Causal language and inferences in observational rotator cuff database studies published from 2013 to 2022

Background: The use of large multi-institutional databases in rotator cuff repair (RCR) research is expanding, but these studies are observational and cannot establish causation. This study examines the prevalence of causal language in clinical RCR database studies published from 2013 to 2022.

Methods: Administrative database and clinical registry studies on RCR published in eight orthopaedic journals from 2013 to 2022 were systematically identified and graded by two reviewers for the presence, absence, or inconsistent use of causal language in both the title/abstract and the full text. Chi-squared analyses were conducted to examine if there was an association between the gradings of articles and both the journal and year of publication.

Results: Of 44 eligible articles, 14 were graded as consistently causal, 16 as inconsistent, and 14 as consistently non-causal. Chi-squared analyses revealed no statistically significant associations between the journal or year of publication and the title and abstract grading (p = 0.626, p = 0.277) or the full text grading (p = 0.374, p = 0.822).

Conclusion: Causal language was present in over two-thirds of observational RCR database studies published from 2013 to 2022. Authors should refrain from using causal language in database studies to prevent misleading readers and misinterpretation of findings.