Predictors of Higher Quality of Systematic Reviews Addressing Nutrition and Cancer Prevention

Abstract

Systematic reviews/meta-analyses (SR/MAs) are considered a reliable source of information in healthcare. We aimed to explore the association of several characteristics of SR/MAs addressing nutrition in cancer prevention and their quality/risk of bias (using assessments from AMSTAR-2 and ROBIS tools). The analysis included 101 SR/MAs identified in a systematic survey. Associations of each specified characteristic (e.g., information about the protocol, publication year, reported use of GRADE, or other methods for assessing overall certainty of evidence) with the number of AMSTAR-2 not met (‘No’ responses) and the number of ROBIS items met (‘Probably Yes’ or “Yes’ responses) were examined. Poisson regression was used to identify predictors of the number of ‘No’ answers (indicating lower quality) for all AMSTAR-2 items and the number of ‘Yes’ or ‘Probably Yes’ answers (indicating higher quality/lower concern for bias) for all ROBIS items. Logistic regression was used to identify variables associated with at least one domain assessed as ‘low concern for bias’ in the ROBIS tool. In multivariable analysis, SR/MAs not reporting use of any quality/risk of bias assessment instrument for primary studies were associated with a higher number of ‘No’ answers for all AMSTAR-2 items (incidence rate ratio (IRR) 1.26, 95% confidence interval (CI) 1.09–1.45), and a lower number of ‘Yes’ or ‘Probably Yes’ answers for all ROBIS items (IRR 0.76, 95% CI 0.66–0.87). Providing information about the protocol and search for unpublished studies was associated with a lower number of ‘No’ answers (IRR 0.73, 95% CI 0.56–0.97 and IRR 0.75, 95% CI 0.59–0.95, respectively) and a higher number of ‘Yes’ or ‘Probably Yes’ answers (IRR 1.43, 95% CI 1.17–1.74 and IRR 1.28, 95% CI 1.07–1.52, respectively). Not using at least one quality/risk of bias assessment tool for primary studies within an SR/MA was associated with lower odds that a study would be assessed as ‘low concern for bias’ in at least one ROBIS domain (odds ratio 0.061, 95% CI 0.007–0.527). Adherence to methodological standards in the development of SR/MAs was associated with a higher overall quality of SR/MAs addressing nutrition for cancer prevention.

Fast and frugal decision tree for the rapid critical appraisal of systematic reviews

“…Previous studies showed that involving librarians or information specialists may improve the reproducibility of literature searches (Koffel, 2015; Rethlefsen et al, 2015) and may be associated with a lower risk of bias of the SRs as shown in a small sample (Aamodt et al, 2019). Another study showed that including information about the search for unpublished studies is related to the methodological quality of SRs (Storman et al, 2022). The fact that the search strategy item is also included in the Rapid Appraisal FFT additionally highlights the role of a high-quality literature search on the overall quality appraisal of SRs.…”

Reporting quality of abstracts of systematic reviews/meta-analyses: An appraisal of Arab Journal of Urology across 12 years: the PRISMA-Abstracts checklist

“…Urology is a rapidly expanding field, with many new management approaches and innovative technologies. To appraise such advances, compare the effectiveness of procedures and their side effects, or alternatively, to evaluate etiology, estimate prevalence, assess diagnosis or prognosis, systematic reviews/meta-analyses (SRs/MAs) are considered as a reliable source of information and evidence [ 1 ]. SRs/MAs assume the top position pertaining to the quality and dependability of the evidence generated, and are a reliable information source [ 2 ].…”