Racial/Ethnic and Income Disparities in the Prevalence of Comorbidities That Are Associated With Risk for Severe COVID-19 Among Adults Receiving HIV Care, United States, 2014–2019

John K. Weiser, MD, MPH; Yunfeng Tie, PhD; Linda Beer, PhD; Robyn Neblett Fanfair, MD, MPH; Roy Luke Shouse, MD, MPH

Disclosures

J Acquir Immune Defic Syndr. 2021;86(3):297-304. 

In This Article

Methods

Design and Data Collection

The Medical Monitoring Project (MMP) is an annual cross-sectional survey designed to produce nationally representative estimates of the behavioral and clinical characteristics of adults with diagnosed HIV in the United States. This analysis presents estimates of adults receiving HIV care drawn from this sample. MMP data collection constitutes routine public health surveillance and was thus determined by the CDC to be nonresearch. This activity was conducted consistent with the applicable federal law and CDC policy.[29] When required, participating states or territories obtained local institutional review board approval to collect data. All participants provided informed consent. MMP uses 2-stage sampling in which, during the first stage, 16 states and one territory, including 6 separately funded metropolitan areas within selected states, were sampled from all states, the District of Columbia, and Puerto Rico. During the second stage, simple random samples of people with diagnosed HIV aged ≥18 years were drawn annually for each participating area from the National HIV Surveillance System, a census of people with diagnosed HIV in the United States. People sampled during the 2016–2018 data collection cycles were recruited for a phone or face-to-face interview and medical record abstraction at their usual place of HIV medical care. Data were collected beginning in June of each cycle until the following May. Medical record data included selected laboratory test results and all diagnoses recorded in the problem list or assessment section of medical encounters during a 2-year, retrospective observation period ending on the interview date. For the 2016–2018 data collection cycles, this period spanned June 2014–May 2019.

All sampled areas and separately funded jurisdictions within states participated in the MMP, including California (including Los Angeles County and San Francisco), Delaware, Florida, Georgia, Illinois (including Chicago), Indiana, MI, Mississippi, New Jersey, New York (including New York City), North Carolina, Oregon, Pennsylvania (including Philadelphia), Puerto Rico, Texas (including Houston), Virginia, and Washington. Annual response rates for adults with diagnosed HIV were 44%–45%. During the 2016–2018 data collection cycles, 12,310 sampled people were interviewed (Figure 1). Of these, 12,189 reported receiving HIV medical care or had medical record documentation of at least one element of HIV care (eg, an HIV laboratory result or antiretroviral prescription) during the past 24 months. We excluded 627 participants without a medical record abstraction containing at least one diagnosed condition. The analytic data set included 11,562 participants who received HIV medical care in the past 24 months and had diagnosis data abstracted (referred to as "adults receiving HIV care").

Figure 1.

Sample disposition for this analysis, the Medical Monitoring Project, 2016–2018 data collection cycles.

Measures

Sociodemographic variables included age, race/ethnicity (non-Hispanic American Indian/Alaska Native, Asian, Black, Native Hawaiian/other Pacific Islander, White, or multiracial, and Hispanic of any race), household income above or below the federal poverty threshold during the past 12 months, and health insurance or coverage type during the past 12 months [any private insurance, public insurance alone, Ryan White HIV/AIDS Program (RWHAP) assistance alone, unspecified, and no insurance/coverage]. We ascertained whether facilities identified by participants as their usual place of HIV care received any funding from the RWHAP.[30] Clinical variables included diagnosed comorbidities identified by the CDC on July 15, 2020, as associated with an increased risk of severe illness from COVID-19, defined as requiring hospitalization (obesity, CKD, diabetes, COPD, immunocompromised from solid organ transplantation, heart disease, and sickle cell disease)[16] as well as diagnosed AIDS-defining illnesses.[31] We ascertained obesity from recorded diagnoses of obesity, without consideration of body mass index. Because medical records often do not differentiate type 1 and type 2 diabetes,[32] we measured diagnosed diabetes, irrespective of type. Diagnosed heart disease included atherosclerotic heart disease, congestive heart failure, cardiomyopathy, pulmonary hypertension or cor pulmonale, and arrhythmia. Other clinical variables included CD4+ T-lymphocyte cell count (CD4 count) and HIV RNA level.

Outcomes

The primary outcome was having ≥1 diagnosed comorbidity associated with an increased risk for severe illness from COVID-19 recorded in the medical record during the 24 months before being interviewed. Because the risk increases with age[33] and possibly with immunocompromised state secondary to HIV, secondary outcomes included having ≥1 diagnosed comorbidity along with combinations of older age (≥50 or ≥65 years) and poorly controlled HIV, defined as lowest CD4 count <200 cells/μL, any HIV RNA ≥200 copies/mL, or diagnosis of an AIDS-defining illness during the past 24 months.

Statistical Analysis

Data were weighted based on known probabilities of selection at the state or territory and person levels, adjusted for nonresponse, and poststratified to known population totals by age, race/ethnicity, and sex from the National HIV Surveillance System. This design allows inference to all adults with diagnosed HIV in the United States.[27]

We estimated the prevalence, with 95% confidence intervals (CIs), of having ≥1 diagnosed comorbidity associated with an increased risk for severe illness from COVID-19 by race/ethnicity, income, type of health insurance, and RWHAP facility funding. We used logistic regression with predicted marginals to estimate prevalence differences (PDs), with corresponding 95% CIs and t test P values, for having ≥1 comorbidity by race/ethnicity, income, health insurance type, and RWHAP facility funding.

Because the median age of Whites was higher than for other race/ethnicities and older adults are more likely to have diagnosed comorbidities, we adjusted the association between race/ethnicity and having ≥1 diagnosed comorbidity for age. To assess whether the association between race/ethnicity and having ≥1 diagnosed comorbidity differed by income level, we added poverty as an interaction term to the race/ethnicity model. For income, health insurance type, and RWHAP-funding status, we reported unadjusted measures of association. We considered a PD ≥5 percentage points to be meaningful from a public health perspective.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....