A systematic review in physical therapy synthesizes evidence on rehabilitation interventions, exercise programs, manual therapy techniques, and functional outcomes to guide physiotherapy practice. The field's reliance on complex, multi-component interventions where therapist skill, patient adherence, and dosage parameters all influence outcomes creates methodological challenges that require discipline-specific approaches to evidence synthesis.
Cochrane Rehabilitation, established as a dedicated Cochrane field in 2016, recognized that rehabilitation interventions differ fundamentally from pharmacological treatments. A drug trial can standardize the intervention perfectly: 50mg of atenolol is 50mg of atenolol regardless of who prescribes it. A physical therapy intervention involves therapist expertise, patient engagement, progression decisions, and environmental factors that vary between every delivery.
The TIDieR checklist (Template for Intervention Description and Replication, Hoffmann et al., 2014) was developed partly in response to rehabilitation research's chronic under-reporting of intervention details. Physical therapy systematic reviewers frequently find that included studies provide insufficient information about exercise parameters, progression criteria, supervision levels, and adherence monitoring.
This intervention complexity means that read about heterogeneity in rehabilitation meta-analyses often reflects genuine clinical variation rather than methodological flaws. Two studies evaluating "exercise for knee osteoarthritis" might test fundamentally different interventions: supervised high-intensity resistance training three times weekly versus a home-based stretching program with monthly phone check-ins.
The World Physiotherapy (formerly World Confederation for Physical Therapy) and national physiotherapy associations increasingly require evidence-based practice competencies, driving demand for high-quality systematic reviews that can actually inform clinical decisions despite this intervention heterogeneity.
Physical therapy evidence is distributed across medical, rehabilitation, sport science, and allied health databases. Missing any major source risks incomplete evidence retrieval.
PEDro (Physiotherapy Evidence Database) is the discipline's flagship resource. Maintained by the University of Sydney, PEDro indexes over 55,000 randomized trials, systematic reviews, and clinical practice guidelines in physiotherapy. Importantly, PEDro independently rates the methodological quality of every indexed trial using the PEDro scale, providing pre-assessed quality ratings that save reviewers time during screening.
CINAHL (Cumulative Index to Nursing and Allied Health Literature) captures rehabilitation nursing, occupational therapy, and physiotherapy research not indexed in PubMed. Its controlled vocabulary includes physiotherapy-specific terms absent from MeSH.
PubMed/MEDLINE and Embase cover the biomedical rehabilitation literature. SPORTDiscus captures sport science and exercise physiology research relevant to performance rehabilitation and injury prevention reviews.
AMED (Allied and Complementary Medicine Database) indexes complementary therapy research that overlaps with physical therapy (acupuncture, manual therapy, yoga interventions).
The Cochrane Rehabilitation field register supplements CENTRAL with rehabilitation-specific trial identification. Build your multi-database read about search strategy using our practical search strategy builder, which supports translation across database-specific vocabularies.
Physical therapy outcomes span impairment-level measures (range of motion, muscle strength), activity-level measures (functional tasks, walking speed), and participation-level measures (return to work, quality of life), reflecting the International Classification of Functioning, Disability and Health (ICF) framework endorsed by the World Health Organization.
Common outcome measures and their meta-analytic considerations include:
Pain scales: Visual Analogue Scale (VAS) and Numeric Pain Rating Scale (NPRS) are the most frequently reported. Although both measure pain intensity, they are not interchangeable in meta-analysis. The minimal clinically important difference (MCID) for VAS in musculoskeletal conditions is approximately 15-20mm on a 100mm scale, and meta-analytic results should be interpreted against this threshold.
Functional outcome tools: The Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire, Oswestry Disability Index (ODI) for low back pain, Western Ontario and McMaster Universities (WOMAC) index for osteoarthritis, and Functional Independence Measure (FIM) for neurological rehabilitation each have established psychometric properties and MCIDs.
Performance-based measures: Timed Up and Go (TUG), 6-Minute Walk Test (6MWT), grip strength dynamometry, and gait speed measured via instrumented walkways provide objective functional data. These measures typically show smaller effect sizes than self-report instruments, a phenomenon systematic reviewers must acknowledge.
When studies use different measures for the same construct (e.g., multiple knee function questionnaires), the standardized mean difference (Hedges' g) allows pooling. Calculate this using our our interactive effect size calculator and visualize results with our convenient forest plot generator.
The FITT principle (Frequency, Intensity, Time, Type) defines exercise dosage parameters, and systematic reviews must extract and report these details to allow clinicians to replicate effective programs. However, Holden et al. (2021) found that only 34% of exercise intervention randomized controlled trials report all four FITT components.
This reporting gap creates a fundamental problem for systematic reviewers. When a meta-analysis finds that "exercise reduces pain in knee osteoarthritis" with a pooled effect of SMD = -0.5, clinicians need to know: what type of exercise? How often? At what intensity? For how long? Without this information, the systematic review fails its primary purpose of informing practice.
Subgroup analysis by exercise parameters is the recommended approach when sufficient studies exist. Separate analyses for strength training versus aerobic exercise versus flexibility programs, supervised versus unsupervised delivery, and different frequency/duration protocols provide clinically actionable findings.
The Consensus on Exercise Reporting Template (CERT) (Slade et al., 2016) provides a standardized checklist for exercise intervention reporting that systematic reviewers can use during understanding data extraction to systematically capture dosage information from included studies.
Physical therapy randomized controlled trials face inherent risk of bias challenges related to blinding and intervention standardization.
Patient blinding is frequently impossible. A patient receiving manual therapy knows they are receiving hands-on treatment, not a sham. Some trials use sham interventions (detuned ultrasound, sub-therapeutic laser), but these are not always credible. The Cochrane RoB 2 tool handles this through the "deviations from intended interventions" domain, where reviewers judge whether knowledge of allocation affected behavior.
Therapist blinding is almost never possible. The treating physiotherapist knows which intervention they are delivering. The key question for bias assessment is whether this knowledge led to co-interventions or differential encouragement.
Assessor blinding is achievable and should be expected. Studies using blinded outcome assessors for performance-based measures receive better ratings on this domain.
The PEDro scale provides an alternative quality assessment with 11 items specifically designed for physiotherapy trials. Many rehabilitation journals prefer PEDro scores alongside Cochrane RoB 2. Our risk of bias visualizer supports RoB 2 assessment, while PEDro scores are available directly from the PEDro database.
Need expert support for your physical therapy systematic review? Research Gold provides professional evidence synthesis services with experience in rehabilitation outcomes, exercise intervention synthesis, and musculoskeletal meta-analysis. get a free heterogeneity analysis estimate to discuss your project.
Physical therapy meta-analyses require careful interpretation that goes beyond statistical significance to address clinical significance.
The GRADE framework assessment for rehabilitation evidence frequently encounters rating down for risk of bias (blinding issues) and inconsistency (intervention heterogeneity). This means that even well-conducted rehabilitation systematic reviews often produce moderate or low certainty evidence. Apply GRADE assessments using our free GRADE evidence tool.
Minimal clinically important difference (MCID) contextualization should be standard. Rather than simply reporting a statistically significant pooled effect, compare the mean difference against the established MCID for the outcome measure. A statistically significant 5-point improvement on the ODI is clinically meaningless when the MCID is 10 points.
forest plots explained should be organized by clinically meaningful subgroups (acute vs. chronic conditions, supervised vs. unsupervised programs, high vs. low dose) rather than presenting a single overall effect that obscures clinically important variation.
Prediction intervals are particularly valuable in rehabilitation meta-analyses because they show the range of effects expected in future settings, acknowledging that intervention delivery will inevitably vary from the studies in the review.
Leading journals for physical therapy systematic reviews include:
Journal of Physiotherapy (formerly Australian Journal of Physiotherapy), which publishes systematic reviews with meta-analysis and requires PRISMA compliance plus PEDro quality assessment.
Physical Therapy (APTA journal), which prioritizes reviews with clear clinical implications and includes the our prisma flow chart generator tool as a mandatory element.
British Journal of Sports Medicine (BJSM) publishes rehabilitation and sport injury systematic reviews. Archives of Physical Medicine and Rehabilitation covers neurological and musculoskeletal rehabilitation reviews.
Cochrane Database of Systematic Reviews through the Cochrane Rehabilitation field publishes the most rigorous rehabilitation evidence syntheses.
Before submission, verify your manuscript against the our guide to prisma 2020 checklist and structure your research question with the our free pico framework builder.
Physical therapists and rehabilitation researchers possess strong clinical reasoning skills but may lack formal training in meta-analytic statistics or systematic review methodology. Professional support is most valuable when your review requires complex subgroup analyses by exercise parameters, involves mixed outcome measures requiring careful standardization, or faces tight publication deadlines.
Research Gold has supported rehabilitation systematic reviews published in the Journal of Physiotherapy, Physical Therapy, and BJSM. Our methodologists understand the challenges of synthesizing complex interventions and can help navigate the exercise dosage reporting gap. request a free forest plot generation quote or explore our tiered pricing structure for transparent cost information.
Rehabilitation journals accept both formats. Understand when to write a case report versus a systematic review in rehabilitation based on your evidence and goals.
Rehabilitation reviews must document screening agreement. Our inter-rater reliability guide covers kappa calculation, interpretation thresholds, and what to do when agreement is low.
Physical therapy meta-analyses often include studies with varied intervention protocols. Leave-one-out analysis helps identify whether any single study disproportionately influences your pooled result.