Bayesian Logistic Regression Analysis of Sex Differences in Patients with Transient Ischemic Attack (TIA) or Minor Strokes
Conference
65th ISI World Statistics Congress
Format: IPS Abstract - WSC 2025
Keywords: bayesian modeling, logistic regression, medical, sex-specific mortality
Abstract
Understanding sex differences in stroke outcomes is crucial for personalized medical interventions, as strokes are among the top causes of death in Austria. This study employs Bayesian logistic regression models to analyze stroke data, focusing on patients with Transient Ischemic Attack (TIA) or minor strokes, to detect potential sex differences in clinical outcomes. For this end, data of the Austrian Stroke Unit Registry is analysed consisting of a total amount of more than 200000 stroke patients. Bayesian logistic regression models are applied to assess the impact of sex on short and long term clinical outcomes, namely the MRS and mortality at the time of release from the stroke unit and follow up 90 days after release from the stroke unit. Prior distributions were carefully selected based on existing literature and expert opinion to ensure robust and informative posterior estimates. The study also controls for relevant medical covariates, such as age, risk factors etc. as well as eveluating the predictability based on ABCD2 and ABCD3I Scores.