(P-107) Evaluating the Use of Prescription Sequence Symmetry Analysis as a Pharmacovigilance Tool: A Scoping Review

Background: The (prescription) symmetry sequence analysis (PSSA) method is a pharmacovigilance tool employed to rapidly identify drug-induced adverse event signals and potential prescribing cascades. However, no prior review has comprehensively described key aspects of the PSSA methodology, reporting of results, and the full breadth of positive significant signals.

Objectives: To scope published PSSA studies and associated index drug – marker drug dyad signals and describe the utility and methodological advances of PSSA to characterize potential prescribing cascades in pharmacovigilance studies.

Methods: We searched Embase, PubMed, Google Scholar, and Web of Science to identify PSSA studies published by July 2020. We included original research studies that used and/or validated PSSA methodology and only included studies where the index exposure and the marker outcome were both defined as the prescription of a medication. We extracted information regarding general study characteristics, significant signals (adjusted sequence ratio [aSR] > 1), subgroup/sensitivity analyses, and reporting of findings. Signals were further classified using fourth-level Anatomical Therapeutic Chemical (ATC) codes.

Results: We identified 53 studies using/validating the PSSA design. We identified 340 unique index drug – marker drug signals, representing 281 unique index chemical class – marker chemical class dyads. Common signals were identified for index drugs acting upon the nervous system (36%), cardiovascular system (22%), and blood and blood forming organs (15%), and marker drug signals were related to the nervous system (23%), alimentary tract and metabolism (23%), cardiovascular system (19%), and genitourinary system and sex hormones (13%). A negative control was used in 23% of studies (n=11), and a positive control was used in 13% of studies (n=6). Of the 47 PSSA studies that aimed to evaluate/detect signals, 55% used stratified analysis to further characterize their results. Five studies (11%) used a logistic regression model to identify predictors of prescribing sequence order, rather than reporting separate aSRs among strata.

Conclusions: The PSSA methodology is useful in detecting and evaluating drug-induced adverse events and potential prescribing cascades due to its efficient design and self-controlled nature. However, the PSSA design is also susceptible to time-varying confounding and background trends in medication utilization. Therefore, consideration should be given to the design of PSSA studies, including the use of negative and/or positive controls, additional stratified analyses, choice of exposure time windows, and reporting of results.