A scoping review of real-world evidence generation methods to estimate the safety of type 2 diabetes medications on fracture risk.

Abstract

Observational studies provide real-world evidence (RWE) about the long-term effectiveness and rare side effects of medications. However, conflicting results from RWE studies exist regarding the relationship between type 2 diabetes mellitus (T2DM) medications and fracture risk. The objective of this scoping review was to summarize RWE methods employed to estimate the effect of T2DM medication on fracture risk. We searched MEDLINE (OVID), Embase (OVID), and CINAHL (EBSCO) databases for full-text articles published between January 2000 and December 2022 in English and conducted in adults. Studies were included if: (1) the primary exposure was T2DM medication(s), (2) the primary outcome was fracture, and (3) the exposure and outcome were defined using healthcare administrative data. Study design, exposure and outcome measures, and analytical approaches were abstracted and summarized. We identified 50 eligible studies (40 cohort, 9 case-control, 1 self-controlled case series). Most cohort (60%) and all case-control studies defined medication exposure using an intention-to-treat approach. One cohort study considered time trends; 40% of studies (19 cohort and 1 case-control) reported an exposure-outcome risk window, although few (2 cohort studies) justified biological plausibility, and only 46% of studies defined how fractures were measured. We identified large variation in RWE methods used to investigate T2DM medications and fracture risk. Few studies considered biological plausibility or transitions in therapy in their design or provided detail to be able to replicate. Our findings highlight the need for guidance in methodological design to study long-term pharmacotherapy for chronic conditions and a culture shift to improve transparent reporting.