Modeling Premature Adult Mortality From Non-Communicable Diseases in The United States

Abstract

We model mortality from cardiovascular diseases (CVD), cancer, diabetes, and chronic respiratory diseases (CRD) among individuals aged 30–70 years in the United States using quarterly time-series data covering the period 2000 to 2024. We adopt an autoregressive integrated moving average (ARIMA) framework to examine both the long-run trajectory and short-run fluctuations in premature adult mortality. Quarterly data obtained from the World Bank are employed, with mortality from CVD, cancer, diabetes, or CRD specified as the dependent variable, while autoregressive (AR) and moving average (MA) terms capture the inherent temporal dependence in mortality dynamics. Parameter estimation is conducted using the conditional least squares (CLS) technique. Results reveal a positive and statistically significant AR(1) coefficient of 0.987144, indicating strong persistence and inertia in mortality patterns, whereby current mortality outcomes are heavily influenced by their past realizations. In contrast, the MA(4) coefficient is negative and statistically significant at –0.697195, suggesting that short-run shocks to mortality, such as transitory health crises or policy interventions are gradually corrected over time. This dynamic adjustment mechanism reinforces the tendency of the series to revert toward its long-run path. The estimated ARIMA(1,1,4) model satisfies both covariance stationarity and invertibility conditions, confirming its econometric soundness and suitability for forecasting. The adjusted R-squared value of 0.835857 indicates that approximately 83.6% of the variation in premature adult mortality is explained by the model’s dynamic structure, underscoring its strong explanatory power. Out-of-sample forecasts extending to 2050 project a sustained and gradual decline in mortality rates, with levels converging toward approximately 4% by mid-century. These projections are consistent with ongoing improvements in healthcare access, disease prevention strategies, medical technology, and public health interventions in the United States. The study recommends sustained and targeted investments in non-communicable disease prevention, early diagnosis, and health system resilience to consolidate and accelerate the observed long-term decline in premature adult mortality.