Selecting the best research to read is one aspect of evidence-based practice. Randomised controlled trials estimate the difference in outcomes between an intervention and a comparator for a sample of participants. But are all trials free from bias? The short answer is “No”.
A recent Research Note published in the Journal of Physiotherapy discusses how flaws in the design, conduct and reporting of trials can introduce bias that distorts the size of the effect estimates. The Research Note offers guidance for clinicians and educators to choose the best trials to read to inform their practice and teaching. Some resources for researchers to account for bias in systematic reviews and minimise bias in designing and reporting trials are also provided.
There are many sources of bias in trials. The Research Note focuses on how trial participants are allocated to groups (randomisation and concealment), blinding of key people involved in the trial (participants, therapists and assessors) and the completeness of follow-up. Tools have been developed for evaluating the risk of bias in trials, including the Physiotherapy Evidence Database (PEDro) scale and Cochrane risk of bias tool (version 1 and 2). The Research Note summarises the content, validity and reliability of each tool. Guidance is also given on how to interpret the summary score of the PEDro scale.
Resources to minimise risk of bias in trials and systematic reviews are available for researchers. Consensus statements and checklists have been developed to assist researchers to plan (eg, the Standard Protocol Items Recommendations for Interventional Trials (SPIRIT) statement) and report the results of trials (eg, the Consolidated Standards of Reporting Trials (CONSORT) statement), and to judge the certainty of evidence in systematic reviews (eg, the Grading of Recommendations, Assessment, Development and Evaluations (GRADE)).
While the focus of the Research Note is the risk of bias in randomised controlled trials evaluating the effects of interventions, there are research designs that answer other important questions that are also at risk of bias. Some common tools for each study type are suggested.
The ability to quickly identify trials that are relatively free from bias from ones that are not is an important skill for physiotherapists to master. One strategy is to memorise important sources of bias, and another is to use evidence resources that include some pre-appraisal of trials.
Moseley AM, Pinheiro MB. Research Note: Evaluating risk of bias in randomised controlled trials. J Physiother; epub ahead of print 10 March 2022