Methodologic Quality and Statistical Reporting of Physical Therapy Randomized Controlled Trials Relevant to Musculoskeletal Conditions

Publication date: January 2018
Source:Archives of Physical Medicine and Rehabilitation, Volume 99, Issue 1
Author(s): Gabrielle Zoldan Gonzalez, Anne M. Moseley, Christopher G. Maher, Dafne Port Nascimento, Lucíola da Cunha Menezes Costa, Leonardo Oliveira Costa
ObjectivesTo investigate the methodologic quality and statistical reporting of reports of trials indexed on the Physiotherapy Evidence Database (PEDro) classified in the musculoskeletal subdiscipline, and to analyze the characteristics of the trials that can predict trial report quality.DesignCross-sectional study based on a collection of randomized controlled trials. We randomly selected 19% of trials coded as musculoskeletal from PEDro. Methodologic quality was assessed using the PEDro scale. We assessed aspects of the trial using 9 items from the Consolidated Standards of Reporting Trials (CONSORT) statement. We performed multivariate linear regression analysis models to predict the total PEDro score.SettingNot applicable.ParticipantsNot applicable.InterventionsNot applicable.Main Outcome MeasuresNot applicable.ResultsA total of 1404 articles were included in the analysis. The mean total PEDro scale score was 5.27±1.63 points, which reflects low methodologic quality. There was a slight improvement in the quality of articles over time. The characteristics that predicted the total PEDro scale score were endorsement of the CONSORT statement, sample size calculation, lower number of primary outcomes, evaluation of electrotherapy as intervention, if the trial reported the research design in the title, reporting of participant flow diagram, years since publication (most recent trials), and trials published in English.ConclusionsThe quality of the trials in musculoskeletal physical therapy is suboptimal. The use of reporting checklists (eg, CONSORT statement) should be mandatory in all journals. Journal reviewers and journal editors should also use the CONSORT statement during the review process.



http://ift.tt/2pNXe7l