Effectiveness of a theory-driven mHealth intervention in promoting post-surgery rehabilitation adherence in patients who had anterior cruciate ligament reconstruction : A randomized clinical trial

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY-NC-ND 4.0 https://creativecommons.org/licenses/by-nc-nd/4.0/ Effectiveness of a theory-driven mHealth intervention in promoting post-surgery rehabilitation adherence in patients who had anterior cruciate ligament reconstruction : A randomized clinical trial © 2023 Elsevier Accepted version (Final draft) Lee, Alfred S. Y.; Shu-Hang Yung, Patrick; Tim-Yun Ong, Michael; Lonsdale, Chris; Wong, Thomson W. L.; Siu, Parco M.; Hagger, Martin S.; Chan, Derwin K. C. Lee, A. S. Y., Shu-Hang Yung, P., Tim-Yun Ong, M., Lonsdale, C., Wong, T. W. L., Siu, P. M., Hagger, M. S., & Chan, D. K. C. (2023). Effectiveness of a theory-driven mHealth intervention in promoting post-surgery rehabilitation adherence in patients who had anterior cruciate ligament reconstruction : A randomized clinical trial. Social Science and Medicine, 335, Article 116219. https://doi.org/10.1016/j.socscimed.2023.116219 2023 1 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Effectiveness of a Theory-Driven mHealth Intervention in Promoting Post-Surgery Rehabilitation Adherence in Patients Who Had Anterior Cruciate Ligament Reconstruction: A Randomized Clinical Trial Alfred S. Y. Lee1, 2 Patrick Shu-Hang Yung3 Michael Tim-Yun Ong3 Chris Lonsdale4 Thomson W. L. Wong5 Parco M. Siu2 Martin S. Hagger6, 7 Derwin K. C. Chan 1, 2 1Centre for Child and Family Science, The Education University of Hong Kong, Hong Kong, China 2Division of Kinesiology, School of Public Health, The University of Hong Kong, Hong Kong, China 3Department of Orthopaedics and Traumatology, The Chinese University of Hong Kong, Hong Kong, China 4Institute for Positive Psychology and Education, Australian Catholic University, North Sydney, Australia 5Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hong Kong, China 2 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE 6SHARPP Lab, Psychological Sciences, University of California, Merced, USA 7Faculty of Sport and Health Sciences, University of Jyväskylä, Jyväskylä, Finland The project was funded by grants [#16172201] from the Health and Medical Research Fund awarded to the corresponding author. Correspondence concerning this article should be addressed to Derwin K. C. Chan, Centre for Child and Family Science, The Education University of Hong Kong. Room B1-2/F-35, 10 Lo Ping Road, Tai Po, New Territories, Hong Kong. Phone: +852 29487071. Fax: +852 29487160 Email: [email protected]. Article Citation: Lee, A. S. Y., Yung, P. S. H., Ong, M. T. Y. L., C., Wong, T. W. L., Siu, P. M., Hagger, M. S., & Chan, D. K. C. (2023). Effectiveness of a theory-driven mHealth intervention in promoting post-surgery rehabilitation adherence in patients who had anterior cruciate ligament reconstruction: A randomized clinical trial. Social Science & Medicine, 335, 116219. https://doi.org/10.1016/j.socscimed.2023.116219 3 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Highlights • This study examined the effects of a mHealth intervention on the recovery of patients with an ACL rupture. • The smartphone application prevented further decline of orthopedic outpatients’ motivation and adherence to treatment. • The smartphone application fell short in promoting recovery. 4 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Abstract Rationale: Patients with anterior cruciate ligament (ACL) reconstruction often have poor adherence to post-surgery rehabilitation. Objective: This study applied the integrated model of self-determination theory and the theory of planned behavior to examine the effects of a smartphone-delivered intervention on the recovery outcomes of patients with an ACL rupture during post-surgery rehabilitation period. Additionally, we explored the effects of the intervention on participants with different beliefs toward rehabilitation at baseline. Method: The randomized control trial recruited 96 eligible participants (Mage = 27.82, SD = 8.73; female = 39%) who underwent ACL reconstruction surgery. Participants were randomly assigned to an intervention group (n = 41), which received standard post-surgical treatment (usual-care) and smartphone application (“ACL-Well”), or a usual-care control group (n = 55). The primary outcomes were recovery outcomes from ACL surgery measured by knee muscle strength and laxity, and subjective knee evaluation completed 4-month post-intervention. Secondary outcomes were the psychological and behavioral outcomes measured at baseline, at 2and 4-month post-intervention. Results: ANCOVA indicated no significant between-group differences in primary outcomes: knee muscle strength, knee laxity and subjective knee evaluation, F(1, 27 to 55) = 0.01 to 1.36, p = .25 to .99, η2 = .01 to .03. For the secondary outcomes, growth mixture modelling revealed self-determined treatment motivation declined significantly over the intervention period in the control group (M slope = -.39 to -.12, p =.01 to .04), but not in the intervention group (M slope = -.19 to -.08, p = .06 to .38). Conclusions: The smartphone application fell short in promoting orthopedic outpatients’ 5 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE recovery outcomes. Yet, it shows some promises as a mean to maintain patients’ motivation and adherence to treatment. Keywords: mHealth; Integrated model; Motivation; Social cognition beliefs; Treatment adherence, ACL; Self-determination theory; Theory of planned behavior 6 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE After the reconstruction surgeries of anterior cruciate ligaments (ACL), patients are typically prescribed extensive self-administered home-based rehabilitation (e.g., strength and flexibility training) for six to twelve months. Adherence to rehabilitation during this long treatment period is associated with better recovery outcomes, but is often problematic as rates of nonadherence can reach 70% (Essery et al., 2017). The development of efficacious interventions that promote patients’ rehabilitation adherence is therefore warranted and likely to have a substantive impact on recovery rates. Hence, this study aims to examine the effectiveness of a theory-driven mHealth intervention in promoting post-surgery rehabilitation for patients who ruptured and reconstructed their anterior cruciate ligaments (ACL). Our proposed intervention comprised a smartphone application, “ACL-Well”, and incorporated content targeting change in constructs from an integrated theoretical model (Hagger & Chatzisarantis, 2009). The model is based on two prominent theories of motivation from psychology and behavioral science: self-determination theory (Deci & Ryan, 1985) and the theory of planned behavior (Ajzen, 1991). A key prediction of the model is that long-term behavioral adherence and adaptive health/ recovery outcomes occur more likely when individuals possess high self-determined motivation (i.e., a motivational pattern characterized by a pattern of high autonomous motivation and low controlled motivation), positive attitudes (instrumental and affective evaluations of the behavior), positive subjective norms (perceived social appropriateness of the behavior), positive perceived behavioral control (PBC; perceived capacity to perform the behavior), and high intention of performing the behavior (Hagger & Chatzisarantis, 2009; Teixeira et al., 2020). The integrated model provides a comprehensive explanation on the determinants of rehabilitation adherence and the process involved (Hagger & Chatzisarantis, 2009), and the psychological pathways of the model have been supported by empirical evidence in the context of 7 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE rehabilitation from ACL surgery (Chan et al., 2017; Lee et al., 2020). Intervention studies have independently applied the behavior change strategies of SDT (e.g., the provision of clear rationale and meanings of following the treatment) and TPB (e.g., promoting the benefits of engaging in the behaviors) to facilitate adaptive behavioral patterns in various health contexts, including weight management (LaRose et al., 2022), physical activities (Ha et al., 2018), and HIV/AIDS prevention (Siuki et al., 2019). As far as we know, there have not been any theory-driven interventions that applied either the concepts of SDT, TPB, or an integration of both theories to enhance patients’ adherence to post-surgery rehabilitation. Using the integrated model to develop a mHealth intervention to facilitate patients’ adherence to post-surgery rehabilitation would offer novel insights valuable to research and practice. In this study, we aim to apply this integrated model to develop a mobile phone app, “ACLWell”, to promote better post-surgery recovery outcomes among ACL surgery patients. Using a 4month randomized controlled design, we examined the efficacy of the app on patients’ postsurgery recovery outcomes, rehabilitation adherence, and changes in the psychological constructs from the integrated model. We also explored the effects of the app on participants with different beliefs toward rehabilitation at baseline. It is hypothesized that: (H1) The intervention group would have better recovery outcomes (i.e., knee muscle strength, knee laxity and subjective knee evaluation) compared to the control group at follow-up. (H2) The control group who received ‘usual care’ post-intervention and did not receive the “ACL-Well” app would have significant declines in the behavioral and psychological 8 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE outcomes (i.e., rehabilitation adherence and psychological factors of the integrated model). (H3) The intervention group who received the app would have no significant declines in the behavioral and psychological outcomes. Method Participant The study is a registered clinical trial (HKUCTR-2761) which received ethical approval from the Institutional Review Board of the first author’s institution [Blinded for review]. Patients who had ACL reconstruction (N = 124) were recruited to the study from the orthopedic clinic of a major public hospital in Hong Kong. Patients were recruited during their first post-operation hospital consultation after discharge. Patients were eligible for inclusion in the study if they: (1) were adults aged between 18 and 60 years, (2) had received ACL reconstruction surgery in the previous 2 weeks, and (3) were regular smartphone users. The clinic specialists referred eligible patients to the lead researcher, who provided them with a written information sheet regarding the study and the opportunity to participate. Patients were recruited between August 15, 2017, and August 31, 2018, with follow-up data collection completed on January 4, 2019. A statistical power analysis specifies a power level of 80%, an alpha level of .05, and a medium-to-large effect size of .32 (Sonnery-Cottet et al., 2019), indicating a minimum sample size of 78 patients was needed to find effects in ANOVA. Assuming an attrition rate of 20%, we determined at least 94 participants needed to be recruited for the study. Finally, 96 eligible participants (Mage = 27.82, SD = 8.73, range = 18 to 53; female = 39%) provided informed consent. All the participants were Chinese from Hong Kong. On average, participants ruptured their ACL 8.91 (SD = 15.75) months 15 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE type of clinical outcomes. Systematic reviews and meta-analyses have shown that the impact of rehabilitation adherence on clinical outcomes is relatively inconsistent and indeterminate across studies. For example, recovery outcomes of patients may vary across different types of rehabilitation exercise (Zaffagnini et al., 2015), so the types of exercises that patients adopt during the rehabilitation program might have introduced additional error variance in the effects of the intervention on clinical outcomes. This is plausible because orthopedic surgeons and physiotherapists tailor the type and intensity of the rehabilitation exercises they prescribe to patients, which may have affected adherence independent of the intervention (Zaffagnini et al., 2015). On a different note, the lack of significance of the intervention effects on the clinical outcomes might be due to a relatively short follow-up (i.e., 4 months post-surgery) in our study. During the first four months post-surgery, the intensity of the rehabilitation exercises (including those in the “ACL-Well”) for patients with ACL reconstruction surgery were generally relatively mild (Roi et al., 2006; Shaw, 2002), and so the variance of the clinical outcomes might be less dependent on rehabilitation adherence and other psychological factors of ACL-patients’ rehabilitation. Therefore, a longer follow-up may be able to detect the growth trajectories of the recovery progress when patients are ready to pick up more intensive rehabilitation exercises, and their recovery would be more responsive to the effort and frequency of their rehabilitation. Intervention Effects on Secondary Outcomes According to the exploration analyses on secondary outcomes, the intervention was effective in maintaining self-reported rehabilitation adherence, but only among patients with lower rehabilitation adherence at baseline. The corresponding intervention effect on patients with higher initial rehabilitation adherence was not statistically significant. The effect of the intervention on orthopedic patients’ rehabilitation adherence seems to depend on patients’ initial rehabilitation 16 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE adherence after ACL surgery. These findings are consistent with previous intervention research demonstrating that prior adherence is an important consideration when evaluating the effects of behavior change interventions (Evers et al., 2012). This implies that the intervention is beneficial to prevent further decline of rehabilitation adherence among patients who experience difficulties performing their rehabilitation. Few interventions based on behavioral theory have been applied to promote behavior change in ACL-patients’ recovery and adherence, and even fewer have utilized mHealth techniques to deliver the intervention to patients with ACL rupture. Consequently, the current research adds value to the literature by demonstrating the development of a theory-based intervention using a smartphone app and its efficacy in preventing the decline of adherence and motivation to rehabilitation. The mHealth intervention maintained patients’ self-determined treatment motivation, characterized by high autonomous and low controlled motivation. In this study, behavioral strategies for promoting self-determined motivation included providing meaningful rationales for rehabilitation, acknowledging patients’ perspectives and providing support and encouragement. The effectiveness of these behavioral strategies derived from SDT has been evidenced in other health settings (Teixeira et al., 2020), and we have now extended their applications in a clinical setting among orthopaedic patients. The results support using the “ACLWell” app and smartphone delivery in maintaining self-determined treatment motivation among orthopaedic patients during their rehabilitation period. Despite the supportive findings of self-determined motivation and rehabilitation adherence, the intervention effects on patients’ social cognition variables were inconsistent. There were no statistically significant effects of the intervention on post-intervention social cognition constructs, and intention. One possible reason for this finding was that the behavioral strategies we applied 17 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE within this mHealth intervention in relation to some of the TPB variables (e.g., provision of information about others’ approval for subjective norms) were not well suited to this clinical setting. Existing interventions that have applied the concepts of TPB were shown to be effective but these interventions were primarily conducted in non-clinical settings and preventive settings (Tyson et al., 2014), which may be different from the setting and sample of our study. Study Limitations and Future Directions This research has several conceptual and methodological shortcomings. First, self-report measures were the current study’s primary assessment type. Participants’ responses to these measures may be influenced by self-serving bias, consistency tendency, and other methodological artifacts (Chan et al., 2020). Future studies should consider using non-self-report measures, such as patients’ attendance to physiotherapy clinics, and implicit association tests to measure patients’ psychological and behavioral patterns of their rehabilitation (Chan et al., 2018). Second, we did not record any data on the number of times and the total duration participants used “ACL-Well” during the intervention period. This information could provide essential information about how the usage and rehabilitation adherence are related to the effectiveness of the mHealth intervention (Vriend et al., 2015). Future research should include this quantitative information, together with qualitative data about the extent the patients study the health information and how they respond to daily supportive pop-up messages. Third, the current study did not assess some potential confounding factors (e.g., the grade of the ACL tears, the socio-economic status of the participants, and leg dominance). Future research might consider taking these factors into account when testing the mHealth intervention comprehensively. Fourth, the small sample size might account for nonsignificant intervention effects on the outcomes, and future research should recruit larger sample sizes to increase the likelihood of detecting significant intervention effects. In the original proposal, 18 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE the sample size was first calculated based on the ANOVA. However, ANCOVA appeared to be more appropriate for this study in the later stages because the analysis would account for the essential covariates (i.e., age, sex, months for post-ACL-rupture and meniscus injury). This may have lowered the power of the current study. Finally, the high attrition rate at follow-up was notable in the current study. Although our analysis accounted for the potential confounding effects of missing data, the high attrition rate might have revealed further non-adherence to ACL reconstruction rehabilitation exercises (Chan et al., 2017). More research is warranted to investigate not only patients’ adherence to rehabilitation but also their adherence to follow-up. 19 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Reference Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179–211. Bowerman, S.J., Smith, D.R., Carlson, M., & King, G.A. (2006). A comparison of factors influencing ACL injury in male and female athletes and non-athletes. Physical Therapy in Sport, 7, 144–152. Chan, D.K.C., Keatley, D.A., Tang, T.C., Dimmock, J.A., & Hagger, M.S. (2018). Implicit versus explicit attitude to doping: Which better predicts athletes’ vigilance towards unintentional doping? Journal of Science and Medicine in Sport, 21, 238–244. Chan, D.K.C., Lee, A.S.Y., Hagger, M.S., Mok, K.-M., & Yung, P.S.-H. (2017). Social psychological aspects of ACL injury prevention and rehabilitation: An integrated model for behavioral adherence. Asia-Pacific Journal of Sports Medicine, Arthroscopy, Rehabilitation and Technology, 10, 17–20. Chan, D.K.C., Lonsdale, C., Ho, P.Y., Yung, P.S., & Chan, K.M. (2009). Patient motivation and adherence to postsurgery rehabilitation exercise recommendations: the influence of physiotherapists' autonomy-supportive behaviors. Archives of Physical Medicine and Rehabilitation, 90, 1977–1982. Chan, D.K.C., Stenling, A., Yusainy, C., Hikmiah, Z., Ivarsson, A., Hagger, M.S., et al. (2020). Editor's choice: Consistency tendency and the theory of planned behavior: A randomized controlled crossover trial in a physical activity context. Psychology & Health, 35, 665– 684. Collins, N.J., Misra, D., Felson, D.T., Crossley, K.M., & Roos, E.M. (2011). Measures of knee function: International Knee Documentation Committee (IKDC) Subjective Knee 20 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Evaluation Form, Knee Injury and Osteoarthritis Outcome Score (KOOS), Knee Injury and Osteoarthritis Outcome Score Physical Function Short Form (KOOS‐PS), Knee Outcome Survey Activities of Daily Living Scale (KOS‐ADL), Lysholm Knee Scoring Scale, Oxford Knee Score (OKS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Activity Rating Scale (ARS), and Tegner Activity Score (TAS). Arthritis Care and Research, 63, S208–S228. Cvjetkovic, D.D., Bijeljac, S., Palija, S., Talic, G., Radulovic, T.N., Kosanovic, M.G., et al. (2015). Isokinetic testing in evaluation rehabilitation outcome after ACL reconstruction. Medical Archives, 69, 21. Deci, E.L., & Ryan, R.M. (1985). The general causality orientations scale: Self-determination in personality. Journal of Research in Personality, 19, 109–134. Essery, R., Geraghty, A.W., Kirby, S., & Yardley, L. (2017). Predictors of adherence to homebased physical therapies: a systematic review. Disability and Rehabilitation, 39, 519– 534. Evers, A., Klusmann, V., Schwarzer, R., & Heuser, I. (2012). Adherence to physical and mental activity interventions: coping plans as a mediator and prior adherence as a moderator. British Journal of Health Psychology, 17, 477–491. Fu, C.L.A., Yung, S.H.P., Law, K.Y.B., Leung, K.H.H., Lui, P.Y.P., Siu, H.K., et al. (2013). The effect of early whole-body vibration therapy on neuromuscular control after anterior cruciate ligament reconstruction: a randomized controlled trial. The American Journal of Sports Medicine, 41, 804–814. 21 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Ha, A.S., Lonsdale, C., Lubans, D.R., & Ng, J.Y. (2018). Increasing students’ physical activity during school physical education: Rationale and protocol for the SELF-FIT cluster randomized controlled trial. BMC Public Health, 18. Hagger, M., Keatley, D., & Chan, D.K.C. (2014). CALO-RE taxonomy of behavior change techniques. In R.C. Eklund, & G. Tenenbaum (Eds.), Encyclopedia of Sport and Exercise Psychology: SAGE Publications, Inc. Hagger, M.S., & Chatzisarantis, N.L. (2009). Integrating the theory of planned behaviour and self‐determination theory in health behaviour: a meta‐analysis. British Journal of Health Psychology, 14, 275–302. LaRose, J.G., Leahey, T.M., Lanoye, A., Bean, M.K., Fava, J.L., Tate, D.F., et al. (2022). Effect of a Lifestyle Intervention on Cardiometabolic Health Among Emerging Adults: A Randomized Clinical Trial. JAMA Network Open, 5. Lee, A.S.Y., Yung, P.S.-H., Mok, K.-M., Hagger, M.S., & Chan, D.K.C. (2020). Psychological processes of ACL-patients' post-surgery rehabilitation: A prospective test of an integrated theoretical model. Social Science and Medicine, 244, 112646. Levesque, C.S., Williams, G.C., Elliot, D., Pickering, M.A., Bodenhamer, B., & Finley, P.J. (2006). Validating the theoretical structure of the Treatment Self-Regulation Questionnaire (TSRQ) across three different health behaviors. Health Education Research, 22, 691–702. Muthén, L.K., & Muthén, B. (2017). Mplus user's guide: Statistical analysis with latent variables, user's guide: Muthén & Muthén. Roi, G., Nanni, G., & Tencone, F. (2006). Time to return to professional soccer matches after ACL reconstruction. Sport Sciences for Health, 1, 142–145. 22 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Shaw, T. (2002). Accelerated rehabilitation following anterior cruciate ligament reconstruction. Physical Therapy in Sport, 3, 19–26. Siuki, H.A., Peyman, N., Vahedian-Shahroodi, M., Gholian-Aval, M., & Tehrani, H. (2019). Health education intervention on HIV/AIDS prevention behaviors among health volunteers in healthcare centers: An applying the theory of planned behavior. Journal of Social Service Research, 45, 582-588. Teixeira, P.J., Marques, M.M., Silva, M.N., Brunet, J., Duda, J., Haerens, L., et al. (2020). A Classification of Motivation and Behavior Change Techniques Used in SelfDetermination Theory-Based Interventions in Health Contexts. Motivation Science, 6, 438–455. Tyson, M., Covey, J., & Rosenthal, H.E. (2014). Theory of planned behavior interventions for reducing heterosexual risk behaviors: A meta-analysis. Health Psychology, 33, 1454. Vriend, I., Coehoorn, I., & Verhagen, E. (2015). Implementation of an app-based neuromuscular training programme to prevent ankle sprains: a process evaluation using the RE-AIM Framework. British Journal of Sports Medicine, 49, 484–488. Zaffagnini, S., Grassi, A., Serra, M., & Marcacci, M. (2015). Return to sport after ACL reconstruction: how, when and why? A narrative review of current evidence. Joints, 3, 25. 23 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Table 1. Clinical Assessments Performance Knee Muscle Strength Intervention (n = 26) Control (n = 30) LSI% M [95%CI] LSI% M [95%CI] ANCOVA F P Ext PT/BW 60°/s 69.28 [59.73, 79.91] 75.14 [67.14, 81.60] 0.85 .36 Flex PT/BW 60°/s 84.51 [75.58, 93.04] 84.18 [77.23, 90.46] 0.01 .95 Ext PT/BW 180°/s 74.61 [63.15, 87.35] 81.76 [73.48, 87.83] 1.36 .25 Flex PT/BW 180°/s 90.51 [80.44, 100.96] 90.81 [83.33, 99.48] 0.01 .96 Intervention (n = 16) Control (n = 23) M [95%CI] mm M [95%CI] mm Knee Laxity 2.77 [1.44, 4.14] 2.72 [1.88, 3.76] 0.01 .96 Subjective Knee Evaluation Intervention (n = 32) Control (n = 39) M scores [95%CI] M scores [95%CI] 70.10 [65.55, 74.69] 70.09 [66.11, 74.01] 0.01 .99 Note. LSI = Limb Symmetry Index. Ext = Extension. Flex = Flexion. PT = Peak Torque. BW = Body Weight. ANCOVAs were adjusted for age, sex, months for post-ACL-rupture and meniscus injury. 24 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Figure 1. CONSORT Flow Diagram Assessed for eligibility (N = 124) Excluded (N = 28)  Not meeting inclusion criteria (N = 28) • Age < 18 (N = 3) • ACL reconstruction surgery > 2 weeks (N = 25) Analysed (N = 41) Completed the 4th month assessments  Questionnaire (N = 32)  Biodex (N = 26)  KT-1000 (N = 16) Discontinued intervention  Lost to follow-up (N = 8)  Re-injured (N = 1)  Unable to get to physiotherapy (N = 6)  Unable to get to clinic (N = 16) Allocated to intervention group (N = 41) Completed the 4th month assessments  Questionnaire (N = 39)  Biodex (N = 30)  KT-1000 (N = 23) Discontinued intervention  Lost to follow-up (N = 14)  Relocation (N = 2)  Unable to get to physiotherapy (N = 9)  Unable to get to clinic (N = 13) Allocated to control group (N = 55) Analysed (N = 55) Allocation Analysis Follow-Up Randomized (N = 96) Enrollment 31 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE data were imputed using the full-information maximum likelihood method (Muthén & Muthén, 2017). Data files, analysis scripts, and outputs for this study are available online. 32 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Appendix VI Preliminary analyses and descriptive statistics of study variables at the baseline No significant difference was found between intervention and control group in terms of the baseline characteristics, t(94) = -1.46 to 1.62, p = .11 to .86. Baseline characteristics of participants are presented in Table S1. Regarding the dropout analyses, we found no significant difference between the participants who complete (n = 72) and those who did not complete (n = 24) the 4th month survey in terms of gender, age, months of post-ACL-rupture, meniscus injury, or the study variables at baseline, t(94) = -0.13 to 1.45, p = .15 to .90. According to the results of Little’s missing completely at random (MCAR) test (i.e., χ2 = 262.73, df = 300, p =.94), our data failed to reject the null hypothesis of MCAR (Little & Rubin, 2019). The results provided some evidence that no clear pattern existed in the missing data. Descriptive statistics, zero-order correlation, reliability estimates, skewness and kurtosis are presented in Table S2. Table S1. Baseline characteristics Variables Intervention Group Control Group Independent t-test t p Sexa Male 25 (61%) 34 (62%) 0.18 .86 Female 16 (39%) 21 (38%) Ageb 27.56 (8.13) 28.00 (9.18) 0.24 .81 Time of ACL Injury (Months Ago)b 7.89 (14.02) 9.65 (16.97) 0.53 .60 Meniscus injuryb Yes No 15 (38%) 22 (55%) 29 (53%) 21 (38%) 1.62 .11 Sporting Experience (Years)b 12.22 (5.68) 11.69 (7.91) -0.35 .73 Rehabilitation Adherenceb 12.15 (1.59) 11.62 (1.80) -1.49 .14 Self-Determined Treatment Motivationb 2.06 (1.20) 1.98 (1.21) -0.32 .75 Social Cognition Beliefs in Rehabilitationb 6.05 (0.73) 6.09 (0.65) 0.38 .70 Intention in Rehabilitationb 6.42 (0.81) 6.32 (0.75) -0.60 .55 Note. aData expressed as N (%); bData expressed as M (SD). 33 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Table S2. Descriptive statistics of secondary outcomes 1 2 3 4 1. Rehabilitation Adherence 1 2. Self-Determined Treatment Motivation .271** 1 3. Social Cognition in Rehabilitation .57*** .15 1 4. Intention in Rehabilitation .62*** .22* .75*** 1 Mean 11.83 2.01 6.04 6.35 SD 1.72 1.20 0.67 0.78 McDonald’s Omega .78 .83 .91 .98 Skewness -0.51 0.71 -0.88 -1.21 Kurtosis -0.45 0.51 0.89 1.11 Note. * p < .05, ** p < .01, *** p < .001 34 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Appendix VII Sensitivity analyses Sensitivity analyses involved repeating the analysis after removing the four covariates (i.e., age, sex, months for post-ACL-rupture and meniscus injury). There were still no significant differences between intervention and control group in primary outcomes (F(1, 37 to 69) = 0.19 to 2.29, p = .21 to .89). In relation to the treatment of missing data, we conducted another set of sensitivity analysis. In particular, five imputed data sets were combined to generate a final imputed dataset. No significant differences between the intervention and control group were found in the primary outcomes (F(1, 37 to 69) = 0.01 to 1.01, p = .32 to .98). Therefore, the main findings of our intervention (i.e., ANCOVAs) was shown to be robust against patients’ background and patterns of missing data. 35 Running Head: MOBILE INTERVENTION FOR REHABILITATION ADHERENCE Appendix VIII Fit indices and estimates of the growth trajectories in each class Classes solution fit indices Growth factors for each class n ABIC ENT Group Class n (%) M Intercept [95% CI] M Slope [95% CI] Hypotheses Adherence 2 1077.45 .82 Intervention low 14 (34%) 10.50 [9.54, 11.46] .04 [-.17, .24] H2 supported; H3 partially supported high 27 (66%) 12.99 [12.41, 13.60] -.50 [-.79, -.21] Control low 22 (40%) 9.90 [9.02, 10.78] -.25 [-.44, -.06] high 33 (60%) 12.68 [12.03, 13.32] -.15 [-.28, -.03] Self-Determined Treatment Motivation H2 supported; H3 partially supported 2 867.82 .88 Intervention low 34 (83%) 3.70 [3.17, 4.22] -.08 [-.14, -.01] high 7 (17%) 5.59 [4.65, 6.53] -.11 [-.31, .09] Control low 49 (89%) 3.68 [3.44, 3.93] -.12 [-.21, -.03] high 6 (11%) 6.15 [5.06, 7.23] -.39 [-.56, -.21] Social Cognition H2 supported; H3 not supported 1 619.27 1.00 Intervention 41 (100%) 6.04 [5.85, 6.23] -.12 [-.17, -.08] Control 55 (100%) 6.07 [5.93, 6.22] -.07 [-.12, -.01] Intention 2 715.79 .97 Intervention low 5 (12%) 4.60 [4.20, 4.99] .26 [.07, .44] H2 and H3 partially supported high 36 (88%) 6.62 [6.48, 6.75] -.19 [-.25, -.13] Control low 8 (15%) 4.98 [4.66, 5.29] .12 [-.03, .26] high 47 (85%) 6.56 [6.43, 6.68] -.16 [-.23, -.09] Note. ABIC= Adjusted-Bayesian Information Criterion. ENT = Entropy. CI = Confidence Interval.