Optimal Staffing in Community Health Centers to Improve Quality of Care: A Generalized Production Function Approach

Background: Community health centers (CHCs) are important safety-net providers. In 2016, 1367 federally-funded CHCs served 26 million patients or 1 in 12 Americans. Meanwhile, researchers have documented significant changes in CHC workforce, with more advanced clinicians (APCs) (i.e., nurse practitioners, physician assistants, certified nurse midwives) each year. However, little is known about how this changing workforce is affecting the quality of care at CHCs.

Objectives: This study focuses on four research questions 1) how do staffing configurations and capital contribute to the production of quality of care?; 2) whether there are trade-offs between quality of care and quantity of services?; 3) how different professions substitute or complement each other in the CHC production process?

Data: This study used the Uniform Data System (UDS) from 2014 to 2016. All CHCs that receive Section 330 grants under the Public Health Service Act are required to report annually to the Bureau of Primary Health Care in the UDS. These reports include information on staffing, utilization, quality of care, revenue, and patient characteristics of the community health center. The UDS data was linked with the Internal Revenue Service (IRS) database on Form 990 Return of Organization Exempt from Income Tax. IRS Form 990 provides information on capital expenditures and stock of CHCs. The final database contained 3,147 center-year observations from 1,184 health centers.

Methods: We relied on Diewart’s flexible functional form to estimate a system of two generalized linear production functions, using quality of care and volume of services as outputs, and we showed its applicability to our empirical setting. Five categories of labor inputs, including primary care physicians (PCP), APCs, nurses, other medical support staff, and administrative and enabling staff, and capital were included in the production function. Other organizational and patient characteristics were controlled. The quality of care index was defined as the average percent of diabetic patients who have A1C level controlled and hypertensive patients who have blood pressure controlled. A revenue function was estimated to examine the complementarity and substitutivity of labor and capital inputs.

Findings: In the production of quality of care, APCs had almost the same marginal productivity as PCPs. Each additional FTE of APC and PCP improve the quality index by 0.24 percentage points at sample mean. On average, an additional APC provides 1,440 visits annually, 25 percent lower than the productivity of an additional PCP who provides 1,795 visits annually. The Hicks elasticity of substitution between APC and PCP estimated from the revenue function was -2.023, indicating that APC and PCP are q-substitutes in CHC production process.

Conclusion: Advanced practice clinicians and primary care physicians are similarly effective in improving quality of care measures. However, in a fee-for-service environment, PCPs are preferred by CHCs since they provide a greater volume of services and much higher revenues. However, as outcome-based alternative payment models become more diffused, the trade-off between these two labor inputs needs to be reconsidered.