Practical and policy implications of using different rural-urban classification systems: a case study of inpatient service utilization among Veterans Administration users.

CONTEXT Several classification systems exist for defining rural areas, which may lead to different interpretations of rural health services data. PURPOSE To compare rural classification systems on their implications for estimating Veterans Administration (VA) utilization. METHODS Using 7 classification systems, we counted VA health care enrollees who lived in each category, and number admitted to VA hospitals or non-VA hospitals under Medicare. For dual VA-Medicare enrollees over age 65, we compared VA and private sector hospitalizations on numbers of admissions and bed-days of care. We compared VA enrollees' relative proportions across rural to urban categories for each classification system and evaluated discordance between systems at the veterans-integrated service networks (VISN) level. FINDINGS Enrollment and inpatient utilization counts for rural veterans vary considerably from one classification system to another, though the systems generally agree that admission rates, length of stay, and reliance on the VA for care are lower for rural veterans. Among older dual VA and Medicare enrollees, rural residents rely on non-VA facilities more, though this effect also varies widely depending on the classification scheme. VISNs vary greatly in the proportions of patients who are rural residents, and in the degree to which classification systems are discordant in designating patients as rural. CONCLUSIONS Decisions about allocating VA health care resources to target "rural" patients may be affected greatly by the rural classification system chosen, and the impact of this choice will affect some hospital networks much more than others.