gms | German Medical Science

Background/research question: Implications for research (IfR) sections in systematic reviews (SRs) aim to highlight research gaps, guide future research, and reduce uninformative research. Description of IfR is required in Cochrane Reviews and the Cochrane Handbook (chapter 15.6.3) suggests considering PICO(S) and GRADE factors in the IfR section. Additionally, the PRISMA 2020 reporting guideline mandates reporting of IfR (Item 23d, [1]).

We aimed i) to assess whether SRs included an IfR section and, if yes, ii) to examine whether IfR considered PICO(S) elements, GRADE factors [2], or other factors.

Methods: We conducted a meta-research study based on an existing sample of SRs on COVID-19 treatment identified in the L·OVE COVID-19 database by May 2021 (PROSPERO: CRD42021240423 [3]). SRs had to address treatment measures for COVID-19 and report at least one health-related outcome. After pilot-testing (5 SRs), predefined IfR variables, i.e. consideration of PICO(S), GRADE factors, other factors, were extracted by three reviewers. Additionally, we used weekly consensus discussions to ensure the high quality of the extraction process.

Results: We included 326 SRs for this analysis. Our analysis focused on the 284 SRs (87.1%) with an IfR section from which 201 (70.8%) referred to PRISMA and 66 (23.2%) assessed the certainty of evidence using GRADE.

Commonly addressed PICO(S) factors (n=284) were 'intervention' (242, 85.2%), 'study design' (207, 72.9%), 'population' (195, 68.7%), and 'outcome' (127, 44.7%).

GRADE factors were infrequently used to derive IfR recommendations (n=284): 'risk of bias' (14, 4.9%), 'imprecision' (8, 2.8%), 'inconsistency' (7, 2.5%), 'publication bias' (3, 1.1%) and 'indirectness' (1, 0.4%).

Seventeen from 284 SRs (6.0%) recommended 'better reporting' of future studies. Other IfR recommendations were 'efficacy and safety' (84, 29.6%), 'adequate sample size / power' (49, 17.3%), and 'standardization of procedures in clinical trials' (17, 6.0%).

Conclusion: Almost nine out of ten SRs included an IfR section and seventy percent of those used PRISMA. However, many IfR sections were insufficient, especially for GRADE factors.

To increase value of future research and to guide clinically relevant future research, IfR recommendations in SRs should be reported in a more explicit, structured format, and consider PICO(S) elements and GRADE factors.

Competing interests: Es liegen keine Interessenkonflikte vor.