The protocol of this meta-epidemiological study is available in the Open Science Framework (OSF). We reported the study according to the Guidelines for Reporting Meta-epidemiological Methodology Research.18

We included I) randomized controlled trials (RCTs) of II) individuals with health conditions treated by III) telehealth interventions in physiotherapy, IV) considering any comparison intervention, and V) clinical health outcome. Telehealth in physiotherapy studies was considered if it accounted for at least 50% of the total intervention. We classified the telehealth intervention components as follows: I) education, II) psychological, III) exercise, IV) physical activity, V) multicomponent interventions, and VI) others, following the classifications used in previous studies.19

We considered studies to be equity-focused trials if they directly investigated the intervention's effectiveness on an individual’s health in the trial with respect to any of the PROGRESS-Plus factors (e.g., a culturally sensitive pain neuroscience and exercise program delivered via telehealth for low-income black women). We considered studies to be health equity-oriented if they included any form of exploratory analysis (e.g., subgroup or interaction analyses) related to any PROGRESS-Plus factors. We classified countries according to the World Bank classification.20 We have not included protocols, conference abstracts, commentaries, editorials, or letters. Also, health conditions that are extremely prevalent in certain factors, for example, gender, such as fibromyalgia, breast cancer, and others, were not considered due to an unclear distinction between equity and inequality within the study.

One review author searched the Cochrane Library (CENTRAL), the Physiotherapy Evidence Database (PEDro), and PubMed from 2020 to November 2023. We chose this period because the CONSORT-Equity 2017 guidelines were published in 2017, outlining standards for reporting key points in health equity studies.5 We consider our data threshold a reasonable timeline for disseminating and implementing information related to health equity.

The Cochrane Library (CENTRAL), PEDro database, and PubMed database are considered the most comprehensive sources for searching for indexed physical therapy trials and systematic reviews.21 We used a search strategy with study design filter22 (e.g., randomized controlled trial), relevant telehealth terms (e.g., telerehabilitation)19 and physiotherapy terms (e.g., physical therapy modalities) (Supplementary file 1). Studies published in languages other than English, Spanish, or Portuguese were translated by DeepL software.23

A pair of review authors independently screened title and abstract citations in Covidence.24 Pairs of review authors evaluated potential full-text articles identified independently to determine the selection criteria. Disagreements between the reviewers were resolved through discussion or consensus with a third reviewer. To facilitate the screening, we used a step-by-step process with examples from a previous study modified for our purpose15 (Supplementary file 2).

We developed a standardized spreadsheet to collect the data and performed a pilot test on a random sample of 10 RCTs. Data were extracted by one review author and two reviewers, and 30% of randomly selected manuscript reports were verified to avoid extraction errors.25 Data correction was made [n = 12 (0.5%)] of the data verified, and any discrepancies between reviewers were resolved by consensus meetings. We extracted information on the characteristics of the authors, study/methods, participants, intervention, technology platform, outcome measure, subdiscipline, results/findings, and data related to PROGRESS plus factors.

We assessed potential dimensions of health inequalities using the PROGRESS plus factors.2 The PROGRESS plus factors were composed of the following factors: (I) Place of residence; (II) Race/ethnicity/culture; (III) Occupation, (IV) Gender, (V) Religion, (VI) Education, (VI) Socioeconomic status, (VIII) Social capital and networks (Measures neighbourhood or community trust, ability to rely on neighbours for support) and IX) minority groups or people in vulnerable situations. The acronym Plus factor covers factors such as age, disability, sexual orientation, people at a temporary disadvantage, and any additional factors specific to the context.2 The last item, minority groups, is an additional item proposed in this study due to being a relevant contextual factor with poor attention in physiotherapy.

First, we assessed whether the trial mentions, evaluates, or defines health equity in the objective, background, methods, or discussion sections. We evaluated the inclusion of PROGRESS plus factors in the description of participants (e.g., baseline characteristics), in primary or subgroup analyses (e.g., advantaged or disadvantaged groups), in interaction analyses (e.g., gender), and in practical study implications.

We analyzed all the data descriptively. Telehealth equity-oriented trials were evaluated through the hypothesis-testing approach with trial characteristics demonstrated in the Harvest plot.15,26 The hypothesis-testing approach specifies three social gradient hypotheses ‘positive gradient’, ‘negative gradient’ and ‘no social gradient’.26 A ‘positive social gradient’ proposes that an intervention is more effective or beneficial for an outcome in more advantaged groups.26 A ‘negative gradient’ suggests that an intervention is more effective or beneficial for an outcome in more disadvantaged groups.26 The ‘no social gradient’ represented by the null hypothesis proposes that an intervention has no differential effects or similar benefits between groups in subgroups or interaction analyses.26 Trial characteristics consisted of intervention type, sample size, and type of outcome measurement (i.e., objective, self-reported or observation measures). We considered the outcome analyzed in the short-term follow-up (closest to 3 months after randomization). For studies that conducted subgroup and interaction analyses, we preferred interaction analysis. We performed additional analysis that considered funding status (yes/no) and country income level (LMICs and HICs).

Overall, the PROGRESS Plus factors most frequently mentioned in the background were place of residence (n = 29, 40%), followed by gender/sex (n = 21, 29%) and race/ethnicity/culture (n = 19, 26%). The most frequently reported baseline characteristics were gender/sex (n = 76, 85%), education (n = 52, 58%), and social capital (n = 44, 49%). No PROGRESS Plus factors were consistently reported across all studies (Table 2).

Telehealth benefits were observed in 28 trials (62%), with outcomes primarily assessed via self-reported measures (n = 31, 68%). Place of residence was considered in 14 trials (31%),29,30,36,43,47,53,54,61,104–109 with six36,47,53,104,107,108 demonstrated significant outcomes favoring telehealth compared to no intervention,53 unmatched105,108,109 and matched treatment36,47,104,106,107 for adherence, function, function capacity, knowledge assessment, pressure pain threshold and HbA1c levels.

Race/ethnicity/culture was considered in 11 (24%) trials,31,32,34,37,40–42,45,46,51,59,110 with five31,37,41,42,45 demonstrated significant outcomes favoring telehealth compared to no intervention,41 minimal intervention,31 matched42,45 and unmatched treatment34,110 for pain intensity, function, health related quality of life, moderate levels of physical activity (MVPA) or BMI. Occupation was considered in eight (18%) trials,48–50,52,55–57,62 of which five48,50,52,56,62 demonstrated significant outcomes favoring telehealth compared to no intervention48,52,56 and matched treatment50,62 for MVPA, lost work time, COPD-specific knowledge, sedentary time or pain intensity.

Gender/sex was considered in 11 (24%) trials,27,28,32,35,38,39,59,104,111,112 with seven27,28,35,38,104,112 reporting significant outcomes favoring telehealth compared to no intervention,28,38 minimal111 and matched treatment27,35 for pain intensity, balance or MVPA. Education was considered in only one (2%) trial34 that reported significant outcomes favoring telehealth compared to unmatched treatment34 for the MVPA outcome.

Social capital and networks were considered in two (4%) trials,33,58 with only one58 (50%) demonstrating significant outcomes favoring telehealth compared to no intervention58 for the exercise knowledge outcome. Two (5%) trials44,60 considered people or groups in vulnerable situations (e.g., refugees and people with HIV), both demonstrating significant outcomes favoring telehealth compared to matched44 and unmatched treatment60 for MVPA or weight control strategies. Religion and socioeconomic status were not considered in any trials. Fig. 2 presents the results, and supplementary file 7 provides additional details on telehealth equity-focused trials.

Fig. 2.

Effectiveness telehealth equity-focused trials using the PROGRESS plus factors.

Telehealth effectiveness was observed in 15 (30%) trials, and 33 (67%) trials considered more than one PROGRESS plus factor. Gender/sex factors were considered in 33 (67%) trials, with 11 (22%) trials66,71,72,77–80,97,103,113,114 favoring improved clinical outcomes. The primary methods were linear mixed-effects models and generalized and multivariate linear regression analyses. Outcome measures were primarily self-reported (n = 19, 39%) or objective (n = 16, 32%).

Place of residence was analyzed in four (8%) trials,74,78,79,87 three78,79,87 which demonstrated significant outcomes compared to minimal intervention,78 matched79 and unmatched treatment87 favoring telehealth regarding MVPA and mobility. No social gradient was observed across trials, indicating similar effectiveness between groups in subgroup or interaction analyses.

Race/ethnicity/culture was analyzed in nine (18%) trials70,82,86,89,90,93,97–99 Among these, four trials86,90,93,97 demonstrated significant outcomes favoring telehealth over no intervention,93 matched90 and unmatched treatment86,97,98 for outcomes such as MVPA, number of trigger areas or HbA1c. Only one trial97 reported a negative gradient, indicating a decrease in inequity. This trial indicated that participants from various ethnic groups (African American, Native American/American Indian, and Hispanic) were less likely to discuss the number of asthma triggers compared to non-Hispanic White participants. Eight trials (89%) reported no social gradients.

Occupation was analyzed in nine (18%) trials68,70,71,77,80,89,93,94,102 Five trials68,71,77,79,93 demonstrated significant outcomes favoring telehealth over no intervention,77,80,93 minimal intervention71 and matched treatment68 regarding MVPA, depression, perceived stress. One trial68 reported a positive gradient, indicating increased inequities; it suggested that nurses with fixed work hours were more likely to increase their MVPA levels. Eight trials (89%) reported no social gradients.

Gender/sex was analyzed in 39 (79%) trials.63–67,71–73,75–82,84,85,88,89,91,92,94–103,113–119 Eleven trials,66,71,72,77–80,97,103,114 demonstrated significant outcomes favoring telehealth over no intervention,77,80,103 minimal intervention,71,78,114 matched66,72,79,113 and unmatched treatment97 for outcomes such as knowledge assessment, function, MVPA, goal adjustment, depression, perceived stress, mobility, asthma trigger identification or adherence to lifestyle behavioral changes. One trial,103 reported a positive gradient, suggesting that males are more likely to adhere to diet-related lifestyle behavior changes. Two trials78,97 reported negative gradients: one trial78 suggested that female cancer survivors are more likely to increase MVPA, while the other trial97 indicated that females are less likely to discuss the number of asthma triggers. The remaining 35 trials (71%) reported no social gradient. Religion was not analysed in any trials.

Education was analyzed in 15 (30%) trials.66,67,70,71,75,77,80,83,86,89,91–94,102,103,114 Seven trials71,77,80,86,93,103,114 demonstrated significant outcomes favoring telehealth over no intervention,77,80,93,103 minimal intervention71 and unmatched treatment86 for outcomes including knowledge, function, MVPA, adherence to lifestyle behavior changes, goal adjustment, improved depression or perceived stress. One trial103 reported a positive gradient, indicating that individuals with secondary or higher secondary education were more likely to adhere to smokeless tobacco and betel nut cessation. The remaining 14 trials (93%) reported no social gradient.

Socioeconomic status was analyzed in seven (14%) trials.69,70,77,86,90,94,101 Three trials77,86,90 demonstrated significant outcomes favoring telehealth over no intervention,77 matched90 and unmatched treatment86 for outcomes such as MVPA, depression, adherence to lifestyle behavior changes for smoking and betel nut cessation, or perceived stress; however, no social gradients were reported.

Ten (20%) trials analyzed social capital and networks.64,69,71,77,80,86,93,94,97,103 Eight trials demonstrated significant outcomes favoring telehealth over no intervention,77,80,93,103 minimal intervention,71 matched69 and unmatched treatment86,97 for outcomes such as MVPA, depression, and the number of trigger areas. One trial69 reported a negative gradient, suggesting that mothers' partners who utilized collaborative planning and supportive communication were more likely to increase MVPA. Another trial103 reported a negative gradient, indicating that nuclear families were more likely to adhere to smokeless tobacco and betel nut cessation. Other trials reported no social gradient. One (2%) trial analyzed minority groups or populations in vulnerable situations, demonstrating significant outcomes favoring telehealth over no intervention for depression.116 This trial reported no social gradient. Fig. 3 presents the results, while supplementary file 8 provides additional details on telehealth equity-oriented trials.

Fig. 3.

Effectiveness telehealth equity-oriented trials using the PROGRESS plus factors.

Overall, 77 (82%) highlighted one or more PROGRESS-Plus factors in their sections on practical and research implications. The most frequently mentioned factors were place of residence (36 trials, 46%), gender/sex (33 trials, 43%), and race/ethnicity/culture (24 trials, 31%) (Table 2).

Organizations funded 82% of the trials; place of residence (14 trials, 28%), gender/sex (11 trials, 22%), and occupation (8 trials, 16%) were the most frequently addressed factors in equity-oriented telehealth trials. In high-income countries, place of residence (14 trials, 28%), gender/sex, and race/ethnicity/culture (11 trials, 22% each) emerged as the most frequently addressed factors. Detailed information is provided in Supplementary files 9–10.

Most telehealth equity trials come from high-income countries and focus predominantly on factors such as gender/sex, education, and race/ethnicity/culture. Gender/sex was the most reported factor among the studies' baseline characteristics. However, none of these factors were consistently documented across all trials. This gap is important because comprehensive reporting of these factors is essential for robust evidence synthesis and informed clinical decision-making.

Nearly half of the trials were classified as equity-oriented, with 24% primarily addressing race/ethnicity/culture and gender/sex. Among these equity-oriented trials, 2% to 12% showed positive or negative social gradients. For example, a telephone health coaching intervention for female cancer survivors in regional or remote areas increased moderate-to-vigorous physical activity compared to their male counterparts,78 similarly, a web-based program favored nurses with fixed rather than rotating shift schedules.68 These examples underscore the need for interventions designed to be accessible, acceptable, and effective for disadvantaged populations.2,122

Randomized trials can contribute meaningfully to health equity when they involve socially disadvantaged populations or examine differential effects across groups.125 Besides, most studies can be equity-relevant without an explicit purpose5 as conducting studies in a specific population or context and running effect-modification analyses considering factors such as gender, ethnicity, or socio-economics. In this case, both a robust claim of interaction and external validity of the results will provide additional information for stakeholders' decision-making, especially when population characteristics and context are taken into account. Our study provides an equity lens on telehealth equity in physiotherapy trials and a ground for data-based discussion in the community.

Health equity research was highlighted by several organizations, particularly within the field of musculoskeletal telehealth.17 The clinical fields most frequently studied were musculoskeletal, cardiothoracic, and continence or women's health. Approximately 2.41 billion people could benefit from rehabilitation treatment, of which 1.71 billion people have musculoskeletal conditions.126 Future telehealth equity studies should broaden their scope across different clinical fields and consistently report key PROGRESS-Plus factors.

Half of the included telehealth trials conducted effect-modification analyses. Researchers often conduct effect modification analyses to assess whether intervention effects vary according to other variables, such as age, disease severity, or clinical setting.127 This is critical, as health inequities typically arise from an interplay of multiple PROGRESS factors rather than a single, stand-alone factor.2 Exploring evidence regarding the digital divide is essential to ensure that benefits reach the populations most in need, rather than being limited to the wealthiest sociodemographic groups.11,124 Furthermore, addressing these factors helps overcome the early-stage challenges of expanding telehealth interventions by mitigating the consequences of differential effects in access and usage.43,124 Incorporating these considerations into effect modification analyses would prevent policymakers from overlooking strategies that promote inclusivity, benefit target groups, and enhance the capacity for global health responses.124

Our study's strengths include highlighting evidence for telehealth equity in physiotherapy and providing an overview of platform technologies. Further, we have used the PROGRESS plus framework, which is widely used and encouraged by organizations and collaborations. Our study has limitations related to study design and search strategy. Due to limited time, we did not include systematic reviews that may take an equity perspective. As we did not assess the credibility of the effect-modification analyses in this study, caution is warranted when interpreting the trial results to avoid leading patients to suboptimal care. A further limitation concerns our decision to exclude conditions disproportionately prevalent in specific groups. While this approach improved methodological clarity, it may have inadvertently excluded trials where these conditions act as important markers of social vulnerability (e.g., breast cancer in women, prostate cancer in men).

Future telehealth equity research requires a minimum level of alignment and compliance among researchers, publishers, and funders. Regardless of whether a trial is equity-focused or equity-oriented, researchers should use the CONSORT-Equity 2017 checklist to guide planning, design, and reporting of their findings. External validity is a critical factor in healthcare decision-making, particularly regarding the applicability of evidence across local, regional, and international contexts. While the growing interest in effect modification analysis is understandable, researchers conducting equity-oriented trials should utilize the ICEMAN tool to ensure they meet the criteria for robust analysis. This approach will facilitate more credible claims of effect modification without compromising optimal patient care.

Publishers should require adherence to reporting checklists to improve documentation of PROGRESS factors, which are considered relevant social determinants of health. Funding organizations should require a clear distinction between equity-focused and equity-oriented designs within healthcare research. Transparent reporting enables interested parties to identify, monitor, and allocate research funding to minorities and other disadvantaged or neglected populations.