This population-based retrospective cohort study evaluated how the availability of COVID-19 vaccines related to differences in COVID-19 mortality among people with HIV (PWH) in Florida. The authors compared two study intervals: the pre-vaccine period (March 1, 2020–April 30, 2021) and the vaccine availability period (May 1, 2021–December 31, 2021). The investigation aimed to describe overall mortality trends and to assess subgroup differences across demographic and community-level characteristics.
Primary data came from the Florida Department of Health’s enhanced HIV/AIDS Reporting System (eHARS). The initial dataset included 129,770 PWH. After applying inclusion and exclusion criteria and linking to community data, the analytic cohort consisted of 122,527 PWH (94.4% of the initial sample). Included individuals were aged 18 years or older, residing in Florida by December 31, 2019, and alive during at least part of the study period. Vital status and causes of death were verified through routine linkages to FDOH Vital Records, the Social Security Death Master File, and the National Death Index. The deidentified dataset used for analyses was linked in June 2022.
Exclusions comprised a small number of individuals over age 90 (n = 131) removed for potential data-quality concerns and 4,340 people with invalid or missing ZIP Code Tabulation Area (ZCTA) information after linkage to community-level social vulnerability data. Additionally, 2,682 decedents who died outside Florida or lacked a valid ZCTA at death were excluded from analyses that required geographic linkage.
Researchers linked individual records to ZIP Code Tabulation Areas (ZCTAs) to incorporate community-level measures. The primary community-level exposure was the Social Vulnerability Index (SVI) at the ZIP code level. Rural-Urban Commuting Area (RUCA) codes were also incorporated to capture urban–rural context. Individual-level covariates available in eHARS included age, sex, race/ethnicity, place of birth, HIV transmission mode, and viral suppression status. The study reports adjusted analyses that account for these variables when evaluating subgroup differences in COVID-19 mortality.
The authors used a competing risks modeling approach to compare COVID-19 mortality among PWH between the pre-vaccine and vaccine availability periods. The pre-vaccine period was defined as March 1, 2020 through April 30, 2021. The vaccine availability period was defined as May 1, 2021 through December 31, 2021. Analyses examined overall mortality rates per 100,000 person-months and estimated adjusted hazard ratios (HRs) with 95% confidence intervals for subgroup comparisons, including race/ethnicity and SVI categories.
Overall COVID-19 mortality rates among PWH increased from the pre-vaccine period to the vaccine availability period. Reported rates were 22.3 per 100,000 person-months before vaccine availability and 27.5 per 100,000 person-months during the vaccine availability period. The authors note that increased mortality during the vaccine availability period was driven in part by the Delta variant surge in mid-to-late 2021.
Racial and ethnic disparities in COVID-19 mortality persisted across periods but showed variation by group. Compared with Non-Hispanic White (NHW) PWH, Non-Hispanic Black (NHB) PWH had higher adjusted hazards of COVID-19 mortality both before and during vaccine availability: HR 2.06 (95% CI 1.45–2.91) in the pre-vaccine period and HR 1.60 (95% CI 1.13–2.27) during vaccine availability. For Hispanic PWH, the adjusted hazard relative to NHW was elevated in the pre-vaccine period (HR 2.24, 95% CI 1.55–3.23) but attenuated during vaccine availability (HR 1.03, 95% CI 0.68–1.57).
In the final adjusted model, residence in ZIP Code-level areas with high overall Social Vulnerability Index (SVI) was associated with increased COVID-19 mortality hazard during the vaccine availability period (HR 1.75, 95% CI 1.04–2.97). The study incorporated RUCA codes to consider urban–rural context, and the authors discuss prior findings that vaccination sites were more concentrated in urban counties, potentially influencing access in rural areas. The analyses indicate that community-level social vulnerability remained an important correlate of mortality patterns among PWH during the vaccine availability period.
The study documents higher overall COVID-19 mortality among PWH during the vaccine availability period compared with the earlier pandemic period, with important heterogeneity across subgroups. Persistent elevated hazards among NHB PWH and the association with high SVI areas underscore ongoing inequities. The authors emphasize that these are observational comparisons and do not estimate the causal effect of individual vaccination uptake.
Limitations noted by the authors include necessary exclusions for missing ZCTA data and deaths occurring outside Florida, and the restricted availability of individual-level vaccination data in the surveillance dataset (the study does not provide direct measures of vaccination status). The authors recommend tailored vaccination strategies and structural interventions to advance equity in pandemic response for PWH, highlighting that improving vaccine accessibility and addressing community-level vulnerabilities remain essential.
Further details on analytic code (SAS programs) and variable definitions are publicly available in the study’s GitHub and Zenodo repositories, and data requests for deidentified surveillance data are subject to Florida Department of Health data-use procedures as described by the authors.