First Psychotropic Medication Redemption Is Not a Reliable Measure of Abortion's Mental Health Burden: A Reanalysis and Review of the Methodological Limits and Interpretive Boundaries of Steinberg et al. (2026)

Le résumé fourni par la source

Abstract Background: Steinberg et al. (2026) used Danish national registries to examine first psychotropic medication redemptions before and after first-trimester medication and procedural abortions. The study provides a large population-based analysis of treatment initiation, but it excludes many women with prior psychotropic medication use, censors follow-up at the first qualifying redemption, and reports post-abortion risks only in separate time intervals rather than also reporting a cumulative 0-5-year estimate. Methods: This re-examination evaluates cohort selection, outcome scope, the pre-abortion reference period, depletion of the event-free risk set, comparison-group limitations, multiple-testing procedures, and effect heterogeneity. It also presents a limited aggregate-data reanalysis in which the published fully adjusted interval-specific IRRs are summarized over the first five years using duration weights of 1, 1, and 3 years. The reporting issue is compared with a prior reanalysis of Danish psychiatric-contact data. Results: The descriptive 0-5-year IRR indices were 1.02 for any psychotropic medication after medication abortion and 0.99 after procedural abortion; 1.00 and 1.05, respectively, for antidepressants; and 1.23 and 1.05, respectively, for antianxiety medications. The antianxiety pattern after medication abortion was elevated in every reported interval through five years. Valid confidence intervals for these five-year summaries cannot be calculated from the published aggregate estimates because person-time and covariance information are unavailable. In the earlier Danish study, cumulative reanalysis increased the reported overall odds ratio for first psychiatric contact from 1.12 at nine months to 1.49 at twelve months. Conclusions: A first psychotropic medication redemption is not a reliable measure of total mental health burden. Interval-specific estimates are useful for identifying timing, but they should be reported alongside model-estimated cumulative results over clinically meaningful periods. Exact cumulative 0-5-year estimates require reanalysis of the individual-level Danish registry data. Keywords: induced abortion; psychotropic medication; mental health services; psychiatric treatment; registry study; recurrent events; selection; effect heterogeneity

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.