The E¤ect of Rurality on Mental and Physical Health

Any time researchers choose to conduct a study on any aspect of rural health care, they are faced with a di¢ cult methodological choice regarding the oper- ational de?nition of rural to use in their study. While this seems to many who are not familiar with rural research to be a straightforward question (and often naive rural researchers expect to ?nd a single answer that is commonly agreed upon by experts in rural research), to date there is no single answer that is based on a consensus and supported by scienti?c evidence. Coburn, et al. (2007) states that ?there is no single, universally preferred de?nition of rural that serves all policy purposes.? One goal of this paper is to provide a method useful in informing the choice of rural de?nitions for a speci?c research study. This methods paper presents a systematic evaluation of the impact of the choice of rural de?nition on results. It was stimulated by the need to select a rural de?nition to use in related papers on the impact of community resources on mental and general health outcomes in other research by the authors. A na- tional dataset, the Community Tracking Survey, 2000-2001, includes individual level observations from household interviews. We merge it with county level data re?ecting community resources, and we use econometric methods to ana- lyze this multi-level data, accounting for individuals from the same family being included in the dataset. The e¤ect of using four di¤erent de?nitions of rural available for use in county level analysis is presented. A statistical analysis of the impact of the choice of a rural de?nition on outcomes and on the esti- mates and signi?cance of explanatory variables in the model is presented and is used to inform the selection of the de?nition to use in other research. Dif- ferences in results for mental health, physical health, and utilization of health care variables are evaluated. The choice of a rural definition is presented and justi?ed using the methodological analyses presented in this paper. Strengths and weaknesses of using county-level community characteristics as compared to data from larger geographic areas, such as Health Services Areas, or from smaller geographic areas, such as census tracts or zip codes, are discussed. Fi- nally, the need for a methods study to guide the use of multi-level geographic data to re?ect community characteristics within health care studies is proposed.