The authors concluded that exercise programmes based on Tai Chi had better results than mixed programmes, but without clear differences. Several potential limitations in the search, review process and statistical analysis of this review mean that the reliability of the authors' conclusion is uncertain.

To evaluate the effectiveness of physical exercise programmes to reduce pain in patients with hip and knee osteoarthritis.

MEDLINE, Science Citation Index, Scopus and SPORTDiscus were searched between February and March 2010. Search terms were reported.

Randomised controlled trials (RCTs) and quasi-experimental studies of physical exercise programmes of at least four weeks in adults diagnosed with lower limb osteoarthritis (American College of Rheumatology criteria) were eligible for inclusion. Studies had to measure pain assessed by the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC).

Included were studies of programmes that contained strength exercises, Tai Chi, aerobic, hydrotherapy and mixed exercise (aerobic resistance, strength and flexibility). Control groups received education, attention control, placebo or no intervention. Intervention duration ranged from six to 72 weeks, frequency ranged from one to seven sessions per week and session duration (where reported) was between 20 and 90 minutes. Interventions were delivered largely in clinics or in patients' homes. There were no details of patient characteristics.

The authors did not state how many reviewers were involved in study selection.

Study quality was assessed using the Delphi List (Verhagen et al. 1998) of randomisation, allocation concealment, specification of inclusion criteria, similarity of baseline characteristics, blinding, provision of point estimate and variability and use of intention-to-treat analysis. Studies were rated as yes, no and don't know.

The authors did not state how many reviewers carried out the quality assessment.

Data were extracted to enable calculation of mean change (pre-post) scores for the interventions of interest, along with 95% confidence intervals (CI) where standard deviations were not reported. Effect sizes for each type of exercise programme (Tai Chi, aerobic, hydrotherapy, and mixed) were calculated and interpreted as small (up to 0.2), medium (0.2 to 0.5) or large (>0.5).

The authors did not state how many reviewers carried out the data extraction.

A narrative synthesis of individual effect sizes was presented. Effect sizes, grouped by type of exercise programme, were pooled in meta-analyses. Subgroup analyses were reported in terms of specific characteristics of the exercise (such as subtypes and frequency/duration of sessions).

Thirty-three studies (2,679 participants) were included in the review. Study quality was mixed. It seemed that not all the included studies were part of the assessment. Twenty-seven studies reported randomisation. All studies specified inclusion criteria. Most studies were similar at baseline. Sixteen studies used intention-to-treat analysis. Blinding of outcome assessment was present in 19 studies. Allocation concealment and participant and provider blinding were not met in most of the studies.

A large pooled effect size was noted for Tai Chi programmes (10 studies with effect sizes that ranged from 0.19 to 1.67) Medium pooled effect sizes were noted for strength programmes (10 studies with effect sizes that ranged from 0.01 to 1.12; three sessions per week achieved effective improvement), aerobic exercise (two studies with effect sizes 0.03 and 1.06) and hydrotherapy (five studies with effect sizes that ranged from 0.19 to 0.72). Mixed exercise programmes (six studies) showed a small pooled effect. The authors reported a statistically significant difference in effect sizes between Tai Chi and mixed exercise programmes (p<0.05).

Subgroup results were reported in the paper.

Exercise programmes based on Tai Chi had better results than mixed programmes, but without clear differences.

The review question was clear and inclusion criteria were potentially reproducible. The search strategy appeared limited in scope with regard to data sources and search dates. There was no apparent search for unpublished material, so studies may have been missed and publication bias could not be ruled out. The authors did not report their review processes, so errors and bias may have been present. Study quality was assessed using established criteria, although blinding was likely to be inappropriate to the interventions studied and the included study designs were unclear (the authors concluded that few randomised studies existed, yet most of the studies met this quality criterion). Clinical variation was evident, but there was no assessment of statistical heterogeneity and no mention of the type of meta-analysis employed. Confidence intervals were not presented for the pooled results and it was not clear how many studies were included in each pooled analysis.

This was a poorly-reported review and the reliability of the authors' conclusion is uncertain.

Practice: The authors did not state any implications for practice.

Research: The authors stated that future research should explore the effectiveness of combined exercise programmes in patients with osteoarthritis.

Ministerio de Educacion, Spain; FEDER (Fonds Européen de Développement Régional); Autonomous Government of Extremadura, Spain.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.