Forecasting U.S. health spending by disease, type of care, age and sex to 2040

Healthcare spending in the U.S. is expected to increase in the future. However, currently only aggregated estimates of total spending exist. In order to design policies to address the changing needs of the health system and potentially curb rising spending, estimates need to be disaggregated by demographics and disease. We estimate future healthcare spending by disease, type of care, age and sex using a component model. Changes in the population structures, disease prevalence, utilization patterns, and costs are each forecasted separately and combined to produce estimates of total spending.

We estimate spending and utilization by age and sex for 165 disease and injury categories, four key risk factors and nine non-disease categories such as well person care and well pregnancy care. In addition, our estimates are further disaggregated into eight types of service, such as inpatient care and prescribed pharmaceuticals. Survey, administrative and claims data is synthesized to produce a comprehensive set of annual estimates for 1996 through 2013. We combine these estimates with population estimates and prevalence rates from the Global Burden of Disease 2013 project.

We use population forecasts and forecasts of disease-specific prevalence by age and sex from the Global Burden of Disease Health Forecast project. For our baseline model we assume that the utilization per prevalent case will be constant over time and annualized rate of change of price will be constant. We forecast each component out to 2040 and calculate total spending by type of service, disease, age and sex.

Our preliminary estimates suggest that total health spending will reach five billion by the year 2037. The segment of the population over the age of 65 will consume an increasing proportion of total health spending which will reach 42.9% in 2040. Total spending on Alzheimer’s disease will climb from the thirteenth most expensive disease in 2015 to the third most expensive disease in 2040.

Understanding the changing profile of diseases and demographics in healthcare spending is essential for planning for future health resource allocation or intervention. Our model provides estimates at the level of detail necessary for policy. Additionally, it will let us test different scenarios in which interventions change the trajectory of disease prevalence, utilization, or cost, and examine the effect on total health spending.