This is a new application from the University of North Carolina at Chapel Hill that proposes to assess the successes and shortcomings of state programs that provide scholarships, loans, and loan repayment incentives to primary care providers, in exchange for required service in needy areas. It also proposes to identify how these programs can be designed and operated to maximize the availability of primary providers in underserved areas. Specifically, the project would: 1) identify and describe current State service-contingent programs across the country; 2) describe and contrast the criteria states use to identify communities eligible to receive obligated providers; describe how the communities where obligated providers actually serve differ from all eligible communities; and contrast the communities where obligated providers work to the communities where other recently trained, non-obligated physicians work; 3) describe the default rates, buy-out rates, and retention duration of providers obligated to state programs and the factors affecting these program outcomes; 4) test whether factors known to be important to the satisfaction and retention of National Health Service Corps providers are also relevant to providers obligated to state programs; and 5) recommend strategies to maximize the impact of state scholarship, loan, and loan-repayment programs. The application proposes to accomplish these aims through three phases of work. In phase one, the investigators will gather information on all state service-contingent programs from available reports and direct telephone contact with states. They will then select approximately 20 states for further study, analyzing patterns and trends in program design, program-specific rates of provider buy-out and default, and multivariate relationships between provider characteristics and the probability of buy-out and default. Phase two will describe and contrast the criteria that these 20 states use to identify communities eligible for assistance and the processes they use to assign providers to communities. The investigators will compare state community-eligibility criteria and provider placement processes to federal criteria and processes. Also during phase two, they will develop logistic and Poisson multiple regression models to estimate the characteristics of counties that predict 1) eligibility to receive a service-obligated provider, 2) likelihood of receiving any provider, if eligible, and 3) the number of providers received by eligible counties. Data from the Area Resource File will be combined with data obtained from state programs to test these models. In phase three, the investigators plan to obtain questionnaire responses from 1,000 current and past state program participants and from 1,000 recently-trained non-program generalist physicians working in the same 20 states. Physicians will be sampled from program beneficiary databases that states will be asked to provide and from the American Medical Association's (AMA) Physician Masterfile. Survey results would describe and contrast the work, personal lives, satisfaction, and retention of state program participants; identify predictors of retention; and identify the long-term effects of program involvement on physicians' careers. Retention duration will be predicted using Cox proportional hazard models. Ordinary least squares regression will be used to estimate correlates of physician satisfaction, level of involvement with their placement communities, and other outcomes.
