Retention in Care Among People Living with HIV: The Syndemic Effect of Substance Abuse, Depression, and Type 2 Diabetes and the Impact of the COVID-19 Pandemic

It has been over forty years since the first cases of acquired immune deficiency syndrome (AIDS) were reported in 1981. Despite major progress in human immunodeficiency virus (HIV) treatment, prevention, care, and support services to mitigate the global impact of the HIV/AIDS epidemic, retention in HIV care (RIC), an important indicator of HIV prevention and control programs, remains a major challenge. According to the latest estimates from the United States (US), Less than 60% of Americans aged ≥13 years with diagnosed HIV were retained in care in 2019. Previous research has found that mental health conditions and chronic illnesses are common comorbidities among people living with HIV (PLHIV) and are associated with poor HIV outcomes, including reduced rates of RIC. Importantly, substance abuse (SA), depression, and type 2 diabetes (DM2) often co-occur among PLHIV, but little is known about the cumulative and interactive effects of these conditions on RIC. Syndemic (or synergistic epidemics) theory has become a well-established framework for the explanation of how adverse health conditions interact together within social and structural inequalities, which, cumulatively affect health outcomes. Furthermore, the recent emergence of the coronavirus disease 2019 (COVID-19) pandemic has been accompanied by widespread social, economic, and healthcare system disruptions, exposing, and often exacerbating health inequalities and adding to the challenges faced by PLHIV when seeking HIV care services. While syndemic theory has been widely applied in HIV research, this framework has been rarely utilized for RIC. Therefore, guided by syndemic theory, this dissertation has two objectives: 1) to examine the syndemic effect of SA, depression, and DM2 on RIC and 2) to evaluate the impact that the COVID-19 pandemic has had on RIC. These two objectives comprise the two manuscripts of this dissertation, respectively. This dissertation uses a retrospective cross-sectional study among PLHIV who were seen for HIV care at the Penn State Health (PSH) Milton S. Hershey Medical Center (HMC) in Central Pennsylvania (PA). Patients were included in the analysis if they: 1) initiated HIV care before March 2019 (study baseline), 2) had ≥1 HIV medical care encounter during the pre-COVID-19 year (3/1/2019 to 2/28/2020), 3) were ≥18 years old at baseline, and 4) identified as being active in care at the end of the first year of the pandemic (2/28/2021). RIC, the primary outcome of the two objectives, was defined according to the Health Resources and Services Administration HIV/AIDS Bureau (HRSA/HAB) and the Centers for Disease Control and Prevention (CDC). Patients were considered to be retained in care if they had ≥2 HIV medical care encounters at least 90 days apart within the 12-month measurement year (the pre-COVID-19 year for objective 1 and the first year of the COVID-19 pandemic for objective 2). An HIV medical care encounter was defined as documentation of at least one provider visit (either in-person or virtual), one CD4 cell counts test, or one viral load (VL) test. The predictor variables, including demographics, HIV-related factors, and syndemic factors (SA, depression, and DM2) were captured from two datasets: Path To Health dataset (PaTH) and QuickData dataset (QuickData), the latter is a database used by the HIV clinical program. For the first objective, the study duration was restricted to the pre-COVID-19 year to avoid the potential effects of the COVID-19 pandemic on the study outcome (RIC). I examined four aims in the first manuscript. First, I evaluated the co-occurrence of the syndemic factors (SA, depression, and DM2) among the study sample. Second, I investigated the association between each of the syndemic factors and RIC. Third, I assessed the additive effects of the syndemic factors on RIC. And fourth, I tested whether the syndemic factors interact to exacerbate RIC. To test the syndemic effect of SA, depression, and DM2 on RIC (third and fourth aims), I used two approaches. In the first approach (aim 3), I applied the traditional method of testing syndemic factors as cumulative adversities by using the sum score specification (the number of syndemic factors), for which a syndemic index (range from 0= no syndemic factors to 3=all three syndemic factors) was created. The syndemic index was evaluated once as a categorical variable (with 0 as the reference category), and once as a continuous exposure, which evaluates the linear dose response by regressing the outcome on the syndemic index. In the second approach (aim 4), I used the recommended method of applying a series of linear regression models and logistic regression models with product terms to test for the synergistic/interaction effect of SA, depression, and DM2 on both additive and multiplicative scales, respectively. The second objective had three aims: 1) to assess factors associated with RIC during the first year of the pandemic (during the pandemic), 2) to evaluate determinants of changes in RIC during the pandemic (lost RIC and improved RIC) compared to the one year preceding the pandemic (pre-pandemic), and 3) to test the syndemic effect of SA, depression, and DM2 on RIC during the pandemic. I first performed descriptive analysis of the study sample and then conducted multiple backward logistic regression to assess factors associated with RIC during the pandemic (aim 1) and to evaluate determinants of lost RIC and improved RIC during the pandemic compared to the pre-pandemic RIC (aim 2). Unadjusted and adjusted odds ratios (OR) and 95% confidence intervals (CI) were reported for all predictor variables. For aim 3, I used the same two approaches of the syndemic index and the interaction effect (previously described) for testing the syndemic effect of SA, depression, and DM2 on RIC during the pandemic. In the first manuscript, 440 (70.9%) of the total sample (n=621) were retained in care during the pre-COVID-19 year, and 327 (52.7%) had ever been diagnosed with at least one syndemic factor (SA, depression, or DM2). There was a positive association between depression and SA (adjusted odds ration [aOR]= 2.07; 95% CI: 1.32, 3.24; p=0.001) and between depression and DM2 (aOR=1.75; 95% CI: 1.12, 2.74; p=0.014) among the study sample (aim 1). None of the syndemic factors was significantly associated with RIC (aim 2). In both unadjusted and adjusted regression analyses, there was a dose response pattern of the syndemic index with being retained in care (aOR for 1 syndemic factor vs. none: 1.02, 95% CI: 0.69, 1.52; 2 syndemic factors: 1.45, 95% CI: 0.80, 2.62; 3 syndemic factors: 1.61, 95% CI: 0.43, 5.99) but no group reached statistical significance (aim 3). Interactions on both additive and multiplicative scales were not significant, demonstrating no syndemic effect of SA, depression, and DM2 on RIC during the pre-COVID-19 year (aim 4). In the second manuscript, 321 (51.7%) of the study sample (n=621) were retained in care during the pandemic. Factors associated with RIC during the pandemic included age (p=0.002), education (p=0.04), and viral load suppression (VLS) in the pre-pandemic year (p=0.05) (aim 1). Determinants of lost RIC during the pandemic included age (older age ≥50: aOR=0.56, 95% CI 0.34, 0.91; p=0.02 compared to younger age <50) and race/ethnicity (non-Hispanic black: aOR=1.95, 95% CI 1.04, 3.66; p=0.40 compared to non-Hispanic white) (aim 2). While determinants of improved RIC during the pandemic included only age (older age ≥50: aOR=2.28, 95% CI 1.23, 4.22; p=0.01 compared to younger age <50). HIV mode of transmission (MSM aOR=1.79, 95% CI 0.97, 3.32; p=0.06 compared to other modes of transmission), and VLS at baseline (VLS aOR=3.48, 95% CI 0.93, 12.99; p=0.06 compared to no VLS) were associated with improved RIC at p<0.10 (aim 2). The number of syndemic factors among the study sample was significantly associated with RIC in a dose-response relationship in the unadjusted model; the odds of RIC increased by 42% for each additional syndemic factors (OR = 1.42, 95% CI 1.16, 1.75; p=0.001). But the association was not significant after adjusting for age, sex, education level, race and ethnicity, preferred language, HIV mode of transmission, VLS at baseline, and RIC pre-pandemic (aim 3). Significant interactions on both additive and multiplicative scales were found between SA and depression on being retained in care during the pandemic, demonstrating synergistic effect of these two health conditions on RIC (aim 3).

Files

Metadata

Work Title Retention in Care Among People Living with HIV: The Syndemic Effect of Substance Abuse, Depression, and Type 2 Diabetes and the Impact of the COVID-19 Pandemic
Access
Open Access
Creators
  1. Tarek B. Eshak
Keyword
  1. HIV
  2. Retention in Care
  3. Syndemic
  4. Substance Abuse
  5. Depression
  6. Type 2 Diabetes
  7. COVID-19
License In Copyright (Rights Reserved)
Work Type Dissertation
Acknowledgments
  1. Thank you to Christine Quimby for her important contributions to data cleaning.
Publisher
  1. Tarek B. Eshak
Publication Date June 25, 2022
Deposited June 25, 2022

Versions

Analytics

Collections

This resource is currently not in any collection.

Work History

Version 1
published

  • Created
  • Updated
  • Updated
  • Updated Acknowledgments Show Changes
    Acknowledgments
    • Thank you to Christine Quimby for her important contributions to data cleaning.
  • Added Creator Tarek B. Eshak
  • Added Tarek_Eshak_DrPH_Dissertation_ScholarSphere_Submission.pdf
  • Updated Description, License Show Changes
    Description
    • It has been over forty years since the first cases of acquired immune deficiency syndrome (AIDS) were reported in 1981. Despite major progress in human immunodeficiency virus (HIV) treatment, prevention, care, and support services to mitigate the global impact of the HIV/AIDS epidemic, retention in HIV care (RIC), an important indicator of HIV prevention and control programs, remains a major challenge. According to the latest estimates from the United States (US), Less than 60% of Americans aged ≥13 years with diagnosed HIV were retained in care in 2019. Previous research has found that mental health conditions and chronic illnesses are common comorbidities among people living with HIV (PLHIV) and are associated with poor HIV outcomes, including reduced rates of RIC. Importantly, substance abuse (SA), depression, and type 2 diabetes (DM2) often co-occur among PLHIV, but little is known about the cumulative and interactive effects of these conditions on RIC.
    • Syndemic (or synergistic epidemics) theory has become a well-established framework for the explanation of how adverse health conditions interact together within social and structural inequalities, which, cumulatively affect health outcomes. Furthermore, the recent emergence of the coronavirus disease 2019 (COVID-19) pandemic has been accompanied by widespread social, economic, and healthcare system disruptions, exposing, and often exacerbating health inequalities and adding to the challenges faced by PLHIV when seeking HIV care services. While syndemic theory has been widely applied in HIV research, this framework has been rarely utilized for RIC. Therefore, guided by syndemic theory, this dissertation has two objectives: 1) to examine the syndemic effect of SA, depression, and DM2 on RIC and 2) to evaluate the impact that the COVID-19 pandemic has had on RIC. These two objectives comprise the two manuscripts of this dissertation, respectively.
    • This dissertation uses a retrospective cross-sectional study among PLHIV who were seen for HIV care at the Penn State Health (PSH) Milton S. Hershey Medical Center (HMC) in Central Pennsylvania (PA). Patients were included in the analysis if they: 1) initiated HIV care before March 2019 (study baseline), 2) had ≥1 HIV medical care encounter during the pre-COVID-19 year (3/1/2019 to 2/28/2020), 3) were ≥18 years old at baseline, and 4) identified as being active in care at the end of the first year of the pandemic (2/28/2021). RIC, the primary outcome of the two objectives, was defined according to the Health Resources and Services Administration HIV/AIDS Bureau (HRSA/HAB) and the Centers for Disease Control and Prevention (CDC). Patients were considered to be retained in care if they had ≥2 HIV medical care encounters at least 90 days apart within the 12-month measurement year (the pre-COVID-19 year for objective 1 and the first year of the COVID-19 pandemic for objective 2). An HIV medical care encounter was defined as documentation of at least one provider visit (either in-person or virtual), one CD4 cell counts test, or one viral load (VL) test. The predictor variables, including demographics, HIV-related factors, and syndemic factors (SA, depression, and DM2) were captured from two datasets: Path To Health dataset (PaTH) and QuickData dataset (QuickData), the latter is a database used by the HIV clinical program.
    • For the first objective, the study duration was restricted to the pre-COVID-19 year to avoid the potential effects of the COVID-19 pandemic on the study outcome (RIC). I examined four aims in the first manuscript. First, I evaluated the co-occurrence of the syndemic factors (SA, depression, and DM2) among the study sample. Second, I investigated the association between each of the syndemic factors and RIC. Third, I assessed the additive effects of the syndemic factors on RIC. And fourth, I tested whether the syndemic factors interact to exacerbate RIC. To test the syndemic effect of SA, depression, and DM2 on RIC (third and fourth aims), I used two approaches. In the first approach (aim 3), I applied the traditional method of testing syndemic factors as cumulative adversities by using the sum score specification (the number of syndemic factors), for which a syndemic index (range from 0= no syndemic factors to 3=all three syndemic factors) was created. The syndemic index was evaluated once as a categorical variable (with 0 as the reference category), and once as a continuous exposure, which evaluates the linear dose response by regressing the outcome on the syndemic index. In the second approach (aim 4), I used the recommended method of applying a series of linear regression models and logistic regression models with product terms to test for the synergistic/interaction effect of SA, depression, and DM2 on both additive and multiplicative scales, respectively.
    • The second objective had three aims: 1) to assess factors associated with RIC during the first year of the pandemic (during the pandemic), 2) to evaluate determinants of changes in RIC during the pandemic (lost RIC and improved RIC) compared to the one year preceding the pandemic (pre-pandemic), and 3) to test the syndemic effect of SA, depression, and DM2 on RIC during the pandemic. I first performed descriptive analysis of the study sample and then conducted multiple backward logistic regression to assess factors associated with RIC during the pandemic (aim 1) and to evaluate determinants of lost RIC and improved RIC during the pandemic compared to the pre-pandemic RIC (aim 2). Unadjusted and adjusted odds ratios (OR) and 95% confidence intervals (CI) were reported for all predictor variables. For aim 3, I used the same two approaches of the syndemic index and the interaction effect (previously described) for testing the syndemic effect of SA, depression, and DM2 on RIC during the pandemic.
    • In the first manuscript, 440 (70.9%) of the total sample (n=621) were retained in care during the pre-COVID-19 year, and 327 (52.7%) had ever been diagnosed with at least one syndemic factor (SA, depression, or DM2). There was a positive association between depression and SA (adjusted odds ration [aOR]= 2.07; 95% CI: 1.32, 3.24; p=0.001) and between depression and DM2 (aOR=1.75; 95% CI: 1.12, 2.74; p=0.014) among the study sample (aim 1). None of the syndemic factors was significantly associated with RIC (aim 2). In both unadjusted and adjusted regression analyses, there was a dose response pattern of the syndemic index with being retained in care (aOR for 1 syndemic factor vs. none: 1.02, 95% CI: 0.69, 1.52; 2 syndemic factors: 1.45, 95% CI: 0.80, 2.62; 3 syndemic factors: 1.61, 95% CI: 0.43, 5.99) but no group reached statistical significance (aim 3). Interactions on both additive and multiplicative scales were not significant, demonstrating no syndemic effect of SA, depression, and DM2 on RIC during the pre-COVID-19 year (aim 4).
    • In the second manuscript, 321 (51.7%) of the study sample (n=621) were retained in care during the pandemic. Factors associated with RIC during the pandemic included age (p=0.002), education (p=0.04), and viral load suppression (VLS) in the pre-pandemic year (p=0.05) (aim 1). Determinants of lost RIC during the pandemic included age (older age ≥50: aOR=0.56, 95% CI 0.34, 0.91; p=0.02 compared to younger age <50) and race/ethnicity (non-Hispanic black: aOR=1.95, 95% CI 1.04, 3.66; p=0.40 compared to non-Hispanic white) (aim 2). While determinants of improved RIC during the pandemic included only age (older age ≥50: aOR=2.28, 95% CI 1.23, 4.22; p=0.01 compared to younger age <50). HIV mode of transmission (MSM aOR=1.79, 95% CI 0.97, 3.32; p=0.06 compared to other modes of transmission), and VLS at baseline (VLS aOR=3.48, 95% CI 0.93, 12.99; p=0.06 compared to no VLS) were associated with improved RIC at p<0.10 (aim 2).
    • The number of syndemic factors among the study sample was significantly associated with RIC in a dose-response relationship in the unadjusted model; the odds of RIC increased by 42% for each additional syndemic factors (OR = 1.42, 95% CI 1.16, 1.75; p=0.001). But the association was not significant after adjusting for age, sex, education level, race and ethnicity, preferred language, HIV mode of transmission, VLS at baseline, and RIC pre-pandemic (aim 3). Significant interactions on both additive and multiplicative scales were found between SA and depression on being retained in care during the pandemic, demonstrating synergistic effect of these two health conditions on RIC (aim 3).
    License
    • https://rightsstatements.org/page/InC/1.0/
  • Published
  • Updated

Version 2
published

  • Created
  • Updated Keyword Show Changes
    Keyword
    • HIV, Retention in Care, Syndemic, Substance Abuse, Depression, Type 2 Diabetes, COCID-19
    • HIV, Retention in Care, Syndemic, Substance Abuse, Depression, Type 2 Diabetes, COVID-19
  • Published
  • Updated Work Title Show Changes
    Work Title
    • RETENTION IN CARE AMONG PEOPLE LIVING WITH HIV: THE SYNDEMIC EFFECTS OF SUBSTANCE ABUSE, DEPRESSION, AND TYPE 2 DIABETES AND THE IMPACT OF THE COVID-19 PANDEMIC
    • Retention in Care Among People Living eith HIV: The Syndemic Effect of Susbstance Abuse, Depression, and Type 2 Diabetes and the Impact of the COVID-19 Pandemic
  • Updated Work Title Show Changes
    Work Title
    • Retention in Care Among People Living eith HIV: The Syndemic Effect of Susbstance Abuse, Depression, and Type 2 Diabetes and the Impact of the COVID-19 Pandemic
    • Retention in Care Among People Living with HIV: The Syndemic Effect of Susbstance Abuse, Depression, and Type 2 Diabetes and the Impact of the COVID-19 Pandemic
  • Updated Work Title Show Changes
    Work Title
    • Retention in Care Among People Living with HIV: The Syndemic Effect of Susbstance Abuse, Depression, and Type 2 Diabetes and the Impact of the COVID-19 Pandemic
    • Retention in Care Among People Living with HIV: The Syndemic Effect of Substance Abuse, Depression, and Type 2 Diabetes and the Impact of the COVID-19 Pandemic
  • Updated