Neighborhood as a Social Determinant of Health in Surgery

Introduction

Social determinants of health (SDOH) defined broadly are the nonclinical factors that impact a person's health. According to the World Health Organization, SDOH consist of the conditions in which people are born, grow, learn, work, and age. Alongside schools and workplaces, neighborhoods represent one of the major contexts in which people lead a considerable portion of their lives. Place and location can determine the quality of schools, access to services, job, transportation, as well as exposure to pollution crime, noise, and other environmental features. It follows that residential location can exert a decisive influence on variations in a person's health.

Attempts to link residential location to health outcomes long predate modern medicine. As earlier as the 5h century BCE, Hippocrates' treatise On Airs, Waters, and Places laid the foundation for one of the earliest understandings of the relationships between places, health, and disease. It is Jon Snow (a surgeon!), however, who is frequently cited as an early pioneer who literally drew the connection between neighborhoods and health by means of a map in the 19th century. By developing a precise dot map of cases of and deaths from cholera in 1854, Snow pinpointed the source of cholera to the now infamous Broad Street pump in the Soho neighborhood of London.

Today the study of “contextual effects” has exploded in large part due to the ease with which neighborhood variables as proxied by various census measures could be attached to existing individual-level data. As a result, this is forcing us to clearly articulate why and how context might be important and, more generally why multilevel thinking (and multilevel analysis) is critical not only to fully understand causation but also to identify promising polices to improve health.

Commonly Used Measures of Neighborhoods

The term “neighborhood” generally refers to the place where a group of people live. This, of course, can be defined at several scales starting with an individual building all the way up to regional delivery region. Given the increasing body of literature about neighborhoods, it seems reasonable to assume a consensus definition has been proposed. However, this is far from the case. Although there have been many proposed definitions of what constitutes a neighborhood, the focus of how these neighborhoods can be operationalized in research design has not received a formal or exhaustive assessment. In other words, how to meaningfully delineate boundaries to assess features contained within them and the implications of their features for health of individuals who live, work, and play within these areas remains much to the researcher's discretion. Here, we provide an overview and describe some commonly used examples in public health research. Neighborhood measures are herein organized in ascending order from small geographic area to largest.

Census block group —Developed and used by the US Census Bureau, census block group is a subdivision of a census tract. Census block groupings are the smallest geographic unit for which the US Census Bureau publishes sample data. Typically, census block groups have a population of 600 to 3000 people.

Census tract —Developed and used by the US Census Bureau, census tracts have an average of approximately 4000 people and are designated to be a homogenous spatial units with respect to population characteristics, economic status, and living conditions.

ZIP code and ZCTAs —A zone improvement plan code or “ZIP code” is one of the more commonly used geographic level exposures in research. The ZIP code is classically used by the United States Postal Service and crudely defines geographic areas into a five-digit code. Each ZIP code may include upward of 30,000 people. Alternatively, ZIP Code Tabulation Areas (ZCTAs) were created by the US Census Bureau based on census blocks and for the most part coincide with ZIP codes. The ZIP+4 code is a basic five-digit code with four digits added as an extra identifier. It helps to identify a geographic segment within the five-digit delivery area, such as a city block or a group of apartments.

County —Administrative or political subdivision of the state in which their boundaries are drawn. Today, 3142 counties and county equivalents carve up the United States, ranging in number from 3 for Delaware to 254 for Texas.

For the purposes of research, there are trade-offs with each of these scales. While census block groups have the most granularity, most datasets have privacy provisions that prevent that level of granularity. Moreover, at that scale, there may not be enough surgical patients to acquire adequate statistical power. On the other end, county is widely available in several datasets and carries far more people. County-level measures may oversimplify the differences that may exist within a county. As such, ZIP code, ZCTAs, and census tract are the most used scales that manage these granularity and scale trade-offs.