What level macro? Choosing appropriate levels to assess how place influences population health

It is increasingly recognized that characteristics of different spatial levels (e.g., neighborhoods, states, nations) are associated with population health and may present an opportunity for intervention and health promotion. However, it is often unclear which spatial level may be most suitable to pursue research questions concerned with the production of population health. In this paper we present a discussion about the key conceptual and methodologic considerations relevant to the choice of various spatial levels that may be helpful for ecologic or multilevel analysis that are concerned with understanding the production of health of populations. Although researchers often implicitly mean neighborhoods when discussing how ‘place’ affects health in the contemporary population health literature, we present historical origins and current definitions regarding a range of commonly-operationalized spatial levels including neighborhoods, cities, counties, metropolitan areas, states, and nations, along with a discussion of the strengths of the use of each level when considering the production of health. Reflecting the preponderance of literature in the area, we use U.S. based examples to illustrate the concepts. However, the concepts addressed here are meant to illuminate considerations about relevant spatial units in different national contexts. We also discuss the how the choice of unit affects methodological aspects of any particular study, including construct validity of measures, as well as internal validity, external validity, and statistical conclusion validity of the study. Although some methodological issues, such as causal inference from observational neighborhood studies, have been discussed in the literature with respect to place and health, other issues such as generalizability and policy relevance have received less attention. By articulating theoretical and empirical justifications for using alternate definitions of place, we strive to broaden the conceptualization of how place affects health that may be useful for future research.