This review has been carried out following the PRISMA 2009 guidelines [34]. The review protocol was registered with an international registration database [PROSPERO, Registration Number: CRD42016047846].

As with most of the recent systematic reviews on the topic [2,26,27,33,35], the following literature searches were performed to identify all possible studies that could help answer the research question. MEDLINE, Physiotherapy Evidence Database (PEDro), Scopus, PsycINFO, Web of Knowledge, CINAHL, SPORTDiscus, Directory of Open Access Journals (DOAJ), Cochrane, Embase, Academic Search Complete, Fuente Académica, and Consejo Superior de Investigaciones Científicas (CSIC) were searched. In addition, the search was performed in a relevant bibliographic database from the University of Málaga.

The initial search was carried out in June 2016, and was completed with a new search to update the review in October 2016. The following combinations of keywords were used: telerehabilitation, telerehabilitation physiotherapy, post-surgery telerehabilitation, musculoskeletal disorders, systematic review, telemedicine (mobile health or health, mobile or mHealth or telehealth or eHealth), telemedicine physiotherapy program. The limits of searches were studies published between 2000 and 2016.

The main steps related to the search phase are reported in Figure 1 using the PRISMA flow diagram [36]. After the application of the selected keywords, the entire set of records was analyzed to identify eventual duplication of articles retrieved from different sources; the remaining articles were then assessed in full text for eligibility so as to identify all those matching the inclusion criteria (see Multimedia Appendix 1).

Two authors (JMPB and RMV) independently screened the titles and abstracts of all records retrieved using the database search strategy. The full text was obtained if further information was required to determine eligibility, or if uncertainty prevailed between authors. For trials published in a language other than English or Spanish, a translated version of the abstract was sourced to determine eligibility.

Disagreements between authors were initially resolved via discussion, and then by consultation with a third reviewer (FJBL).

Eligibility criteria were based on the PICOS framework [37], as follows:

All types of study designs were considered: randomized clinical trials (RCT), clinical trials (CT), case reports, controlled clinical trial, and pilot study. Articles that were limited to describing the feasibility and fit-out of telerehabilitation interventions were excluded.

For all eligible trials, data extraction was independently completed by 2 authors (JMPB and MJEP), and was cross-checked for consistency by a third author (RMV). The primary authors of eligible trials were contacted when information was considered to be missing for either the quality assessment or data extraction process.

An important step in conducting a systematic review is to assess the methodological quality of each included trial. In addition, reporting methodological quality provides clinicians with information about whether the results of clinical trials should influence their clinical practice. A valid way of assessing the methodological quality of clinical trials is therefore essential [38].

Two independent reviewers [JMPB and RMV] completed the checklist based on the PEDro scale. The methodological quality and risk of bias were evaluated using the PEDro scale [38,39] based on the Delphi list [40]. It is considered a useful tool for carrying out the assessment methodology in scientific research.

The PEDro scale scores 10 items: random allocation, concealed allocation, similarity at baseline, subject blinding, therapist blinding, assessor blinding, >85% follow-up for at least one key outcome, intention-to-treat analysis, between-group statistical comparison for at least one key outcome, and point and variability measures for at least one key outcome. Items are scored as either present (1) or absent (0) and a score out of 10 is obtained by summation [38]. The scale includes an additional item (eligibility criteria) to evaluate the external validity, but the score is not counted.

According to Moseley et al, studies with a PEDro score ≥5 will be considered at low risk of bias and high methodological quality [41].

A study with a PEDro score of ≥6 is considered to have level 1 evidence (6-8=good, 9-10=excellent) and a study with a score of ≤5 is considered to have level 2 evidence (4-5=acceptable, <4=poor) [42].

Levels of evidence help us target the search at the type of evidence that is most likely to provide a reliable answer. They have been designed so that they can be used as a shortcut for busy clinicians, researchers, or patients to find the likely best evidence [43]. Grades of recommendation describe the strength and therefore value of the evidence relative to how rigorous the study was (see Tables 1 and 2) [44].

This review confirms the strong evidence in favor of telerehabilitation among patients undergoing total knee and hip arthroplasty and the limited evidence in the upper limb (moderate and weak evidence).

To the best of our knowledge, this is the first review focused on telerehabilitation research after surgical procedures on orthopedic conditions. This systematic review applied a qualitative evaluation to provide a wider picture of currently available evidence.

First, we will discuss the contributions of the first systematic reviews on the topic. Second, we discuss the generalizations, previous results, and future recommendations. Third, we discuss if results are extrapolated to the upper limb (the results of this review show poor-quality methodology and moderate and weak evidence). Finally, we discuss about the inherent difficulties in conducting telerehabilitation research and future research recommendations.

Regarding the first aspect noted in the discussion, the first systematic reviews, contributions concluded that better-quality studies are needed as well as studies on the use of telerehabilitation in routine care. Telerehabilitation research is generally not very good and there are many reviews that criticize this [2,26,27,33,35]. In our review, 60% of the included studies are of poor methodological quality with weak evidence.

Regarding the second aspect noted in the discussion, the most recent systematic reviews provide statements on the effects of telerehabilitation interventions. Two recent studies provide statements such as, “there is a strong positive effect for patients following orthopedic surgery” [1] and “there is unequivocal evidence that the management of musculoskeletal conditions via real-time telerehabilitation is effective in improving physical function, disability, and pain” [33]. We agree with these statements, but only and exclusively for some pathologies. Our systematic review shows that there is still insufficient evidence on upper limb surgeries’ telerehabilitation interventions.

Regarding the third aspect noted in the discussion: does this statement transfer to upper limb such as shoulder arthroscopy, carpal tunnel release surgery, hand surgery, or shoulder arthroplasty? Could this be extrapolated to the rehabilitation process in fractures or surgery interventions in upper and lower limbs? These unresolved clinical questions reaffirm the need to identify the available evidence in post-surgical rehabilitation with telerehabilitation interventions.

For this systematic review, we seek to find evidence of post-surgical telerehabilitation programs, with special emphasis on programs that can be integrated into clinical practice.

In our review, telerehabilitation research in the upper limb (shoulder joint replacement, proximal humerus fractures, carpal tunnel release, and cuff rotator tears) presents moderate and weak levels of evidence. Notable is the judgment that none of the telerehabilitation studies in the upper limb included in this review present a high level of evidence and recommendation. There is still a very small database for telerehabilitation studies after a musculoskeletal surgery that provides useful data on clinical outcomes, especially in conditions other than the replacement of joints in the lower limb. Therefore, conclusive evidence on the efficacy of telerehabilitation for treatment after an orthopedic surgery, regardless of pathology, was not obtained.

Research background has been used to discuss the strengths of telerehabilitation and the opportunities for future interventions and policies. Regarding the final aspect noted in the discussion, what are the inherent difficulties in conducting telerehabilitation research?

During the search, we observed a number of studies that provide descriptions of telerehabilitation interventions of low methodological quality. No validated clinical outcomes, too small a sample size, and a lack of comparison group are frequently found. Moreover, differences in telerehabilitation interventions, treatment period, and follow-up, create doubts in identifying whether the telerehabilitation gives comparable or better results.

A frequent problem in studies of telerehabilitation is the lack of blinding of therapists and patients. There is evidence that in clinical trials where allocation is not concealed and assessors, therapists, and participants are not blinded, a larger effect of intervention is reported than in higher quality trials with adequate blinding procedures [65].

It may be that there are good-quality studies the publication of which has been delayed; however, our findings are aligned and consistent with the most recent revisions regarding the need for future research needs to have stronger and more solid studies.

One of the biases identified is that telerehabilitation groups have more frequent contact with health professionals and with the intervention (especially in videoconferencing and phone contact), so they are likely to receive additional services. This creates biases whether the positive results are related to a more elaborate program than really with the interventions method.

How could this be addressed in future research? As blinding of patients and therapists is not possible in telerehabilitation interventions, several methodological aspects are fundamental for future research.

Telerehabilitation interventions should be conceptualized, coded, classified, and grouped in a similar way to physiotherapy technique codes, enabling identification in detail when the effect is due to the type of intervention. Comparison group must be the actual best evidence treatment for the same condition that allows identification of whether telerehabilitation offers better or comparable outcomes. Telerehabilitation frequency must be the same as the control group to avoid biases related to a more elaborate program. Greater homogeneous is needed especially in terms of type, duration, and intervention follow-up for each specific pathologies. Studies that show negative results should be published, avoiding publication biases. Large sample size and improvement in study quality (allocated and evaluator blinding) must be addressed. Orthopedic conditions and musculoskeletal injuries different to replacement joints in lower limbs need quality research.

High-methodological-quality studies should be conducted to confirm that telerehabilitation shows clinically relevant outcomes after surgery in orthopedic and musculoskeletal injuries, especially in upper limbs. Telerehabilitation appears to be an effective alternative to face-to-face service delivery after hospital discharge of patients following total knee arthroplasty and hip replacement. Clinical outcomes are comparable and not inferior. Despite some limitations, there seem to be clear benefits from physiotherapy at a distance regardless with the telerehabilitation technique it offers (videoconferencing, phone intervention, asynchronous video exercise programs, or gaming). Future challenges include identifying whether positive results are due to the type of intervention or the increased frequency and intensity that telerehabilitation allow.

Future research recommendations for telerehabilitation should include high-quality studies with clear conclusions and statements that could improve health interventions and health policies.