Analysis of Conclusiveness over Time of Cochrane Systematic Reviews in Physiotherapy: A Methodological Study 0413

Abstract

Objectives

To determine the proportion of conclusive Cochrane reviews relevant to the field of physiotherapy and to investigate whether conclusiveness has been increasing over time.

Design

Methodological study.

Setting

PEDro database.

Participants

Two hundred Cochrane systematic reviews were randomly selected, with 50 from each of the 4 publication periods: 2000-2005, 2006-2010, 2011-2015, and 2016-2020.

Interventions

Not applicable.

Main Outcome Measures

Two independent reviewers extracted: year of publication, country, Cochrane Review Group, participants, intervention, comparators, primary outcomes, number of randomized controlled trials, total cumulative number of patients enrolled, and the need for further studies. Certainty of evidence was assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE), ranging from very low to high. GRADE data from primary outcomes were used if available, otherwise reviewers applied the GRADE approach to remaining outcomes after a set of prespecified criteria. To analyze change over time, chi-square tests of independence were used for GRADE proportions, criteria for downgrades, study conclusiveness, and reported need for further studies.

Results

Outcomes rated very low significantly increased over the 23 period from 14% in 2000-2005 to 34% in 2016-2020 (P<.001). Low-rated outcomes trended downward over the same period and outcomes rated moderate or high remained consistent. Proportion of high-rated outcomes never exceeded 5% per time period and composed 3% of the entire sample. Risk of bias was the most prevalent reason for downgrade, composing 45% of all outcome downgrades. Inconsistency demonstrated a time-wise increase from <1% in 2000-2005 to 9% in 2016-2020 (P<.001). Study conclusiveness remained unchanged (P>.05) and consisted of 4% of all studies in the sample. Proportion of conclusive studies was highest in 2016-2020 with 4 out of 50 (8%). Studies reporting no further research needs were also unchanged (P>.05) and did not exceed 14% of studies per time period, or 9% overall.

Conclusions

The proportion of Cochrane reviews deemed conclusive has not changed over time and remains a very small proportion of those published. Outcomes with high certainty of evidence are also few and remain unchanged. Further research is necessary in the physiotherapy field to provide higher quality evidence and improve conclusiveness in systematic reviews.

Disclosures

none.