Updated Estimates of Chronic Conditions Affecting Risk for Complications from Coronavirus Disease, United States

We updated estimates of adults at risk for coronavirus disease complications on the basis of data for China by using recent US hospitalization data. This update to our previous publication substitutes obesity for cancer as an underlying condition and increases adults reporting any of the conditions from 45.4% to 56.0%.

were considered to be at risk for hospitalization for COVID-19 because of an underlying condition. New measures included the risk factors of ever smoking 100 cigarettes; sedentary lifestyle, defined as no leisure time physical activity in the past month; and inadequate fruit and vegetable consumption, defined as consuming the combination <5 times/day on the basis of responses to 5 questions. A measure of the number of risk factors (0-3) was also created. Demographic measures were the same as in the previous report (1) with the addition of census region (Northeast, Midwest, South, or West) (4).
Stata version 14.1 (StataCorp LP, https://www. stata.com) was used for analysis to account for the complex sample design of the Behavioral Risk Factor Surveillance System. Point estimates and 95% CIs are reported by using the weights, stratum, and PSU variables supplied in the dataset (3). Missing values were excluded from analysis.
The sample was described previously (1). Prevalence rates of the separate conditions were 8.5% for cardiovascular disease, 6.6% for chronic obstructive pulmonary disease, 9.1% for asthma, 10.8% for diabetes, 32.4% for hypertension, and 30.1% for obesity. We provide updated results (Table 1, https://wwwnc. cdc.gov/EID/article/26/9/20-2117-T1.htm) based on US hospitalizations, along with the results previously published (1), with the difference being the substitution of obesity for cancer as an underlying condition. Overall, 56.0% (95% CI 55.7%-56.4%) of respondents had >1 underlying conditions using the updated definition, compared with 45.4% using the previous definition, an increase of 23% over the previous rate. For the updated measure, 49.2% of employed or selfemployed adults reported an underlying condition, and 18.7% of employed adults reported >2 underlying conditions. These results compare with 26.6% of all adults reporting >2 underlying conditions. Although the percentage of adults with any of the conditions increased with age (Table 1), 60.7% of those with underlying conditions were <60 years of age when using the new definition, compared with 53.4% when using the earlier definition.
Prevalence rates for the risk factors were 40.4% for ever smoking, 26.6% for sedentary lifestyle, and 84.1% for inadequate fruit and vegetable consumption. Adults with each risk factor (or more risk factors) were more likely than those without the risk factor (or fewer) to report any underlying condition (Table 1). Among ever smokers, 40.7% currently smoke, resulting in a current smoking rate of 16.5%. Updated state rates ranged from 45.6% in the District of Columbia to  (6). Taken together, these results suggest that risk stratification based on age or number of underlying conditions might be considered as a means of more safely phasing in returning to work. All the underlying conditions used in our updated measure are conditions for which behavioral risk factors have been well established (7,8). In unadjusted results each selected risk factor was associated with increased likelihood of reporting any of the 6 underlying conditions. In addition, the results for increasing number of risk factors indicate a stepwise increase in the percentage of adults reporting any of the underlying conditions with each additional risk factor (Table 1). These results suggest the potential for lowering risk for COVID-19 hospitalizations by reducing any or all of these 3 risk factors. In addition, being older, male, or African American also increased the likelihood of reporting an underlying condition, which are all groups that hospitalization data (2) suggested were disproportionately affected by COVID-19. These results suggest that observation might be caused by increased rates of underlying conditions among these groups. In addition, living in either the Midwest or South increased the likelihood of reporting an underlying condition compared with living in the West. This result is consistent with studies showing that obesity rates are also highest in these regions (9).
Our study does not address possible differences in contracting the disease, only the risk for hospitalization among those with COVID-19, based on US results for underlying conditions (2). Because only noninstitutionalized adults were surveyed, 1.3 million adults in nursing homes (10) were excluded, which almost certainly underestimates those with underlying conditions who were included in hospitalization data. Data are self-reported, and reliability and validity can vary for different measures tested (11). However, as long as a respondent was told they had a chronic condition, validity has been shown to be high. Age groups used for analysis did not match those used for weighting data, but that factor should have a minimal effect on results. Low response rates could introduce bias but, as noted, validity appears high for most measures used in this study.
We estimate 56.0% of US adults are at risk for needing hospitalization for COVID-19 because of underlying conditions, representing a 23% increase from the 45.4% earlier estimates, which excluded obesity. These underlying conditions are, in turn, associated with modifiable risk factors, including ever smoking, being sedentary, and inadequate fruit and vegetable consumption. These results suggest the potential for policies for opening businesses based on risk stratification of the population and for possible improvement of risk status through lifestyle change. A national focus on, and support for, a health promotion campaign would be timely.
Collection, analysis, and interpretation of data for this study were supported by the Centers for Disease Control and Prevention Grant/Cooperative Agreement no. 1U58DP006069-01.

About the Author
Ms. Adams is a consultant at On Target Health Data LLC, Suffield, CT. Her primary research interest is chronic diseases (including dementia) and their risk factors.