Government loan quebec student

INTRODUCTION

In 1970 the National Health Service Corps (NHSC or Corps) was created as part of the Emergency Health Personnel Act. The goal of this agency is to provide health personnel, most often physicians, to communities that are deemed "medically underserved." The Corps has provided an important subsidy to medical students by providing access to full scholarships covering tuition and fees in exchange for subsequent service in locations designated by the Corps.

This paper investigates two issues not examined in previous evaluations of the NHSC. First, although most studies have found that the NHSC physicians are more likely than non-enrollees to leave the community in which they initially locate, the techniques used commonly fail to recognize that the decision to enroll in the NHSC and the decision of where to locate in subsequent years may be endogenous. The results of this study are consistent with previous findings that participation in the Corps decreases the tendency to stay in the original location, even when controlling for self-selection into the program. Second, a broader measure of programmatic success is proposed. If enrollees do not remain in their initial practice locations but subsequently locate in other communities with low access to primary health care, then the program has increased access to health care in medically underserved communities. An accurate measure of the program's effectiveness should incorporate this benefit.

Data obtained from the American Medical Association [1997] allow the geographical positioning of every physician in the United States for 1981,1986,1991, and 1996. By comparing locational choices of NHSC enrollees with those of non-enrollees, the extent to which these federal programs decrease inequities through subsequent actions (that is, after the completion of the subsidy) can be determined. Two main questions are pursued in this paper:

1. To what extent does the use of the broader metric of program success proposed here-subsequent service to undeserved communities-affect the evaluation of the program's success? If the evaluation of the program takes notice of the externalities inherent in these types of programs, is a more favorable evaluation of the program obtained?

2. To what extent does the endogeneity of program enrollment bias measures of the programmatic effect? How does the unobserved heterogeneity influence the inferences drawn from the data?

Results indicate that measuring success as subsequently practicing in an underserved community yields a positive assessment of the Corps. Furthermore, there is little evidence of endogeneity between practicing in an underserved location and the decision to enroll in the Corps. That is, it is not some unobserved preference that is affecting both, but a true programmatic effect. Nevertheless, evidence suggests that index site retention is endogenous with NHSC enrollment, and estimates of the NHSC effect on same-site retention become more negative when accounting for self-selection. Therefore, NHSC service decreases the tendency to remain in the initial practice location.

BACKGROUND

NHSC was created in 1970 to help balance geographical access to primary health care. This program works in the following manner. Medical students enroll in the program early in their medical school training and receive a full scholarship for tuition. For every year the student receives a scholarship, the student must serve one year at a location designated by the NHSC. The enrollee is presented with a list of approved practice locations from which to choose. Physicians express a preference for particular locations, and medical organizations at those locations select from those expressing an interest. After fulfilling the contractual commitment, physicians are free to enter the private workforce. In addition to the scholarship program, the Corps began offering a loan repayment option in 1987 in which physicians enroll upon graduation and the Corps repays a portion of the medical school debt for every year of service.1

Numerous papers have examined the determinants of physician distribution: the Newhouse, Williams, Bennett, and Schwartz (NWBS) [1982a, 1982b] studies (and the subsequent Newhouse [1990] article) combine physician distribution with standard economic location theory. They argue that location theory predicts that physicians will be less likely to locate in smaller areas due strictly to demand considerations. They argue against the conventional thinking at the time that market failures were commonplace in the physician labor supply market.

Miller, Dixon, and Fendley [1986], focusing on the West South Central states, utilize a human capital approach to find that, while there was a surplus of health care professionals in a considerable number of markets, there were also many rural areas that experienced shortages. One reconciliation of the Miller, Dixon, and Fendley [1986] results with the NWBS findings is that the quality of care, the quantity demanded, the effort per physician (in full time equivalent units, say), or some combination differs between different types of areas. The findings of Miller, Dixon, and Fendley [1986], then, support the case of market failure, not on the basis of "counts" but of "unmet need."

Numerous studies have examined physician retention. For example, Horner, Samsa, and Ricketts [1993] follow a cohort of almost 2,000 North Carolina physicians, examining how their characteristics predict whether they will move to rural or urban areas. They also investigate the hazard rate of terminating their employment. Comparisons of mean tenure suggest that hazard rates differ little between rural and urban physicians; however, controlling for physician characteristics, the propensity for leaving a community is 28 percent greater for physicians in rural communities than for physicians in urban communities. They do not, however, control for the initial selection of area type. Ricketts et al. [1996] examine migration patterns of obstetricians-gynecologists into and out of rural counties. They find that county characteristics, such as population growth, affect the relative flows into or out of the county. Furthermore, young physicians are more likely to migrate than older ones.

Pathman, Konrad, and Ricketts [1994] find that Corps physicians who are graduates of public medical schools are more likely to leave the initial placement than NHSC graduates of private schools. This effect does not exist in the non-Corps cohort. Singer et al. [1998] examine the retention of physicians at Community Health Centers. They present differences between the Corps and non-Corps physicians, such as race (Asians less likely to be Corps, African-Americans more likely) and specialty status (specialists less likely). Unfortunately, they analyze the retention of NHSC physicians separately from the retention of non-NHSC physicians, and hence are not able to measure explicitly the NHSC programmatic effect. By comparing the empirical distribution of retention rates, they (implicitly) conclude that NHSC physicians are less likely to remain in their positions than are non-enrollees. Mofidi et al. [2002] and Porterfield et al. [2003] survey NHSC alumni to analyze factors predicting index site retention.

The literature, then, in general, tends to conclude that NHSC physicians leave the community rather quickly after fulfilling their commitment. On this basis, the program seems to be supported only by a "bandage" effect, where the "bandage" justification is the increase in access while the physician is contractually obligated to serve. That is, there is no evidence that NHSC physicians continue to provide access to underserved populations after their obligation is completed, so the only impact of the Corps is the contemporaneous (during-contract) effect.

Some literature has examined the efficacy of government policy to encourage physician location into communities with low access to physicians. Bolduc, Fortin, and Fournier [1996] and Bolduc, Fortin, and Gordon [1997] examine the location choice of Quebec physicians in response to various policy options, such as reimbursement rates. They find that substantial redistribution into underserved communities is possible using market incentives. Rabinowitz et al. [2000] investigate the effect of multiple physician characteristics on the propensity to practice in a medically underserved location. They obtain a statistically significant odds ratio estimate of 2.2 for NHSC enrollment, but do not control for self-selection into the Corps. Therefore, we cannot conclude whether the Corps program increases the likelihood of practicing in underserved locations or whether Corps enrollees are ex ante more likely to practice in underserved communities.