Language of publication has a small influence on the quality of reports of controlled trials of physiotherapy interventions

Objectives: To investigate whether the methodological quality is influenced by language of publication in reports of randomized controlled trials and controlled clinical trials of physiotherapy interventions. Study Design and Setting: Bibliometric and methodological quality data from all reports of trials indexed on the Physiotherapy Evidence Database (PEDro) up to February 2011 were extracted. Descriptive statistics on the total PEDro score and the 11 individual PEDro items were calculated for each language of publication and for all non-English-language reports combined. Regression models were calculated to predict the total PEDro score and the presence of each of the 11 items of the PEDro scale using the language of publication as an independent variable. Results: A total of 13,392 reports of trials were used for this study, 12,532 trials published in English and 860 published in other languages. Overall methodological quality was better for English reports than reports written in other languages (β = 0.15, 95% confidence interval = 0.04, 0.25). Specifically, reporting was better for items relating to random allocation, concealed allocation, and blinding of assessors, worse for more than 85% follow-up and intention-to-treat analysis, and no different for eligibility criteria and source specified, baseline comparability, blinding of subjects and therapists, reporting of between-group statistical comparisons, and reporting of point measures and measures of variability. Conclusion: Language of publication is associated with the methodological quality of reports of physiotherapy trials. Although English reports are more likely to have better methodological quality than reports written in other languages, the magnitude of this influence is small.