Research on neighborhood effects on health in the United States: A systematic review of study characteristics

- Multi-level neighborhoods and health literature has grown over the past 20 years.
- Observational cross-sections, census-based boundaries, and two-level designs were dominant.
- BMI/obesity and neighborhood SES were the most common outcomes and exposures, respectively.
- Making causal inferences and modeling complex and dynamic relationships are future priorities.
- Future research should inform interventions that improve health and reduce disparities.

Neighborhood effects on health research has grown over the past 20 years. While the substantive findings of this literature have been published in systematic reviews, meta-analyses, and commentaries, operational details of the research have been understudied. We identified 7140 multi-level neighborhoods and health papers published on US populations between 1995 and 2014, and present data on the study characteristics of the 256 papers that met our inclusion criteria. Our results reveal rapid growth in neighborhoods and health research in the mid-2000s, illustrate the dominance of observational cross-sectional study designs, and show a heavy reliance on single-level, census-based neighborhood definitions. Socioeconomic indicators were the most commonly analyzed neighborhood variables and body mass was the most commonly studied health outcome. Well-known challenges associated with neighborhood effects research were infrequently acknowledged. We discuss how these results move the agenda forward for neighborhoods and health research.