A call for improving the internal validity and the reporting of manual therapy trials self-labelled as pragmatic : A methodological review

Objectives: This study provides further data from a review assessing methodological characteristics of pragmatic randomised controlled trials (pRCTs) published in manual therapy (MT). In this second part, we aim to assess the report, the risk of bias (RoB), and the relationship between these items and the pragmatism scores of the self-labelled pRCTs in the MT field. Study design and setting: We searched MEDLINE and the Cochrane Central Register of Controlled Trials for pRCTs in the MT field from inception to January 2024. Two independent reviewers screened the trials using several CONSORT extensions and assessed them using the Cochrane Risk of Bias tool. We performed a descriptive analysis using frequencies and percentages and a relation analysis between the trials' pragmatism, their reporting, and their RoB. Results: We included 39 self-labelled MT pRCTs. Compliance with CONSORT items was higher than 70 % in one-third of the included trials (13/39) but varied across items. Performance and detection bias were the main threats to internal validity (we rated 90 %, 35/39, and 77 %; 30/39 of trials at high risk of bias, respectively). Selective reporting bias was unclear in almost half of the sample (46 %; 18/39). No relation was found between the highly pragmatic attitude and good reporting except for CONSORT item 25 (Sources of funding and other support) (p = 0.006). No relation was found between the RoB and the pragmatic attitude of the studies. The percentage of compliance with CONSORT items was higher in the trials with low RoB. Conclusion: Pragmatic trials in MT have significant methodological limitations, and their reporting is suboptimal. Nonetheless, trials with less risk of bias had higher compliance with CONSORT items.