The Comparative Retention of National Health Service Corps and Other Rural Physicians

Abstract

To contrast the retention of physicians serving National Health Service Corps (NHSC) Scholarship Program obligations in rural settings to that of non-NHSC physicians working in the same or similar practices, and to identify promising retention-enhancing strategies. Cohort study. Four hundred twelve primary care physicians initially identified during an earlier study as working in a national stratified random sample of 178 externally subsidized rural clinics in 1981. Thirty-six percent were serving obligations to the NHSC, nearly all through the NHSC's Scholarship Program. The NHSC and non-NHSC inception cohorts (those first coming to their 1981 [or "index"] practices from May 1979 through December 1981) were created from within the entire group for use in most analyses. In 1990, physicians were resurveyed to learn of their backgrounds, experiences in their index practices, and their subsequent career moves. By 1984 and in each year thereafter, fewer NHSC than non-NHSC physicians of the entire respondent cohort remained (1) in their index practices, (2) in their index communities, and (3) in practice in any rural county (P less than .001). In the inception cohort, fewer NHSC than non-NHSC physicians were retained within all three settings by the third year after their initial dates of employment (P less than or equal to .01). After 8 years of employment, group retention rates for NHSC and non-NHSC inception cohort physicians were 12% vs 39% in the index practice and 29% vs 52% in nonmetropolitan practice. Physicians in both NHSC and non-NHSC groups who left their index practices generally left rural practice altogether. When compared to non-NHSC physicians working in comparable rural settings, the retention of rural NHSC physicians is seen to be poor and only partially explained by fixed physician, practice, or community variables. Long-term retention of NHSC providers is now receiving much needed attention at the federal level.