Investigators analyzed hospitalizations for pneumonia and sepsis drawn from the Healthcare Cost and Utilization Project State Inpatient Databases across seven U.S. states, covering 2018–2021. The sample included 3,949,906 hospital admissions identified in that period. The authors reported the study design as retrospective and adult-focused; the abstract did not report detailed patient-level demographics beyond race categories or antiretroviral treatment data. Hospital-level clustering was incorporated into the analytical approach.
The study team compared crude in-hospital mortality by HIV status and race, then estimated adjusted associations using multilevel logistic regression. Models included adjustments for demographics, comorbidities, acute organ failures present on admission, do-not-resuscitate (DNR) status, surgical status, and hospital-level clustering. The investigators explicitly tested effect modification by race through an HIV × race interaction term and reported a significant interaction (P = 0.0033). The abstract did not provide additional modeling diagnostics or model fit statistics.
Crude in-hospital mortality was lower among people with HIV than among patients without HIV: 8.4% versus 10.2%. The absolute risk difference was 1.8% (95% CI, 1.6–2.0). Those unadjusted percentages reflect population averages across the included states and the two diagnostic categories (pneumonia and sepsis); the abstract does not disaggregate crude rates by specific race groups in the text.
Data indicated heterogeneity in the association between HIV and mortality when stratified by race. After multivariable adjustment, HIV infection was associated with increased odds of in-hospital death for patients identified as White (adjusted odds ratio [aOR], 1.09; 95% CI, 1.03–1.16) and for patients identified as Asian (aOR, 1.53; 95% CI, 1.13–2.08). The adjusted association between HIV and mortality was not statistically significant among patients identified as Black or Other race, according to the abstract. The authors also reported that racial disparities differed depending on HIV status when they evaluated race-specific comparisons separately among patients with and without HIV.
The study team observed that average, population-level estimates conceal important heterogeneity across racial groups. Investigators reported a significant HIV × race interaction, which they interpreted through an intersectionality lens: that is, the interplay of race and HIV status affected pneumonia- and sepsis-related mortality in nonuniform ways. The abstract notes that an intersectionality framework could provide insight into structural mechanisms behind these disparities among people with HIV, though the abstract did not specify which structural factors were measured or how they were operationalized in the dataset.
The abstract presents key findings but leaves several questions open. The source did not report, within the abstract, granular patient-level demographic breakdowns (for example, age distribution by race and HIV status), measures of HIV disease control such as CD4 count or viral load, or antiretroviral therapy status. The abstract also did not provide details about how race was categorized at the point of care or whether ethnicity and social determinants of health were included in adjusted models. Investigators adjusted for comorbidities and acute organ failures present on admission, DNR status, and surgical status, but the abstract does not list the specific comorbidity definitions or which organ failures were coded. Finally, the authors noted differences in racial disparities by HIV status and recommended applying an intersectionality framework to further examine structural drivers; the abstract did not specify next analytic steps, causal mechanisms, or policy interventions.
The study contributes a large, multistate hospital-based analysis indicating that the relationship between HIV and in-hospital mortality for pneumonia and sepsis varies by race, with statistically significant adjusted increases in mortality among White and Asian patients but not among Black or Other race groups. Key methodological elements and primary effect estimates (sample size, crude mortality, interaction P value, and select race-specific adjusted odds ratios) are reported in the abstract. Additional patient-level clinical detail and operational definitions were not provided in the abstract and would require consultation of the full manuscript for deeper evaluation.