Public Health's Share of US Health Spending

Himmelstein and Woolhandler’s finding that public health’s share of US health spending is falling was based on a study of National Health Expenditure Accounts (NHEA) data.¹ The NHEA data identify what ostensibly counts as public health spending across the whole of the United States. However, as has been shown in recent years, this figure is likely a significant overestimate of actual spending on public health because it includes state and local expenditures for many activities focused on individual health care rather than population-level health (e.g., behavioral health care, disability-related health care, safety net clinical care services, and other community health care services).^2,3 The NHEA estimates of public health spending have conflated population health activities with certain types of individual clinical care services for many years, calling into question the reliability of trend analysis using the NHEA’s Public Health Activity estimates. Data from membership organizations of state and local health departments have shown a more substantial decrease in public health spending, and NHEA components related to safety net care have shown growth.^4,5 Taken together, the actual decrease in public health spending may be coming from a lower level, and the drop may be more substantial than Himmelstein and Woolhandler reported. Precise and meaningful analysis will remain elusive until fundamental inconsistencies in the definition of public health activity are resolved.

NHEA data on public health spending are based on data collected by the US Census Bureau’s divisions of state and local finance.⁶ However, there is discordance between definitions underlying the Centers for Medicare and Medicaid Services (CMS) public health activity estimate, which purportedly only counts population-based services with the actual data points from the Census that serve as the basis of the CMS estimates. The Census data encompass all public spending on health, excluding hospitals and third-party reimbursements for clinical services (e.g., Medicaid).^3,7 As a result, the NHEA data include many expenditures on clinical care for individuals that arguably should not be included in the national public health activity estimate.

Over the past two years, a research team from the Johns Hopkins Bloomberg School of Public Health and the de Beaumont Foundation has manually recoded Census administrative expenditure data from states across the country to parse population-focused governmental public health spending from spending on individual health care services. We estimate that slightly less than 40% of the Census’ reports of state government nonhospital health expenditure is for population-focused public health spending. Because this Census data serves as the basis for the CMS public health activity estimate, this translates to a significantly lower estimate, with much of the current estimate going toward direct clinical care services that should be counted elsewhere in the NHEA.⁸

Reaching consensus on a clear delineation of what is and what is not a public health expenditure would enable better data on trends in public health spending to improve the basis for analyses like Himmelstein and Woolhandler’s. In turn, this will provide a foundation to better assess public health spending and make more informed resource allocation decisions.