COVID-19 and the Social Determinants of Health

The Social Determinants of Health (SDOH) is an index that considers different characteristics of the population within an area as a predictor of the health needs of persons living in the area. The COVID-19 pandemic has generated attention to this methodology, with some public health professionals specifically questioning if the number of COVID-19 cases in an area is related to the needs predicted by the SDOH indices.

Social Determinants of Health are conditions in the places where people live, learn, work, and play that affect a wide range of health risks and outcomes ( CDC ).

A  story map  and  web mapping applications  on the social determinants of health were created last year for Western New York. These applications use data from the American Community Survey (ACS), 2017 and 2018, 5 yr. estimates. The study region was 8 counties in Western New York. Analysis was completed for census tracts and zip code (tabulation) areas. Additionally, the SDOH indices were based on counts and concentration values. Information on previous work may be found  here . COVID-19 confirmed cases were obtained from an  Erie County website  on Oct 22, Oct 31, Nov 8, and Nov 16, 2020. Dates used throughout this story are those indicated as the last update date for the case counts.

The SDOH index in an area is a function of the characteristics of the entire region. Past efforts derived the SDOH indices for zip codes in Western New York (8 counties). Since COVID-19 cases count data (zip code summary level) are only available for Erie County, the SDOH indices were recalculated to represent high needs and low needs zip codes in Erie County only.

Not all zip codes are included in the SDOH analysis. Those with a very small population were removed from the study due to low reliability or high margin of error on census estimates.

Of the zip codes included in Erie County, the total confirmed cases on Oct 21, 2020 was 12,345.

• 467 new cases were reported by Oct 30, a 4% increase from the previous week.

• 796 new cases were reported by Nov 7, a 6% increase from the previous week.

• 1,248 new cases were reported on Nov 15, a 9% increase from the previous week,

bringing the total confirmed cases to 14,856.

The SDOH index may be used to group zip codes into highest needs, high needs, and low needs categories. The highest needs zip codes have an SDOH index that is larger than the index in 95% of the zip codes in the region. Low needs zip codes are in the bottom 50% with respect to the SDOH index.

The average number of confirmed cases in the highest needs zip codes is over 7 times greater than the average number of confirmed cases in the low needs zip codes, until the recent surge in cases.

This past week saw the rate dropped to just under 7 times greater.

The average number of confirmed cases in the high needs zip codes is nearly 4 times greater than the average number of confirmed cases in the low needs zip codes.

This past week indicates the rate is closer to 3.5 times greater.

The statistics indicate that zip codes with high and highest needs have more than 3/4 of the COVID - 19 cases but represent less than half of the zip codes in the county and contain about 2/3 of the population in the county.

Just over 10% of the population is in the highest needs zip codes but more than 20% of the confirmed COVID-19 cases. The recent surge in cases has seen the share of cases in the highest needs zip code drop to just under 20% of the confirmed COVID-19 cases.

The COVID-19 rates per 1,000 people in high needs areas are nearly double the rates in the low needs zip codes. The COVID-19 rates per 1,000 people in the highest needs areas are nearly triple the rates in the low needs zip codes.

However, the recent surge in cases has slightly reduced these relationships.

The chart shows this relationship between needs based on the SDOH index for a community and the COVID-19 cases on all four dates.

Maps show the spatial relationship between the SDOH index and COVID-19 cases. The first map shows the relationship between the SDOH index and the number of COVID-19 cases as of Oct 21. Zoom, pan, and click zip codes for pop-up information on specific zip codes.

The larger circles are over the red and pink areas, which are those zip codes with higher needs.

This next map shows the relationship between the SDOH index and the number of COVID-19 cases per 1,000 persons. Many of the highest rate locations are in the city of Buffalo where many of the high needs areas are located.

The number of confirmed cases and cases per 1,000 people (on Oct 21) are shown below for all of the high needs zip codes. Zip codes are sorted in descending order based on the SDOH index. The closer a zip code is to the bottom of this list, the more it has in common with low needs zip codes.

The SDOH index is based on three factors; Social, Economic, and Housing and Transportation. Variables associated with each factor are explained and the COVID-19 cases (on Oct 21) are summarized for zip codes in Erie County, NY.

High needs zip codes, based on the social factor, have 77% of all cases and the rate of cases in the highest need zip codes is nearly three times higher than the low needs zip codes.

High needs zip codes, based on the economic factor, have just under 3/4 of all cases and the rate of cases in the highest need zip codes is three times higher than the low needs zip codes. The rate in the high needs zip codes is double the rate in the low needs zip codes.

High needs zip codes, based on the housing and transportation factor, have 4/5 of all cases and the rate of cases in the highest need zip codes is more than three times higher than the low needs zip codes. The rate in the high needs zip codes is double the rate in the low needs zip codes.

Confirmed case counts on Oct 30, Nov 7, and Nov 15, 2020 have been added to the analysis. On the left side of the map below circle size represents the number of new confirmed cases by zip code between 10/21 and 10/30. Similarly, circles on the right side are new confirmed cases by zip code between 10/30 and 11/07. Swipe the center bar to see more or less of either map. Zoom, pan, and view pop-up information for these maps.

The new data confirms the patterns shown with the earlier data. The average number of confirmed cases in the highest needs zip codes are twice as many as in the high needs zip codes and nearly 7 times more than the number in low needs zip codes.

The low needs zip codes contain 32% of the population and 30% of the households but less than 25% of the confirmed COVID-19 cases. The highest needs zip codes contain 12% of the population and 11% of the households but around 20% of the confirmed COVID-19 cases.

On average, a zip code in Erie County had :

• 8 new cases between Oct 21 - Oct 30 and,

• 13 new cases between Oct 30 - Nov 7 and,

• 21 new cases between Nov 7 - Nov 15.

During the first period, only one of the 6 (17%) zip codes with over 20 new cases is a low needs zip code. During the second period, 4 of the 14 (29%) zip codes with 20 or more new cases are low needs zip codes. The recent spike resulted in 8 of the 25 (32%) zip codes with 20 or more new cases are low needs zip codes.

The Social Determinants of Health model does not consider age and occupation variables. However, these characteristics may play a role in the spread of the coronavirus as well as in the severity of response. Additional Census data were obtained from the ACS to examine age and occupation to confirmed COVID-19 Cases.

Census data from the American Community Survey (2018, 5 yr. estimates) on age and occupation have been added to data on confirmed COVID-19 cases on 4 dates (10/21, 10/30, 11/07, and 11/15) for zip codes in Erie County NY.  

Table B01001: SEX And AGE was used to create 3 age categories, namely

• Age 5 - 17 to represent school age children
• Age 18 - 24 to represent young adults, typically college age, and
• Age 60 plus - to represent the population that seems to suffer most from the pandemic.

Additionally, subject table S2401 was used to obtain information related to specific occupations. The universe for these occupation statistics is Civilian population, 16 years and over. The occupation data:

• Healthcare practitioners and technical occupations in Management, business, science, and arts occupations (S2401_C01_015)
• Healthcare support occupations in Service occupations (S2401_C01_019), and 
• Protective service occupations in Service occupations (S2401_C01_020)

are combined into a single front line, first responders occupation under the assumption that these people will be more exposed to the virus on a regular basis. Please note that some zip codes do not have reliable counts of detailed occupations.

The two charts below compare the normalized values for the number of front line, first responders to the SDOH Index and to the normalized number of COVID-19 cases on Oct 21.

The relationship between occupation and needs is weaker than the relationship between front line occupations and COVID-19 cases. There is a large cluster of points in the low cases, below average number of residents who are front line, first responders. The relationship weakens outside of this quadrant.

The strongest relationships with the SDOH Index and COVID-19 cases is with the young adult age group.

The importance of the presence of young adults within a zip code and the number of COVID-19 cases may be seen in the summary table and charts below. Throughout each time period, zip codes with greater than average number of young adults have the highest average number of confirmed COVID-19 cases. Zip codes with fewer than average number of young adults have the lowest average number of confirmed COVID-19 cases. As mentioned in the media, this age group may be more likely to frequent bars and house parties.

The stronger relationship between the COVID-19 cases and Young Adult population is evident when considering the zip codes with higher than average number of cases per 1,000 people. As shown below, nearly 60% of all of the zip codes with higher than average number of cases per 1,000 people have higher than average number of young adults.

These maps show the spatial relationship between each of the four new variables and the number of confirmed cases. COVID-19 Cases are shown for two dates, Oct 21 and Nov 15.

A dashboard of this dataset for the first three dates may be found  here .