Two researchers (CC and PPQ) independently searched the PubMed, Embase, Web of Science, and CENTRAL databases from inception to March 31, 2025. The search was limited to English-language literature. Randomized controlled trials investigating the effectiveness of postoperative telerehabilitation treatment for patients with hip fractures were included (Multimedia Appendix 1).

Additional articles were identified by reviewing the reference lists of relevant reviews. The following keywords and the corresponding MeSH (Medical Subject Headings) terms were used: (telerehabilitation OR telerehabilitations OR tele-rehabilitation OR tele-rehabilitations OR remote rehabilitation OR remote rehabilitations OR rehabilitation, remote OR rehabilitations, remote OR virtual rehabilitation OR virtual rehabilitations OR rehabilitation, virtual OR rehabilitations, virtual) AND (hip fractures OR fractures, hip OR intertrochanteric fracture OR intertrochanteric fractures OR trochanteric fractures OR fractures, trochanteric OR femur trochlear fracture OR femur trochlear fractures OR trochlear fracture, femur OR trochlear fractures, femur OR fractures, femur trochlear OR femoral trochlear fracture OR femoral trochlear fractures OR fracture, femoral trochlear OR fractures, femoral trochlear OR trochlear fracture, femoral OR trochlear fractures, femoral OR subtrochanteric fracture OR subtrochanteric fractures).

The inclusion criteria were as follows: (1) studies involving participants diagnosed with a hip fracture who were in a stable condition (no further fractures were observed or reported) after surgical intervention; (2) studies involving telerehabilitation as the primary intervention following surgery, including the use of SMS text messaging, telephones, smartphones, and internet-based and software-based methods; (3) studies that reported at least one outcome of interest, including the HHS, TUG test, SPPB score, BI, MBI, and FIM; and (4) randomized controlled trials.

The exclusion criteria were (1) only study protocol or conference abstract available and (2) unavailability of the full text.

The search results were imported into EndNote X9 (Clarivate Analytics) for duplicate removal. To ensure the accuracy of this process, 2 authors (CC and PPQ) manually verified the deleted duplicates. The titles and abstracts of the remaining records were then screened independently by these authors (CC and PPQ). Full texts were retrieved for further review if deemed eligible by either author. Any disagreements between the reviewers were resolved through discussion with a third author (CL).

Two authors (CC and PPQ) independently extracted data from the included studies. The extracted data encompassed the following: basic study information (the first author, publication year, and country), participant baseline characteristics (sample size, age, and gender), details of the telerehabilitation and control interventions, and outcome measures (including follow-up period and measurement tools).

Two authors (CC and PPQ) independently evaluated the quality of the included studies using version 2 of the Cochrane risk-of-bias tool for randomized trials [20]. The risk-of-bias assessment consisted of the following domains: (1) randomization process, (2) deviations from intended interventions, (3) missing outcome data, (4) measurement of the outcome, and (5) selection of the reported results. The overall risk of bias was classified as low risk of bias, some concerns, and high risk of bias.

The Cochrane Collaboration’s Review Manager software (version 5.3) was used for all statistical analyses. The overall estimate of the treatment effect was calculated using the means and SDs of outcome scores for continuous data in the telerehabilitation intervention and control groups. Short- and long-term effects were analyzed by comparing the differences in outcome scores between the 2 groups at the intervention end point and at the last follow-up time point, respectively. For studies using the same measurement tool, we calculated a pooled estimate of the mean differences (MDs) with 95% CIs. In cases in which different measurement tools were used, we used standardized MDs (SMDs).

To investigate potential clinical and methodological heterogeneity across the included studies, the following factors were summarized and compared: (1) patient characteristics (age and baseline scores of the outcomes of interest), (2) characteristics of telerehabilitation (duration, frequency, and delivery mode), (3) type of control intervention, (4) follow-up period for the outcomes of interest, and (5) risk of bias. Where low clinical and methodological heterogeneity was found across the studies in a given meta-analysis, a fixed-effects model was used; otherwise, a random-effects model was applied. Statistical heterogeneity was quantified using the I² statistic. I² values greater than 50% were considered to indicate substantial heterogeneity [21]. When substantial statistical heterogeneity was present and a sufficient number of studies were available, subgroup analyses were performed based on the abovementioned factors contributing to clinical or methodological heterogeneity. Furthermore, a sensitivity analysis was planned to assess the impact on the pooled results provided that a sufficient number of studies were available. This was performed by excluding studies with a high risk of bias or those that differed substantially from others in the factors considered for subgroup analyses.

The initial database search yielded a total of 1045 records, of which 6 (0.6%) met the eligibility criteria and were included. A further 2 studies were identified by checking the reference lists of relevant reviews [11], resulting in a total of 8 studies for the systematic review and meta-analysis. The study selection process is illustrated in the PRISMA flow diagram (Figure 1).

The 8 included studies comprised 740 patients, including 569 female and 161 male individuals. The sample sizes varied from 31 to 232 in the individual studies.

The treatment strategies for postoperative telerehabilitation for patients with hip fractures included strengthening, balance, coordination, stretching exercises, BADL training, introduction of walking aids, and safety education. Most included studies provided telerehabilitation through mobile apps or software [5,22-28] supervised by professionals remotely, one study used a DVD supervised via telephone calls [27], and one study used a telephone call as the primary strategy [28]. The control interventions comprised home-based exercise through written materials [5,25,26], in-person rehabilitation [23,24], follow-up assessment through telephone [22] and health nutrition education [27], and no intervention [28]. The duration of telerehabilitation varied from 35 minutes to 6 months. The follow-up periods varied from 3 weeks to 9 months [23,25-28]. Table 1 presents the main characteristics of the included studies [5,22-28].

Four studies [23,24,26,28] had a low risk of bias, 3 [22,25,27] had some concerns, and 1 [5] had a high risk of bias. Overall, the randomization process was the most problematic domain as 3 studies poorly described the method used [5,22,25]. The results of the methodological quality assessment are provided in Figure 2 [5,22-28].

HHS was measured in 2 studies involving a total of 156 participants at the intervention end point [5,22]. Low clinical heterogeneity was observed in patient characteristics (mean age range 72-77 years; comparable baseline mean HHS between 34.9 and 43.6). However, substantial clinical heterogeneity was found regarding the duration of telerehabilitation (6 months [5] vs 3 months [22]). A random-effects meta-analysis yielded a statistically significant, positive result (MD 7.42, 95% CI 5.61-9.23; P<.001; Figure 3 [5,22]) with low statistical heterogeneity (I²=3%; P=.36; Figure 3 [5,22]).

SPPB was measured in 4 studies (n=430 participants) at the intervention end point [22-24,27]. Considerable clinical heterogeneity was observed in baseline SPPB scores (mean scores 4.71-5.9 [22,27] vs 2.69-3.21 [23,24]). Furthermore, both the duration of telerehabilitation (3 months [22-24] vs 6 months [27]) and the type of control intervention (telephone assessment [22,27] vs in-person rehabilitation [23,24]) varied across studies. A random-effects meta-analysis showed a significant positive result at the intervention end point (MD 1.34, 95% CI 1.14-1.55; P<.001; Figure 4 [22-24,27]) with low to moderate statistical heterogeneity (I²=33%; P=.21; Figure 4 [22-24,27]). SPPB was measured in 2 studies (n=292 participants) at follow-up [23,27]. In addition to the aforementioned clinical heterogeneity, great methodological heterogeneity was noted in the follow-up period (3 months [27] vs 9 months after the intervention [23]). The random-effects analysis remained significant at follow-up (MD 1.17, 95% CI 1.00-1.34; P<.001; Figure 4 [22-24,27]) with low statistical heterogeneity (I²=0%; P=.72; Figure 4 [22-24,27]).

The TUG test was measured in 4 studies (n=156 participants) at the intervention end point [22,24-26]. Considerable clinical heterogeneity was found regarding the patient characteristics (mean age 65.7-67.3 years [26] vs 75.17-82.1 years [22,24,25]; mean baseline TUG score 21.4-25.6 [26] vs 39.7-45.2 [25] vs 66.53-99.72 [24]; data unavailable for the study by Zhang et al [22]). The duration of telerehabilitation (3-6 weeks [25,26] vs 3 months [22,24]) and type of control intervention (written materials [25,26] vs telephone assessment [22] vs in-person rehabilitation [24]) also varied. A random-effects meta-analysis demonstrated a significant positive result at the intervention end point (MD –8.45, 95% CI −11.28 to −5.62; P<.001; Figure 5) with low statistical heterogeneity (I²=0%; P=.53; Figure 5 [22,24-26]). The TUG test was measured in 2 studies (n=63 participants) at follow-up [25,26]. Low methodological heterogeneity was observed in the follow-up period (3-4 weeks after the intervention [25,26]) and risk of bias (low to moderate risk of bias [25,26]). However, due to the considerable clinical heterogeneity noted above (in patient age and baseline function), a random-effects model was applied. This analysis yielded a significant positive result at follow-up (MD −7.66, 95% CI −13.78 to −1.53; P=.001; I²=0%; P=.37; Figure 5 [22,24-26]).

Because various BADL measurement tools were used across the included studies, SMDs were used as the effect size measure. Zhang et al [22] reported BADL outcomes separately for the hip replacement and internal fixation groups. BADLs were measured in 5 studies (n=354 participants) at the intervention end point [5,22-25]. Considerable clinical heterogeneity was observed regarding patient characteristics (baseline dependence level: maximal dependence [5,22] vs mild to moderate dependence [23-25]), duration of telerehabilitation (3 weeks [25] vs 3 months [22-24] vs 6 months [5]), and type of control intervention (written materials [25] vs telephone assessment [5,22] vs in-person rehabilitation [23,24]). In addition, substantial methodological heterogeneity was observed in the risk of bias (high [5] vs some concerns [22,25] vs low [23,24]). A random-effects meta-analysis yielded a significant positive result with high heterogeneity (SMD 1.65, 95% CI 0.78-2.51; P<.001; I²=91%; P<.001; Figure 6 [5,22-25]). Subgroup analyses were conducted based on baseline dependence level, duration of telerehabilitation, and type of control intervention. For studies involving patients with a high baseline dependence level [5,22], the analysis showed a positive result with high heterogeneity (SMD 1.87, 95% CI 0.59-3.16; P=.004; I²=89%; P<.001; Figure 6 [5,22-25]). For studies with a lower baseline dependence level [23-25], the result was also positive with high heterogeneity (SMD 1.42, 95% CI 0.04-2.81; P=.04; I²=94%; P<.001; Figure 6 [5,22-25]). Subgroup analysis by duration of telerehabilitation was not possible due to an insufficient number of studies. For studies with telephone-based assessment as a control intervention [5,22], the result was positive with high heterogeneity (SMD 1.87, 95% CI 0.59-3.16; P=.004; I²=89%; P<.001; Figure 6 [5,22-25]). For studies with in-person rehabilitation as the control [23,24], the result was similarly positive with high heterogeneity (SMD 1.90, 95% CI 0.16-3.63; P=.03; I²=95%; P<.001; Figure 6 [5,22-25]). A sensitivity analysis was conducted by (1) excluding 2 studies with intervention durations shorter [25] or longer [5] than 3 months and (2) excluding 1 study with a high risk of bias [5]. Pooling the 3 remaining studies with a 3-month intervention duration [22-24] using a random-effects model yielded a positive result with low statistical heterogeneity (SMD 1.13, 95% CI 0.76-1.50; P<.001; I²=0%; P=.82; Figure 6 [5,22-25]). Pooling the 4 studies with low to moderate risk of bias [22-25] yielded a positive result with high heterogeneity (SMD 1.36, 95% CI 0.52-2.21; P=.002; I²=88%; P<.001). BADLs were measured in 3 studies (n=257 participants) at follow-up [23,25,28]. Considerable clinical heterogeneity was observed regarding baseline dependence level (mild to moderate dependence [23,25] vs minimal dependence [28]), duration of telerehabilitation (2.5-3 weeks [25,28] vs 12 weeks [23]), delivery mode of telerehabilitation (telephone based [28] vs software based [23,25]), and type of control intervention (no intervention [28] vs in-person rehabilitation [23] vs written materials [25]). Moreover, substantial methodological heterogeneity was found in the follow-up period (3 weeks [25] vs 5.5 months [28] vs 9 months [23] after the intervention). A random-effects model yielded a positive result with moderate heterogeneity at follow-up (SMD 0.43, 95% CI 0.05-0.81; P=.03; I²=48%; P=.15; Figure 7 [23,25,28]). A sensitivity analysis excluded 1 study [28] that differed substantially from the others in design (baseline dependence level, type of telerehabilitation, and type of control intervention). A random-effects model for the remaining 2 studies [23,25] yielded a positive result with low heterogeneity (SMD 0.66, 95% CI 0.26-1.05; P=.001; I²=0%; P=.41; Figure 7 [23,25,28]).

Our review found that postoperative telerehabilitation was effective in improving short- and long-term hip function, functional mobility, and ability to perform BADLs in patients with hip fractures compared to conventional rehabilitation. However, current evidence was weak due to the limited number and insufficient quality of the included studies and the heterogeneity across studies.

Our findings align with previous evidence that telerehabilitation is beneficial to short-term recovery of functional mobility and ability to perform BADLs at posttreatment compared to usual care [12]. Moreover, our review found that hip function was also improved after telerehabilitation. In addition, our findings revealed the effect of telerehabilitation on long-term recovery (3 weeks to 9 months after the intervention) in functional mobility and the ability to perform BADLs. The effectiveness in improving BADLs showed substantial heterogeneity across the included studies in the postintervention analysis. Although subgroup analyses were conducted based on the type of control intervention and the baseline dependence level, they failed to reduce the heterogeneity. A potential source of heterogeneity may be the duration of telerehabilitation as the statistical heterogeneity decreased dramatically when 2 studies with telerehabilitation shorter (3 weeks) [25] and longer (6 months) [5] than 3 months were excluded. Another potential source of heterogeneity may be differences in adherence to the telerehabilitation programs. Although only 2 of the 5 studies in the BADL analysis reported adherence rates for telerehabilitation, the substantial difference (87% [25] vs 15% [24]) suggests considerable variability in compliance across studies.

This systematic review revealed that telerehabilitation strategies, including exercises for strengthening the lower extremities, balance, coordination, stretching, and BADL training, introduction of walking aids, and safety education are commonly used for patients with hip fractures. According to clinical practice guidelines, postoperative progressive resistance exercises and balance training are strongly recommended, and BADL training and complication prevention are suggested for older patients with hip fractures [29]. However, it is essential to standardize these training programs to mitigate the risk of adverse events [30,31].

Regarding the delivery mode of telerehabilitation, most studies used software- or mobile app–based telerehabilitation with remote supervision by professionals, and only 2 studies conducted before 2015 used DVD-based or telephone-based telerehabilitation. Recently, the application of virtual reality has become an increasingly common approach in telerehabilitation, with superior long-term benefits for patients with total knee replacements compared with traditional rehabilitation [32]. Although all these remote rehabilitation approaches could facilitate adherence to postoperative management, reduce costs, and decrease the burden on therapists and caregivers, the efficacy of different delivery strategies could vary. Future studies are needed to compare the cost-effectiveness ratio between different delivery modes of telerehabilitation.

Our findings may have several clinical implications. First, substantial clinically important improvements in functional mobility, as measured using the SPPS [33] and TUG test [34,35], were observed following telerehabilitation compared to usual care at the postintervention and follow-up time points. However, the overall improvement in HHS (7.35 points) fell below the minimal clinically important improvement threshold (15.9-18 points) [36]. For BADLs, a large short-term effect size (SMD 1.65) was found, but this decreased to a small to moderate effect at the long-term follow-up (SMD 0.43). Furthermore, no included studies demonstrated a clinically relevant improvement for BADLs (21 points on the FIM [37] and 9.8 points on the BI [38]). These findings indicate that telerehabilitation may be more beneficial for functional mobility than for hip function or BADLs. Second, we noted that most included studies were conducted in countries with established research environments (Israel, Spain, Portugal, and the United States), with China being the only included country with an emerging research landscape. Therefore, the generalizability of telerehabilitation requires further investigation, specifically in terms of its feasibility and effectiveness in low- and middle-income countries. Third, the included studies exhibited variable follow-up periods (3 weeks to 9 months), which impedes the investigation of long-term effects. Future studies could adopt a more standardized follow-up schedule (eg, every 3 months after the intervention). Fourth, adherence to telerehabilitation was poorly reported in the included studies, leaving its influence on the effect size unclear. As a key proposed benefit of telerehabilitation is the facilitation of prolonged rehabilitation training—and studies have demonstrated higher compliance in telerehabilitation groups, potentially leading to greater improvement [22,26]—detailed adherence data are crucial for interpreting effectiveness. Future studies should report adherence and compliance rates in detail and develop effective supervision approaches to ensure the efficacy of telerehabilitation.

This study has several limitations. First, it only included literature published in English, which may have omitted relevant evidence from articles published in other languages, potentially introducing language bias and limiting the generalizability of the findings. Second, the small number of included studies and the substantial heterogeneity observed influenced the strength of the evidence. Third, no statistical analysis for publication bias was conducted due to the small number of included studies.

Telerehabilitation is effective for short- and long-term functional recovery in postoperative patients with hip fractures compared to conventional rehabilitation. However, the evidence is weak due to the limited number of included studies and the high heterogeneity across studies. Future high-quality studies with larger sample sizes are needed to investigate the effectiveness of telerehabilitation. These studies should report the intervention strategies, delivery modes, and adherence rates of interventions in detail; incorporate comprehensive outcome measures for hip function (eg, HHS), functional mobility (eg, SPPS and TUG test), and ability to perform BADLs (eg, BI, MBI, and FIM); and apply a relatively standard follow-up period to investigate the long-term effect of telerehabilitation.