PhD Oral Exam - Ariany Marques Vieira, Health and Exercise Science

When studying for a doctoral degree (PhD), candidates submit a thesis that provides a critical review of the current state of knowledge of the thesis subject as well as the student’s own contributions to the subject. The distinguishing criterion of doctoral graduate research is a significant and original contribution to knowledge.

Once accepted, the candidate presents the thesis orally. This oral exam is open to the public.

Abstract

This thesis aims to identify the main methodological questions around Rapid Reviews (RRs) methods and undertake methodological studies to explore the impact of time-saving methods on review results. Study 1: An eDelphi study and consensus meeting were conducted, involving experts and evidence synthesis knowledge users. From an initial list, participants rated (low, medium or high importance) and ranked each item’s importance to improve the time-efficiency of RRs. Items rated as high by ≥75% of participants progressed to the next round, and the final list was concluded during the consensus meeting. Study 2: This methodological study used Cochrane cardiac rehabilitation reviews to assess how database selection impacts study inclusion and outcomes. By examining where each included study was indexed and re-running meta-analyses, we evaluated whether treatment effects varied based on different database combinations. Study 3: This methodological study compared single-review and peer-review (two independent reviewers with a third for discrepancies) approaches for screening titles and abstracts. We assessed the percentage of missed studies, sensitivity, specificity, time, and costs for each method. Results: Study 1 identified seven highly important methodological questions. Three items on search strategy, two on study selection, one on quality/bias assessment, and one on data extraction. Study 2 found that Embase plus CENTRAL was the best database combination. When considering the estimated effects on mortality, when combining the major databases in pairs (MEDLINE, Embase, or CENTRAL), only 38% of results were identical to all databases combined. This percentage increased to 66% when combining three databases. Study 3 found that a single review approach missed 4% of inclusions (sensitivity was 0.84, and specificity was 0.86) and took half the time and costs of peer-review study selections. Conclusion: Search strategy is an important methodological question and based on our results, using at least three databases is recommended for a meta-analysis, but one large database may suffice depending on the review context. Regarding study selection, a single review approach can be useful when time is short. This thesis sets a research agenda to optimise RRs and has the potential to influence global literature, establish best practices, and offer replicable methods for researchers.