant data. e SD calculated from published SE. f Higher scores indicate less disability. effect remains unclear. The evidence was also very low certainty: outcomes could reflect the different aspects of the patient experience there were only three studies reporting this outcome, and one of the that are being measured, differences in population, or limitations in studies27 showed a very marked baseline between-group difference, the quality and quantity of the body of research for this outcome. which could have influenced the result. Physical performance measures were highly variable, encom- The meta-analysis for self-reported physical function also showed passing strength, dexterity and range of motion. However, the find- worthwhile benefit in favour of using mobile apps, albeit with low- ings were remarkably consistent, with only one of 12 physical certainty evidence; this was considered worthwhile for regular performance outcomes demonstrating greater improvement in the smartphone users. For others, the SMD of –0.92 equated to 7.8 points mobile app group. This is consistent with previous literature,42 sug- on the Knee Injury and Osteoarthritis Outcome Score, which was gesting that mechanisms underlying the improvements in pain and close to the estimated minimum important difference of 8 reported in disability with exercise in chronic musculoskeletal pain may be due a recent systematic review.40 This benefit was greater in the subgroup to factors other than improvements in physical performance, such as of studies with knee osteoarthritis populations. In contrast, a recent psychological and/or neurophysiological changes.42 systematic review, which looked at a broader range of technology to support exercise for people with knee osteoarthritis,11 found no clear Only two studies evaluated the effect of mobile apps on adherence between-group difference in self-reported physical function. This to exercise.28,31 One showed no benefit for people with ankle could indicate that the use of mobile apps is more effective than other sprains28 and the other found a clear benefit on adherence in women types of technology in achieving this outcome. However, further with knee osteoarthritis.31 Two of three studies with adherence data research addressing the methodological concerns of the body of ev- for the experimental intervention group found that adherence idence to date would be needed to confirm this conclusion. exceeded the threshold for satisfactory adherence to exercise of 80% proposed by Bailey et al.4 Despite adherence to exercise being the- Pain interference (the degree to which an individual’s engagement orised as a key mechanism influencing overall health outcomes,4 less in activities, including physical activities, is restricted) and physical than one-quarter of studies reported this outcome. Furthermore, no function (the ability to perform physical activities) both measure studies measured self-reported or objectively measured physical ac- aspects of physical ability. However, pain interference also includes tivity levels. From a biopsychosocial perspective, physical activity cognitive, emotional and recreational activities, and more specifically participation is key to successful management of chronic disease, measures the impact of pain, whereas physical function may also be rather than being an end result of medical treatments that resolve impacted by factors other than pain.41 The apparent difference in the disease or symptoms.43 Measurement of adherence to exercise and results for pain interference and self-reported physical function physical activity levels is strongly recommended in future studies of interventions to support behaviour change in musculoskeletal pain Subgroup SMD conditions, including mobile apps. Future app design incorporating Study (95% CI) adherence-reporting features, either through self-report or incorpo- rating data from wearable devices, could enhance research data Other conditions collection and act as an additional behaviour change-focused inter- Blanquero³7 vention (eg, through feedback on performance). Van Reijen²9 Subtotal There was very low certainty evidence of negligible between- group difference for fear avoidance beliefs and benefit in favour of Knee OA mobile apps for work participation from single studies, and health- Alasfour³¹ related quality of life showed mixed results. Psychosocial factors Chitkar³³ (such as anxiety, depression, illness beliefs, self-efficacy, catastroph- Thiengwittayaporn³º ising and coping strategies) are also important predictors of the Subtotal course of musculoskeletal pain17,44 and associated with lower adherence.45 Further elaborating the effect of mobile apps on these –2 –1 0 1 2 factors may reveal important mechanisms of action. Favours exp Favours con Using intervention functions listed on the Behaviour Change Wheel to help understand the mechanisms of action46 of mobile apps Figure 7. Forest plot of the results of random effects meta-analysis on the effect of was an important secondary aim of this review. ‘Training’ in perfor- intervention on self-reported physical function using post-intervention scores, mance of the exercises (provided in all mobile app and control in- demonstrated via SMD (95% CI). Higher mean values indicate greater disability. terventions), ‘enablement’ (eg, providing social support through OA = osteoarthritis. encouragement and setting goals) and ‘environmental restructuring’ (eg, alarms and reminders) were used frequently in the mobile apps. In contrast, ‘persuasion’, ‘education’ and ‘incentivisation’ were used only in single studies and no studies incorporated ‘modelling’ (which
32 Thompson et al: Mobile apps supporting exercise for musculoskeletal pain Table 5 Results for pain intensity and pain interference measures. Study Outcome measure Post-intervention mean (SD) Between-group difference mean (95% CI) Abadiyan 202132 Progressive isoinertial lifting evaluation (lb) Exp Con –6.5 (–17.0 to 4.0)a Alasfour 202031 Five-times sit to stand test (n) –0.2 (–1.0 to 0.60)a Blanquero 202037 Nine-hole peg test (s) 60.2 (17.6) 66.7 (15.3) Grip strength (kg) 2.6 (1.1) 2.8 (1.3) 3.0 (–0.9 to 6.9)a Choi 201936 Pinch strength (kg) 25 (5) 28 (8) 2.8 (–4.3 to 9.9)a Shoulder flexion ROM (deg) 22.9 (8.8) 0.4 (–1.9 to 2.7)a Lee 201727 Shoulder abduction ROM (deg) 5.6 (3.7) 20.1 (14.6) 4.0 (–2.2 to 10.2)a Thiengwittayaporn 202130 Shoulder external rotation ROM (deg) 146 (15.0) 5.2 (3.8) 2.0 (–5.3 to 9.3)a Shoulder internal rotation ROM (deg) 151 (18.0) 142 (14.0) 6.0 (–0.4 to 12.4)a Neck flexion strength (kg) 47 (15.0) 149 (16.0) –0.9 (–2.1 to 0.3)a Neck extension strength (kg) 11.7 (2.6) 41 (15.0) –2.6 (–7.5 to 2.3)a Knee flexion ROM (deg) 13.2 (5.1) 12.6 (2.9) 4.8 (–2.5 to 12.1)a 25.9 (6.9) 15.8 (5.9) 3.1 (0.5 to 5.7)a 21.1 (9.3) 129.0 (6.5) 125.9 (5.6) ROM = range of motion. Higher scores indicate better performance. Between-group difference adjusted so positive difference indicates greater benefit in experimental group. a Calculated from published means (SD). Table 6 Results for adherence and psychosocial outcomes. Study Outcome measure Post-intervention mean (SD) Between-group difference mean (95% CI) Exp Con 25 (8 to 42)a Not calculable Adherence % of prescribed exercises completed (self-report) 85 (14) 60 (33) Not calculable Alasfour 202031 % of prescribed exercise sessions completed 119 (44) Not measured –2 (–15 to 11)a Lee 201727 % of prescribed exercise time completed 134 (89) Not measured % of prescribed exercises completed (self-report) 64 (45) –18 (–33 to –3) Van Reijen 201628 66 (45) 0.78 (–4.6 to 6.1) Psychosocial outcomes Time to return to work (d) 76 (33) 3.6 (–3.7 to 11.0) Blanquero 202037 FABQ – Physical activity (0 to 24) 12.0 (5.7) 94 (32) Lee 201727 FABQ – Work (0 to 42) 20.7 (6.4) 11.2 (5.6) 17.1 (9.3) FABQ = Fear Avoidance Beliefs Questionnaire. Between-group difference adjusted so positive difference indicates greater benefit in experimental group. a Calculated from published means (SD). Table 7 Results for health-related quality of life. Study Short Form-36 subscale (0 to 100) Post-intervention mean (SD) Between-group difference mean (95% CI) Chitkar 202133 Physical functioning Exp Con 9 (6 to 12)b Lee 201727 Role: physical 0 (0 to 0)b Bodily pain 42 (7)a 32 (7)a 4 (0 to 9)b General health 13 (0)a 13 (0)a 1 (–2 to 4)b Vitality 50 (8)a 46 (8)a 10 (6 to 13)b Social functioning 29 (6)a 28 (6)a 1 (–4 to 6)b Role: emotional 67 (7)a 57 (7)a 13 (1 to 26)b Emotional well-being 41 (10)a 40 (10)a 0 (0 to 0)b Physical functioning 30 (25)a 17 (25)a –5 (–19 to 10)c Role: physical 59 (0)a 59 (0)a 5 (–19 to 29)c Bodily pain 90 (19) 94 (10) –1 (–14 to 13)c General health 78 (28) 73 (23) –7 (–30 to 16)c Vitality 68 (15) 69 (16) 5 (–12 to 22)c Social functioning 50 (31) 57 (12) 10 (–7 to 27)c Role: emotional 57 (19) 52 (16) 4 (–35 to 44)c Emotional well-being 81 (21) 72 (14) 1 (–17 to 19)c 73 (41) 69 (43) 65 (20) 64 (18) Between-group differences adjusted so positive differences indicate greater benefit in the experimental group. a Standard deviation calculated from published standard error. b Between-group difference calculated from published means and standard deviations as calculated from standard errors. c Calculated from published means (SD). may be appropriate in this setting), ‘coercion’ or ‘restriction’ (which This systematic review adhered to the recommendations for conducting and reporting systematic reviews.20 Focusing on mobile are perhaps not appropriate). Incorporating ‘education’, ‘incentivisa- app interventions specifically, rather than all digital health in- tion’ and ‘persuasion’ intervention functions, such as through the use terventions, enabled a more focused exploration of the unique effects of notifications, badges or rewards, could provide alternative avenues of apps. The broad population reflected clinical practice and opti- to improve motivation46 and address barriers such as poor self- mised external validity and generalisability. It was judged appropriate efficacy and low self-motivation known to negatively impact adher- to group all musculoskeletal pain conditions together, as the clinical ence in musculoskeletal pain conditions.45 Modelling, such as the course and prognostic factors are likely to be similar.3 Narrowing the incorporation of story-telling with real-life examples, could also definition of exercise to include only clinician-prescribed exercise or enhance motivation.46
Research 33 physical activity tailored to the individual and their condition was Footnotes: a Covidence systematic review software, Veritas Health intended to mirror health professionals’ clinical practice, rather than broader public health interventions. Narrowly defined eligibility Innovation, Melbourne, Australia criteria for comparators (ie, including only similar exercise/physical b Review Manager V.5.3, The Nordic Cochrane Centre, Copenha- activity interventions and excluding usual care, inactive controls or other non-exercise programs) ensured that between-group differ- gen, Denmark ences could be attributed to the mobile app. The comprehensive search strategy minimised the risk of studies being missing; however, eAddenda: Figures 4, 6 and 8, and Appendices 1 to 6 can be found at least one non-English language study was omitted due to inability to obtain a translation. The ROB2 has been criticised due to poor online at https://doi.org/10.1016/j.jphys.2022.11.012 clinimetric properties;47,48 however, it is still preferred due to its measurement of risk of bias rather than methodological quality or Ethics approval: Nil. adherence to reporting guidelines. ROB2 was assessed independently by two reviewers, followed by discussion of discrepancies, to improve Competing interests: Nil. accuracy of risk of bias results. The emergent nature of this field may be subject to time-lag bias, increasing the possibility that useful Sources of support: Nil. studies are as yet unavailable.26 This systematic review also needs to be considered in the context of changes in technology over time and Acknowledgements: Tania Celeste peer reviewed the search the rapidly evolving nature of research in this area; therefore, it would be good to update this review in the near future. strategy using the PRESS checklist. Categorising the behaviour change content of the mobile apps Provenance: Not invited. Peer reviewed. using behaviour change intervention functions15 provided an over- view of their behaviour change content and was appropriate for the Correspondence: Mark Merolli, Physiotherapy Department, The level of detail likely to be found in published articles at this time. This review did not assess the quality of operationalisation of the behav- University of Melbourne, Australia. Email: [email protected] iour change intervention functions, only their presence or absence. Therefore, the behaviour change categorisation undertaken should be References interpreted cautiously, as a broad, preliminary exploration of behav- iour change content. Future research could investigate other aspects 1. Organization WH. Musculoskeletal conditions. https://www.who.int/news-room/ of mobile app-based behaviour change content, such as behaviour fact-sheets/detail/musculoskeletal-conditions. Published 2021. Accessed June 25, change techniques. Future mobile app interventions need to be better 2021. described in terms of their behaviour change components, to enable future reviews to determine the mechanisms of action associated 2. Cieza A, Causey K, Kamenov K, Hanson SW, Chatterji S, Vos T. Global estimates of with the best outcomes (eg, intervention design could be described in the need for rehabilitation based on the Global Burden of Disease study 2019: a terms of behaviour change models such as the Capability, Opportu- systematic analysis for the Global Burden of Disease Study 2019. Lancet. 2020;396: nity, Motivation, Behaviour (COM-B) model).46 There is also potential 2006–2017. to explore how unique features of mobile apps, such as synchroni- sation with wearable devices, may assist with delivery of in- 3. Lin I, Wiles L, Waller R, Goucke R, Nagree Y, Gibberd M, et al. What does best terventions directed at behaviour change. practice care for musculoskeletal pain look like? Eleven consistent recommenda- tions from high-quality clinical practice guidelines: systematic review. Br J Sport Based on these findings, it is recommended that mobile apps be Med. 2020;54:79–86. considered as an option to deliver therapeutic exercise and physical activity programs for people with musculoskeletal pain conditions. 4. Bailey DL, Holden MA, Foster NE, Quicke JG, Haywood KL, Bishop A. Defining Initial evidence suggests that mobile apps may help achieve worth- adherence to therapeutic exercise for musculoskeletal pain: A systematic review. while improvements in pain intensity, self-reported physical function Br J Sport Med. 2020;54:326–331. and work participation for regular smartphone users, and potentially worthwhile improvements in these outcomes for others. There were 5. Argent R, Daly A, Caulfield B. Patient involvement with home-based exercise no signals that pain interference, physical performance, fear avoid- programs: can connected health interventions influence adherence? JMIR mHealth ance beliefs, adherence and health-related quality of life are detri- uHealth. 2018;6:e47. mentally affected by using mobile apps compared with other modes of delivery. Due to the low to very low quality of evidence for these 6. Meade LB, Bearne LM, Sweeney LH, Alageel SH, Godfrey EL. Behaviour change tech- outcomes, these conclusions may change as further research be- niques associated with adherence to prescribed exercise in patients with persistent comes available. Future design of mobile apps and their incorpora- musculoskeletal pain: Systematic review. Br J Health Psychol. 2019;24:10–30. tion into therapeutic interventions should harness mobile apps’ potential to deliver more behaviour change features in order to 7. Chehade MJ, Yadav L, Kopansky-Giles D, Merolli M, Palmer E, Jayatilaka A, et al. achieve the best possible outcomes for people with musculoskeletal Innovations to improve access to musculoskeletal care. Best Pract Res Clin Rheu- pain conditions. matol. 2020;34:101559. What is known on this topic: Exercise programs are a high- 8. World Health Organization. WHO Guideline Recommendations on Digital In- value, low-cost intervention for improving symptoms and func- terventions for Health System Strengthening. Geneva: WHO; 2019. tion in musculoskeletal pain conditions, provided that sufficient adherence is achieved. Adherence to exercise can be improved 9. Bunting JW, Withers TM, Heneghan NR, Greaves CJ. Digital interventions for pro- by behaviour change interventions. moting exercise adherence in chronic musculoskeletal pain: a systematic review What this study adds: Mobile apps supporting therapeutic and meta-analysis. Physiotherapy. 2020;15:15. exercise or tailored physical activity programs for musculoskel- etal pain conditions may help in reducing pain intensity and 10. Schäfer AGM, Zalpour C, Von Piekartz H, Hall TM, Paelke V. The efficacy of elec- improving physical function. These effects are particularly tronic health-supported home exercise interventions for patients with osteoar- evident in trials involving people with knee osteoarthritis. Mobile thritis of the knee: Systematic review. J Med Internet Res. 2018;20. apps utilised a limited range of behaviour change intervention functions. 11. Chen T, Or CK, Chen J. Effects of technology-supported exercise programs on the knee pain, physical function, and quality of life of individuals with knee osteoar- thritis and/or chronic knee pain: A systematic review and meta-analysis of ran- domized controlled trials. J Am Med Informat Assoc. 2021;28:414–423. 12. Yang Q, Van Stee SK. The comparative effectiveness of mobile phone interventions in improving health outcomes: Meta-analytic review. JMIR mHealth uHealth. 2019;7. 13. Van Stee SK, Yang Q. The effectiveness and moderators of mobile applications for health behavior change. In: Technology and Health. Academic Press. 2020:243–270. 14. Richardson M, Khouja CL, Sutcliffe K, Thomas J. Using the theoretical domains framework and the behavioural change wheel in an overarching synthesis of systematic reviews. BMJ Open. 2019;9:e024950. 15. Michie S, van Stralen MM, West R. The behaviour change wheel: a new method for characterising and designing behaviour change interventions. Implement Sci.6:42. 16. Geuens J, Swinnen TW, Westhovens R, de Vlam K, Geurts L, Abeele VV. A review of persuasive principles in mobile apps for chronic arthritis patients: Opportunities for improvement. JMIR mHealth uHealth. 2016;4. 17. Martinez-Calderon J, Flores-Cortes M, Morales-Asencio JM, Luque-Suarez A. Which Psychological Factors Are Involved in the Onset and/or Persistence of Musculo- skeletal Pain? An Umbrella Review of Systematic Reviews and Meta-Analyses of Prospective Cohort Studies. Clin J Pain. 36:626–637. 18. Hewitt S, Sephton R, Yeowell G. The Effectiveness of Digital Health Interventions in the Management of Musculoskeletal Conditions: Systematic Literature Review. J Med Internet Res. 22:e15617. 19. McKay FH, Slykerman S, Dunn M. The APp Behavior Change Scale: Creation of a scale to assess the potential of apps to promote behavior change. JMIR mHealth uHealth. 2019;7. 20. Page MJ, McKenzie JE, Bossuyt P, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Med Flum. 2021;57:444–465. 21. McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. An evi- dence based checklist for the Peer Review of Electronic Search Strategies (PRESS EBC). Evid Based Libr Inf Pract. 2010:149–154. 22. Van Lummel RC, Walgaard S, Pijnappels M, Elders PJ, Garcia-Aymerich J, van Dieën JH, et al. Physical performance and physical activity in older adults: Asso- ciated but separate domains of physical function in old age. PLoS ONE. 2015;10. 23. Sterne JACSJ, Page MJ, Elbers RG, Blencowe NS, Boutron I, Cates CJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:I4898.
Search
