COVID-19 and Neighborhood Deprivation in Utah, Spring 2020

By the spring of 2020, COVID-19 was known to be hitting racial and ethnic minority groups and essential workers hardest. What had not been examined was the role of place — the social and economic disadvantage of the area someone lives in, known as deprivation. By July 9, 2020, Utah had reported 27,356 confirmed cases, and the Utah Department of Health set out to measure how infection, hospitalization and testing varied with it.

The headline: compared with Utahns in the least deprived areas, those in the most deprived had about three times the odds of a confirmed infection.

Measuring deprivation

Utah divides itself into 99 small statistical areas, the smallest units with enough data to report reliably. Each has a score on the state's Health Improvement Index (HII), from 72 to 160, built from nine indicators:

1. median family income
2. income disparity — households under $10,000 against those at $50,000 or more
3. home ownership
4. unemployment
5. families below the poverty threshold
6. single-parent households with children under 18
7. adults 25 and older with under nine years of schooling
8. adults 25 and older with at least a high school diploma
9. people below 150% of the poverty threshold

Areas were sorted into quintiles, from very low (least deprived) to very high (most deprived). The least deprived are concentrated in urban and suburban northern Utah — Salt Lake, Davis and Wasatch counties, for example. The most deprived are mostly in rural central and southern Utah, the western Salt Lake City metropolitan area, and parts of other cities such as Ogden and Logan.

Every confirmed case reported between March 3 and July 9, 2020 was placed in its area by address. The team calculated rates and odds for each fifth against the least deprived, and also age-weighted versions that adjust for each area's age mix using the 2000 U.S. Census standard population.

What they found

• Infection climbed steadily with deprivation. The jump in incidence from high to very high was as large as the whole climb from very low to high — a sign, the authors suggest, that extreme deprivation may compound transmission.
• Hospitalization also rose, but less steeply: 1.6 times the odds in the most deprived areas. The surprisingly high figure for average areas may reflect people there being more likely to seek care and to have insurance than in more deprived areas.
• Testing varied least: the testing rate in the most deprived areas was about 25% higher than in the least deprived, while the share of tests coming back positive rose from 5.0% to 12.0%.
• Age-weighting generally made the gaps larger, because more deprived areas have younger populations — younger Hispanic and Pacific Islander families, for example.

Who lives in the most deprived areas

The higher-risk sectors counted were food preparation and serving, building and grounds cleaning and maintenance, construction and extraction, production, and transportation and material moving.

Why place matters

Risks can cluster in the same neighborhoods. People in deprived areas may be more likely both to work where they can be infected and to live in crowded homes where the rest of the household can catch it. They may also be unable to isolate — because of work, crowded housing or having no car. A New York City study found the same pattern among pregnant women: lower odds of infection in higher-income neighborhoods, higher odds where households were more crowded.

What the authors recommend

Public health agencies should use deprivation and other social determinants of health to find where COVID-19 falls hardest, and to target those places:

• more and easier testing, contact tracing and places to isolate;
• preventive care and disease management;
• prevention guidance sent to clinics, community centers and businesses in the area;
• materials that are linguistically and culturally appropriate, in the first language of the communities at risk;
• partnerships with organizations that can reach those communities.

Limits

1. Cases with no or mild symptoms go unrecognized, so incidence and testing may be underestimated.
2. The index is a composite; it is hard to say which of its parts drive the results, or whether it predicts better than any one of them.
3. Small areas do not always match communities: a transient student population can raise an area's score, and a low-scoring area can still hold underserved groups such as American Indian communities.
4. Only area-level survey variables were available; others, such as overall housing density, were not.
5. Population figures predate the pandemic, and measures like unemployment may have changed.
6. The job sectors do not cover all high-risk work, and a person's role (frontline worker or manager) might describe risk better.

Sources

Based on Lewis NM, Friedrichs M, Wagstaff S, et al., "Disparities in COVID-19 Incidence, Hospitalizations, and Testing, by Area-Level Deprivation — Utah, March 3–July 9, 2020," MMWR Morbidity and Mortality Weekly Report volume 69, number 38, Centers for Disease Control and Prevention; a work of the United States government in the public domain. The report is inconsistent in three places, and this page leaves those figures out: its text and Table 1 disagree over which incidence rates per 100,000 are the age-weighted ones; its text and Table 3 give different shares of food-insecure residents in the most deprived areas (22.6% and 26.6%); and its summary dates the period to June 9 rather than July 9.