Transition Dynamics of County Obstetric Service Absence: Population Exposure and Status Changes in the United States, 2015–2023

Vollständiger Abstract

Worum geht es in dieser Arbeit?

ABSTRACT Objective To quantify persistence and transition dynamics of county obstetric service absence and assess whether birth volume and physician supply add predictive information beyond prior‐year status. Study Setting and Design National longitudinal county–year panel study (2015–2023). Obstetric service absence was defined using a 2‐year confirmation rule, with contemporaneous status used for the 2015 baseline year to reduce misclassification. We estimated annual conditional transition probabilities (entry and exit) and modeled current‐year absence as a function of prior‐year absence and year fixed effects, with and without contemporaneous birth volume and physician supply. Data Sources and Analytic Sample All US counties ( N = 3144) observed annually. Obstetric service availability was identified from Centers for Medicare & Medicaid Services Provider of Services files. Residence‐based births, physician supply, and bed capacity were obtained from the Area Health Resources Files. Population denominators for women aged 15–44 were drawn from the American Community Survey. Principal Findings Obstetric service absence was highly persistent. County prevalence was stable (34.2% in 2015 and 34.2% in 2023). Population exposure declined: the share of women aged 15–44 residing in counties without services decreased from 6.5% to 5.3%, and the share of births to residents of such counties decreased from 6.6% to 5.7%. Annual transitions were rare (entry ≤ 0.4%; exit ≤ 0.9%), and conditional stay probabilities were ≥ 0.991 for both presence and absence. In lagged‐status models, prior‐year absence strongly predicted current absence (0.994); adding birth volume and physician supply did not materially change the estimate (0.993). Conclusions On policy‐relevant timescales, obstetric service absence behaves as a high‐inertia system state. Rare transitions and dominance of prior status suggest slow adjustment, with implications for closure prevention and targeting support near volume and workforce feasibility thresholds.