The forecasted prevalence of comorbidities and multimorbidity in people with HIV in the United States through the year 2030: A modeling study

Abstract

Background: Estimating the medical complexity of people aging with HIV can inform clinical programs and policy to meet future healthcare needs. The objective of our study was to forecast the prevalence of comorbidities and multimorbidity among people with HIV (PWH) using antiretroviral therapy (ART) in the United States (US) through 2030.

Methods and findings: Using the PEARL model-an agent-based simulation of PWH who have initiated ART in the US-the prevalence of anxiety, depression, stage ≥3 chronic kidney disease (CKD), dyslipidemia, diabetes, hypertension, cancer, end-stage liver disease (ESLD), myocardial infarction (MI), and multimorbidity (≥2 mental or physical comorbidities, other than HIV) were forecasted through 2030. Simulations were informed by the US CDC HIV surveillance data of new HIV diagnosis and the longitudinal North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD) data on risk of comorbidities from 2009 to 2017. The simulated population represented 15 subgroups of PWH including Hispanic, non-Hispanic White (White), and non-Hispanic Black/African American (Black/AA) men who have sex with men (MSM), men and women with history of injection drug use and heterosexual men and women. Simulations were replicated for 200 runs and forecasted outcomes are presented as median values (95% uncertainty ranges are presented in the Supporting information). In 2020, PEARL forecasted a median population of 670,000 individuals receiving ART in the US, of whom 9% men and 4% women with history of injection drug use, 60% MSM, 8% heterosexual men, and 19% heterosexual women. Additionally, 44% were Black/AA, 32% White, and 23% Hispanic. Along with a gradual rise in population size of PWH receiving ART-reaching 908,000 individuals by 2030-PEARL forecasted a surge in prevalence of most comorbidities to 2030. Depression and/or anxiety was high and increased from 60% in 2020 to 64% in 2030. Hypertension decreased while dyslipidemia, diabetes, CKD, and MI increased. There was little change in prevalence of cancer and ESLD. The forecasted multimorbidity among PWH receiving ART increased from 63% in 2020 to 70% in 2030. There was heterogeneity in trends across subgroups. Among Black women with history of injection drug use in 2030 (oldest demographic subgroup with median age of 66 year), dyslipidemia, CKD, hypertension, diabetes, anxiety, and depression were most prevalent, with 92% experiencing multimorbidity. Among Black MSM in 2030 (youngest demographic subgroup with median age of 42 year), depression and CKD were highly prevalent, with 57% experiencing multimorbidity. These results are limited by the assumption that trends in new HIV diagnoses, mortality, and comorbidity risk observed in 2009 to 2017 will persist through 2030; influences occurring outside this period are not accounted for in the forecasts.

Conclusions: The PEARL forecasts suggest a continued rise in comorbidity and multimorbidity prevalence to 2030, marked by heterogeneities across race/ethnicity, gender, and HIV acquisition risk subgroups. HIV clinicians must stay current on the ever-changing comorbidities-specific guidelines to provide guideline-recommended care. HIV clinical directors should ensure linkages to subspecialty care within the clinic or by referral. HIV policy decision-makers must allocate resources and support extended clinical capacity to meet the healthcare needs of people aging with HIV.

Fig 1

Schematic representation of (A) the PEARL model and (B) the risk factors and comorbidities with high prevalence in people with HIV using ART. (A) PEARL model simulating people with HIV using ART in the United States. Footnotes: HIV Surveillance data was sourced from the US Centers for Disease Control and Prevention’s HIV Surveillance Reports, available at https://www.cdc.gov/hiv/library/reports/hiv-surveillance.html. The NA-ACCORD data was available after the collaboration approved our submitted concept sheet (https://naaccord.org/collaborate-with-us). (B) Schematic of the risk factors and comorbidities with high prevalence in people with HIV using ART. Footnotes: ART = antiretroviral therapy (HIV treatment). CD4 = CD4 T-lymphocyte cell count. Details on the mathematical functions represented by the arrows between the risk factors and mental and physical comorbidities can be found at PEARLHIVmodel.org.

Fig 2

Forecasted^a number of PWH using ART in the US and forecasted prevalence of mental and physical comorbidities and multimorbidity among PWH using ART in the US (A) overall and (B) by subgroup^b (A) Overall. (B) By subgroup^b Footnotes: ≥1 Ment. = anxiety and/or depression (i.e., ≥1 of the mental comorbidities included) ≥2 Phys. = physical multimorbidity, defined as ≥2 physical comorbidities ≥2 Any = mental or physical multimorbidity, defined as ≥2 physical or mental comorbidities ≥1 Ment. and 2 Phys. = mental comorbidity and physical multimorbidity, defined as ≥1 mental comorbidity and ≥2 physical comorbidities. ^aAlthough these estimates are all PEARL forecasts, 2010 was during the calibration period (where observed NA-ACCORD data were available to inform the estimates) and 2020 and 2030 were forecast periods (without observed NA-ACCORD data). ^bNote that the y axes are different across the subgroups to allow visualization of the number of comorbidities within each year. ART, antiretroviral therapy; Black/AA, Black/African American; NA-ACCORD, North American AIDS Cohort Collaboration on Research and Design; PWH, people with HIV; US, United States.

Fig 3

Forecasted prevalence (and shaded 95% uncertainty ranges) of individual comorbidities among PWH using ART (A) overall, (B) among the 15 subgroups, (C) among the subgroup with the oldest median age in 2030, and (C) among the subgroup with the youngest median age in 2030. (A) Forecasted prevalence (and shaded 95% uncertainty ranges) of comorbidities among all PWH to the year 2030. (B) Forecasted prevalence (and shaded 95% uncertainty ranges) of individual comorbidities, within the 15 subgroups. Footnotes: CKD, stage ≥3 chronic kidney disease; ESLD, end-stage renal disease; MI, myocardial infarction. The 95% credibility interval is estimated as the 2.5% and 97.5% range of results from running the simulation 200 times.

Fig 4

Forecasted absolute percentage point change (blue = decrease, red = increase) in the prevalence of individual comorbidities from 2020 to 2030, by subgroup. Footnotes: ^bThe y axes are different across the subgroups to allow visualization of the number of comorbidities within each year.

Fig 5. The relative difference of the proportion with physical multimorbidity in 2030 [outcome] comparing scenarios in which comorbidity incidence was decreased by 25% (down arrow scenario) and increased by 25% (up arrow scenario) to assess the influence of estimated probabilities on prevalence estimates.

Relative difference when probability of the incidence of a comorbidity was decreased by 25% and increased by 25%, compared to the baseline scenario (no modification to the probability of the incidence of a comorbidity). Footnotes: ↑relative difference = (% with physical multimorbidity _{increase scenario}—% with physical multimorbidity _{baseline scenario}) % with physical multimorbidity _{baseline scenario} ↓relative difference = (% with physical multimorbidity _{decrease scenario}—% with physical multimorbidity _{baseline scenario}) % with physical multimorbidity _{baseline scenario} Physical multimorbidity = ≥2 physical comorbidities.