Investigating the effectiveness of an artificial intelligence-enhanced physical therapy and home exercise program on knee pain: a pilot study

Introduction

Knee osteoarthritis (OA) is one of the most prevalent chronic musculoskeletal conditions and a leading cause of pain, disability, and reduced quality of life among adults worldwide. The global prevalence of knee OA was estimated at approximately 365 million cases in 2019, with age-standardized prevalence increasing by 7.5% between 1990 and 2019 (1). A meta-analysis of 73 studies estimated a global knee OA prevalence of 23% in adults older than 40 years, with a lifetime risk of symptomatic knee OA of 45% by age 85 years (2). The disease is characterized by progressive cartilage degeneration, subchondral bone remodeling, synovial inflammation, and altered joint biomechanics, now understood as a whole-joint disorder rather than simply a cartilage-centric “wear-and-tear” disease (2,3). These pathological changes contribute to chronic pain, stiffness, and functional limitation affecting activities of daily living and physical activity participation (4). As population aging and obesity rates increase, the global burden of knee OA continues to rise, with projections suggesting prevalence may increase by 43.8% by 2035, placing substantial strain on healthcare systems and underscoring the need for effective and accessible conservative management strategies (5).

Exercise-based physical therapy (PT) is a cornerstone of nonoperative management for knee OA and is strongly recommended by clinical practice guidelines from the American College of Rheumatology, Osteoarthritis Research Society International, and other major professional societies (6,7). Systematic reviews of randomized trials demonstrate that exercise significantly reduces pain, improves physical function and quality of life, with benefits typically occurring at least 2 to 6 months after formal treatment and is overall cost effective compared to arthroplasty surgical intervention (8). Individually supervised exercises provide greater pain reduction than group-based exercises (2). However, despite its established clinical value, access to conventional in-person PT remains limited for many patients. Common barriers include transportation challenges, time constraints, financial burden, workforce shortages, and prolonged wait times for outpatient appointments (9-12). These barriers may delay initiation of therapy, limit adherence, or result in incomplete treatment courses, thereby reducing the real-world effectiveness of PT for knee OA. Even for those who do participate in PT, traditional programs lead to difficulty with sustained adherence to home exercise programs (HEPs) following supervised PT, a key component to maintain therapeutic benefit (9,10).

In response to these challenges, digital health and virtual PT models have emerged as potential strategies to improve access to musculoskeletal care (13). The coronavirus disease of 2019 (COVID-19) pandemic further accelerated the adoption of tele-rehabilitation by disrupting routine in-person services and necessitating remote care delivery (14,15). Prior studies suggest that telehealth-based PT can be feasible and acceptable to patients while also improving pain and function in selected musculoskeletal populations (16,17). Recent evidence suggests that technology-enabled remote delivery of exercise, diet, and education interventions may be cost-effective (7). However, existing digital PT solutions vary substantially in their technological approach and clinical workflow. Many rely primarily on synchronous video visits, while others incorporate wearable sensors, high-speed cameras, or virtual and augmented reality systems. Although these technologies may enhance supervision or engagement, they often increase cost, technical complexity, and equipment requirements, potentially limiting scalability and long-term adoption (17,18).

Recent advances in artificial intelligence (AI), particularly computer vision-based technologies, offer an alternative approach to virtual rehabilitation. Vision-based AI systems can evaluate joint motion and exercise performance using standard device cameras, eliminating the need for wearable sensors or specialized hardware (19). ViFive (Sunnyvale, CA) is a digital PT platform that leverages vision-based AI to provide real-time exercise feedback and engagement tracking using commonly available devices. This sensorless design distinguishes ViFive from many existing tele-rehabilitation and AI-assisted platforms to track movement that require additional equipment or continuous synchronous supervision. Prior validation studies have demonstrated the ability of vision-based AI to accurately assess joint range of motion in musculoskeletal applications (19), though clinical outcome data in patients with knee OA remain limited.

Beyond traditional clinical endpoints such as pain and function, patient-centered factors are increasingly recognized as important in the conservative management of knee OA. Patients’ interest in pursuing surgical intervention may reflect symptom burden, perceived treatment effectiveness, and confidence in conservative care pathways. While such measures are exploratory and not formally validated outcomes, they may provide complementary insight into the feasibility, sustainability, and patient-perceived value of emerging digital PT interventions.

Given the heterogeneity of existing virtual PT platforms and the limited evidence specific to low-cost, sensorless AI-enhanced programs for knee OA, preliminary investigation is warranted before undertaking larger comparative trials. Accordingly, this study was designed as a pilot, hypothesis-generating investigation to evaluate the feasibility, adherence, and preliminary clinical outcomes associated with an AI-enhanced virtual PT and HEP for knee OA. The primary objective of this study is to evaluate the effectiveness of a virtual, AI-enhanced PT program, such as ViFive, in reducing knee pain and improving functional outcomes and quality of life in adults diagnosed with knee OA. Secondary exploratory aims included evaluating changes in participants’ interest in surgical management and sustained engagement with a HEP following an initial period of guided virtual therapy. Furthermore, the study aims to explore the feasibility and sustainability of engaging participants in a HEP following the initial guided virtual PT sessions. By addressing these uncertainties, this pilot study seeks to inform the design and justification of future controlled studies evaluating sensorless, AI-based virtual PT in knee OA care.

Methods

From January to December of 2023, consecutive patients were recruited from an academic outpatient clinic, once they were diagnosed with knee OA and recommended to have PT. Given the pilot nature of this investigation, a time-limited trial for 1 year was chosen was set to provide a referendum on this technology prior to engaging in a larger, long-term study. Inclusion criteria included patients 18–65 years old with knee pain and a diagnosis of knee OA assessed via Kellen-Lawrence grade on radiographs and clinical evaluation by a Physical Medicine and Rehabilitation as well as Sports Medicine Board Certified Specialist physician (E.Y.R.). Exclusion criteria included the patient having undergone a recent (within 3 months) knee surgery within the start time of the study and patients outside of the aforementioned age range. Patients who fit the inclusion criteria were given an option to choose the ViFive (Sunnyvale, CA) virtual rehab or an in-person PT. If the patients chose the in-person PT, they were not able to participate in the study and were not followed. During the first 12 weeks, participants met weekly via video call with a coach, who was a licensed doctor of physical therapy (DPT). During these calls, exercises that were planned for the patient’s upcoming module were introduced and practiced by participants, coaches provided support and recommendations for symptom management such as exercise modifications, and educational topics were covered. Specific topics in health education were directed by the therapist and often involved the benefits of physical exercise. In addition, feedback was solicited regarding patients’ pain, ability to perform exercises, and questions with the platform as well as next steps for the patient’s progression were answered by the coach. Between calls, participants were expected to complete a HEP minimum three times per week using the ViFive app that utilized AI-based visual tracking of their movements and recorded engagement data. The exercise program included dynamic stretching, balance, and resistance training for the hips and lower extremities. The plan consists of ten modules that patients progressed through after the DPT’s review the patients’ progress and cleared them for the next module. The progress of patients depends on patients’ ability to perform the exercises, effectively and without discomfort. The actual content of the modules was designed by the DPTs and is individually tailored to each individual patient; however, the modules were crafted based on standard PT principles such as improving range of motion and strength of the muscles surrounding the knee. The difficulty of resistance exercises was progressed over the 12-week period using body weight exercises alone without additional equipment. Coaches accessed participants’ engagement data weekly to serve as a source of accountability and identify exercises that needed modification or review during video calls. The principal investigator monitored safety and finalized the decision for the progression of exercise program.

During the latter 12 weeks, participants were given access to the same ViFive virtual HEP, but no additional weekly meetings with their coach was scheduled. Each participant was asked to complete a questionnaire from the Knee Injury and Osteoarthritis Outcome Scale (KOOS) (Figure 1), Visual Analog Scale (VAS) for pain quantification, and asked, “Are you considering surgical options to manage your knee OA symptoms? Yes or no?” to measure interest in progression to surgery at baseline and at 3 and 6 months.

Figure 1 Example KOOS. KOOS, Knee Injury and Osteoarthritis Outcome Scale.

Statistical analysis

Wilcoxon signed-rank tests were used to analyze changes in all patient-reported outcome (PRO) measures from baseline to 3 and 6-month follow-up. All analyses were conducted in RStudio version 2023.12.1+402 (Boston, MA) using a two-sided level of significance of 0.05. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional review board of Stanford University (No. 63794) and informed consent was obtained from all individual participants.

Results

Thirty-four adult participants were initially recruited, and twenty-three participants completed all required questionnaires for the study (Figure 2). During the first 12 weeks of the study period, 3 participants sought a surgical treatment, three participants converted to an in-person PT, and one participant did not respond when the research team attempted to contact them. Patients were not asked further questions about their decision to unenroll and not participate in the ViFive protocol to respect patient’s autonomy and privacy. From the week 13 to week 24, 4 participants did not respond when the research team attempted to contact them. Given that the analysis was completed at the 12- and 24-week marks, patients that dropped out or were excluded from the study were not included for analysis at these different time points. Demographic data of age and gender from each time frame (baseline, 3 months and 6 months) are listed in Table 1 respectively to illustrate the impact of dropouts. Participant ages ranged from 26–65 years old. 18 female and 9 male participants finished 3-months follow-up and 15 female and 8 male participants completed the study with 6-months follow-up. The severity of knee OA was diagnosed with the Kellgren-Lawrence (K/L) grading system at baseline as well as at 3 and 6 months (Table 1).

Figure 2 Study participant flow chart. PT, physical therapy.

From baseline to 3 months follow-up as well as from baseline to 6 months follow-up, significant improvements in all PROs were observed. The median VAS decrease was 1 point (P<0.001) at 3 and 6 months, the median KOOS pain increase from baseline was 12.5 points and 16.7 points at 3 and 6 months respectively (P<0.001 comparing post intervention score to baseline at 3 and 6 months), the median KOOS symptom increase was 3.6 points at 3 months and sustained at 6 months (P<0.001 comparing post intervention score to baseline at 3 months and P=0.051 at 6 months from baseline), the median KOOS activity of daily living (ADL) increase was 12.4 points at 3 months and slightly decreased to an 11.8 point increase at 6 months (P<0.001 comparing post intervention score to baseline at 3 and 6 months), the median KOOS sport and recreation increase was 22.5 points and sustained at 6 months (P<0.001 comparing post intervention score to baseline at 3 and 6 months), and the median KOOS quality of life increase was 18.8 points at 3 months and a 16.7 point increase from baseline at 6 months (P<0.001 comparing post intervention score to baseline at 3 and 6 months) (Table 2). There was no significant increase in patient-reported interest in surgery and a significant decrease in patient report interest in surgery at 6 months (P=0.18 at 3 months and P=0.04 at 6 months) (Table 2).

Discussion

Key findings

Our findings support the effectiveness of a visual AI-based virtual rehab intervention in improving pain, function, quality of life, and activity participation in patients with knee pain over a 6-month period. The study is notable for being the first to investigate a low-cost, sensor less, AI-enhanced PT option specifically for patients with knee OA. Importantly, improvements in pain and function observed at the 3-month follow-up were sustained at the 6-month mark, suggesting patient adherence to the HEP beyond the initial guided sessions.

Strengths and limitations

The strengths of this pilot study include its innovative approach to utilizing AI technology in PT, which addresses common barriers associated with traditional in-person therapy. Overall, majority of participants (87.1%) remained with the AI assisted intervention after onboarding with the app. The technology seemed to be well accepted to provides a user interface that allowed users to understand and navigate the system well (Figure 3). It is noted however that this specific finding is subjected to selection bias as patients designated themselves as being a part of the ViFive group instead of the in-person option. The prospective design and use of validated outcome measures enhance the reliability of the findings. The ViFive system provided AI assisted interactive daily session combined with a weekly virtual meeting with a coach reviewing compliance may have improved the PROs (Figure 4A,4B). The participants used commercially available tablet with a camera to run the app. For other clinical use, a desktop or a laptop computer with a camera can be used. However, the study also has limitations, including a small sample size, lack of control group and a higher-than-expected dropout rate, which may affect the generalizability of the results. Patients who dropped out of the study were not asked follow-up questions regarding their decisions to drop out of the study. Therefore, it is not clear what factors lead to this decision. The conductors of this study also instituted an age-restriction <65 years of age within the study to increase the likelihood of usability and adherence to the technology as a way to better isolate the platform’s efficacy. Given the pilot nature of this study, the designers felt that the age-restrictions and patient-selected method for opting into the study was most appropriate. With this and the ability for patients to drop out of the ViFive protocol introduced selection bias for patients who likely would interact with and adhere to a virtual, AI-based PT program as compared with the general public. This element of selection bias and age restriction may limit the generalizability of the conclusions made within this manuscript.

Figure 3 Example dashboard detailing a participant’s HEP using the ViFive application. HEP, home exercise program.

Figure 4 Example AI-derived visual outputs that participants would interact with during exercises guided by the ViFive Program. (A) An example exercise with AI-assistance demonstrating a knee angle measurement. (B) Demonstration of knee flexion exercise with overlay of an example figure performing the exercise. This image is published with the participant’s consent. AI, artificial intelligence.

Comparison with similar research

This study aligns with emerging research that highlights the potential of digital health interventions in musculoskeletal care. Previous studies have shown that technology can impact multiple areas of the orthopedic care cycle. Applications have been developed using both guidelines for practitioners seeing patients with orthopedic concerns allow surgeons to guide post-operative rehabilitation, bridge gaps in telehealth, as well as virtual therapy (17,18,20). These technology solutions can effectively reduce pain and improve function in various patient populations and overall mitigate some of the burdens that patients with musculoskeletal concerns face each day. This study uniquely focuses on a low-cost, sensorless, AI-enhanced platform, distinguishing it from other telehealth PT interventions that often require specialized equipment.

While telerehabilitation has been previously identified as a viable approach to conditions such as knee arthritis, ViFive goes beyond traditional telerehabilitation platforms, simple video chats with therapists to guide exercises, through its AI-based platform (21,22). ViFive arms licensed therapists with additional information provided by the AI-based information such as range of motion and consistency of HEP adherence. These factors allow therapists to understand how the patient’s body is moving without tactile feedback and difficult perspective through video chat, which has been previously highlighted as a concern therapists have with other telerehabilitation platforms (23). Many telerehabilitation studies monitor patients just during the duration of the therapy (8–12 weeks), however this study is unique as it follows patients to the delivery of a HEP and for 6 months (22). This highlights that ViFive and other AI-based PT strategies should be studied further to demonstrate their sustained benefits.

Previous studies have listed that the minimal clinically important difference (MCID) for the KOOS is 12 points overall. Additionally, for the subsets, 12 points for Pain, 10 points for ADL, 9 points for sports and recreation and 14 for quality of life, however this is primarily listed in the surgical literature rather than the non-operative intervention literature (24). Our study demonstrated that the median of participants who completed the intervention was greater than the MCID for the KOOS suggesting that this may be a viable option for patients. However the reported change in VAS scores were less than the historic MCID for VAS, which is 1.37 (25). This highlights that ViFive and other AI-based telerehabilitation platforms are viable treatments for those with knee OA leading to long term results and adherence to HEP, a key concern in many telerehabilitation programs (22). These findings seem consistent with improvements in functional outcome scores seen in other telerehabilitation studies within a similar demographic for knee OA as well (26).

Of note, while interest in pursuing surgery is not a validated outcome measure, it is a positive indicator that patients’ desire surgery decreased, which can be a goal of patients who are trying to avoid more invasive procedures. The findings contribute to the growing body of literature supporting the use of virtual PT as a viable alternative to traditional methods.

Explanations of findings

The significant improvements in pain and function may be attributed to the structured nature of the virtual program, which included personalized coaching and a tailored home exercise regimen. The AI technology facilitated real-time feedback and monitoring of exercise performance, enhancing participant engagement and adherence to the HEP. The sustained improvements at the 6-month follow-up suggest that the initial guided sessions may have prepared participants for independent exercise, fostering long-term adherence and continued benefits. However, the presented study is a pilot study with a focus on hypothesis generation and suggests that these findings should be confirmed with a more robust study in the future.

Implications and actions needed

The findings of this pilot study have important implications for the management of knee OA, particularly in enhancing access to care for patients facing barriers to traditional PT by access and time. Virtual, AI-enhanced PT programs could be a potential additional conservative therapy option for patients, offering a cost-effective and accessible alternative. Patients may use this path as an initial treatment option while they are waiting for the in-person PT. Future research should focus on larger, randomized controlled trials to further validate these findings and explore the long-term effects of virtual therapy for patients of all ages. Also, a longer duration follow-up can be considered. Additionally, efforts should be made to improve technological literacy among older adults to maximize the usability and effectiveness of digital health interventions in this demographic as this would likely lead to increased success amongst older adults. Future studies could also consider investigating a direct comparison of AI-sensorless interventions to in-person PT as well as other virtual PT options such as those that use augmented reality or video-based PT counseling sessions for participants of all ages.

Conclusions

This pilot study highlights that an AI-based virtual PT option such as ViFive, may help some patients’ functional outcomes for those with knee OA. This study highlighted that after 24-week program patients aged 18–65 had sustained functional benefits which may be explained by increased adherence to a HEP beyond the initial 12-week PT regimen within the ViFive program. This study was not without its limitations, however the aim of the of authorship was to demonstrate that this type of technology was clinically viable to help patients with knee OA. Further studies are warranted to include a wider array of patients and a direct comparison of in-person PT to an AI-based virtual PT.

Acknowledgments

The authors appreciate the support from the Higgins Family Foundation.

Footnote

Data Sharing Statement: Available at https://jmai.amegroups.com/article/view/10.21037/jmai-2025-1-166/dss

Peer Review File: Available at https://jmai.amegroups.com/article/view/10.21037/jmai-2025-1-166/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jmai.amegroups.com/article/view/10.21037/jmai-2025-1-166/coif). M.W.K. serves as an unpaid editorial board member of Journal of Medical Artificial Intelligence from June 2025 to May 2027. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional review board of Stanford University (No. 63794) and informed consent was obtained from all individual participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Yang G, Wang J, Liu Y, et al. Burden of Knee Osteoarthritis in 204 Countries and Territories, 1990-2019: Results From the Global Burden of Disease Study 2019. Arthritis Care Res (Hoboken) 2023;75:2489-500. [Crossref] [PubMed]
2. Duong V, Oo WM, Ding C, et al. Evaluation and Treatment of Knee Pain: A Review. JAMA 2023;330:1568-80. [Crossref] [PubMed]
3. Katz JN, Arant KR, Loeser RF. Diagnosis and Treatment of Hip and Knee Osteoarthritis: A Review. JAMA 2021;325:568-78. [Crossref] [PubMed]
4. Dantas LO, Salvini TF, McAlindon TE. Knee osteoarthritis: key treatments and implications for physical therapy. Braz J Phys Ther 2021;25:135-46. [Crossref] [PubMed]
5. Chen J, Chen X, Wang T, et al. Global burden of knee osteoarthritis from 1990 to 2021: Trends, inequalities, and projections to 2035. PLoS One 2025;20:e0320115. [Crossref] [PubMed]
6. Kolasinski SL, Neogi T, Hochberg MC, et al. 2019 American College of Rheumatology/Arthritis Foundation Guideline for the Management of Osteoarthritis of the Hand, Hip, and Knee. Arthritis Care Res (Hoboken) 2020;72:149-62. [Crossref] [PubMed]
7. Marriott KA, Birmingham TB. Fundamentals of osteoarthritis. Rehabilitation: Exercise, diet, biomechanics, and physical therapist-delivered interventions. Osteoarthritis Cartilage 2023;31:1312-26.
8. Sharma L. Osteoarthritis of the Knee. N Engl J Med 2021;384:51-9. [Crossref] [PubMed]
9. Dijkstra F, van der Sluis G, Jager-Wittenaar H, et al. Facilitators and barriers to enhancing physical activity in older patients during acute hospital stay: a systematic review. Int J Behav Nutr Phys Act 2022;19:99. [Crossref] [PubMed]
10. Zalewski K, Alt C, Arvinen-Barrow M. Identifying barriers to remaining physically active after rehabilitation: differences in perception between physical therapists and older adult patients. J Orthop Sports Phys Ther 2014;44:415-24. [Crossref] [PubMed]
11. Maresova P, Krejcar O, Maskuriy R, et al. Challenges and opportunity in mobility among older adults - key determinant identification. BMC Geriatr 2023;23:447. [Crossref] [PubMed]
12. Passalent LA, Landry MD, Cott CA. Wait times for publicly funded outpatient and community physiotherapy and occupational therapy services: implications for the increasing number of persons with chronic conditions in ontario, Canada. Physiother Can 2009;61:5-14. [Crossref] [PubMed]
13. Gupta L, Najm A, Kabir K, et al. Digital health in musculoskeletal care: where are we heading? BMC Musculoskelet Disord 2023;24:192. [Crossref] [PubMed]
14. Costa F, Janela D, Molinos M, et al. Telerehabilitation of acute musculoskeletal multi-disorders: prospective, single-arm, interventional study. BMC Musculoskelet Disord 2022;23:29. [Crossref] [PubMed]
15. Gilbert AW, Booth G, Betts T, et al. A mixed-methods survey to explore issues with virtual consultations for musculoskeletal care during the COVID-19 pandemic. BMC Musculoskelet Disord 2021;22:245. [Crossref] [PubMed]
16. Tenforde AS, Borgstrom H, Polich G, et al. Outpatient Physical, Occupational, and Speech Therapy Synchronous Telemedicine: A Survey Study of Patient Satisfaction with Virtual Visits During the COVID-19 Pandemic. Am J Phys Med Rehabil 2020;99:977-81. [Crossref] [PubMed]
17. Havran MA, Bidelspach DE. Virtual Physical Therapy and Telerehabilitation. Phys Med Rehabil Clin N Am 2021;32:419-28. [Crossref] [PubMed]
18. Carulli C, Rossi SMP, Magistrelli L, et al. Can Artificial Intelligence Help Orthopaedic Surgeons in the Conservative Management of Knee Osteoarthritis? A Consensus Analysis. J Clin Med 2025;14:690.
19. Meng Y, Lee M, Clark HJY, et al. Evaluation of range of motion using vision artificial intelligence (AI) in musculoskeletal medicine. J Med Artif Intell 2024;7:22.
20. Rossi SMP, Panzera RM, Sangaletti R, et al. Problems and Opportunities of a Smartphone-Based Care Management Platform: Application of the Wald Principles to a Survey-Based Analysis of Patients’ Perception in a Pilot Center. Healthcare 2024;12:153.
21. Tore NG, Oskay D, Haznedaroglu S. The quality of physiotherapy and rehabilitation program and the effect of telerehabilitation on patients with knee osteoarthritis. Clin Rheumatol 2023;42:903-15. [Crossref] [PubMed]
22. Plavoukou T, Iosifidis M, Papagiannis G, et al. The Effectiveness of Telerehabilitation in Managing Pain, Strength, and Balance in Adult Patients With Knee Osteoarthritis: Systematic Review. JMIR Rehabil Assist Technol 2025;12:e72466. [Crossref] [PubMed]
23. Saaei F, Klappa SG. Rethinking Telerehabilitation: Attitudes of Physical Therapists and Patients. J Patient Exp 2021;8:23743735211034335. [Crossref] [PubMed]
24. Migliorini F, Maffulli N, Schäfer L, et al. Minimal clinically important difference (MCID), substantial clinical benefit (SCB), and patient-acceptable symptom state (PASS) in patients who have undergone total knee arthroplasty: a systematic review. Knee Surg Relat Res 2024;36:3. [Crossref] [PubMed]
25. Angst F, Aeschlimann A, Stucki G. Smallest detectable and minimal clinically important differences of rehabilitation intervention with their implications for required sample sizes using WOMAC and SF-36 quality of life measurement instruments in patients with osteoarthritis of the lower extremities. Arthritis Rheum 2001;45:384-91. [Crossref] [PubMed]
26. Azma K. Efficacy of tele-rehabilitation compared with office-based physical therapy in patients with knee osteoarthritis: A randomized clinical trial. J Telemed Telecare 2018;24:560-5. [Crossref] [PubMed]