Document Type

Report

Publication Date

10-2021

Abstract

Health mostly happens outside of the hospital and the doctor’s office. This reality is called “the social determinants of health” (SDOH). SDOH’s impact can be understood a number of ways. This analysis shows that great local healthcare doesn’t preclude poor population health, exhibited by high infant mortality and low life expectancy rates.

The analysis found that while Ohio and Cuyahoga County have world-class healthcare, they also have third-world health outcomes.

• Ohio’s life expectancy (76.8) ranks 12th worst nationally, well below Midwestern peers. Cuyahoga County’s life expectancy (77) ranks 41st out of 89 counties in the State.

• Black Cuyahoga County residents live shorter lives (73.6) than Whites (78.3) Hispanics (82.7) and Asians (89).

• When it comes to infant mortality rates, Ohio’s infant mortality (6.97) ranks 9th worst in Ohio.

• Rates vary dramatically by race, with Black Ohioans (14.3) having more than double the infant mortality rates of Whites (5.1) and Hispanics (5.8).

• Cuyahoga County’s infant mortality (9) is tied for second last for those Ohio counties in which the figure was calculable. The infant mortality rate for Blacks in Cuyahoga County is quadruple that of Whites and nearly triple that of Hispanics.

The analysis intended to go beyond showing that health disparities occur, using a novel SDOH “big” dataset to shed insight on how they occur. In doing so, the analysis conceptualized SDOH as either being upstream, or influenced by structural factors like class and race; midstream, or influenced by neighborhood factors like residential segregation, environmental toxins, and individual behavior; and downstream, or the prevalence of chronic disease and psychosocial stress.

Two machine learning models, or Random Forest, were run. The first model calculated the 20 highest risk factors for infant death for all counties in the Unites States. The second model calculated the 20 factors with the highest predictive power on life expectancy for all census tracts in Cuyahoga County. The results explain what factors predict high infant mortality rates between counties in the U.S. and what factors predict life expectancy rates between neighborhoods within Cuyahoga County. These factors are varied and range from the percent of knowledge workers in a neighborhood, the proximity to five star Yelp establishments, foreclosures, sheriff sales, car volume and proximity and volume, and ground-level Ozone.

In all, the analysis calls for a methodological and conceptual approach in which SDOH researchers and practitioners are “swimming upstream” to address the root causes of health disparities from a policy standpoint, while continuing to tackle midstream and downstream factors through behavioral- and neighborhood-based intervention. To do this, the authors suggest implementing more data science practices into the social science field, an interdisciplinary movement termed “computational social science”.

Share

COinS