Pilot study on comparisons between the effectiveness of mobile video-guided and paper-based home exercise programs on improving exercise adherence, self-efficacy for exercise and functional outcomes of patients with stroke with 3-month follow-up: A single-blind randomized controlled trial

Objective: To compare the effectiveness of mobile video-guided home exercise program and standard paper-based home exercise program. Methods: Eligible participants were randomly assigned to either experimental group with mobile video-guided home exercise program or control group with home exercise program in a standard pamphlet for three months. The primary outcome was exercise adherence. The secondary outcomes were self-efficacy for exercise by Self-Efficacy for Exercise (SEE) Scale; and functional outcomes including mobility level by Modified Functional Ambulatory Category (MFAC) and basic activities of daily living (ADL) by Modified Barthel Index (MBI). All outcomes were captured by phone interviews at 1 day, 1 month and 3 months after the participants were discharged from the hospitals. Results: A total of 56 participants were allocated to the experimental group (n=27) and control group (n=29). There were a significant between-group differences in 3-months exercise adherence (experimental group: 75.6%; control group: 55.2%); significant between-group differences in 1-month SEE (experimental group: 58.4; control group: 43.3) and 3-month SEE (experimental group: 62.2; control group: 45.6). For functional outcomes, there were significant between-group differences in 3-month MFAC gain (experimental group: 1.7; control group: 1.0). There were no between-group differences in MBI gain. Conclusion: The use of mobile video-guided home exercise program was superior to standard paper-based home exercise program in exercise adherence, SEE and mobility gain but not basic ADL gain for patients recovering from stroke.


Introduction
Stroke, also known as cerebrovascular accident, is an acute disturbance of focal or global cerebral function with signs and symptoms lasting more than 24 h or leading to death presumably of vascular origin. 1 It is the third leading cause of death in Hong Kong after cancer and heart disease. More than 3000 people died in Hong Kong each year for this condition. 2 The most widely recognized impairment caused by stroke is motor impairment, which restricts function in muscle movement or mobility. 3,4 It was shown that patients with stroke who had as much practice as possible within 6 months after stroke onset could achieve the best functional outcome. [5][6][7] In Hong Kong, under the service model of Hospital Authority, patients would be referred to ambulatory services, such as geriatric day hospital, or domestic physiotherapy services to continue their stroke rehabilitation after discharge from hospital. However, some patients could not attend ambulatory services due to various di±culties, such as transportation and absence of carers. The low frequency of domestic physiotherapy service also reduced the e®ectiveness of rehabilitation of patients. From our local data, around a quarter of home-care patients have not received any further rehabilitation training from ambulatory or domestic physiotherapy services. Among this population, there were approximately 56% of them rated to be Modi¯ed Functional Ambulatory Category (MFAC) 2-5, who may have had higher potential of improvement in stroke outcomes if appropriate rehabilitation training was given. 8 Home-based exercise program would be a good choice for¯lling this post-discharge gap of rehabilitation continuation.
Traditionally, for continuation of exercise training program, physiotherapists would prescribe home exercises programs in paper-based format. 9,10 However, evidence showed that home exercise prescription in paper-based format does not lead to better adherence to a home exercise program compared to having no written and pictorial instructions for patients with stroke less than four months. 10 The recent increasing accessibility of smart technology 11 o®ers an opportunity to advance the mode of delivery of home exercise program by mobile devices such as video-guided exercise on electronic tablets 12 and video-guided exercise on mobile apps. 13 However, the e®ects of mobile video-guided exercise programs were controversial. A study showed that home exercise programs lmed on an electronic tablet, with an automated reminder, were not superior to standard paperbased home exercise programs in terms of adherence, motor function, or satisfaction for patients recovering from stroke. 12 In contrast, another study showed that people with musculoskeletal conditions who adhere better to their home exercise programs in video-based format are provided with an app with remote support compared to paper handouts; although the clinical importance of this added adherence is unclear. 13 Since one of the major outcomes to evaluate the e®ectiveness of the mode of delivery of home exercise program is adherence, 10,12,13 it is worthwhile to study the factor a®ecting adherence of mode of delivery of home exercise program. It was suggested that one of the barriers to adherence of home exercise is low self-e±cacy. 14,15 According to the theory of self-e±cacy, self-e±cacy is de¯ned as a person's con¯dence in their ability to perform a task. 16 Thus, self-e±cacy plays an important role in maintaining the exercises behavior after stroke [17][18][19] and improving self-e±cacy for exercise (SEE) could in°uence long-term exercise behavior as well as the early stages of exercise adoption. [20][21][22][23] Another factor a®ecting adherence of home exercise was the delivery mode of training program. 13 Studies showed that visual information brings more bene¯ts to patients than verbal information alone. [24][25][26][27] Adherence of home exercise program has positive correlation with physical function and physical performance of patients. 15,28 It is worthwhile for physiotherapists to investigate the mode of delivery to enhance exercise adherence and self-e±cacy of post-discharge home exercise for patients with stroke. The objective of the study is to compare the e®ectiveness of videoguided exercise program and standard paper-based home exercise program on adherence of exercise, self-e±cacy and functional outcomes in patients with stroke within 3-month follow-up.

Methods
A randomized, controlled, assessor-blinded clinical trial was conducted between July 2018 and June 2019. Participants were recruited from the inpatient Stroke Rehabilitation Program in the Department of Medicine and Geriatrics of Tai Po Hospital and Shatin Hospital. The two hospitals provided multidisciplinary inpatient stroke rehabilitation to more than 1000 stroke patients yearly. All those stroke patients were diagnosed with acute stroke and were transferred from all the three acute hospitals in New Territories East Cluster. The hypothesis was that participants prescribed with video-guided home exercise will demonstrate higher adherence of exercise, better self-e±cacy and better functional outcomes when compared with the participants in paper-based exercise prescription group.
Participants were eligible for inclusion if their principle diagnosis was stroke, participants or their carers have smart devices such as smart phones or tablets that are able to scan QR code and connect to the Internet as well. All the participants or carers could read Chinese. They were excluded if (i) could not follow gesture and instructions; (ii) MFAC upon discharge was below 2 or above 5, (iii) no smart device for video and (iv) refusal.
This was a single-blinded randomized study. The study procedure°owchart is presented in Appendix B. Before enrolment and randomization, potential participants were screened by their case therapists to ensure they met the inclusion criteria and none of the exclusion criteria. Eligible participants were then randomly assigned to either Intervention Group or Control Group in a 1:1 ratio. Investigators who were responsible for data collection were blinded to the group allocation. A randomization list was developed by personnel who were not involved in this study and who would not have any contact with the study participants. The details of the list were unknown to any of the investigators and study coordinators, and were contained in a set of sealed, sequentially numbered envelopes. Each enrolled participant was allocated to the next sequential number on the list. All research study personnel, including those who collect data and assess outcomes, were blinded to the group assignment. The envelopes and the randomization list were not revealed to any of the study personnel until completion of recruitment and data collection.
Research team adopted per-protocol analysis method and provided two brief sessions to all physiotherapists of the stroke units to inform the process of subject recruitment of potential participants. Case physiotherapists would inform research team for subject recruitment once potential participants have discharge plan, usually 1-2 weeks before discharged. Case physiotherapists provided brief information about the study to potential participants or their carers before their discharge from hospitals. Research team approached the participants and gave detailed information about the study and veri¯es their interest in participating once verbal consent was obtained. The Joint Chinese University of Hong Kong -New Territories East Cluster Clinical Research Ethics Committee (The Joint CUHK-NTEC CREC) approval was obtained prior to the commencement of the study. This trial design was registered prospectively with the ClinicalTrials.gov Protocol Registration and Results System (Clin-icalTrials.gov ID: NCT03509363). Written informed consent was obtained from all participants.
Pre-discharge training sessions lasting for 10-15 min were provided to participants, and their carers if any, of both groups in order to make them familiar with the selected home exercises and the technique of using mobile phone to scan QR for intervention group. Participants allocated to the intervention group were prescribed a set of exercise video with QR code provided in home exercise pamphlets and they had to perform the prescribed exercises under the guidance of the videos. On the other hand, participants in control group were given instructions for their home exercise program in a traditional pamphlet includes photographs and instructions of exercise demonstration. The content of home exercise program in both groups was the same and was based on the recommendations from the National Stroke Foundation Clinical Guidelines, included nine arm control exercises, six leg control exercises, six truck control exercises and four mobility exercises ( Table 1).
The doses for the participants were prescribed by their case physiotherapists according to the needs and abilities of participants. The total number of home exercise varied from 3 to 5 and the treatment frequencies of each exercise varied from daily to 3 times per day so the course length varied from 10 to 30 min daily. Participants' needs and the abilities were de¯ned as amount of assistance and the mobility by participants' pre-discharge MFAC.
All participants were screened by their case physiotherapists according to exercise prescription guidelines of American College of Sports Medicine 29 for any contraindication for exercise. Only patients who were medically¯t for exercise were recruited. Exercise-related adverse events such as chest tightness or pain, dizziness, and/or trip, stumbles, or falls might happen during exercise but which was not as higher risk than the usual prescription of home exercises. Suitability of participating home exercise program will be assessed by physiotherapists based on environmental risk, fall risk, and competence of participants or carers in performing exercise with participants.
All participants' data related to the research were collected in data collection sheets by case therapists of the patients. The data in the data collection sheets were transferred to a database in Excel format that could be accessed only by the research team members. Therefore, we could judge which participants were included in the analysis. Data on patients' demographics, gender, age, site of lesion, side of stroke, type of stroke, education level, experience of using mobile device, present of complication such as neglect and dysphasia, Outcome measures including self-reported exercise adherence, self-e±cacy for exercise (SEE-C), MFAC, and Chinese version of the Modi¯ed Barthel Index (MBI) were assessed on phone follow-up basis by a blinded-assessor at 1 day, 1 month, and 3 months after the participants were discharged from hospitals. For between group comparisons, since participants had di®erence baseline functional status, baseline functional status a®ects MBI and MFAC but not SEE and Adherence, therefore MBI gain and MFAC gain were used to compare the improvement of functional status for fair comparisons. In addition, since we directly compared Adherence and SEE that 3 time points were tested; while we compared MBI gain and MFAC gain that only 2 time points were tested i.e., 1-month gain (1-month minus baseline) and 3-month gain (3-month minus baseline).

Self-reported exercise adherence
Since there has no well-developed measures that capture self-reported adherence to prescribed but unsupervised home-based rehabilitation exercises; and since we interviewed the participants by phone, some details could not be recalled. So we adopt the concept visual analogue scale to ask the adherence by percentage (0-100%) as a total e®ect of adherence of number of session, daily frequent, repetition, set and quality of movement of the whole review period. The exercise adherence was measured by asking participants to report their percentage of exercise completion between 3 period of time i.e., from discharge to 1 day post-discharge, from 1 day post-discharge and 1-month post-discharge and from 1-month post-discharge to 3-month post-discharge during phone follow-up. Log sheets were not used since it was reported that log sheet that needs to be¯lled out regarding the completion of each exercise would serve as a reminder and a motivational track record for the patient and assists patients in improving their adherence. 31 Chinese version of self-e±cacy for exercise (SEE-C) An original English version of the SEE scale was designed to test people's con¯dence to continue exercising in the face of barriers to exercise. 32 This scale has a range of total scores from 0 to 90. A higher score indicates higher SEE. Estimates of the reliability and validity of the nine item SEE scale have been widely tested and shown to be valid for use in various settings, with internal consistency (alpha ¼ 0:93) and validity with e±cacy expectations signi¯cantly related to exercise activity, and factor loadings all greater than 0.50. [33][34][35] The Cronbach's alpha coe±cient of Chinese version of the SEE scale was 0.75 35 and Pearson's correlation revealed a statistically signi¯cant correlation between perceived health and SEE-C score (r ¼ À0:17, p ¼ 0:019). 35

Modi¯ed functional ambulatory category (MFAC)
The MFAC was a 7-point Likert Scale (1-7) that was used to classify a patient's walking capacity. Gait was divided into seven categories, ranging from no ability to walk and requires manual assistance to sit or is unable to sit for 1 minute without back or hand support (MFAC 1) to the ability to walk independently on level and nonlevel surfaces, stairs, and inclines (MFAC 7). 36 The inter-rater reliability of the MFAC (intraclass co-e±cient [ICC]) was 0.982 (0.971-0.989), with a kappa coe±cient of 0.923 and a consistency ratio of 94% for stroke patient 37 and the ICC of the MFAC in patients with hip fractures is 0.96, with a construct validity of r ¼ 0:81 on the Elderly Mobility Scale (EMS). 36 Participants will be asked to describe their current mobility status via phone follow-up.

Chinese version of the modi¯ed barthel index (MBI-C)
MBI was used to assess patients' basic activities of daily living (ADL) in this study. MBI measures the participant's performance on 10 functional items including self-care, continence, and locomotion. The values assigned to each item was based on the amount of physical assistance required to perform the task and added to give a total score ranging from 0 to 100 (0 ¼ fully dependent, 100 ¼ fully independent) with higher score indicating higher levels of physical function. 38 There was no subtotal score because there was no subscale. 38 The internal consistency reliability coe±cient for MBI was 0.90. 38 The Chinese version of MBI has been found to have good validity and reliability for assessing stroke patients. 39 A study showed that the correlations between the performance-based ADL and the interview-based ADL were r greater than 0.97 for the total score and r greater than 0.85 for most of the individual items. 40 Although the randomization design of the study reduced the biases which could compromise the outcomes, we analyzed the confounding variables such as age, gender, type of stroke, side of hemiplegia, present of complication such as neglect and dysphasia, 41 follow-up physiotherapy such as ambulatory day hospital and domiciliary physiotherapy 42 availability of carer, 43 and experience of smart device 44 to indicate any signi¯cant association in the study.
Pair t-test and Chi-Square test were used to analysis the between-groups di®erences of baseline characteristics. Chi-Square test was used to analysis the between-groups of exercise adherence at all three time points i.e., baseline, 1-month, and 3-months. Independent t-test was used to analysis the betweengroups of SEE, MFAC gain, and MBI gain at all three time points i.e., baseline, 1-month, and 3months. Pair t-test was used to analysis the withingroup di®erences between baseline to 1-month and baseline to 3-months of MFAC and MBI. The sig-ni¯cance level was set at 0.05 for between-group analysis. All of these analyses were performed with SPSS version 18.0. Those participants with missing data due to loss of contact in phone interviews in three consecutive working days were omitted.

Results
There was a total of 115 stroke patients who were screened by physiotherapists for this study from August 2018 to March 2019. Of these, 55 participants were excluded, in which, 15 could not follow gesture and instruction; 18 had mobility below MFAC 5 or above MFAC 5; 17 had no smart device for video watching and 5 refused. The remaining 60 participants were recruited; with 28 randomly allocated to the experience group and 32 randomly allocated to the control group (Fig. 1). There were no adverse events recorded as a result of participation in this trial. Of the 60 participants, there were four withdrawals. One hospitalized due to convulsion, one hospitalized due to pneumonia and two were lost of contact.
A total of 56 participants completed the study without missing data. The groups appeared well matched and no between-group di®erence in age, gender, time since stroke, type of stroke, side of hemiplegia, complication, site of lesion, follow-up physiotherapy, type of carer, experience in smart device, baseline SEE and baseline MFAC but slight di®erence in baseline MBI. Mean age of participants was 69.8 years (SD 14.9). The mean time since stroke was 39.5 (SD 15.3) ( Table 2).

Discussion
The main¯nding of the study was that mobile video-guided home exercise program was superior to standard paper-based home exercise programs in terms of exercise adherence, SEE and mobility gain but not basic ADL gain for patients recovering from stroke. This¯nding contrasted to the study of Emmerson et al. which showed that home exercise programs¯lmed on an electronic tablet was not superior to standard paper-based home exercise programs in terms of adherence, motor function, or satisfaction for patients recovering from stroke. 12 There are several possible explanations for thesē nding, the adherence of home exercise may depend on the type of exercise and interest of patients. It was suggested that mobility decline is an essential concern in chronic stroke patients. 42 Emmerson's study 12 only provided upper limb exercises that patients may have less interest to participate than the exercises in this study which consisted for arm, leg, trunk and mobility. We also found that there was a close relationship between exercise adherence and SEE. During the 3 months follow-up period, the SEE slightly increased (from 52.5 to 62.2 in video group) with the exercise adherence (from 74.1% to 75.6%) in experimental group; whereas the SEE was slightly decreased (from 48.1 to 45.6) with exercise adherence (from 64.1% to 55.2%) in control group ( Table 3). The present results echoed with the theory of self-e±cacy, the stronger the individual's self-e±cacy, the more likely it is that people will initiate and persist with a given activity 16,45 that SEE was positive proportion to exercise adherence in both groups. The study also found that video-guided home exercise program could improve exercise adherence and SEE even without any regular encouragement, or automated reminder as in previous studies by Lambert et al. 13 and Emmerson et al. 12 The higher mobility gain of patients in experimental group than control group could be explained by adherence of home exercise program that has positive correlation with physical function and physical performance of patients. 28,29 No di®erence in MBI gain in experimental group when compared to control group means the video-guided exercise program improves mobility or ambulation (as measured by MFAC) but its e®ects could not be translated to improvement of ADL (as measured by MBI). This could be explained by speci¯city principle 46 that in training the home exercise program in this study was motor control and mobility orientated but not ADL orientated.
The limitation of the study included information bias introduced by outcome collection by telephone interview; selection bias to Chinese population since all the videos and pamphlets were in Chinese; and confounding bias by confounding variables such as, by chance, the experimental group had signi¯cant higher baseline MBI than that of control group. However, the e®ects of the confounding bias have been minimized by randomization and statistical analysis. The intervention¯delity monitoring is another limitation of the study; a predischarge training session was used to minimize the limitation. Further research studies with large sample size would allow comparing the most e®ective groups of the video-guided home exercise. In addition, as mentioned in the introduction session, it seems that adherence was a®ected by both self-e±cacy and delivery mode of training program. The relationship between self-e±cacy and delivery mode of training program such as video-guided element is still unclear and needs further study.

Conclusion
The use of mobile video-guided home exercise program was superior to standard paper-based home exercise program in exercise adherence, SEE and mobility gain but not basic ADL gain for patients recovering from stroke.

Con°ict of Interest
The authors declared no potential con°icts of interest with respect to the research, authorship, and/or publication of this paper.