Journal of Physiotherapy 61 (2015) 117–124 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Research Rehabilitation that incorporates virtual reality is more effective than standard rehabilitation for improving walking speed, balance and mobility after stroke: a systematic review Davide Corbetta a, Federico Imeri b, Roberto Gatti c a Rehabilitation Department, San Raffaele Hospital, Milan; b Private Practice; c Laboratory of Analysis and Rehabilitation of Motor Function, Neuroscience Division, San Raffaele Hospital, Milan, Italy KEY WORDS ABSTRACT Virtual reality exposure therapy Question: In people after stroke, does virtual reality based rehabilitation (VRBR) improve walking speed, Stroke rehabilitation balance and mobility more than the same duration of standard rehabilitation? In people after stroke, Walking does adding extra VRBR to standard rehabilitation improve the effects on gait, balance and mobility? Postural balance Design: Systematic review with meta-analysis of randomised trials. Participants: Adults with a clinical diagnosis of stroke. Intervention: Eligible trials had to include one these comparisons: VRBR replacing some or all of standard rehabilitation or VRBR used as extra rehabilitation time added to a standard rehabilitation regimen. Outcome measures: Walking speed, balance, mobility and adverse events. Results: In total, 15 trials involving 341 participants were included. When VRBR replaced some or all of the standard rehabilitation, there were statistically significant benefits in walking speed (MD 0.15 m/s, 95% CI 0.10 to 0.19), balance (MD 2.1 points on the Berg Balance Scale, 95% CI 1.8 to 2.5) and mobility (MD 2.3 seconds on the Timed Up and Go test, 95% CI 1.2 to 3.4). When VRBR was added to standard rehabilitation, mobility showed a significant benefit (0.7 seconds on the Timed Up and Go test, 95% CI 0.4 to 1.1), but insufficient evidence was found to comment about walking speed (one trial) and balance (high heterogeneity). Conclusion: Substituting some or all of a standard rehabilitation regimen with VRBR elicits greater benefits in walking speed, balance and mobility in people with stroke. Although the benefits are small, the extra cost of applying virtual reality to standard rehabilitation is also small, especially when spread over many patients in a clinic. Adding extra VRBR time to standard rehabilitation also has some benefits; further research is needed to determine if these benefits are clinically worthwhile. [Corbetta D, Imeri F, Gatti R (2015) Rehabilitation that incorporates virtual reality is more effective than standard rehabilitation for improving walking speed, balance and mobility after stroke: a systematic review. Journal of Physiotherapy 61: 117–124] ß 2015 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 recent approach that may enable simulated practice of functional tasks at a higher dosage than traditional therapies.7,8 It consists of Several studies have assessed stroke survivors’ opinions about techniques that allow sensory experimentation through the the conditions that facilitate activity and participation in daily interaction between humans and informatics technologies.9 Virtual life.1,2 Over 70% of the respondents in these studies rated the ability reality has been defined as the ‘use of interactive simulations created to ‘get out and about’ in the community as very important.1 However, nearly 40% of people who experience a stroke are either with computer hardware and software to present users with unable to walk or limited to walking within their immediate opportunities to be engaged in environments that appear and feel environment.1 Because of this limited walking ability, they cannot similar to real-world objects and events’.10 The key features of all participate in community activities, which leads to a reduced virtual reality applications are the sense of ‘presence in’ and ‘control quality of life.3,4 An objective of rehabilitation after stroke is to over’ the simulated environment.11 The sense of ‘presence in’ return the survivors to social and working activities. consists of the feeling of being in an environment, even if not The high repetition of task-oriented exercises5 has been physically present in that environment; the sense of ‘control over’ described as being important for locomotion recovery. In particu- involves the possibility of interaction with the environment and lar, the repetition of tasks connected to locomotion has been objects.12 These two aspects distinguish virtual reality from other shown to be effective in many aspects such as improving walking distance and speed in people exhibiting motor deficits following forms of visual imaging such as watching videos or television. VRBR stroke.6 Virtual reality based rehabilitation (VRBR) is a relatively attempts to simulate real-world activities, which may provide more involving tasks when compared to standard rehabilitation. The use of virtual reality encourages a higher number of exercise http://dx.doi.org/10.1016/j.jphys.2015.05.017 1836-9553/ß 2015 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/).
118 Corbetta et al: Virtual reality for stroke rehabilitation repetitions.13,14 It is also described as promoting motor learning Box 1. Inclusion criteria through immediate feedback about performed tasks15 that are related to real-life activities16,17 such as navigation18 and road Design crossing.19,20 Randomised or quasi-randomised controlled trial Previous systematic reviews have reported a moderate advan- Population tage obtained from VRBR on body functions of the upper limb21,22 Adults (> 18 years old) and lower limb,23 and on activities (especially those related to the Clinical diagnosis of stroke (ischaemic or haemorrhagic) lower limb22,24,25) when compared to standard rehabilitation in No other pathological conditions affecting lower limbs (eg, people with stroke. A Cochrane systematic review26 published in 2015 concluded that there was insufficient evidence to draw musculoskeletal disorders) conclusions about the effectiveness of VRBR in improving gait Intervention speed in people with stroke. More trials have been published since these earlier reviews conducted their searches, allowing for meta- VRBR using head mounted devices or conventional analyses of more outcomes and more specific comparisons. workstation (eg, monitor, keyboard) A lot of interactive gaming consoles are available and used in Any intensity rehabilitation units27,28 but virtual reality programs designed Duration exceeding a single treatment session specifically for rehabilitation purposes are still expensive and, thus, Outcomes not frequently used in clinical contexts. The development of a body Walking speed of evidence about VRBR for the functional recovery of people after Balance stroke may further assist the clinician in the choice of rehabilitation Mobility approach. The aim of this work was to systematically review Comparisons published studies of the efficacy of VRBR versus standard rehabili- VRBR replacing some or all standard rehabilitation tation in subjects presenting motor limitation following stroke. VRBR added to standard rehabilitation versus standard Studies performing VRBR of walking, balance and/or mobility were rehabilitation alone included in the review, assuming that a post-stroke physiotherapy program that targets deficits in balance may be also effective in VRBR = virtual reality based rehabilitation. restoring independent functional walking.29 In fact, impaired balance seems to be related to a decreased locomotor function.30,31 assessment was achieved by assigning a judgment of ‘low risk’ of This review therefore sought to answer the following questions: bias when bias was considered unlikely to have seriously altered the results, ‘high risk’ of bias when the potential for bias seriously 1. In people after stroke, does VRBR improve walking speed, weakened confidence in the results, or ‘unclear risk’ when there balance and mobility more than the same duration of standard was some doubt about the effect of bias on the results. It was rehabilitation? applied for seven specific domains: sequence generation, alloca- tion concealment, blinding of participants and personnel, blinding 2. In people after stroke, does adding extra VRBR to standard of outcome assessment, incomplete outcome data, selective rehabilitation improve the effects on gait, balance and mobility? outcome reporting and ‘other issues’. Considering the nature of the interventions in the included studies, blinding of the Method participants and personnel would have been impractical, so only outcome assessor blinding was considered. Identification and selection of studies Participants To be eligible, studies had to have examined adults aged over In August 2014, the Cochrane Central Register of Controlled Trials (from 1929), PubMed (from 1950), Embase (from 1980), 18 years and with a clinical diagnosis of ischaemic or haemorrhagic CINAHL (from 1982) and PEDro (from 1929) databases were stroke, as defined by the World Health Organization.35 Confirma- electronically searched. A modified sensitivity maximising version tion of the clinical diagnosis using imaging was not compulsory. of the Cochrane Highly Sensitive Search Strategy32 was combined with the subject-specific search in order to identify randomised Intervention trials that tested VRBR to train stroke survivors who had motor Eligible studies evaluated VRBR that replaced, or was in addition deficits that impaired locomotion and balance. Four key terms – ‘stroke’, ‘virtual reality’, ‘walking’ and ‘postural balance’ – were to, standard rehabilitation to improve gait, balance and/or mobility used to generate a list of search terms, which were combined into a in people after stroke. If the total regimen exceeded a single session, search strategy adapted to each database (Appendix 1 on the any duration of VRBR was acceptable. The VRBR had to meet the eAddenda). definition of Schultheis 2001: an advanced form of human-computer interface that allows the user Reference lists of identified studies and published reviews were to ‘interact’ with and become ‘immersed’ in a computer-generated manually checked for additional trials. References retrieved by the environment in a naturalistic fashion.36 electronic search were compared for duplicate entries using the ‘find duplicates’ facility of reference management software33 and The VRBR consisted of either a single type of exercise (eg, were manually crosschecked. Two review authors (DC and FI) walking while watching videos or moving in a virtually reproduced independently selected potentially eligible articles based on the setting) with various aims (eg, increasing walking speed, improv- titles and abstracts. Full-text copies of these articles were assessed ing gait and balance) or in a combination of different types of against the inclusion criteria presented in Box 1. Disagreements exercises (eg, weight shifting toward the paretic side, propriocep- were solved by discussion, with a third reviewer (RG) consulted if tive neuromuscular facilitation, or muscle strengthening). Trials the disagreement persisted. Eligible studies underwent data that compared different types of VRBR without a comparison group extraction by two reviewers (DC and FI) who worked indepen- were not included. dently and used a piloted, standardised data collection form. Outcome measures Assessment of characteristics of included studies The primary outcome was walking speed evaluated with Quality objective measures (eg, the 6-minute walk test, the 10-metre walk The quality of the included studies was analysed with the test, or instrumental gait analysis devices).37 The secondary outcomes were: measures of balance, assessed with functional Cochrane Collaboration’s tool for assessing risk of bias.34 The scales such as the Berg Balance Scale,38 and mobility, evaluated with performance measures such as the Timed Up and Go test.39 Data were extracted for the end of the intervention period and at
Researc_)T1]D$erFIG[(iguh 119 the longest follow-up point reported in each of the included studies. Any statements about adverse events were also noted. Data analysis Results from comparable trials were pooled using RevMan Figure 1. Flow of studies through the review. software.40 For the primary outcome (walking speed), data in m/s aPapers may have been ineligible for failing to meet more than one eligibility were directly obtained from each article or they were converted to criterion. m/s from the reported test description and results. For example, the velocity for performing the 6-minute walk test was calculated was unclear and half of the trials did not properly report the by dividing the distance covered in metres by 360 seconds (total allocation procedure. The majority of the studies reported that the duration of the test), or the gait speed reported as m/min in the outcome assessors were blinded. Seven studies reported with- study of Jaffe and colleagues41 was converted to m/s. For secondary drawals45,47–51,56 and provided the reasons for these dropouts. All outcomes, measures were similar among included studies; trials were analysed on a per protocol basis. There were no marked therefore, all results were expressed as mean differences on the differences in quality between the studies that had the same same scale. Change scores and their standard deviations were used duration of treatment in both the experimental and control to compute pooled effect estimates. The pooled results from the groups41,43–45,47,49–52,54,56 and the studies that added VRBR to meta-analyses were therefore expressed as weighted mean standard rehabilitation in order to produce a greater amount of differences (MD) with 95% CI, in the original units of the treatment in the experimental group.42,48,53,55 measurement. Four authors of the included studies were contacted through emails for data not reported in their papers.42–45 Two Participants authors42,43 replied and provided the unreported data. The The included studies involved 341 participants: 169 were remaining unreported measures of variability were estimated through the use of reported variances with an appropriate randomised to receive VRBR and 172 to receive standard rehabilita- correction, as suggested in the Cochrane Handbook.34 In one tion. The mean age of the participants in the included studies ranged study45 with non-parametric distribution of data, mean changes from 53 to 65 years. About 44% of the participants were female. and their relative measures of variability were estimated with the Table 2 reports the characteristics of the participants in the included method proposed by Hozo et al.46 Heterogeneity was assessed by studies. The majority of the studies enrolled subjects who had had an visual inspection of the forest plot and consideration of the I2 episode of ischaemic stroke more than 6 months before enrolment statistic in conjunction with the chi-square test.34 into the study. Where reported, they had preserved ability to walk with or without an assistive device44,45,47–49,52–54 or the ability to Results maintain an upright posture.42,50 Flow of studies through the review After screening the search results, 15 studies were identified for inclusion in the review.41–45,47–56 Hand searching did not identify any additional papers. The flow of studies through the review is shown in Figure 1. Characteristics of included studies The included studies took place in seven countries: eight trials took place in Korea,47–49,52–56 two in the USA,41,44 one in Taiwan,51 one in Singapore,43 one in Brazil,42 one in Spain50 and one in Italy.45 Quality Intervention The individual items achieved by each of the included In the experimental groups of 11 studies,41,43–45,47,49–52,54,56 studies are presented in Table 1. The quality of the trials was VRBR was integrated into or was used in place of standard good, although in three out of 15, the randomisation procedure rehabilitation, resulting in an equal total treatment time between Table 1 Methodological quality of included studies. Study Random Concealed Assessor Dropouts Reasons for Selective reporting Type of allocation allocation blinding (%) withdrawals bias analysis Barcala et al (2013)42 Cho et al (2012)48 LR LR LR 0 - LR P Cho et al (2013)47 LR Unclear Unclear 8 Yes LR P Cho et al (2014)49 LR 12 Yes LR P Jaffe et al (2004)41 LR LR LR 6 Yes LR P Jung et al (2012)52 LR Unclear LR 0 Unclear P Kang et al (2012)56 Unclear LR 0 - LR P Kim et al (2009)53 LR HR LR <1 - LR P Llore´ ns et al (2015)50 LR Unclear LR 0 Yes LR P Mirelman et al (2009)44 LR LR 9 - LR P Morone et al (2014)45 LR LR LR 0 Yes LR P Park et al (2013)54 LR Unclear LR <1 - LR P Rajaratnam et al (2013)43 Unclear LR 0 Yes LR P Song et al (2014)55 LR LR Unclear 0 - LR P Yang et al (2008)51 Unclear LR LR 0 - LR P LR LR Unclear 16 - LR P Unclear LR Yes Unclear Unclear Unclear HR = high risk of bias, LR = low risk of bias, P = per protocol analysis, Unclear = unclear risk of bias.
120 Corbetta et al: Virtual reality for stroke rehabilitation Table 2 Summary of the included studies. Study Participants Intervention Outcome measures Barcala et al (2013)42 n = 20 Exp = Conventional rehabilitation (60 min PT, 2/wk x 5 wk) Balance = BBS, TUG Cho et al (2012)48 Age (yr) = 64 (SD 14) + balance training on Nintendo WBB (30 min, 2/wk x 5 wk) Postural stability = COP oscillations Gender = 9 M, 11 F ADL Independence = FIM Cho et al (2013)47 Con = Conventional rehabilitation (60 min PT, 2/wk x 5 wk) n = 22 Balance = BBS, TUG Age (yr) = 64 (SD 8) Exp = Conventional rehabilitation (30 min PT + 30 min OT, 5/wk x 6 wk) Postural stability = PSV Gender = 14 M, 8 F + balance training on Nintendo WBB (30 min, 3/wk x 6 wk) Balance = BBS, TUG n = 14 Con = Conventional rehabilitation (30 min PT + 30 min OT, 5/wk x 6 wk) Gait kinematics = Spatiotemporal gait Age (yr) = 65 (SD 5) Gender = 7 M, 7 F Exp = Conventional rehabilitation (30 min PT + 30 min OT + 20 min FES, parameters (including walking speed) 5/wk x 6 wk) and treadmill walking in virtual outdoor environment Cho et al (2014)49 n = 30 (30 min, 3/wk x 6 wk) Balance = BBS, TUG Age (yr) = 65 (SD 6) Postural stability = PSV Gender = 15 M, 15 F Con = Conventional rehabilitation (30 min PT + 30 min OT + 20 min FES, Gait kinematics = Spatiotemporal gait 5/wk x 6 wk) and treadmill walking training (30 min, 3/wk x 6 wk) Jaffe et al (2004)41 n = 20 parameters (including walking speed) Age (yr) = 61 (SD 10) Exp = Conventional rehabilitation (30 min PT + 30 min OT + 20 min FES, Gender = 12 M, 8 F 5/wk x 6 wk) and treadmill walking in a virtual outdoor Gait endurance = 6MWT environment (30 min, 3/wk x 6 wk) Gait kinematics = Spatiotemporal gait Jung et al (2012)52 n = 21 Kang et al (2012)56 Age (yr) = 62 (SD 7) Con = Conventional rehabilitation (30 min PT + 30 min OT + 20 min FES, parameters (including walking speed) Gender = 13 M, 8 F 5/wk x 6 wk) and treadmill walking training (30 min, 3/wk x 6 wk) Others = Obstacle clearance test, Balance test Adverse events n = 30 Exp = Stepping over virtual objects on treadmill (60 min, 3/wk x 2 wk) Age (yr) = 56 (SD 7) Con = Stepping over real foam objects in a hallway (60 min, 3/wk x 2 wk) Balance = TUG Gender = 16 M, 14 F Balance self-efficacy = ABC scale Exp = Treadmill walking in a virtual outdoor environment (30 min, Kim et al (2009)53 n = 24 5/wk x 3 wk) Balance = TUG Age (yr) = 53 (SD 9) Walking speed = 10MWT Gender = 13 M, 11 F Con = Treadmill walking training (30 min, 5/wk x 3 wk) Gait endurance = 6MWT Others = FRT Llore´ ns et al (2015)50 n = 20 Exp = Conventional rehabilitation (PT 5/wk x 4 wk) + treadmill walking Mirelman et al (2009)44 Age (yr) = 57 (SD 11) with optic flow (30 min, 3/wk x 4 wk) Balance = BBS Gender = 9 M, 11 F Postural stability = PSV Con 1 = Conventional rehabilitation (PT 5/wk x 4 wk) + treadmill training Walking speed = 10MWT n = 18 (30 min, 3/wk x 4 wk) Gait kinematics = Spatiotemporal gait Age (yr) = 61 (SD 9) Gender = 15 M, 3 F Con 2 = Conventional rehabilitation (PT 5/wk x 4 wk) + stretching added parameters (including walking speed) ROM exercises (30 min, 3/wk x 4 wk) Others = MMAS Morone et al (2014)45 n = 47 Park et al (2013)54 Age (yr) = 60 (SD 10) Exp = Conventional rehabilitation (40 min PT, 4/wk x 4 wk) + VR exercises Balance = BBS, Tinetti POMA, BBA Gender = Unreported for balance and stepping skills (30 min, 4/wk x 4 wk) Walking speed = 10MWT n = 16 Con = Conventional rehabilitation (40 min PT, 4/wk x 4 wk) Balance = BBS Age (yr) = 48 (SD 8) Gait endurance = 6MWT Gender = 11 M, 5 F Exp = Conventional rehabilitation (30 min PT, 5/wk x 4 wk) + stepping task Gait kinematics = Spatiotemporal gait in a 3D virtual environment (30 min, 5/wk x 4 wk) parameters (including walking speed) Con = Conventional rehabilitation (60 min PT, 5/wk x 4 wk) Others = FMA (lower extremity) Exp = Robot training for foot movements in a virtual environment Follow-up = 3 mth (60 min, 3/wk x 4 wk) Balance = BBS Con = Robot training for foot movements without a virtual environment Walking speed = 10MWT (60 min, 3/wk x 4 wk) ADL Independence = BI Others = FAC Exp = Conventional rehabilitation (40 min PT, 2 times/day x 4 wk) + Nintendo WBB (20 min, 3/wk x 4 wk) Walking speed = 10MWT Gait kinematics = Spatiotemporal gait Con = Conventional rehabilitation (40 min PT, 2 times/day x 4 wk) + balance therapy (20 min, 3/wk x 4 wk) parameters (including walking speed) Follow-up = 1 mth Exp = Conventional rehabilitation (60 min PT, 5/wk x 4 wk) + VR-based postural control exercises (30 min, 3/wk x 4 wk) Con = Two administrations of conventional rehabilitation (60 min PT, 5/wk x 4 wk) + (30 PT min, 3/wk x 4 wk) Rajaratnam et al (2013)43 n = 19 Exp = Conventional rehabilitation (40 min PT, 15 sessions) + balance Balance = BBS, TUG Song et al (2014)55 Age (yr) = 62 (SD 9) training on Nintendo WBB or Microsoft Kinect (20 min, 15 sessions) Postural stability = COP oscillations Yang et al (2008)51 Gender = 7 M, 12 F ADL Independence = MBI Con = Conventional rehabilitation (60 min PT, 15 sessions) n = 20 Others = FRT Age (yr) = 63 (SD 14) Exp = Conventional rehabilitation (25 min PT, 5/wk x 3 wk) + VR-based Gender = 11 M, 9 F balance training (25 min, 3/wk x 3 wk) Balance = BBS Others = FI, SI, WDI n = 20 Con = Conventional rehabilitation (25 min PT, 5/wk x 3 wk) Age (yr) = 58 (SD 11) Gait kinematics = walking speed Gender = 10 M, 10 F Exp = Treadmill walking in virtual outdoor environment (20 min, Others = ABC, WAQ, CWT 3/wk x 3 wk) Follow-up = 1 mth Con = Treadmill training simulating stepping obstacles (20 min, 3/wk x 3 wk) 10MWT = 10-metre Walk Test, 6MWT = 6-minute Walk Test, ABC = Activities-specific Balance Confidence, ADL = activities of daily living, BBA = Brunel Balance Assessment, BBS = Berg Balance Scale, BI = Barthel Index, Con = control group, COP = centre of pressure, CWT = Community Walk Test, Exp = experimental group, F = female, FES = functional electrical stimulation, FI = Falling Index, FIM = Functional Independence Measure, FMA = Fugl-Meyer Assessment, FRT = Functional Reach Test, M = male, MBI = Barthel Index, MMAS = Modified Motor Assessment Scale, OT = occupational therapy, POMA = Performance-Oriented Mobility Assessment, PSV = Postural Sway Velocity, ROM = range of motion, SI = Stability Index, TUG = Timed Up and Go, WAQ = Walking Ability Questionnaire, FAC = Functional Ambulatory Category, WDI = Weight Distribution Index, PT = physiotherapy, VR = virtual reality, WBB = Wii Balance Board.
Researcu]GIFe_$g[D(4T)irh 121 the experimental and control groups. Among these studies, the Study MD (95% CI) interfaces most frequently used for walking rehabilitation were Mirelman44 Random virtual reality treadmill training systems.41,47,49,51,52,56 Some Park54 consisted of a treadmill and a wide screen that projected a real- Yang51 world video recording in order to reproduce an immersive virtual Pooled outdoor environment;47,49,51 others used a head-mounted device instead of the monitor.41,52,56 One study used the head-mounted −0.2 −0.1 0 0.1 0.2 device without a treadmill.54 For balance training, one study43 Favours SR (m/s) Favours VRBR used Microsoft Xbox 360 Kinecta, one study45 used Nintendo Wii Fitb and one study50 used an audio-visual system combined with a Figure 4. Weighted mean differences (95% CI) of the effect beyond the end of the motion-tracking system in order to immerse participants in a 3D intervention period of substituting some or all of standard rehabilitation (SR) with virtual environment. In the study of Mirelman and colleagues,44 a virtual reality based rehabilitation (VRBR) on walking speed, pooling data from robotic virtual reality device was used for training movement of three trials (n = 54). the lower extremity. See Figure 3 on the eAddenda for a more detailed forest plot. No In four studies,42,48,53,55 VRBR was added to standard rehabili- important statistical heterogeneity was observed (I2 = 26%). tation, resulting in a greater amount of treatment time in the experimental group. Two of these studies53,55 used the IREX1c Three studies measured walking speed beyond the end of the virtual reality system for rehabilitation of walking and balance. It intervention period.44,51,54 These studies reported data on consisted of a television monitor, a video camera, cyber gloves and 54 participants, 28 of whom received VRBR. The effect of the virtual objects, scenes and a large screen. The other two studies VRBR was well maintained for 1 to 3 months after the end of the only trained balance by using the Wii Fit balance program.42,48 intervention period, with a mean difference of 0.12 m/s (95% CI 0.03 to 0.20), as presented in Figure 4. See Figure 5 on the eAddenda Frequency of interventions varied from 242 to 641 days a week for a more detailed forest plot. No statistical heterogeneity was and lasted from 241 to 6 weeks.47–49 The duration of each training observed (I2 = 0%). session ranged from 20 minutes51 to 1 hour.41–44,50 In addition to the studies that could be meta-analysed, Morone Outcome and colleagues45 measured gait speed over 10 m but only reported Nine studies measured locomotor function: five used the 10- percentage improvement. They reported that at the end of the 4-week intervention period gait speed improved by 35% in the metre walk test,45,50,53,54,56 two used the 6-minute walk test41,44 experimental group and by 27% in the control group. One month and two measured gait velocity.47,51 Balance was assessed using after ceasing the intervention each group improved a further 6%. the Berg Balance Scale in nine studies42,43,45,47–50,53,55 and mobility Although these differences were not statistically significant, they was assessed using the Timed Up and Go test in seven are in the same direction and of a similar magnitude to the meta- studies.42,43,47–49,52,56 Outcomes were assessed immediately after analysed studies of this review. the intervention. Only four studies included follow-up evaluations at 151,54 or 344,45 months after training. Balance Balance was assessed with the Berg Balance Scale in five Does VRBR improve outcomes more than the same duration of standard rehabilitation? studies.43,45,47,49,50 These studies reported data on 130 participants, 67 of whom received VRBR. Replacing some or all of the standard Walking speed rehabilitation with VRBR (for the same total treatment time) Walking speed was obtained from walking measures reported in significantly improved balance, with a mean difference of 2.1 points on the 0-to-56-point Berg Balance Scale (95% CI seven studies41,44,47,50,51,54,56 and converted to the same unit of 1.8 to 2.5), as presented in Figure 6. See Figure 7 on the eAddenda measurement (m/s). These studies reported data on 138 participants, for a more detailed forest plot. No statistical heterogeneity was 65 of whom received VRBR. Replacing some or all of the standard observed (I2 = 0%). rehabilitation with VRBR (for the same total treatment time) significantly improved walking speed, with a mean difference of 0.15 m/s (95% CI 0.10 to 0.19), as presented in Figure 2. re_2iuFIG]D(g[$T) Study MD (95% CI) Mobility Cho47 Random Jaffe41 Mobility was assessed using the Timed Up and Go test in five Kang56 studies.43,47,49,52,56 These studies reported data on 114 participants, Lloréns50 Mirelman44 5ure(_6)[FTiDg$IG] 3 of whom received VRBR. Replacing some or all of the standard Park54 Yang51 Study MD (95% CI) Pooled Cho47 Random Cho49 Lloréns50 Morone45 Rajaratnam43 Pooled −0.50 −0.25 0 0.25 0.50 Favours SR (m/s) Favours VRBR −4 −2 0 2 4 Favours SR (points) Favours VRBR Figure 2. Weighted mean differences (95% CI) of the effect immediately after Figure 6. Weighted mean differences (95% CI) of the effect of substituting some or intervention of substituting some or all of standard rehabilitation (SR) with virtual all of standard rehabilitation (SR) with virtual reality based rehabilitation (VRBR) on reality based rehabilitation (VRBR) on walking speed, pooling data from seven trials the Berg Balance Scale score (0 to 56 points), pooling data from five trials (n = 130). (n = 138).
F$D)8_eIugi([G]r1T 22 F$DT)0rugi([_GI1e]Corbetta et al: Virtual reality for stroke rehabilitation Study MD (95% CI) Study MD (95% CI) Cho47 Random Barcala42 Random Cho49 Cho48 Jung52 Kim53 Kang56 Song55 Rajaratnam43 Pooled −4 −2 0 2 4 Favours SR (points) Favours VRBR −10 −5 0 5 10 Favours SR (s) Favours VRBR Figure 10. Mean differences (95% CI) of the effect of adding extra virtual reality based rehabilitation (VRBR) to standard rehabilitation (SR) on the Berg Balance [_er$(Fig12])TuGIDScale score (0 to 56 points), with no pooling due to heterogeneity (n = 86). Figure 8. Weighted mean differences (95% CI) of the effect of substituting some or Study MD (95% CI) all of standard rehabilitation (SR) with virtual reality based rehabilitation (VRBR) on Barcala42 Random the Timed Up and Go test, pooling data from five trials (n = 114). Cho48 Pooled rehabilitation with VRBR (for the same total treatment time) significantly improved mobility, with a mean difference of −1 −0.5 0 0.5 1 2.3 seconds on the Timed Up and Go test (95% CI 1.2 to 3.4), as Favours SR (s) Favours VRBR presented in Figure 8. See Figure 9 on the eAddenda for a more detailed forest plot. Figure 12. Weighted mean differences (95% CI) of the effect of adding extra virtual reality based rehabilitation (VRBR) to standard rehabilitation (SR) on the Timed Up Substantial statistical heterogeneity was observed (I2 = 84%), and Go test, pooling data from two trials (n = 42). which was mainly due to the magnitude of the effect estimated from the studies of Kang et al56 and Rajaratnam et al.43 Performing Adverse events a sensitivity analysis, through the exclusion of these studies from None of the reports of the included studies made an explicit the overall estimation, the level of heterogeneity became accept- able (I2 = 0%) with a similar estimate of 1.3 (95% CI 1.0 to 1.7). The statement about adverse events, but some implied that no adverse analysis is presented in Appendix 2 on the eAddenda. events occurred. For example, Cho and colleagues stated that a staff member stayed close to each participant during the rehabilitation Adverse events in order to prevent falls.48 One of the study reports included a statement about adverse Discussion events, stating that there were no falls and no undue cardiovascu- lar responses in either group.41 However, some of the other study The meta-analyses in this systematic review identified some reports included statements that implied that adverse events beneficial effects of VRBR on walking speed, balance and mobility would have been mentioned if they had occurred. For example, outcomes in stroke survivors. These analyses are based on 15 eligible Yang et al51 and Kang et al56 stated that a staff member stayed trials with a total of 341 participants, which exceeds the amount of close to each participant during the intervention in order to data relating to clinical mobility outcomes that has been reported in prevent falls. past systematic reviews on VRBR after stroke.21–26,57 Also, this review conducted meta-analyses for clinical mobility outcomes Does adding extra VRBR to standard rehabilitation improve with separate meta-analyses depending on whether the VRBR was outcomes? substituted for, or in addition to, standard rehabilitation. Only one of the past reviews conducted meta-analyses with this distinction, but Walking speed it only analysed walking speed, not balance or mobility.57 Therefore, One study, involving 42 participants, assessed the effect of extra while the results of this new review are consistent with the general finding of the past reviews (ie, that VRBR appears to be beneficial for VRBR on walking speed.53 Although the group that received the people with stroke), some important new insights have been extra VRBR increased walking speed by an average of 0.21 m/s obtained. more than the standard rehabilitation group, this was not statistically significant (95% CI –0.23 to 0.65). This study did not The meta-analyses of those trials where the VRBR was assess outcomes beyond the intervention period. substituted for some or all of the standard rehabilitation (to give the same total treatment time) showed significant improvements Balance in walking speed, balance and mobility. These results indicate that, Balance was assessed with the Berg Balance Scale in four for a given treatment time, VRBR is more beneficial than standard rehabilitation. These findings predict that even greater effects studies.42,48,53,55 These studies reported data on 86 participants, would be seen in the remaining analyses (ie, those where the VRBR 43 of whom received the extra VRBR. These studies were too was provided as extra treatment time added to a standard heterogeneous to be pooled (I2 = 97%), as presented in Figure 10. See rehabilitation regimen). However, this was not clearly observed in Figure 11 on the eAddenda for a more detailed forest plot. these outcomes for several reasons. For walking speed, only one study53 analysed the effect of additional VRBR. The mean estimate Mobility (0.21 m/s) was greater than the effect seen in the earlier meta- Mobility was assessed using the Timed Up and Go test in two analysis (0.15 m/s, Figure 2), but the result was insignificant. The wide confidence interval (95% CI –0.23 to 0.65) means that the studies.42,48 These studies reported data on 42 participants, 21 of potential for a strong benefit in walking speed from additional whom received the extra VRBR. The group that received the extra VRBR improved their mobility on the Timed Up and Go test significantly more than the standard rehabilitation group, with a mean difference of 0.7 seconds (95% CI 0.4 to 1.1), as presented in Figure 12. See Figure 13 on the eAddenda for a more detailed forest plot. No statistical heterogeneity was observed (I2 = 0%).
Research 123 VRBR has not yet been excluded; therefore, further research could people undergoing rehabilitation. Therefore, it appears to be help to refine this estimate. Although four studies reported data for justified to propose VRBR to people who have experienced a stroke balance, the studies were too heterogeneous to be pooled. When in order to promote their recovery of walking speed, balance and mobility was analysed, a significant benefit was observed. mobility. However, the effect (0.7 seconds on the Timed Up and Go test, Figure 12) was smaller than the effect seen in the earlier meta- What is already known on this topic: Problems with analysis (2.3 seconds, Figure 8). This effect may also be too small to walking speed, balance and mobility are common after stroke, be considered clinically worthwhile by many patients; given that but high repetition of task-oriented exercises can improve the time spent doing the additional VRBR in the included studies these sequelae. Virtual reality-based rehabilitation enables was 30 minutes, two to three times per week, for 5 to 6 weeks. simulated practice of functional tasks, with moderate benefits on some upper and lower limb tasks over standard rehabilita- From the analysis of the included studies, it may not be possible tion for people with stroke. to generalise about the efficacy of VRBR in motor recovery of the What this study adds: Substituting some or all of a standard full range of people after stroke. First of all, most of the studies only rehabilitation regimen with virtual reality-based rehabilitation recruited participants with mild motor impairment, as was elicits greater benefits in walking speed, balance and mobility demonstrated by their ability to walk independently and by the in stroke patients. high Berg Balance Scale scores. Furthermore, almost all of the studies recruited people who had a stroke more than 6 months eAddenda: Figures 3, 5, 7, 9, 11 and 13 and Appendices 1 and before study enrolment, with only three studies43,45,55 evaluating 2 can be found online at doi:10.1016/j.jphys.2015.05.017. the VRBR in acute stroke patients. Footnotes: a Microsoft Xbox 360 Kinect, Microsoft Co., An open question is whether the changes induced by VRBR are Redmond, WA, USA, b Nintendo Wii Fit, Nintendo Inc., Japan, clinically relevant. In previous studies, Flansbjer and collea- c IREX1, GestureTek, Toronto, Canada. gues58,59 reported 95% CIs of the smallest real difference as – 0.15 to 0.25m/s for comfortable walking speed, –3.4 to 4.9 points Ethics approval: Not applicable. for the Berg Balance Scale and –3.8 to 2.6 seconds for the Timed Up Competing interests: Nil. and Go Test for individuals with chronic hemiparesis subsequent Source(s) of support: Nil. to stroke. Even though the smallest real difference is not an Acknowledgements: We wish to acknowledge Dr Luciana instrument to define clinical relevance, the fact that the noted Barcala and Dr Bala Rajaratnam for providing unpublished data. effects were smaller than the smallest real difference limits the We thank Francesca Nicastro for the language consult. ability to conclude that these were real improvements. Provenance: Not invited. Peer-reviewed. Correspondence: Davide Corbetta, Rehabilitation Department, The effect of VRBR on walking speed would seem to be San Raffaele Hospital, Milan, Italy. Email: [email protected] maintained from 1 month51,54 to 3 months44 of follow-up. The optimal frequency, intensity, time and type of VRBR are still References unclear. Finally, no adverse events were reported in the included studies, suggesting that VRBR can be considered a safe treatment 1. Lord SE, McPherson K, McNaughton HK, Rochester L, Weatherall M. Community for subjects after stroke. ambulation after stroke: how important and obtainable is it and what measures appear predictive? Arch Phys Med Rehabil. 2004;85:234–239. The effects obtained by VRBR could be due to the multisensory (visual and auditory) feedback provided by virtual reality systems 2. Harris JE, Eng JJ. Goal priorities identified through client-centred measurement in and to the influence of motivational aspects on motor perfor- individuals with chronic stroke. Physiother Can. 2004;56:171–176. mance.25,60,61 These sensory information allow the central nervous system to better control position and orientation of body segments 3. Nadeau SE, Wu SS, Dobkin BH, Azen SP, Rose DK, Tilson JK, et al. Effects of task- adapting to the complex external environment.62 Moreover, You specific and impairment-based training compared with usual care on functional et al63 suggested that treatment using virtual reality facilitates walking ability after inpatient stroke rehabilitation: LEAPS Trial. Neurorehabil cortical reorganisation. The VRBR settings were also used to Neural Repair. 2013;27:370–380. reproduce training activities that closely reproduce real-world tasks, which have been shown to maximise training effects.64 This 4. Nichols-Larsen DS, Clark PC, Zeringue A, Greenspan A, Blanton S. Factors influencing represents one of the most important features of exercises stroke survivors’ quality of life during subacute recovery. Stroke. 2005;36:1480– proposed in neurorehabilitation; they must be highly repetitive 1484. and task oriented in order to facilitate the recovery of functions and activities.65 5. Malouin F, Richards CL, McFadyen B, Doyon J. New perspectives of locomotor rehabilitation after stroke. Med Sci (Paris). 2003;19:994–998. The authors of several of the eligible studies42,47,49,51,53 included statements that the VRBR was motivating and more 6. French B, Thomas L, Leathley M, Sutton C, McAdam J, Forster A, et al. Repetitive task involving than standard rehabilitation, although none of them training for improving functional ability after stroke. Cochrane Datab Syst Rev. directly assessed the attitude of participants toward VRBR. 2007; 4:CD006073. Although this meta-analysis suggests that VRBR improves 7. Kwakkel G, Van Peppen R, Wagenaar R, Wood Dauphinee S, Richards C, Ashburn A, walking speed, balance and mobility in people with stroke more et al. Effects of augmented exercise therapy time after stroke. A meta-analysis. than the same time spent doing standard rehabilitation, further Stroke. 2004;35:1–11. randomised trials with large sample sizes are encouraged. The additional data would help to confirm these results and to improve 8. Merians A, Jack D, Boian R, Tremaine M, Burdea G, Adamovich S, et al. Virtual reality the precision of the estimates. Further trials that apply the VRBR as augmented rehabilitation for patients following stroke. Phys Ther. 2002;82:898–915. extra time added to a standard rehabilitation regimen will help to provide estimates specifically about this use, where the effects on 9. Holden MK. Virtual environments for motor rehabilitation: review. Cyberpsychol walking speed and balance are unclear. Finally, further trials could Behav. 2005;3:187–207. also help to determine the optimal frequency, intensity, time and type of VRBR, as well as identifying what may be causing some of 10. Weiss P, Kizony R, Feintuch U, Katz N. Virtual reality in neurorehabilitation. In: the heterogeneity seen in this review. Selzer M, Cohen L, Gage F, Clarke S, Duncan P, eds. In: Textbook of Neural Repair and Rehabilitation. Cambridge, UK: Cambridge University Press; 2006:182–197. In conclusion, VRBR appears to produce greater benefits in walking speed, balance and mobility for a given amount of 11. Witmer BG, Singer MJ. Measuring presence in virtual environments: a presence rehabilitation time than standard rehabilitation after stroke. VRBR questionnaire. Presence. 1998;7:225–240. did not appear to increase the likelihood of adverse events and it has been reported to increase motivation and involvement of 12. Slater M. A note on presence terminology. PRESENCE – Connect [On-line] http:// s3.amazonaws.com/publicationslist.org/data/melslater/ref-201/ a%20note%20on%20presence%20terminology.pdf.[accessed 29 April 2015]. 13. Rizzo A, Kim G. A SWOT analysis of the field of virtual reality rehabilitation and therapy. Presence. 2005;14:119–146. 14. Bryanton C, Bosse J, Brien M, McLean J, McCormick A, Sveistrup H. Feasibility, motivation and selective motor control: virtual reality compared to conventional home exercise in children with cerebral palsy. Cyberpsychol Behav. 2006;9:123–128. 15. Sveistrup H. Motor rehabilitation using virtual reality. J Neuroeng Rehabil. 2004;1:10. 16. Riva G, Bacchetta M, Baruff M, Borgomainerio E, De-France C, Gatti F, et al. VREPAR projects: the use of virtual environments in psycho-neuro-physiological assess- ment and rehabilitation. Cyberpsychol Behav. 1999;2:69–79. 17. Lee JH, Ku J, Cho W, Hahn WY, Kim IY, Lee SL, et al. A virtual reality system for the assessment and rehabilitation of the activities of daily living. Cyberpsychol Behav. 2003;6:383–388. 18. Johnson DA, Rose FD, Rushton SK, Pentland B, Attree EA. Virtual reality: a new prosthesis for brain injury rehabilitation. Scott Med J. 1998;43:81–83.
124 Corbetta et al: Virtual reality for stroke rehabilitation 19. Weiss PL, Naveh Y, Katz N. Design and testing of a virtual environment to train 44. Mirelman A, Bonato P, Deutsch JE. Effects of training with a robot-virtual reality stroke patients with unilateral spatial neglect to cross a street safely. Occ Ther Int. system compared with a robot alone on the gait of individuals after stroke. Stroke. 2003;10:39–55. 2009;40:169–174. 20. Katz N, Ring H, Naveh Y, Kizony R, Feintuch U, Weiss P. Interactive virtual 45. Morone G, Tramontano M, Iosa M, Shofany J, Iemma A, Musicco M, et al. The environment training for safe street crossing of right hemisphere stroke patients efficacy of balance training with video game-based therapy in subacute stroke with unilateral spatial neglect. Disabil Rehabil. 2005;27:1235–1243. patients: a randomized controlled trial. Biomed Res Int. 2014;2014:580–861. 21. Sveistrup H. Motor rehabilitation using virtual reality. J Neuroeng Rehabil. 2004;1:10. 46. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, 22. Lohse KR, Hilderman CG, Cheung KL, Tatla S, Van der Loos HF. Virtual reality range, and the size of a sample. BMC Med Res Methodol. 2005;5:13. therapy for adults post-stroke: a systematic review and meta-analysis exploring 47. Cho KH, Lee WH. Virtual walking training program using a real-world video virtual environments and commercial games in therapy. PLoS One. 2014;9:e93318. recording for patients with chronic stroke: a pilot study. Am J Phys Med Rehabil. 23. Saposnik G, Levin M, SORCan Working Group. Virtual reality in stroke rehabilita- 2013;92:371–380. tion: a meta-analysis and implications for clinicians. Stroke. 2011;42:1380–1386. 24. Moreira MC, de Amorim Lima AM, Ferraz KM, Benedetti Rodriguez MA. Use of 48. Cho KH, Lee KJ, Song CH. Virtual-reality balance training with a video-game system virtual reality in gait recovery among post stroke patients – a systematic literature improves dynamic balance in chronic stroke patients. Tohoku J Exp Med. 2012; review. Disabil Rehabil Assist Technol. 2013;8:357–362. 228:69–74. 25. Imam B, Jarus T. Virtual reality rehabilitation from social cognitive and motor learning theoretical perspectives in stroke population. Rehabil Res Pract. 49. Cho KH, Lee WH. Effect of treadmill training based real-world video recording on 2014;2014:594–540. balance and gait in chronic stroke patients: a randomized controlled trial. Gait 26. Laver KE, George S, Thomas S, Deutsch JE, Crotty M. Virtual reality for stroke Posture. 2014;39:523–528. rehabilitation. Cochrane Datab Syst Rev. 2011;9:CD008349. 27. National Stroke Foundation. National Stroke Audit Rehabilitation Services. Melbourne, 50. Llore´ ns R, Gil-Go´ mez JA, Alcan˜ iz M, Colomer C, Noe´ E. Improvement in balance 2010. using a virtual reality-based stepping exercise: a randomized controlled trial 28. Wiihabilitation [Internet]. Available from http://wiihabilitation.co.uk.[accessed involving individuals with chronic stroke. Clin Rehabil. 2015;29:261–268. 17 April 2012]. 29. Wu¨ est S, van de Langenberg R, de Bruin ED. Design considerations for a theory- 51. Yang YR, Tsai MP, Chuang TY, Sung WH, Wang RY. Virtual reality-based training driven exergame-based rehabilitation program to improve walking of persons improves community ambulation in individuals with stroke: a randomized con- with stroke. Eur Rev Aging Phys Act. 2014;11:119–129. trolled trial. Gait Posture. 2008;28:201–206. 30. Eng JJ, Tang PF. Gait training strategies to optimize walking ability in people with stroke: a synthesis of the evidence. Expert Rev Neurother. 2007;7:1417–1436. 52. Jung J, Yu J, Kang H. Effects of virtual reality treadmill training on balance and 31. Yavuzer G, Eser F, Karakus D, Karaoglan B, Stam HJ. The effects of balance training on balance self-efficacy in stroke patients with a history of falling. J Phys Ther Sci. gait late after stroke: a randomized controlled trial. Clin Rehabil. 2006;20:960–969. 2012;24:1133–1136. 32. Lefebvre C, Manheimer E, Glanville J. Chapter 6: Searching for studies. In: Higgins JPT, Green S, eds. In: Cochrane Handbook for Systematic Reviews of Interventions 53. Kim JH, Jang SH, Kim CS, Jung JH, You JH. Use of virtual reality to enhance balance Version 5. 1. 0 (updated March 2011). Chichester, UK: The Cochrane Collaboration; and ambulation in chronic stroke: a double-blind, randomized controlled study. 2011. Available from www.cochrane-handbook.org [accessed 29 April 2015]. Am J Phys Med Rehabil. 2009;88:693–701. 33. Endnote X7 Software. 1988-2013 New York, USA: Thomson Reuters. U.S. patent No B, 082, 241. 54. Park YH, Lee CH, Lee BH. Clinical usefulness of the virtual reality-based postural 34. Higgins JPT, Green S, eds. In: Cochrane Handbook for Systematic Reviews of Inter- control training on the gait ability in patients with stroke. J Exerc Rehabil. 2013;9: ventions. Version 5. 1. 0 [updated March 2011]. The Cochrane Collaboration; 2011. 489–494. Available from www.cochrane-handbook.org [accessed 29 April 2015]. 35. Hatano S. Experience from a multicentre stroke register: a preliminary report. Bull 55. Song YB, Chun MH, Kim W, Lee SJ, Yi JH, Park DH. The effect of virtual reality and World Health Organ. 1976;54:541–553. tetra-ataxiometric posturography programs on stroke patients with impaired 36. Schultheis M, Rizzo A. The application of virtual reality technology in rehabilita- standing balance. Ann Rehabil Med. 2014;38:160–166. tion. Rehabil Psychol. 2001;46:296–311. 37. Altenburger PA, Dierks TA, Miller KK, Combs SA, van Puymbroeck M, Schmid AA. 56. Kang HK, Kim Y, Chung Y, Hwang S. Effects of treadmill training with optic flow on Examination of sustained gait speed during extended walking in individuals with balance and gait in individuals following stroke: randomized controlled trials. Clin chronic stroke. Arch Phys Med Rehabil. 201;94:2471–2477. Rehabil. 2012;26:246–255. 38. Berg K, Wood-Dauphinee S, Williams JI. The Balance Scale: reliability assessment with elderly residents and patients with an acute stroke. Scand J Rehabil Med. 57. Rodrigues-Baroni JM, Nascimento LR, Ada L, Teixeira-Salmela LF. Walking training 1995;27:27–36. associated with virtual reality-based training increases walking speed of individ- 39. Mao HF, Hsueh IP, Tang PF, Sheu CF, Hsieh CL. Analysis and comparison of the uals with chronic stroke: systematic review with meta-analysis. Braz J Phys Ther. psychometric properties of three balance measures for stroke patients. Stroke. 2014;18:502–512. 2002;33:1022–1027. 40. Review Manager (RevMan) [Computer program]. Version 5.1. Copenhagen: The 58. Flansbjer UB, Holmback AM, Downham D, Patten C, Lexell J. Reliability of gait Nordic Cochrane Centre, The Cochrane Collaboration, 2011. performance tests in men and women with hemiparesis after stroke. J Rehabil Med. 41. Jaffe DL, Brown DA, Pierson-Carey CD, Buckley EL, Lew HL. Stepping over obstacles 2005;37:75–82. to improve walking in individuals with poststroke hemiplegia. J Rehabil Res Dev. 2004;41:283–292. 59. Flansbjer UB, Blom J, Broga˚ rdh C. The reproducibility of Berg Balance Scale and the 42. Barcala L, Grecco LA, Colella F, Lucareli PR, Salgado AS, Oliveira CS. Visual biofeed- Single-leg Stance in chronic stroke and the relationship between the two tests. back balance training using wii fit after stroke: a randomized controlled trial. J Phys Phys Med Rehabil. 2012;4:165–170. Ther Sci. 2013;25:1027–1032. 43. Rajaratnam BS, Gui Kaien J, Lee Jialin K, Sweesin K, Sim Fenru S, Enting L, et al. Does the 60. Langhorne P, Coupar F, Pollock A. Motor recovery after stroke: a systematic review. inclusion of virtual reality games within conventional rehabilitation enhance balance Lancet Neurol. 2009;8:741–754. retraining after a recent episode of stroke? Rehabil Res Pract. 2013;2013:649–561. 61. Gatti R, Tettamanti A, Lambiase S, Rossi P, Comola M. Improving hand functional use in subjects with multiple sclerosis using a musical keyboard: a randomized controlled trial. Physiother Res Int. 2014 Jul 7. http://dx.doi.org/10.1002/pri.1600. 62. Lord SE, Rochester L, Weatherall M, McPherson KM, McNaughton HK. The effect of environment and task on gait parameters after stroke: a randomized comparison of measurement conditions. Arch Phys Med Rehabil. 2006;87:967–973. 63. You SH, Jang SH, Kim YH, Hallett M, Ahn SH, Kwon YH, et al. Virtual reality-induced cortical reorganization and associated locomotor recovery in chronic stroke: an experimenter-blind randomized study. Stroke. 2005;36:1166–1171. 64. Walker C, Brouwer BJ, Culham EG. Use of visual feedback in retraining balance following acute stroke. Phys Ther. 2000;80:886–892. 65. Veerbeek JM, van Wegen E, van Peppen R, van der Wees PJ, Hendriks E, Rietberg M, et al. What is the evidence for physical therapy poststroke?. A systematic review and meta-analysis. PLoS One. 2014;9:e87987.
Journal of Physiotherapy 61 (2015) 103–105 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Editorial Resources that physiotherapists could use to add value to their research Mark R Elkins Editor, Journal of Physiotherapy Many documents have been developed to help clinical Some physiotherapy trials achieve this recommendation by researchers at various stages of their research. Some are intended using a ‘wait list’ design, where the control group is offered the to help when designing the research, while others are intended to treatment after the trial.5 In other physiotherapy trials, post-trial help ensure that the research is ethically conducted. Another large provision of the intervention is not possible. For example, group of documents are those that are intended to help researchers interventions that are beneficial when administered before surgery to publish their researchT[F_1ID$] with the necessary details for others to may not4F[_$ID]T be applicable postoperatively.6,7 An intervention applied appraise and useD$FI]_9[T it appropriately. The purpose of this editorial is with the intention of shortening the duration of inpatient to encourage physiotherapy researchers to view these documents management may have no subsequent purpose in the control as an opportunity to add value to their research at the various group.8,9 Nevertheless, many trials that could provide the interven- stages before publication. tion after the data collection period do not do so. This item was later clarified with: Readers may be familiar with some of these documents. For example, the Declaration of Helsinki is a set of ethical principles It is necessary during the study planning process to identify post-trial governing research involving humans, which was developed by access by study participants to prophylactic, diagnostic and the World Medical Association.1 Many trial reports include a therapeutic procedures identified as beneficial in the study or access statement that the research was ‘conducted in accordance with to other appropriate care. Post-trial access arrangements or other the Declaration of Helsinki’, but the reported methods sometimes care must be described in the study protocol so the ethical review show that this is not true. Perhaps these researchers read the committee may consider such arrangements during its review. Declaration of Helsinki many years ago and believe that it has not changed. However, it has been ][aF1ID$T_0 mended seven times since the Many interpret this as a requirementDFT$[]5_I to identify whether or not original 1964 version. The most recent update was in 2013, so post-trial access to the intervention will be available, meaning that even researchers who have read it relatively recently may find there is no real onus to do so.10 Nevertheless, researchers could use that it has changed.2 Although fundamental ethical principles the Declaration of Helsinki to remind themselves of the importance may change very infrequently, ethical implications of new of this issue and incorporate post-trial provision of the interven- developments in the world need to be incorporated. One such tion, where possible. In addition to improving the trial’s ethical amendment was made in accordance with the introduction of standard, it may improve the scientific standard by fostering clinical trial registers and the various campaigns to encourage recruitment and minimising loss to follow-up in the control group. prospective registration of randomised trials.3,4 $T]1FID[T_ he Declaration of Helsinki DF$]_[I1Tn2 ow _F]DsI$3[1T tates: Reporting guidelines are another type of resource that physiotherapists could use to add value to their research. These E]F_4ITD1$[ very research study $T1DI[_5F][not just randomised trials]$]DT[F16I_ must be documents are designed to help clinical researchers to publish all registered in a publicly accessible database before recruitment of of the details of a research study that readers will need to the first subject. understand its methods, judge its quality, and appropriately apply the results. There are many reporting guidelines related to specific Yet researchers still submit manuscripts to the Journal of types of research. For example, researchers who are writing a Physiotherapy reporting unregistered studies and claiming consis- manuscript about a randomised trial could use the Consolidated tency with the Declaration of Helsinki. All authors are reminded Standards of Reporting Trials (CONSORT) statement.11 The that to be considered by the Journal of Physiotherapy and many statement is summarised into a checklist, which authors can use other physiotherapy journals, manuscripts reporting randomised to identify where in their manuscript each item of the CONSORT trials starting in 2006 or later must be prospectively registered.3 By statement is reported. The completed checklist can also be following the Declaration of Helsinki’s recommendation to submitted with the manuscript for the benefit of reviewers and prospectively register all types of studies, a researcher would editors. The Journal of Physiotherapy has supported the use of make his/her studies more appealing to editors, reviewers and reporting guidelines for over a decade and recently co-published readers because it proves that the research is not biased b][FT17_ID$ y, eg, an editorial reiterating their importance.12 selective reporting of outcomesDF.]_I[T3$ 3 Researchers may also be surprised to learn that the Declaration of Helsinki states that: Unfortunately, as with the Declaration of Helsinki, published reports of research often reveal that these reporting guidelines In advance of a clinical trial, sponsors, researchers and host country have not been used as they were intended. For example, the Journal governments should make provisions for post-trial access for all of Physiotherapy regularly receives manuscripts that report a participants who still need an intervention identified as beneficial randomised trial and state that ‘the CONSORT statement was used in the trial. to guide reporting’, but the appropriate items still do not appear in the paper. This is a problem but, at least at the editorial stage, it is 1836-9553/ß 2015 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/).
104 Editorial fixable. However, it suggests that some researchers view the Replication (TIDieR).18 It also comes with a checklist containing CONSORT statement as an administrative hurdle. It may be more explicit instructions such as: valuable to view it as a resource that they can use to ensure that their research is reported in enough detail to be useful to clinicians Describe the number of times the intervention was delivered and and other researchers. over what period of time, including the number of sessions, their schedule, and their duration, intensity or dose A less fixable problem is heralded by the statement that ‘the CONSORT statement was used to guide the design of the trial’. The and CONSORT statement is a reporting guide, not a design guide. Although many items on the CONSORT checklist are relevant for If the intervention was planned to be personalised, titrated or consideration at the design stage, researchers who use it in this way adapted, then describe what, why, when, and how. risk missing crucial advice about how to[_TD$IF]6 design their clinical trialD$IFT_[81] well. Consider the issue of loss to follow-up in clinical trials, for That is, it guides a researcher to include all the practical example. In a sample of over 10 000 trials of physiotherapy information that clinicians need to apply the study’s intervention. interventions, more than half of the trials failed to follow-up at least The TIDieR checklist is designed to be used in conjunction with 85% of their original participants.13 However, trials can be designed checklists for any study types that involve an intervention. This to incorporate strategies that minimise loss to follow-up.14 would include the CONSORT and SPIRIT statements for randomised Incorporation of these strategies into the trial design is recom- trials, but could also apply to the reporting guidelines for mended in the guide to preparing a trial protocol, known as the observational studies where the participants are receiving an SPIRIT statement (Standard Protocol Items: Recommendations for intervention, a case report of a patient who was receiving a Interventional Trials).15 On this issue of follow-up, the SPIRIT treatment, systematic reviews of such studies, and so on. In this statement provides helpful discussion of issues such as: choosing way, the TIDieR checklist can help researchers to overcome cursory the trial’s duration to maximise clinically relevant outcome guidance about reporting interventions in these other checklists, measurement while minimising loss to follow-up; and which ensuring that the reporting of the intervention is thorough. strategies have been proven to minimise loss to follow-up. It is too late to do anything about loss to follow-up at the reporting stage. Some other resources have been developed to guide research- Accordingly, the CONSORT statement only advises about clear ers on some specialised research issues outside the Equator reporting of the extent of the follow-up, such as distinguishing Network. A recent example is the second version of the Pragmatic- unavoidable loss to follow-up from investigator-determined Explanatory Continuum Index Summary (PRECIS-2).19 This tool exclusion. Other guidelines that could be used at the design stage helps researchers to understand how different aspects of a trial’s are those intended to help researchers in choosing the best outcome design influence whether that trial estimates efficacy (the effect of measures to use in their trials. As examples, the Outcome Measures an intervention when it is administered exactly as intended, ie, an in Rheumatology (OMERACT) website could be used by researchers explanatory trial) or effectiveness (the effect of the intervention in arthritis (www.omeract.org), and the Initiative on Methods, when applied in everyday clinical practice where factors such as Measurement, and Pain Assessment in Clinical Trials (IMMPACT) poor adherence potentially reduce its effect, ie, a pragmatic website has documents to guide pain research (www.immpact.org). trial).20 Although randomised trials have been used as an example Guides for reporting are not strictly limited to clinical research. above, equivalent documents are also available for other study The Guideline for Reporting Evidence-based practice Educational designs: observational studies, systematic reviews, qualitative interventions and Teaching (GREET) has been developed to guide the studies, diagnostic studies, prognostic studies, economic evalua- reporting of educational interventions for developing foundational tions, and so on. Researchers can find many of these documents on knowledge and skills in evidence-based practice.21 Like TIDieR, the Equator Network website]1F[D_.9IT$ GREET can be used in conjunction with checklists such as CONSORT to ensure thorough reporting of the educational intervention. As well as the main reporting guideline for each study type, the Equator Network also compiles extension documents. These are The final point to make is that these documents are helpful. For companion documents with extra detail that is only pertinent (but example, since the CONSORT statement22 was published, there has still very valuable) to some studies with a particular characteristic. been improved reporting of many of the items it recommends for For example, there is an extension to the CONSORT statement that physiotherapy trial reports.13 Researchers should consider them as is specifically designed for non-pharmacologic interventions.16 valuable resources to improve their research, not just as extra This is particularly relevant to physiotherapy intervention trials, paperwork. Some junior researchers even say the checklists help because it more thoroughly addresses issues such as the difficulty them to overcome difficulty in knowing where to start when of blinding physical interventions and the importance of reporting writing a study protocol or report. Therefore, researchers are the skills, experience and any specific extra training of the encouraged to explore each document that is available in relation physiotherapists who apply the study interventions such as to their current and future studies, to read and use each document manual techniques. Despite its relevance to physiotherapy, few carefully – and most crucially – to ensure that each document is authors mention this document in their published trial reports. used for its intended purpose. As noted above, one of the areas in which the CONSORT eAddenda: Nil. statement is too cursory is in its guidance about what intervention Ethics approval: Not applicable. details to include in the published trial report. Specifically, the Competing interests: Nil. CONSORT statement only states: Source(s) of support: Nil. Acknowledgements: I thank Chris Maher for comments on a Describe the intervention for each group with sufficient details to draft of this editorial. allow replication, including how and when they were assessed.11 Provenance: Invited. Not peer-reviewed. Correspondence: Mark Elkins, Department of Respiratory This lacks detail for a researcher who wants to ensure that they Medicine, Royal Prince Alfred Hospital, Sydney, Australia. Email: include all the necessaryTD7_F]$I[ details[I0]F_D$T2 of an intervention for clinical [email protected] physiotherapists to be able to apply it to their patients. But it isn’t just the CONSORT statement; it is a problem that has affected References several of the reporting guidelines[IF]8$D_T to some extent. This deficiency may explain the recent finding that 89% of published trial reports 1. Rickham PP. Br Med J. 1964;2(5402):177. lack enough detail for the intervention to be replicated.17 To rectify 2. World Medical Association. Declaration of Helsinki - Ethical Principles for Medical this, the Equator Network has recently added an extension document called the Template for Intervention Description and Research Involving Human Subjects. 64th WMA General Assembly, Fortaleza, Brazil, October 2013. http://www.wma.net/en/30publications/10policies/b3/ index.html. [Accessed 13th May 2015]. 3. Costa LOP, et al. Braz J Phys Ther. 2012;16:v–ix.
Editorial 105 4. De Angelis C, et al. N Engl J Med. 2004;351:1250–1251. 18. Hoffmann T, et al. BMJ. 2014;348:g1687. 5. le Fort SM, et al. Pain. 1998;74:297–306. 19. Loudon K, et al. Trials. 2013;14:115. 6. Snowdon D, et al. J Physiother. 2014;60:66–77. 20. Hollis S, Campbell F. BMJ. 1999;319:670. 7. Butler GS, et al. Pat Educ Counseling. 1996;28:189–197. 21. Phillips AC, et al. BMC Med Educ. 2013;13:9. 8. Dowsey MM, et al. Med J Aust. 1999;170:59–62. 22. Moher D, et al. BMJ. 2010;340:c869. 9. Condessa RL, et al. J Physiother. 2013;59:101–107. 10. Blackmer J, Haddad H. CMAJ. 2005;173:1052–1053. Websites 11. Schulz KF, et al. BMC Medicine. 2010;8:19. 12. Chan L, et al. Arch Phys Med Rehabil. 2014;95:415–417. www.equator-network.org 13. Moseley AM, et al. J Clin Epidemiol. 2011;64:594–601. www.immpact.org 14. Brueton VC, et al. Cochrane Datab Syst Rev. 2013;12:MR000032. www.omeract.org 15. Chan A-W, et al. BMJ. 2013;346:e7586. 16. Boutron I, et al. Ann Intern Med. 2008;148:295–309. http://dx.doi.org/10.1016/j.jphys.2015.05.012 17. Glasziou P, et al. Lancet. 2014;383:267–276. People’s Choice Award for 2014 The Editorial Board is pleased to introduce the annual People’s Choice Award, which recognises the paper published in Journal of Physiotherapy that generates the most interest by readers of the journal. The winning paper is chosen based on the number of times that each paper published in a given year is downloaded in the six months after its day of publication. The winning paper is ‘Current evidence does not support the use of Kinesio Taping in clinical practice: a systematic review’ by Patrı´cia do Carmo Silva Parreira and colleagues from the Universidade Cidade de Sa˜o Paulo in Brazil.1 This systematic review examined the 12 randomised trials that provided published data by June 2013 about the effect of Kinesio Taping on pain, disability, quality of life, return to work and global perceived recovery in people with musculoskeletal conditions. These trials provided data on 495 participants. Studies were excluded if they were conducted on healthy participants or only reported data on physical performance (eg, vertical jump test). The 12 trials covered a range of musculoskeletal conditions. Among these trials, Kinesio Taping had no benefit over sham taping or other active treatments to which it had been compared, the benefit was too small to be clinically worthwhile, or the trials were of low quality. Therefore the evidence did not support the use of Kinesio Taping for musculoskeletal conditions. The Journal of Physiotherapy has subsequently received and published two more high-quality randomised trials comparing Kinesio Taping to sham taping: one for swelling after ankle sprain by Nunes and colleagues2 and one for low back pain, which again has Patrı´cia do Carmo Silva Parreira as the first author.3 These two trials further reinforce the findings of the review by demonstrating that Kinesio Taping was no better than sham taping for these conditions. The winning paper also generated the highest activity on social media among the papers published in 2014. The Editorial Board of Journal of Physiotherapy congratulates Patrı´cia do Carmo Silva Parreira and colleagues on their success. References 1. Parreira PdCS. Costa LdCM. Hespanhol Junior LC, Lopes AD, Costa LOP. Current evidence does not support the use of Kinesio Taping in clinical practice: a systematic review. J Physiother. 2014;60:31–39. 2. Nunes GS, Vargas VZ, Wageck B, dos Santos Hauphental DP, da Luz CM, de Noronha M. Kinesio Taping does not decrease swelling in acute, lateral ankle sprain of athletes: a randomised trial. J Physiother. 2015;61:28–33. 3. Parreira PCS, Costa LCM, Takahashi R, Hespanhol Junior LC, da Luz Junior MA, da Silva TM, Costa LOP. Kinesio Taping to generate skin convolutions is not better than sham taping for people with chronic non-specific low back pain: a randomised trial. J Physiother. 2014;60:90–96. http://dx.doi.org/10.1016/j.jphys.2015.05.010
Journal of Physiotherapy 61 (2015) 157 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Trial Protocol StressModEx FDI]T–[$6_ Physiotherapist-led Stress Inoculation Training integrated with exercise for acute whiplash injury: study protocol for a randomised controlled trial Carrie Ritchie T[_FIDa$]7 , Justin Kenardy [FD8]_Tb$I , Rob Smeets T]_9FD[$cI , Michele Sterling ]FI[D10$a_T a Menzies Health Institute Queensland, Centre for National Research on Disability and Rehabilitation Medicine (CONROD), NHMRC CRE in Recovery Following Road Traffic Injury, Griffith University, FI1$DT[]G_ old Coast; b CONROD, University of Queensland, [1]$3FDT_BI risbane, Australia; DFT_9[]$cI Department of Rehabilitation Medicine, Maastricht University and Adelante Centre of Expertise in Rehabilitation Medicine and Audiology, Hoensbroek, The Netherlands Abstract physiotherapists more effective than physiotherapy exercise alone in reducing neck pain_F7DI]$1T[ and disability in individuals with acute Introduction: Whiplash associated disorders are the most whiplash associated disorders? Design: Parallel randomised common non-hospitalised injuries following a road traffic crash. controlled trial with blinded outcome assessment. Participants Up to 50% of individuals who experience a whiplash injury will not and $8D_1IF[]Tsetting: 100 individuals with grade II or III (no fracture/ fully recover and report ongoing pain and disability. Most recovery, dislocation or neurological loss) acute whiplash associated if it occurs, takes place in the first 2-3 months post injury, indicating disorder < 4 weeks duration and at least moderate neck pain- that treatment provided in the early stages is critical to long-term related disability and hyper-arousal symptoms will be recruited for outcome. However, early management approaches for people with the study. Participants will be assessed via DT9$1o[_FI] nline surveys or in- acute whiplash associated disorders are modestly effective. One person at a university research laboratory. Interventions will be reason may be that the treatments have been non-specific and have provided at community physiotherapy practices in Brisbane, Gold not targeted the processes shown to be associated with poor Coast, Toowoomba and Mackay, Queensland, Australia. Interven- recovery, such as post-traumatic stress symptoms. Targeting and tion: Clinical-guideline-recommended supervised physiotherapy modulating these early stress responses in the early management of exercise sessions (10 sessions) integrated with six (once per week) acute whiplash associated disorders may improve health outcomes. SIT sessions. Control: Clinical-guideline-recommended supervised Early aggressive psychological interventions in the form of physiotherapy exercise sessions (10) only. Measurements: Primary psychological debriefing may be detrimental to recovery and are (Neck Disability Index) and secondaryF$DT][4I_ (Acute Stress Disorder Scale; now not recommended for management of early post-traumatic Post-traumatic Stress Diagnostic Scale; Depression, Anxiety and stress symptoms. In contrast, Stress Inoculation Training (SIT) is a Stress Scale; Pain Catastrophisingo Scale; Pain Self-Efficacy cognitive behavioural approach that teaches various general Questionnaire; Coping Strategies Questionnaire; Global impression problemFID]_$T-4[1 solving and coping strategies to manage stress-related of recovery; pain intensity; SF36)0_2F[D]$TI outcomes will be measured at anxiety (ie, relaxation training, cognitive restructuring and positive baseline, 6 weeks, 6 months and 12 months DT_12$I[Fa] fter randomisation. self-statements) and provides important information to injured Analysis: Data analysis will be blinded and by intention_]$D[TIF6 toF]TD$I_6[ treat. individuals about the impact of stress on their physical and Outcomes will be analysed using linear mixed and logistic psychological wellbeing. While referral to a psychologist may be regression models that will include baseline scores as covariates, necessary in some cases where acute stress disorder or other more participants as random effects and treatment conditions as fixed significant psychological reactions to stress are evident, in the case factors. Discussion: This study will be the first to address early of acute whiplash injuries, it is neither feasible nor necessary for a stress responses following acute whiplash injury through a novel psychologist to deliver the early stress modulation intervention to intervention that integrates SIT and physiotherapy exercise. all injured individuals. The feasibility of using other specially trained health professionals to deliver psychological interventions Trial registration: Australian New Zealand Clinical Trials has been explored in conditions such as chronic low back pain, Registry. Registration number: ACTRN12614001036606. Was this chronic whiplash and cancer, but few trials have studied this trial prospectively registered: Yes. Funded by: National Health and approach in acute musculoskeletal conditions with the aim of Medical Research Council (NHMRC) Project Grant. Funder approval preventing the development of chronic pain. As physiotherapy is number: APP1069443. Anticipated completion: 31 DecemberDFT$[]5_I the most common intervention received by individuals with a 2016. Provenance: Invited. Not peer-reviewed. Corresponding whiplash injury, physiotherapists are ideally placed to provide SIT author contact details: Professor Michele Sterling, Menzies Health in conjunction with standard physical rehabilitation. This study Institute Queensland, Griffith University, FID$T_1[G] old Coast, Australia (StressModEx) will target individuals in the acute stage of injury 4022. Email: m.sterling@griffith.edu.au and address the stress responses associated with the accident or injury (event-related distress) with the aim of improving both Full protocol: Available on the eAddenda at doi:10.1016/j.jphys. physical and mental health outcomes. Research ]D$I5qT[1F_ uestion: Is $S]DF[_1T6I IT[IF]D$T_3 2015.04.003 integrated with standard physiotherapy exercise and delivered by http://dx.doi.org/10.1016/j.jphys.2015.04.003 1836-9553/Crown Copyright ß 2015 Published by Elsevier B.V. on behalf of Australian Physiotherapy Association. All rights reserved.
Journal of Physiotherapy 61 (2015) 159 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Critically Appraised Papers Surgery for displaced fracture of the proximal humerus may not result in better outcomes than nonsurgical management Synopsis Summary of: Rangan A, Handoll H, Brealey S, Jefferson L, Keding were health-related quality of life (Short-Form 12 health survey), A, Marton BC, et al. Surgical vs nonsurgical treatment of adults with complications related to the fracture and surgery, and mortality. displaced fractures of the proximal humerus: the PROFHER Results: The primary analysis was completed on 231 participants. randomized controlled trial. JAMA. 2015;313:1037-1047. There were no between-group differences in the mean Oxford Shoulder Scores (mean difference 0.8 points, 95% CI –1.3 to 2.8) Question: Does surgical treatment improve patient-reported over the 2 years or at any time point. Over the 2 years there were outcomes in adults with displaced fractures of the proximal also no significant differences between the groups in the physical humerus? Design: Randomised controlled trial with concealed or mental component scores of the Short-Form 12 survey, in allocation. Setting: Thirty-two acute hospitals in the United fracture and surgical complications (30 participants in the surgical Kingdom. Participants: Adults presenting with a displaced group vs 23 in the nonsurgical group), or in mortality (seven fracture of the proximal humerus involving the surgical neck participants in the surgical group vs five in the nonsurgical group). were included. Patients were excluded if they had an associated Both groups received a median of eight sessions of physiotherapy, dislocation of the injured shoulder joint, if they had co-morbidities with more than 90% of physiotherapy sessions including pre- that precluded surgery or anaesthesia, or if illness was terminal. scribed exercise and advice. Conclusion: Surgical treatment for Randomisation of the 250 patients allocated 125 to the surgical displaced fracture of the proximal humerus followed by physio- group and 125 to the non-surgical group. Interventions: Both therapy rehabilitation did not improve clinical outcomes or reduce groups received rehabilitation care provided by physiotherapists complications compared with simply providing a sling and in inpatient, outpatient and community settings. In addition, the providing physiotherapy rehabilitation. surgical group received either internal fracture fixation or hemi- arthroplasty (humeral head replacement). The nonsurgical group Provenance: Invited. Not peer-reviewed. were given a sling for the injured arm for a suggested period of 3 weeks. Outcome measures: The primary outcome was the Oxford Nicholas Taylor Shoulder Score at 6, 12 and 24 months after injury. The Oxford Section Editor, Journal of Physiotherapy Shoulder Score provides a total score based on the patient’s subjective report of pain and function, ranging from 0 (worst http://dx.doi.org/10.1016/j.jphys.2015.05.003 outcome) to 48 (best outcome). Secondary outcome measures Commentary The decision for conservative or surgical management of Rehabilitation is recommended to be functional and activity proximal humeral fractures is widely discussed, and many based, and so is measurement of shoulder function.4 Rangan and displaced proximal humeral fractures (displacement of > 1 cm colleagues used the patient-reported Oxford Shoulder Score as the and/or 45 FI]3[d_TD$ eg angulation of fracture parts) are typically considered primary outcome. The mean scores ranged from 35.6 (6 months) to for surgical intervention. Current evidence indicates that clinical 40.4 (2 years post-fracture) on a 48-point scale. This difference is outcomes are similar to conservative management and that around their 5D$-]I_[1FT DF$4I[p]_T oint minimal clinically important difference. This complication rates are higher.1,2 The pragmatic trial by Rangan could inform us that little improvement may occur past active and colleagues adds high-quality evidence to this limited body of rehabilitation, but also that (more) responsive and functional knowledge, with similar findings of no significant or clinically measures should be used.4 important differences between surgical and conservative manage- ment groups. Although of great value, the findings may not be [5ID$_TPF] rovenance: Invited. Not peer-reviewed. generalisable to more complex patients. Eighty percent of people screened for eligibility were excluded, mostly due to confounding Alexander TM van de Water comorbidities, mental capacity or associated dislocation. Department of Physiotherapy, Saxion University of Applied Sciences, Physiotherapy aims to help patients with their functional Enschede, The Netherlands recovery after fracture. No evidence-based physiotherapy guide- lines are available for the treatment of people recovering from a References proximal humeral fracture. For the trial by Rangan and colleagues, a general physiotherapy protocol with recommendations in six 1. Fu T, et al. Int J Clin Exp Med. 2014;7:4607–4615. rehabilitation phases was developed and received by both groups.3 2. Li Y, et al. PLoS ONE. 2013;8:e75464. It represents current practice with pain management, staged 3. Handoll H, et al. Health Technol Assess. 2015;19:1–280. progression in mobilisation of the upper limb in general and the 4. Van de Water ATM, et al. BMC Musculoskelet Disord. 2015;16:31. glenohumeral joint, followed by progressive strengthening exer- cises and functional use of the arm. http://dx.doi.org/10.1016/j.jphys.2015.05.008 1836-9553/ß 2015 Australian Physiotherapy Association. Published by Elsevier B.V. All rights reserved.
Journal of Physiotherapy 61 (2015) 159 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Critically Appraised Papers Surgery for displaced fracture of the proximal humerus may not result in better outcomes than nonsurgical management Synopsis Summary of: Rangan A, Handoll H, Brealey S, Jefferson L, Keding were health-related quality of life (Short-Form 12 health survey), A, Marton BC, et al. Surgical vs nonsurgical treatment of adults with complications related to the fracture and surgery, and mortality. displaced fractures of the proximal humerus: the PROFHER Results: The primary analysis was completed on 231 participants. randomized controlled trial. JAMA. 2015;313:1037-1047. There were no between-group differences in the mean Oxford Shoulder Scores (mean difference 0.8 points, 95% CI –1.3 to 2.8) Question: Does surgical treatment improve patient-reported over the 2 years or at any time point. Over the 2 years there were outcomes in adults with displaced fractures of the proximal also no significant differences between the groups in the physical humerus? Design: Randomised controlled trial with concealed or mental component scores of the Short-Form 12 survey, in allocation. Setting: Thirty-two acute hospitals in the United fracture and surgical complications (30 participants in the surgical Kingdom. Participants: Adults presenting with a displaced group vs 23 in the nonsurgical group), or in mortality (seven fracture of the proximal humerus involving the surgical neck participants in the surgical group vs five in the nonsurgical group). were included. Patients were excluded if they had an associated Both groups received a median of eight sessions of physiotherapy, dislocation of the injured shoulder joint, if they had co-morbidities with more than 90% of physiotherapy sessions including pre- that precluded surgery or anaesthesia, or if illness was terminal. scribed exercise and advice. Conclusion: Surgical treatment for Randomisation of the 250 patients allocated 125 to the surgical displaced fracture of the proximal humerus followed by physio- group and 125 to the non-surgical group. Interventions: Both therapy rehabilitation did not improve clinical outcomes or reduce groups received rehabilitation care provided by physiotherapists complications compared with simply providing a sling and in inpatient, outpatient and community settings. In addition, the providing physiotherapy rehabilitation. surgical group received either internal fracture fixation or hemi- arthroplasty (humeral head replacement). The nonsurgical group Provenance: Invited. Not peer-reviewed. were given a sling for the injured arm for a suggested period of 3 weeks. Outcome measures: The primary outcome was the Oxford Nicholas Taylor Shoulder Score at 6, 12 and 24 months after injury. The Oxford Section Editor, Journal of Physiotherapy Shoulder Score provides a total score based on the patient’s subjective report of pain and function, ranging from 0 (worst http://dx.doi.org/10.1016/j.jphys.2015.05.003 outcome) to 48 (best outcome). Secondary outcome measures Commentary The decision for conservative or surgical management of Rehabilitation is recommended to be functional and activity proximal humeral fractures is widely discussed, and many based, and so is measurement of shoulder function.4 Rangan and displaced proximal humeral fractures (displacement of > 1 cm colleagues used the patient-reported Oxford Shoulder Score as the and/or 45 FI]3[d_TD$ eg angulation of fracture parts) are typically considered primary outcome. The mean scores ranged from 35.6 (6 months) to for surgical intervention. Current evidence indicates that clinical 40.4 (2 years post-fracture) on a 48-point scale. This difference is outcomes are similar to conservative management and that around their 5D$-]I_[1FT DF$4I[p]_T oint minimal clinically important difference. This complication rates are higher.1,2 The pragmatic trial by Rangan could inform us that little improvement may occur past active and colleagues adds high-quality evidence to this limited body of rehabilitation, but also that (more) responsive and functional knowledge, with similar findings of no significant or clinically measures should be used.4 important differences between surgical and conservative manage- ment groups. Although of great value, the findings may not be [5ID$_TPF] rovenance: Invited. Not peer-reviewed. generalisable to more complex patients. Eighty percent of people screened for eligibility were excluded, mostly due to confounding Alexander TM van de Water comorbidities, mental capacity or associated dislocation. Department of Physiotherapy, Saxion University of Applied Sciences, Physiotherapy aims to help patients with their functional Enschede, The Netherlands recovery after fracture. No evidence-based physiotherapy guide- lines are available for the treatment of people recovering from a References proximal humeral fracture. For the trial by Rangan and colleagues, a general physiotherapy protocol with recommendations in six 1. Fu T, et al. Int J Clin Exp Med. 2014;7:4607–4615. rehabilitation phases was developed and received by both groups.3 2. Li Y, et al. PLoS ONE. 2013;8:e75464. It represents current practice with pain management, staged 3. Handoll H, et al. Health Technol Assess. 2015;19:1–280. progression in mobilisation of the upper limb in general and the 4. Van de Water ATM, et al. BMC Musculoskelet Disord. 2015;16:31. glenohumeral joint, followed by progressive strengthening exer- cises and functional use of the arm. http://dx.doi.org/10.1016/j.jphys.2015.05.008 1836-9553/ß 2015 Australian Physiotherapy Association. Published by Elsevier B.V. All rights reserved.
Journal of Physiotherapy 61 (2015) 148–154 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Research The economic value of an investment in physiotherapy education: a net present value analysis George Rivers a, Jonathan Foo b, Dragan Ilic c, Peter Nicklen b, Scott Reeves d, Kieran Walsh e, Stephen Maloney b a Faculty of Business and Economics; b Department of Physiotherapy; c Medical Education, Research IF4D_T][$and Quality Unit, School of Public Health $4FaD]_TI[ nd Preventive Medicine, Monash University, Australia; d Centre for Health a]_FD[I$4T nd Social Care Research, Kingston University and StID]$F[1_T George’s University of London; e BMJ Learning, BMJ Group, UK KEY WORDS ABSTRACT Cost-benefit analysis Questions: What is the economic value for an individual to invest in physiotherapy undergraduate Economic evaluation education in Australia? How is this affected by increased education costs or decreased wages? Design: A Education cost-benefit analysis using a net present value (NPV) approach was conducted and reported in Australian Physiotherapy dollars. In relation to physiotherapy education, the NPV represents future earnings as a physiotherapist minus the direct and indirect costs in obtaining the degree. Sensitivity analyses were conducted to consider varying levels of experience, public versus private sector, and domestic versus international student fees. Comparable calculations were made for educational investments in medicine and nursing/ midwifery. Results: Assuming an expected discount rate of 9.675%, investment in education by domestic students with approximately 34 years of average work experience yields a NPV estimated at $784,000 for public sector physiotherapists and $815,000 for private sector therapists. In relation to international students, the NPV results for an investment and career as a physiotherapist is estimated at $705,000 in the public sector and $736,000[D$IF2]_T in the private sector. Conclusion: With an approximate payback period of 4 years, coupled with strong and positive NPV values, physiotherapy education in Australia is a financially attractive prospect and a viable value proposition for those considering a career in this field. [Rivers G, Foo J, Ilic D, Nicklen P, Reeves S, Walsh K, Maloney S (2015) The economic value of an investment in physiotherapy education: a net present value analysis. Journal of Physiotherapy 61: 148–154] ß 2015 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 expressed as the total discounted benefit (value of consumption gained) less the total discounted cost (value of consumption In Australia, the number of students choosing to invest in foregone). This comparison is enabled by applying a discount rate knowledge through higher education is increasing, reaching over to convert future costs and benefits to present values. NPV is half a million in 2012.1 Physiotherapy, in particular, has grown consistent with the ideas of both individual time preference, where rapidly, with a 35% increase in students graduating from Australian consumption today is preferred over consumption tomorrow, and programs between 2004 and 2006.2 These graduating students risk, in that future cash flows are not guaranteed. In relation to have enjoyed a job market in which there is a shortage of physiotherapy education, the NPV measure compares the future physiotherapists.3 Physiotherapy shortages are expected to earnings as a physiotherapist minus the costs of obtaining the degree. continue as demand grows due to an aging population, population growth and rising incidence of chronic disease.4 Despite this Economic analysis of education is not a new concept. Studies growth in the profession, a South Australian survey of 561 phy- have been conducted on the costs of American medical education, siotherapists found that more than 60% of respondents believed as the fees are often prohibitive, with the average student that their remuneration was too low.5 graduating with US$167,000 of debt.7 Using an NPV measure of medical education in 2010, Kahn and Nelling8 found that a medical Cost-benefit analysis, which is often used to evaluate policies and degree is a worthwhile investment up to costs of US$140,000 per projects,6 can be used to understand the economic value of pursuing year of study. However, the calculations in these studies have been physiotherapy education in the context of a lifelong career. In cost- based on broad assumptions, using modelled data in the absence of benefit analysis, time-dependent costs and benefits are expressed in actual data. No study of this type has been conducted on money terms, which are based on the preferences of the individuals physiotherapy in Australia or in any other country. affected. The main measure of the total value of an economic decision in cost-benefit analysis is its net present value (NPV). NPV is Physiotherapy education in Australia is growing; there are 17 accredited courses ranging from 4-year undergraduate http://dx.doi.org/10.1016/j.jphys.2015.05.015 1836-9553/ß 2015 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 149 baccalaureates to 2-year or 3-year Master or Doctor of three independent formulas that model present value and are Physiotherapy degrees.9 Qualified physiotherapists enjoy a based on known data and listed assumptions (Box 1). These three wide scope of practice in both the public and private sector, formulas were combined using the following equation to estimate including sports, cardiorespiratory, neurological and paediat- overall NPV: rics.10 A cross-sectional study of 273 Swedish physiotherapy students by O¨ hman and colleagues11 explored the reasons for NPV ¼ PV future earnings À ðPV direct costs þ PV indirect costsÞ choosing a career in physiotherapy. These included an interest in sports and athletics, in working with people, being influenced by Present values (PV) were calculated for future earnings, as well as interaction with physiotherapists and seeing it as an appealing both the direct and indirect costs, reflecting the full opportunity profession. In that study, none of the students reported having cost of undertaking the investment in physiotherapy education. chosen the profession to make a good living. However, it should be Opportunity costs represent the cost of the option forgone as a considered that higher education in Sweden is free of charge for result of the decision to undertake the physiotherapy degree. In Swedish students, while Australian physiotherapy students con- each formula, future cash flows (C) were weighted by a nominal tribute in excess of AUD8,000 per year.12,13 If the government discount rate (r) to calculate the present value, factoring in risk and funding cuts and fee deregulation occur, these contributions could time preference. Following recommendations from the Office of soon increase to as much as international students pay.14 Best Practice Regulation,16 r-values were adjusted to align with McMeeken15 argues that physiotherapy education is already market expectation for consumer price index of 2.5%. A sensitivity disadvantaged through funding mechanisms, with its low attrition analysis was conducted for low-risk, expected-risk and high-risk rates and the high cost of clinical education. As a highly popular scenarios.17 These r-values were 5.575%, 9.675% and 12.750%, course, this would make physiotherapy a prime target for fee respectively. Public and private salary growth rates (g) of 3.84% and increases in a deregulated education market. 3.95%, respectively, were used to estimate future cash-flow earnings. The calculations were modelled for a domestic student Despite the lack of economic data, Szuster and Carson5 found commencing a 4-year physiotherapy degree in 2012 and com- that 75% of physiotherapists chose their profession in order to be mencing work at 22 years old. financially secure. To better understand this concept of economic incentives in career choice, the present study aimed to model the Present value of direct costs value of physiotherapy education through analysing real survey data using a NPV approach. In doing so, this study was intended to PV direct costs ¼ Cdð1 þ rÞ \" À 1 þ g4# present the financial motivation associated with choosing an rÀg 1 1 þ r investment in physiotherapy education, and how this would be affected by increased education costs or decreased wages. Direct costs are the amounts incurred by students as a direct Therefore, the research questions for this study were: result of undertaking the university degree. This was modelled as a 4-year growing annuity. This includes university fees, required 1. What is the economic value for an individual to invest in textbooks, uniform and documentation necessary to obtain the physiotherapy undergraduate education in Australia? degree. Course fees represent an average of three advertised fees for domestic students for a 4-year undergraduate physiotherapy 2. How is this affected by increased education costs or decreased degree in an Australian university. Other cash flows were wages? calculated based on known figures from a leading Australian physiotherapy program. Net cash flow was calculated to be $8,527 Method for 1 year. A growth rate (g) of 2.5% was applied to account for inflation, aligning with market expectations for consumer price Design index.17 This study utilised a cost-benefit analysis with a NPV approach. All dollar values presented in the paper are Australian dollars unless otherwise stated. The calculation of NPV was made using Box 1. Key values and assumptions used in net present value model. Variable Values Assumptions Discount rate (r) Growth rate (g) Low risk = 5.575% Direct costs increase with consumer price index Direct cash flow (Cd) Expected risk = 9.675% Indirect cash flow (Ci) High risk = 12.75% Working full time in either public or private sector Earnings cash flow (Ce) Consumer price index = 2.5% Taxed at Australian tax rates for 2014 to 2015 Public salary = 3.84% Time (t) Private salary = 3.96% 4-year degree from 2012 to 2015 without Domestic fee = $7,960/yr impeded progress International fee = $29,786/yr Other expenses = $]TFI5D3$_[ 67/yr Commence working at 22 years of age Value of time = $22/hr Total course time = 3504.5 hr Anticipated annual salary increase Public = 3.84% Private = 3.96% Continuing professional development and annual registration costs = $1,159 Cost of living $25,009 per year
150 Rivers et al: Economic value of investment in physiotherapy education Present value of indirect costs at which the present value of total costs equals the present value of future earnings (ie, where NPV equals $0). This calculation utilises PV indirect costs ¼ X4 Ci expected wage growth with experience and inflation, as well as t¼1 ð1 þ rÞt average career length. Values are calculated using the ‘what-if analysis’ function in Excel softwarea. Investment in education for The value of time spent obtaining the physiotherapy degree was careers in medicine and nursing/midwifery were also chosen for represented in dollar terms as indirect costs. Seventy-three third- comparison purposes. University fees and course structures were year Australian physiotherapy students from one university cohort based on advertised values from an Australian university for a were surveyed, being asked: ‘If you had to place a dollar value on 5-year undergraduate bachelor’s medical degree and a 3-year your time, what would you give it?’ and ‘How many hours would undergraduate nursing/midwifery bachelor’s degree. Other direct you estimate in total, that you spent engaged in learning activities costs were taken to be the same as physiotherapy education. and study for the 10 weeks of the semester?’ The response rate was Indirect costs were based on extrapolation from the direct 93% (n = 68). From this, the median value that students placed on observations of physiotherapy students and applied on a pro rata their time was $22 per hour, with an interquartile range of $20 to basis to varying course structures. Future earnings were based on $30 per hour. The median time spent on learning activities was the same Australian Bureau of Statistics source data as physio- 300 hours, with an interquartile range of 245 to 400 hours. This therapy, and expected career length from the Australian Health information, together with timetable information from an Practitioner Regulation Agency survey data.18,23 Wage increases Australian physiotherapy program, was used to calculate the were set at the average of public and private growth. All other financial value of time spent per year attributed to the values and calculations were identical to physiotherapy. physiotherapy degree. Results Present value of future earnings Net present value for domestic students PV future earnings ¼ Xt Ce t¼1 ð1 þ rÞt The results of NPV calculations for investments in education by domestic students are shown in Table 1. They represent the low- Upon obtaining their degree, physiotherapists may take any risk (5.575%), expected-risk (9.675%) and high-risk (12.75%) after- number of career pathways, including a mix of public and private tax future cash flows. For an average of 33.86 years of work as a sector roles, management, research and/or education. In order to physiotherapist, the NPV results range from $541,000 in the public make the results generalisable, we have only modelled for sector to $1,492,000 in the private sector (Table 1). All NPV individuals who chose to work exclusively as physiotherapists measures include an average annual direct cost of attendance of in either the public sector or private sector. Expected full-time $8,527 per annum, and are dependent on years of experience (t), wage in both the public and private sector was calculated from the discount rate (r) and salary growth rate (g) for the public and Australian Bureau of Statistics data.18 Wages are expected to private sectors. Importantly, the NPV results are based on an increase due to two factors: increases in experience and position; individual’s experience and competency level, as future earnings and increases due to inflation over time. Experience-based wage reflect jumps to a new salary scale achieved with promotion at increase was estimated by analysing wage data in 5-year 5-year intervals. increments. Inflation was applied using the 20-year Australian average of annual wage increases of 3.84% for public growth and Net present value for international students 3.96% for private growth. Weekly wages were then converted to an annual amount allowing for 4 weeks of annual leave, 9.5% For international students, the indirect opportunity cost of compulsory employer superannuation contribution and Australian education (time cost) was taken to be identical to domestic Tax Office tax rates for 2014 to 2015.19,20 The amounts for annual students; however, the direct attendance cost is higher at $30,354 professional registration and continuing professional development per annum. The higher direct attendance cost takes into account were estimated and deducted to calculate net cash flow. the higher fees for international students compared to domestic students. For an average of 33.86 years of work, the NPV results The amount of return from the physiotherapy degree is range from $464,000 in the public sector to $1,400,000 in the dependent on how long an individual continues to work. The private sector (Table 2). average years of work (t) were estimated using Service and Workforce Planning 2007 survey data21 and found to be Payback analysis 33.86 years; additional sensitivity analysis was completed for individuals who spend more or less time in the profession. A payback analysis was conducted to determine the number of Calculations using cost of living were based on the 2010 Austra- years it would take a physiotherapy student to recover the direct lian Bureau of Statistics data on household expenditure, which and indirect costs of their education investment offset by cost of were adjusted to the individual and increased with the consumer living. The payback period for domestic students is between price index.22 3.77 years and 4.08 years, depending on the risk to future cash flows and whether they are working in the public or private sector A break-even analysis was conducted to determine the hypothetical minimum requirement for year-1 earnings (t = 1), Table 1 Net present value of investment in physiotherapy education by domestic students who commence work in public or private sectors, for different years of experience (t) and discount rates (r). Years worked Public sector (Salary growth rate g = 3.84%) Private sector (Salary growth rate g = 3.96%) t = 10 Low risk r = 5.575% Expected risk r = 9.675% High risk r = 12.75% Low risk r = 5.575% Expected risk r = 9.675% High risk r = 12.75% t = 20 t = 30 $398,880 $313,231 $264,343 $415,266 $326,524 $275,864 t = 33.86 $886,721 $586,160 $444,724 $919,842 $608,936 $462,578 t = 40 $1,296,862 $743,242 $523,585 $1,347,397 $772,722 $544,813 t = 48 $1,437,528 $784,233 $540,876 $1,492,371 $814,967 $562,093 $1,643,725 $833,893 $558,073 $1,687,475 $862,000 $578,885 $1,883,023 $878,219 $571,172 $1,904,707 $902,221 $590,768
Research 151 Table 2 Net present value of investment in physiotherapy education by international students who commence work in public or private sectors, for different years of experience (t,) and discount rates (r). Years worked Public sector (Salary growth rate g = 3.84%) Private sector (Salary growth rate g = 3.96%) Low risk r = 5.575% Expected risk r = 9.675% High risk r = 12.75% Low risk r = 5.575% Expected risk r = 9.675% High risk r = 12.75% t = 10 $315,315 $234,125 $188,238 $331,702 $247,419 $199,758 t = 20 $803,156 $507,054 $368,619 $836,277 $529,831 $386,473 t = 30 $1,213,297 $664,136 $447,479 $1,263,832 $693,617 $468,708 t = 33.86 $1,353,963 $705,127 $464,771 $1,408,806 $735,862 $485,988 t = 40 $1,560,160 $754,787 $481,968 $1,603,910 $782,894 $502,780 t = 48 $1,799,458 $799,113 $495,067 $1,821,143 $823,116 $514,663 (Table 3). Given the higher direct cost of education for has a lower expected career NPV. With a matched comparison of international students, the payback period is slightly longer: years worked, up to t = 10, public physiotherapy has the higher between 6.37 years and 7.91 years. Importantly, the payback NPV. After 10 years, medical practitioners have the highest NPV period for both domestic and international students was estimated values, followed by private physiotherapy, public physiotherapy, taking into account the individual’s time preference and 5-year then nursing/midwifery (Figure 2). earning increases due to experience. Discussion Break-even analysis Whilst less tangible benefits of undertaking an investment in For domestic students, break-even wages range from $3,280 to physiotherapy education are not considered, this study suggests $8,543, depending on the sector of employment and risk profile that an investment in education as a physiotherapist makes good (Table 3). For international students, direct costs are higher, financial sense. The longer the time spent in the profession, and resulting in higher break-even values ranging from $6,062 to with subsequent increases in experience level, the greater the $16,150. Break-even values in the private sector are consistently benefit. This is consistent with the fact that over 75% of higher than the public sector, indicating that there is less physiotherapists indicated that they chose their profession in percentage growth in wages attributed to experience in the order to be financially secure.5 Assuming an expected discount rate private sector. The starting wage needed to break even is higher for of 9.675%, investment in education by domestic students who those who spend less time in the profession. work for approximately 34 years yields a NPV of around $784,000 for public-sector physiotherapists and $815,000 for private-sector Net present value for public versus private sectors physiotherapists. Lower discount rates would only increase the NPVs of such an investment. Moreover, the small difference in At all discount rates and levels of experience, the NPV for the NPVs for careers in the public and private sectors means very little private sector was found to be higher than in the public sector. financial advantage of one sector over the other. For international However, as shown in Figure 1, when comparing effective rates of students, the NPVs for an investment and career as a physiothera- pay per hour, the public sector is shown to be consistently higher pist are estimated to be $705,000 in the public sector and $736,000 for any age group over 24 years. That is: while salaries are higher in in the private sector. This variation in NPVs for domestic and the private sector, the number of hours worked is also higher, international student cohorts is driven by the difference in the ranging from 1.0 to 5.6 additional hours per week compared to present value of direct and indirect costs of education, which is public-sector therapists. The age groups 55 to 59 years and 65 to around $90,700 for domestic students and nearly double at around 69 years show the largest difference in hours worked per week, $169,800 for international students. with an additional 4.2 and 5.6 private hours worked, respectively. Subsequently, the respective public hourly earnings were higher Anecdotally, many physiotherapists believe that careers in the by approximately $5.70 and $9.70. private sector yield higher returns than careers in the public sector, with the majority of students aspiring to work in private practice.24 Comparison to medicine and nursing/midwifery However, whilst NPV figures are consistently greater in the private sector than in the public sector, effective hourly earnings are For medical professionals, and assuming an average career higher in the public sector than in the private sector. This result is length of 42.01 years, NPV amounts range from approximately not unusual, given that private therapists work longer hours – $661,000 to $2,146,000, based on risk profile (Table 4). The average anywhere between 1.0 and 5.6 hours per week more, depending on career length in nursing/midwifery is 39.64 years, resulting in NPV amounts of approximately $652,000 to $1,445,000, based on risk DT1_e)rug$]GiIF([ profile (Table 4). Despite having an expected career length of almost 6 years more than physiotherapy, nursing/midwifery still 45 Table 3 Earnings ($/hr)40 Present value of costs with payback period and break-even wage after physiotherapy education for domestic and international students. 35 Low risk Expected risk High risk 30 Domestic student cost present value $98,527 $90,659 $85,462 25 Public payback period, (yr) 3.99 4.04 4.08 Private payback period, (yr) 3.77 3.80 3.82 20 Public break-even wage Private break-even wage $3,280 $5,108 $6,564 15 $5,053 $7,062 $8,543 Public International student cost present value $182,092 $169,764 $161,568 Public payback period, (yr) 10 Private payback period, (yr) 6.70 7.30 7.91 Private Public break-even wage 6.37 6.91 7.39 Private break-even wage $6,062 $9,565 $12,409 5 $9,338 $13,224 $16,150 0 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 Age group (yr) Figure 1. Earnings per hour in public and private sectors, by age group.
152 Rivers et al: Economic value of investment in physiotherapy education Table 4 Costs, salary and average career duration for physiotherapy, medicine, and nursing/midwifery, with net present value data for different years of experience (t) with expected risk (r = 9.675%). Direct costs Physiotherapy public sector Physiotherapy private sector Medicine all sectors Nursing/Midwifery all sectors Indirect costs Salary ($/hr), range $30,906 $30,906 $47,529 $20,547 Career length (yr), average $59,753 $59,753 $126,830 $51,862 Net present value, by years worked 27.19 to 40.79 27.73 to 36.18 27.53 to 45.53 26.50 to 32.51 t = 10 t = 20 33.86 33.86 42.01 39.64 t = 30 t = 40 $313,231 $326,524 $325,555 $300,027 t = expected $586,160 $608,936 $676,123 $536,161 $743,242 $772,722 $886,825 $670,345 $833,893 $862,000 $1,008,078 $747,167 $784,233 $814,967 $1,025,512 $745,146 (t = 33.86) (t = 33.86) (t = 42.01) (t = 39.64) [(Figure_2)TD$IG] 1,200,000 financial prospects of a career in medicine, lasting an average of 10 years more than physiotherapy, resulting in $200,000 greater 1,000,000 NPV. Various other models have been used to measure the NPV of medical education in the United States, with values ranging from Net present value ($) 800,000 US$430,000 to US$2,233,000.8 Differences can be attributed to different estimates of career length, different risk and growth 600,000 Physiotherapy Public models and different source data. Importantly, these studies are 400,000 Physiotherapy Private not comparable to Australian data, as university fee structure and 200,000 Medicine wages are inherently different. Nursing/Midwifery 0 Taking an expected risk of future cash flows, this study reveals that the first year earnings required to breakeven with the -200,000 investment in physiotherapy education would only need to be $5,100 in the public sector and $7,100 in the private sector -400,000 5 10 15 20 25 30 35 40 (Table 4). Break-even estimates represent the theoretical limit to 0 Period worked (yr) which wages could be decreased and still break even. Whilst it would be unrealistic to see wages decrease to such low break-even Figure 2. Net present value comparison of physiotherapy, medicine and nursing/ values, these values indicate that regardless of wage changes, midwifery over years worked. physiotherapy is likely to remain a positive financial investment. Coupled with strong and positive NPVs, this makes physiotherapy the age cohort. Physiotherapists may consider whether a lifetime education in Australia for both domestic and international NPV difference of $31,000 is worth working the extra hours, given students a financially attractive prospect. that over 30% of physiotherapists indicated that they worked more hours than they would like.5 In the same cross-sectional study, An international comparison of country profiles in relation to over 50% of doctors agreed that they worked more hours than they the physiotherapy sector shows that Australian education is not would like. Career length should be considered with the higher excessively expensive in terms of direct costs (see Table 5). The direct cost of attendance in Australia reflects 53% of annual salaries, as compared to other countries: ranging from 21 to 152% in Singapore and India, respectively. In a deregulated market with Table 5 Physiotherapy course duration and cost, absolute salary, as salary as a percentage of the cost of education, by country. Country Years of studya Cost of educationa Salaryb Cost % of salary United States of America 3 year PG $87,831 $78,766 112% Australia 4 year UG $31,840 $60,280 53% Canada 2 year PG $15,924 $64,361 25% Netherlands 4 year UG $17,620 $51,976 34% India 4.5 year UG $5,256 $3,458 152% Singapore 3 year UG $6,972 $32,904 21% United Kingdom 3 year UG $51,810 $49,276 105% PG = postgraduate, UG = undergraduate. a Data based on randomly selected higher education programs. b All figures from www.payscale.com converted to AUD in December 2014. Table 6 Sensitivity analysis with variable direct costs of net present value, payback period and break-even analysis using expected time (t), growth rate (g) and risk (r), by sector. Total direct cost Public sector Private sector $50,000 Net present Payback period Break-even wage Net present Payback Break-even $75,000 value (t = 33.86) value (t = 33.86) period wage $100,000 4.76 $6,184 $125,000 $765,138 5.78 $7,592 $795,873 4.49 $8,549 $150,000 $740,138 6.86 $9,001 $770,873 5.45 $10,497 $715,138 8.01 $10,409 $745,873 6.47 $12,444 $690,138 9.02 $11,818 $720,873 7.53 $14,391 $665,138 $695,873 8.54 $16,339
Research 153 increased education fees, economic incentives in physiotherapy information may allow for a better understanding of why would see significant changes. Table 6 presents a sensitivity physiotherapists choose particular specialties, settings, locations, analysis of direct costs ranging from $50,000 to $150,000, revealing or career longevity. significantly lower NPVs assuming expected career length and risk profile. This is associated with longer payback periods ranging What is already known on this topic: The number of from 4.49 to 9.02 years, which in a workforce where 10% of students investing in tertiary education in physiotherapy is physiotherapists under 25 years of age expect to work less than increasing. Most students aspire to work in private practice. 5 years, may deter potential applicants.21 Many physiotherapists believe that their remuneration is too low. The results in this paper are based on real survey data relating What this study adds: An investment in education as a to the indirect costs of education, the direct cost data of Australian physiotherapist makes good sense financially, with a brief physiotherapy degrees and the earnings data from the Australian pay-back period. The small difference in net present values Bureau of Statistics. The use of real survey data allows for a more for careers in the public and private sectors means little accurate estimation of indirect costs and, in doing so, provides a financial advantage of working in one sector over the other. comprehensive estimate of full opportunity costs of the education investment. Estimates of NPV are likely to be conservative due to Footnotes: aMicrosoft, Redmond, USA. the ceiling effects of assumptions around Australian Bureau of Ethics approval: Ethics approval for the attainment of student Statistics data relating to earnings data of wage levels over $2000 financial data was obtained through the Monash University per week. The ceiling effect is most pronounced in medicine, Human Research Ethics Committee, approval number CF14/307 followed by physiotherapy then nursing, as the percentage of the – 2014000115. workforce falling into the ceiling category is 70%, 12% and 5%, Competing interests: None. respectively. Additionally, combined with the ceiling effect Source of support: None. relating to the upper limit of Australian Bureau of Statistics data Acknowledgements: None. of over 49 hours per week, estimations of earnings per hour are Provenance: Not invited. Peer-reviewed. also likely to be conservative. Finally, to improve the accuracy of Correspondence: Stephen Maloney, Department of Physiother- NPV, calculations the majority of cost data used was modelled apy, Monash University, Frankston, Australia. Email: stephen. from real data. However, no data were available on the cost of [email protected] ongoing professional development, and it was therefore taken to be $1,000 per year. Typically, however, there is a wide range of References options for development, including obtaining higher university degrees and attending workshops or online seminars. As a result, 1. Department of Education. Selected Higher Education Statistics - 2012 Student Data. costs are likely to vary from year to year and from individual to http://www.education.gov.au/ individual. selected-higher-education-statistics-2012-student-data [accessed 1/11/2014]. This analysis did not consider the effects of non-linear or 2. McMeeken J, Grant R, Webb G, Krause K-L, Garnett R. Australian physiotherapy alternative career pathways such as mixed public/private career student intake is increasing and attrition remains lower than the university profiles, non-clinical roles and part-time work. It is worth average: a demographic study. Aust J Physiother. 2008;54:65–71. considering the impact of maternity leave and family formation on the financial viability of physiotherapy, given that Health 3. Australian Physiotherapy Association. 2012–13 Pre-budget Submission. Melbourne, Workforce Australia notes hours worked per week is on average Australia; Australian Physiotherapy Association; 2012. approximately 10 hours less for females – a difference most pronounced in the age bracket 35 to 44 years.23 It is hypothesised 4. Schofield DJ, Fletcher SL. The physiotherapy workforce is ageing, becoming more that many females may choose to return to the workforce on a masculinised, and is working longer hours: a demographic study. Aust J Physiother. part-time basis during family formation, which is reflected in that 2007;53:121–126. 87% of part-time physiotherapists are female. This is dispropor- tionate to the general workforce, which is 69% female.18 It is not 5. Szuster F, Carson E. Career study of South Australian medical, dental and physiother- possible, nor was it the intention, for this study to determine the apy graduates. Adelaide, Australia: South Australian Department of Health; 2007. financial viability of all career pathways. However, given that the payback period for domestic students is approximately 4 years, it is 6. Abelson P. Public Economics - Principles and practices. Sydney, Australia: Applied likely that many forms of part-time work or career structures Economics; 2003. would result in a positive investment. 7. Doroghazi R, Alpert JS. A medical education as an investment: financial food for There are many reasons why individuals choose to join a thought. Am J Med. 2014;127:7–11. profession and many factors may influence the decision to stay in or leave that profession. However, the purpose of this study was to 8. Kahn MJ, Nelling EF. Estimating the value of medical education: a net present value focus on the financial incentives of physiotherapy and to inform approach. Teach Learn Med. 2010;22:205–208. decision-making. From a purely economic perspective, an investment in medical education is superior to physiotherapy 9. Australian Health Practitioner Regulation Agency. Approved Programs of Study – when working for longer than 10 years. Nonetheless, it is Physiotherapy, 2013. http://www.ahpra.gov.au/Education/Approved-Programs-of- acknowledged that there are many non-financial factors such Study.aspx?ref=Physiotherapist [accessed 1/11/2014]. as job security, intellectual stimulation, personal preference, and contribution to improving the human condition. Therefore, an 10. Australian Physiotherapy Association. APA National Groups. www.physiotherapy. interest in studying physiotherapy does not necessarily equate to asn.au/APAWCM/The_APA/National_Groups/APAWCM/The_APA/ an interest in similar health-related fields. Studies have been National_Groups/National_Groups. conducted on the motivating factors for physiotherapy in other aspx?hkey=a0e8ec8b-8b21-4432-9ea8-14ecc722da35 [accessed 1/12/2014]. countries, as well as for other professions,11,25,26 but none for physiotherapy in Australia, which presents options for future 11. O¨ hman A, Stenlund H, Dahlgren L. Career choice, professional preferences and research. gender? The case of Swedish physiotherapy students. Adv Physiother. 2001;3:94–107. The NPV approach has provided a new perspective on the return on investment for undertaking physiotherapy education. An 12. Monash University. Bachelor of Physiotherapy (Honours) for 2015. http://www. extension of this approach could be to contrast the NPV with career monash.edu.au/study/coursefinder/course/M3002/ [accessed 1/2/2015]. satisfaction data, to give unique insights into the financial and non- financial incentives to career choices within the profession. Such 13. Swedish Higher Education Authority. Funding. http://english.uka.se/ highereducation/funding.4.4149f55713bbd917563800011054. html#h-StudentfinanceforstudentsstudyinginSweden [accessed 1/2/2015]. 14. Dow C. Reform of the higher education demand driven system (revised). Budget review 2014-2015: Parliament of Australia. 15. McMeeken J. Physiotherapy education – what are the costs? Aust J Physiother. 2008;54:85–86. 16. Office of Best Practice Regulation.Guidance Note - Cost Benefit Analysis. Canberra, Australia: Office of Best Practice Regulation; 2014. 17. National Centre for Social and Economic Modelling.NATSEM HELP repayment Modelling Methodology. Canberra, Australia: National Centre for Social and Eco- nomic Modelling; 2014. 18. Australian Bureau of Statistics. Census of Population and Housing, cat. no. 2940.0. Canberra, Australia: ABS; 2011. 19. Australian Taxation Office. Individual income tax rates. https://www.ato.gov.au/ individuals/income-and-deductions/how-much-income-tax-you-pay/ individual-income-tax-rates/# [accessed 12/11/2014]. 20. Australian Taxation Office. Super. https://www.ato.gov.au/Individuals/Super/ Compulsory-employer-contributions/ [accessed 12/11/2014].
154 Rivers et al: Economic value of investment in physiotherapy education 21. Service and Workforce Planning. Physiotherapy Labour Force - Victoria 2006. Mel- 24. Struber JC. Physiotherapy in Australia - Where to now? Internet J Allied Health Sci bourne, Australia: Victorian Department of Human Services; 2007 Pract. 2003;1:2. 22. Australian Bureau of Statistics. Household Expenditure Survey, Australia: Sum- 25. Eley R, Eley D, Rogers-Clark C. Reasons for entering and leaving nursing: an mary of Results 2009-10: cat. no. 6530.0. Canberra, Australia: ABS; 2011. Australian regional study. Aust J Adv Nurs. 2010;28:6. 23. Health Workforce Australia. National Health Workforce Dataset - Australian 26. McManus I, Livingston G, Katona C. The attractions of medicine: the generic Health Practitioner Regulation Agency (AHPRA). Adelaide, Australia: Health motivations of medical school applicants in relation to demography, personality Workforce Australia; 2013. and achievement. BMC Med Ed. 2006;6:11.
Journal of Physiotherapy 61 (2015) 162 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Clinimetrics The Four Square Step Test Description The Four Square Step Test (FSST) was developed in 2002 to complete the sequence successfully, loses balance, or makes measure the rapid stepping that is often required when changing direction and avoiding obstacles while walking.1 Originally contact with a SPS during the sequence. Subjects are permitted to developed for older adults, the FSST has been widely used in neurological populations3–5 and also in musculoskeletal condi- turn and step forward into each square if they are unable to face tions such as joint replacements, arthritis, osteoporosis, orthotic design for lower extremity trauma and amputees. forward during the entire sequence. The FSST requires a stopwatch and four single point sticks (SPS). The FSST description, scoring instructions, normal age and Using the sticks resting flat on the floor to form a cross, the subject starts in one square and steps in one direction into each of the four population values are freely available on the Internet (eg, http:// squares and then reverses direction back to the start. www.rehabmeasures.org/default.aspx). There are also useful This sequence requires the subject to step forward, backward and sideways to the right and left. The FSST measures the time (in demonstration videos available on YouTube as either FSST or seconds) taken to complete the sequence. The following instruc- tions are given to the subject: ‘Try to complete the sequence as fast 4SST (eg, https://www.youtube.com/watch?v=doaPhh3KJHM). as possible without touching the sticks. Both feet must make contact with the floor in each square. If possible, face forward Reliability, validity and normal values: The FSST has excellent during the entire sequence.’ The sequence is then demonstrated to inter-rater reliability (ICC 0.99)1,5 and retest reliability (0.83 the subject. One the practice trial is completed, to ensure that the to 0.99)1–3 in samples of healthy and impaired older adults, subject knows the sequence, two FSST trials are then completed, with the best time recorded. A trial is repeated if the subject fails to vestibular disorders and stroke. Validity is supported with good correlations with other balance and mobility tests (0.64 to 0.88).1–3 The FSST has been found to discriminate between fall groups of older adults,1 vestibular disorders,2 stroke,3 amputees6 and multiple sclerosis.4 Healthy active adults aged younger than 30 years can complete the FSST in under 6 seconds;5 50 to 65 year olds complete the test in 7.49 seconds (SD 2.34); and healthy people aged 65 to 80 are able to perform the FSST in 10 seconds or less.1 Commentary The FSST is quick to administer, easy to score, requires little and gait aids, although the test is too demanding to be performed space and has no cost. It is widely used, with peer-reviewed using a frame. If a person is unable to complete a successful test, research on the FSST coming from 22 different countries. they are unable to be given a time. The FSST is unique in that it challenges motor planning, Provenance: Invited. Not peer-reviewed. sequencing and recall, whilst simultaneously providing clinicians with the opportunity to measure and observe a person’s clearance Zoe Langford of low obstacles at speed. The ability to clear the trail leg when Austin Health, Melbourne, Australia stepping over an obstacle has been shown to be reduced in healthy populations7 and stroke;8 the FSST enables formal assessment of References this. A failure to complete the FSST has been identified as a risk factor for falling.3 In a population of over 65-year-olds, participants 1. Dite W, et al. Arch Phys Med Rehabil. 2002;83:1566–1571. with > 15 seconds were considered multiple fallers (positive 2. Whitney SL, et al. Arch Phys Med Rehabil. 2007;88:99–104. predictive value 86%, negative predictive value 94%).1 3. Blennerhassett JM, et al. Phys Med Rehabil. 2008;89:2156–2161. 4. Nilsaga˚ rd Y, et al. Clin Rehabil. 2009;23:259–269. Limitations: The FSST combines a physically and cognitively 5. Wilken JM, et al. J Am Acad Orthopaedic Surgeons. 2012;20:S42–S47. demanding task and requires a higher level of skilled physical 6. Dite W, et al. Arch Phys Med Rehabil. 2007;88:109–114. supervision. Participants are tested wearing their usual footwear 7. Heijnen M, et al. Exp Brain Res. 2014;232:2131–2142. 8. Said CM, et al. Phys Ther. 2005;85:413–427. http://dx.doi.org/10.1016/j.jphys.2015.03.005 1836-9553/ß 2015 Australian Physiotherapy Association. Published by Elsevier B.V. All rights reserved.
Journal of Physiotherapy 61 (2015) 163 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Clinimetrics The Senior Fitness Test Summary Rikli and Jones developed the Senior Fitness Test for adults Body Mass Index (BMI) is also recorded as weight/D$IF](T_1[4 height)FD]T_1[I$5 2. aged over 60 years.1–4 It is primarily used to evaluate physical Performance standards for the Senior Fitness Test are based on function in healthy elderly people but is also used for people the results of more than 7000 older adults from the USA. These data with dementia]_8[TD$.FI _[1F]D5$TI It comprises six functional tests of strength, provide excellent normative standards for senior age groups of endurance, balance, agility and flexibility. Each test is scored American men and women ranging from 60 to 94 years of age.1–4 separately on different scales. The scores are not tallied into an Similar normative values have also been reported for a Norwegian overall score. sample where the items were adapted to the European metric and weight system.6,7 The Senior Fitness Test can be performed in people’s homes or The six items comprising the Senior Fitness Test are reliable in clinics and does not require costly tools or technical expertise. with an ICC ranging from 0.8 to 0.98.1–5 Validity of the different The six functional tests include: items of the Senior Fitness Test has been evaluated, where possible, by comparing it with available gold standards. For example, the 1. The Chair Stand Test. This requires people to repeatedly stand up Chair Stand Test has been compared to the one repetition from and sit down on a chair for 30 seconds. The number of maximum leg press (r = 0.78 for men and r = 0.71 for women),1 stands is recorded. This reflects lower body strength. and the Biceps Curl Test has been compared to the combined one repetition maximum for the upper trunk (r = 0.84 for men and 2. The Biceps Curl Test. This requires people to repeatedly lift a 5 lb r = 0.79 for women). Similarly, the 6-minute Walk Test has been (2.27 kg) weight (for women) or an 8 lb (3.63 kg) weight (for compared to time on a treadmill Da][F1_T6$I t 85% of maximal heart rate men) for 30 seconds. The number of lifts is recorded. This (r = 0.82 for men and r = 0.71 for women) and the 2-minute Step reflects upper body strength. Test has been compared to the 1-mile walk time (r = 0.73) and time on a treadmill 1[_TD$6F]aI t 85% of maximal heart rate (r = 0.74). The Chair Sit 3. The 6-minute Walk Test. This is measured in distance (m) and and Reach Test has been compared to the hamstrings flexibility reflects aerobic endurance. The original version of the Senior test, as endorsed by the American Academy of Orthopaedic Fitness Test required people to walk [o]FD12_IT$ n a rectangular ]$cF3DT1[I_ ourse but Surgeons (r = 0.76 for men and r = 0.81 for women). The Back more recent versions use a straight line. If a 6MWT is not feasible Scratch Test and the 2.45-m Up-and-Go test have not been then it is acceptable to replace this test with the 2-minute Step formally validated because there are no gold standards. Instead, test. The number of full steps completed in 2 minutes is they are based on consensus opinion of best overall measures of recorded. shoulder flexibility and combined physiological attributes such as power, speed, agility and balance, respectively.1–4,8–12 4. The Chair Sit and Reach Test. This is measured in distance (cm) and reflects lower body flexibility. 5. The Back Scratch Test. This is measured in distance (cm) and reflects upper body flexibility. 6. The 2.45-m Up-and-Go test. This is measured in time (seconds) and reflects agility and dynamic balance. Commentary The Senior Fitness Test is a practical and suitable set of tests for Birgitta Langhammera and IJ]FD$T_1[ ohan K Stanghelleb clinical use and is appropriate for healthy elderly people and those F2T[]IaD$_ Oslo and Akershus University College / Sunnaas Rehabilitation with dementia. The Senior Fitness Test is simple to use and comes Hospital, Faculty of Health, Physiotherapy ProgrammeF],ID$_3[T [TF_ID]4$ Norway with a thorough instruction manual and a video. In addition, most bSunnaas Rehabilitation Hospital, Faculty of Medicine, University of clinicians would be familiar with all of the items of the Senior Fitness Test. The whole test takes approximately 30 to 40 minutes Oslo$D6,T[]FI_ Norway to perform, which may be a limitation for some time-limited clinicians. It is now available in Polish and Danish as well as the References original English version.4,6–8 1. Jones C, et al. JAPA. 2002;101–110. The Senior Fitness Test has been used in several clinical trials.9– 2. Rikli RE, et al. JAPA. 1999;7:129–161. 12 The normative United States values are widely used and 3. Rikli RE, et al. JAPA. 1999;7:162–181. compare well with the normative data obtained from a Norwegian 4. Jones CJ, et al. J Active Aging. 2002;1:24–30. sample of healthy elderly people.3–5,7 This suggests that the United 5. Hesseberg K, et al. Physiother Res Int. _F]D2T[I$9 015;20:37–44. States normative data may be appropriate for most similar 6. Rikli RE et al. translated by Hansen T. Senior Fitness Test. Physical fitness of older countries. The Senior Fitness Test is recommended for measuring physical fitness in older people with and without cognitive adults – manual and reference values (Fysisk forma˚ en hos eldre-manual og referance impairment and is appropriate for both research and clinical værdie). København: FADL’s Forlag; 2004 purposes.1$DTI[]F_7 7. Langhammer B, et al. Adv in Phys. 2011;13:137–144. 8. Ro´ z˙ aoska-Kirschke A, et al. Med Rehab. 0_12[FTI]$D 006;10:9–16. Provenance: Invited. Not peer-reviewed. 9. Dobek JC, et al. JAPA. [2]FI9D$T_ 007;15:13–25. 10. Wilkin LD, et al. Int J Aging Hum Dev. 2$FD[1]T_0I 010;70:107–118. 11. Purath J, et al. J Am Acad Nurse Practs. 01[]_TD$2FI 009;21:101–107. 12. Taraldsen K, et al. Fysioterapeuten. 2010;1. Viewed 4 March 2015 from http://www. fysioterapeuten.no/xp/pub/mx/filer/0110_Fagartikkel_475376.pdf. http://dx.doi.org/10.1016/j.jphys.2015.04.001 1836-9553/ß 2015 Australian Physiotherapy Association. Published by Elsevier B.V. All rights reserved.
Journal of Physiotherapy 61 (2015) 156 Journal of PHYSIOTHERAPY journal homepage: www.elsevier.com/locate/jphys Appraisal Trial Protocol Understanding how pain education causes changes in pain and disability: protocol for a causal mediation analysis of the PREVENT trial Hopin Lee 1,2, G Lorimer Moseley 1,3, Markus Hu¨ bscher 1,2, Steven J Kamper 4,5, Adrian C Traeger 1,2, Ian W Skinner 1,2, James H McAuley 1,2 1 Neuroscience Research Australia (NeuRA); 2 School of Medical Sciences, University of New South Wales; 3 Sansom Institute for Health Research, University of South Australia, Adelaide; 4 EMGO+ Institute, VU University Medical Centre, Amsterdam, Netherlands; 5 The George Institute for Global Health, University of Sydney, Sydney, Australia Abstract outcomes framework will be used to test single and multiple mediator models. A sensitivity analysis will be conducted to Introduction: Pain education is a complex intervention developed evaluate the robustness of the estimated mediation effects on the to help clinicians manage low back pain. Although complex influence of violating sequential ignorability – a critical assump- interventions are usually evaluated by their effects on outcomes, tion for causal inference. Discussion and significance: Mediation such as pain or disability, most do not directly target these analysis of clinical trials can estimate how much the total effect of outcomes; instead, they target intermediate factors that are the treatment on the outcome is carried through an indirect path. presumed to be associated with the outcomes. The mechanisms Using mediation analysis to understand these mechanisms can underlying treatment effects, or the effect of an intervention on an generate evidence that can be used to tailor treatments and intermediate factor and its subsequent effect on outcome, are optimise treatment effects. In this study, the causal mediation rarely investigated in clinical trials. This leaves a gap in the effects of a pain education intervention for acute non-specific low evidence for understanding how treatments exert their effects on back pain will be estimated. This knowledge is critical for further outcomes. Mediation analysis provides a method for identifying development and refinement of interventions for conditions such and quantifying the mechanisms that underlie interventions. Aim: as low back pain. To determine whether the effect of pain education on pain and disability is mediated by changes in self-efficacy, catastrophisation Trial registration: Australian New Zealand Clinical Trials and back pain beliefs. Design: Causal mediation analysis of the Registry (ACTRN). Registration number: 12612001180808. Was PREVENT randomised controlled trial. Participants and setting: this trial prospectively registered: Yes. Date of trial registration: Two hundred and two participants with acute low back pain from 6 November 2012. Funded by: Australian National Health & Medical primary care clinics in the Sydney metropolitan area. Interven- Research Council (NHMRC). Funder approval number: NHMRC tion: Participants will be randomised to receive either ‘pain grant number APP1047827. Anticipated completion date: Decem- education’ (intervention group) or ‘sham education’ (control ber 2016. Provenance: Not invited. Peer reviewed. Correspon- group). Measurements: All outcome measures (including patient dence: James McAuley, Neuroscience Research Australia (NeuRA) characteristics), primary outcome measures (pain and disability), and University of New South Wales, Barker Street, Randwick, NSW and putative mediating variables (self-efficacy, catastrophisation 2031, Australia. Email: [email protected]. and back pain beliefs) will be measured prior to randomisation. Putative mediators and primary outcome measures will be Full protocol: Available on the eAddenda at doi:10.1016/j.jphys. measured 1 week after the intervention, and primary outcome 2015.03.004 measures will be measured 3 months after the onset of low back pain. Analysis: Causal mediation analysis under the potential http://dx.doi.org/10.1016/j.jphys.2015.03.004 1836-9553/ß 2015 Australian Physiotherapy Association. Published by Elsevier B.V. All rights reserved.
Search
