Trends, disparities and projections of pericardial diseases related mortality in the United States: A 1999–2035 analysis using artificial deep neural network

Abstract

Background: Pericardial Diseases (PD) are a significant cause of morbidity and mortality. However, relevant data for PD-related mortality are limited. This study aims to evaluate the demographic and geographical variations in PD-related mortality in the United States and utilize deep learning model to forecast mortality trends. Methods: PD-related deaths in adults aged ≥25 years were identified using CDC WONDER from 1999 to 2019. Age-adjusted mortality rates (AAMRs) per 100,000 were determined. Joinpoint regression was used to calculate annual percentage change (APC) with 95% Confidence Intervals (CIs). A Long Short-Term Memory (LSTM) recurrent neural network was developed to predict future mortality. Results: < 0.01). Males (2.37) had a consistently higher AAMR than females (1.57). Non-Hispanic (NH) Black or African American had the highest AAMR (2.65), followed by NH White (1.90), Hispanic or Latino (1.62), and NH Asian people (1.57). Regionally, the West (2.11) recorded the highest mortality. The prediction model suggested a gradual increase in AAMR, reaching an estimated 2.54 by 2035. Conclusion: Despite overall decline, the recent and projected rise in PD-related mortality highlights the urgent need to improve equitable healthcare access, particularly among males, NH Black population, and West region, who exhibit persistent disparities.