Briefs Share

Copy or email the link to share Race, Comorbidities, and COVID-19.

COPY

EMAIL

Race, Comorbidities, and COVID-19

Posted on

As coronavirus cases mount, racial minorities appear to be disproportionately affected by the pandemic. Across a sample population of 23 million patients across 16 states, 12% were black and 65% were white, but among the 108,000 of those patients who are documented COVID-19 positive, 22% were black and 46% were white.

Expand
Table 1: Distribution of comorbidities by race for all sampled patients, COVID-19 positive hospitalized patients, and COVID-19 related deaths for patients aged 19-64 years. The “other” category contains the following groups: patients with a race documented as “other” in the EHR; patients without a documented race; and patients with a documented race that is not white or black, which have low sample sizes. The “other” category contains the following groups: patients with a race documented as “other” in the EHR; patients without a documented race; and patients with a documented race that is not white or black, which have low sample sizes.

The root causes of this disparity are not yet fully understood, but recent electronic health record data suggest that comorbidities could play a role. This analysis begins to examine how outcomes compare for patients of different races who share other health characteristics, such as common comorbidities.

In general, black patients were overrepresented among COVID-19 hospitalizations (27%) and deaths (22%) when compared to the broader sample population (12%). When examining only patients with no comorbidities, the gap narrowed. Among patients with no comorbidities, 17% of COVID-19 hospitalizations, 15% of COVID-19 deaths, and 12% of sampled patients were black.

Across all racial groups, patients who were hospitalized or who died had higher rates of comorbidities than the overall sample population. Specifically, the combination of diabetes and hypertension is much more prevalent among hospitalized COVID-19 patients (23%) and among those who died (28%) than in the sample population (5%), regardless of racial group. This pattern of comorbidity prevalence is consistent with past observations.

The data also indicate that black patients experience comorbidities at higher rates in general. For example, 6.7% of black patients ages 19-64 had diabetes and hypertension, compared to 3.6% of white patients. It isn’t yet clear whether the comorbidities themselves lead to worse COVID-19 outcomes for these patients, or whether there are other variables related to both the comorbidities and poor outcomes from COVID-19.

There are 42,388 (39%) COVID-19 patients who do not yet have a known outcome (either survived or died) as of May 4, 2020. Some of these patients may go on to be hospitalized or die, and this should be taken into consideration when examining these data.

The patients in the sample population have all had an interaction with the health system in the past 2 years. Thus, the sample population concentrates on patients with a certain baseline of access to healthcare and a relatively up-to-date, accurate list of diagnoses in their charts. There may be additional health impacts for patients who have limited access to healthcare, and these factors are not accounted for in this analysis.

Further, the analysis does not control for the various factors that might lead to increased comorbidities in specific racial groups. Future studies should explore the effects of variables like geographical location, population density, and socioeconomic status.

The data in this analysis were pooled from 27 health systems representing 240 hospitals that span 16 different states and cover 107 million patients as of May 4, 2020.

Expand
Table 2: Distribution of comorbidities by race for all active patients, COVID-19 positive hospitalized patients, and COVID-19 related deaths for patients aged 65+ years.
Expand
Figure 1: A visual representation of the distribution of comorbidities by race for all sampled patients, COVID-19 positive hospitalized patients, and COVID-19 related deaths for patients aged 19-64 years.
Expand
Figure 2: A visual representation of the distribution of comorbidities by race for all sampled patients, COVID-19 positive hospitalized patients, and COVID-19 related deaths for patients aged 65+ years.
Expand
Table 3: Comorbidities by race compared to national baselines.  

4 Comments

  1. Can you clarify the % of blacks among those without chronic diseases? Seeing that blacks were 12% of the total population but had higher rates of HTN and DM, it doesn’t seem possible that they would also be 12% of the population with no comorbidities.

    Please let me know if this is an error.

    THanks for your good work in putting this information in the public domain

    Jeff Levin-Scherz
    Assistant Prof, Harvard Chan School of Public Health
    jlevin@hsph.harvard.edu

    1. Good question. The percentages are correct. One reason for the similarity might be that nearly ¾ (74.6%) of the sampled population did not have comorbidities. Here are the raw numbers:

      Total N = 27,906,037 for the sampled population:
      • 3,382,960 were black (12%)
      • 18,232,516 were white (65%)
      • 6,290,561 were of another race (23%)

      20,825,809 of the sampled population had no comorbidities:
      • 2,395,071 of those patients who had no comorbidities were black (12%)
      • 13,059,117 of those patients who had no comorbidities were white (63%)
      • 5,371,621 of those patients who had no comorbidities were of another race (26%)

  2. The tested population may deviate significantly from the general population due to CDC screening criteria used to decide whether to test or not. Would the tested NEGATIVE population be a better comparison group vs positive, positive admitted and positive deceased?

    1. COVID-19 testing practices have varied across organizations in terms of availability of tests, criteria for being tested, location and time available for tests, and other factors. Our data scientists were concerned that these inconsistencies would introduce selection bias. Here we controlled for selection bias by focusing on COVID-19 cases resulting in admissions or deaths, since such cases are the most severe.

Leave a Comment

Your email address will not be published. Required fields are marked *. View our comment guidelines.