Sex-related Disparities in Access of Patients with STEMI to PCI Hospitals in Brazil: The Contribution of Blinder–Oaxaca Decomposition Method

Abstract

With regard to discrepancies in the cardiologic assistance, gender differences within predictive models are usually taken as ‘disparities’. However, such differences should be further scrutinised. We aimed to apply the Blinder–Oaxaca method in order to investigate potential sex-related disparities in the time from onset of symptoms to upscale assistance in ST-segment elevation myocardial infarction (STEMI) patients transferred to referential hospitals. All public and private hospitals with 24/7 PCI facilities in the state were included, and 1077 STEMI individuals were prospectively enrolled. We applied the Blinder–Oaxaca decomposition for the predictive model. The study included socioeconomic, clinical and geographic predictors. In ‘crude’ comparison under Student’s t-test, the logarithm of the total time was longer for female than male, reaching a difference of 0.22. In the Blinder–Oaxaca adjusted model, women presented higher total time than men. The total difference was 0.248 (95% CI = 0.051–0.445; p = 0.012), mostly related to the ‘explained’ portion, and the ‘unexplained’ portion reached a tiny fraction of the decomposition of elements, not significantly different from zero. There was no relevant unexplained fraction, also known as discrimination. Hence, the differences are attributable to the influence of the predictors as well as the contrafactual endowments for each sex.