Life Instability and its Effects on Psychosocial Syndemic Problems and HIV-Care Outcomes in People Living with HIV in Care in South Florida

Introduction

Despite significant advancements in prevention and treatment, the HIV epidemic continues to disproportionately affect certain marginalized communities within United States (US) ^1–3. In particular, South Florida contains several “hot-spots” of HIV transmission (e.g., Miami, Fort Lauderdale) ^1,4 that routinely experience poorer HIV prevention and care outcomes compared with other similar US locales ^5,6, potentially due to factors that both facilitate and exacerbate life instability among racial, ethnic, sexual, and gender minorities. Thus, addressing structural inequities (e.g., high rates of unemployment, housing instability) that affect South Florida’s socio-demographically diverse communities is needed to curb the epidemic in this region ^7,8.

Engagement within the HIV-care cascade is impacted by life destabilizing structural and psychosocial factors through various mechanisms. Prior studies have documented a high prevalence of structural life instability among people living with HIV (PLWH) via indicators such as unemployment, low income, housing instability and homelessness, and a history of incarceration [7,9–14]. PLWH who possess greater structural life instability routinely face poorer HIV-related care outcomes such as inconsistent antiretroviral therapy (ART) adherence and viral non-suppression ^15–19. Additionally, the chronic experience of structural life instability has been linked with increased risk for current and future psychosocial concerns such as depression, anxiety, and substance misuse ^20–23 which, in turn, also contribute to poorer HIV-related care outcomes ^24–30. In fact, it is likely that structural life instability and psychosocial syndemic factors have both bidirectional and mutually reinforcing effects on HIV-related care outcomes ^31–34, further signaling the importance of addressing both of these constructs together to support PLWH.

The intersecting relationship between structural life instability, psychosocial inequities, and HIV-related care outcomes can be conceptualized through the lens of syndemics and social stability frameworks. Syndemic theory posits that the interaction of co-occurring biopsychosocial factors drive and amplify morbidity and mortality within both infectious and chronic disease epidemics ³⁵. In the context of HIV, syndemic indicators including poverty, depression, and substance misuse have been found to consistently complicate engagement within the HIV clinical care cascade ^21,36–41. Complementing the risk-focused framing of syndemics, social stability has adopted a strengths-based buffering perspective in its focus on sets of social conditions that are proposed to confer protection from vulnerability to health disparities ⁴². Common conditions that have been implicated in reducing HIV health disparities include housing stability, stable employment, higher income, lacking a history of incarceration, and access to supportive relationships ^43–45. Although these conceptual frameworks have offered insights into biopsychosocial process that may contribute to HIV care inequities ^46,47, more research is needed to evaluate the specific structural and psychosocial mechanisms influencing downstream individual biobehavioral outcomes.

The epidemiology of HIV-related health disparities is well documented; yet, the existing literature has not adequately explored the longitudinal effects of structural life instability and psychosocial syndemic factors within US HIV epidemic “hot-spots”. Although South Florida poses a unique challenge to ending the HIV epidemic, it also presents an opportunity to investigate the underlying mechanisms driving these disparities, particularly among PLWH receiving care in real-world, community HIV treatment settings. As such, this study leveraged data from a demographically diverse cohort of PLWH receiving care at a public clinic in South Florida across two timepoints. The specific study aims were to: (1) examine the effects of structural life instability on subsequent psychosocial syndemic problems, and (2) explore the effects of these structural and psychosocial factors on HIV-related care outcomes, including ART adherence and log RNA viral load among a multiply marginalized sample of PLWH in South Florida.

Methods

Participants and procedures

Participants (N = 241) were PLWH living in South Florida who were engaged in HIV care at a major public HIV clinic housed within an academic medical center. Although this HIV clinic serves over 2000 PLWH yearly, only PLWH who had completed at least two annual psychosocial assessments, baseline (N = 1360) and follow-up (N = 241), between April 2017 and July 2023 were included within this analysis. Eligible participants were: (a) 18 years or older, (b) engaged in HIV-care at UHealth or Jackson Memorial Hospital in Miami, (c) able to speak and understand English, Spanish, or Haitian Creole, and (d) had completed at least two psychosocial assessment batteries within the clinic between April 2017 and July 2023. Excluded participants were PLWH who were unable to provide informed consent due to cognitive or acute psychiatric issues or those who were currently incarcerated. All potentially eligible participants were approached between HIV-care appointments by trained study staff stationed at the clinic to assess study eligibility. Eligible participants were then invited to complete the baseline psychosocial assessment via self-report or interviewer administered in either English, Spanish, or Haitian Creole. Surveys were collected via hand-held tablets with responses stored on the HIPAA compliant and confidential REDCap platform ^48,49. Follow-up psychosocial assessments were completed either in person at the clinic or via phone call with trained study staff members. Informed consent was obtained from all interested participants prior to the start of each psychosocial assessment. All study procedures were approved by the IRB at the University of Miami.

Measures

Sociodemographic information.

Age, race, ethnicity, sex assigned at birth, gender identity, and sexual orientation were collected.

HIV RNA viral load.

Continuous HIV RNA viral load was extracted from participants electronic medical records. The log of the continuous HIV RNA viral load was calculated and utilized for analyses purposes.

ART adherence.

ART adherence was collected via self-report using Wilson and colleagues’ 3-item adherence measure ⁵⁰. ART adherence for the past month was calculated based on participant responses on three questions: number of missed dose days, frequency of optimal adherence (0 = never to 5 = always), and overall self-rating (0 = very poor to 5 = excellent) in the past 30 days. Each item response was coded as a successive measure of overall adherence (i.e., each 1–5 option corresponded with an adherence percentage of 0%, 20%, 40%, 60%, 80% or 100%). An overall past 30-day ART adherence score was created by linearly transforming the item responses to a scale of 0 (missed all doses) to 100 (perfect adherence) using participant responses on the 3-items ^21,44,50. An overall adherence mean summary for the three individual items was calculated and used for analyses.

Structural life instability.

An additive index of structural life instability (0–6) was created by taking the sum of 6 observed structural life instability conditions. All 6 life instability factors were dichotomized so that participants were positive (1) or negative (0) for that indicator.

1. Educational attainment. Educational attainment was assessed using a one-item question asking participants to identify the highest level of education achieved. Response options included: 8^th grade or lower, partial high school, high school graduate, partial college, college graduate, partial graduate school, and graduate school degree. Participants were positive for this structural life instability indicator if they reported having less than a high school degree or GED equivalent.

2. Housing instability. Housing instability was assessed via a one-item question asking participants about experiences of homelessness (such as sleeping in a car, public place not intended for sleeping, homeless shelter, single room occupancy or welfare hotel or motel) or temporary/transitional housing in the past 12 months. Participants were deemed positive for this syndemic indicator if they endorsed either homelessness or past 12-month temporary/transitional housing.

3. Under-employment. Employment was evaluated by asking participants about their current employment status. Response options included: full-time work, part-time work, full- or part-time school, neither work nor school, on disability, or “other”. Write-in responses for participants endorsing “other” were reviewed and recategorized under the most relevant response option. Participants were deemed positive for this syndemic indicator if they endorsed anything other than “full-time work”, “part-time work”, or “work and school.”

4. Incarceration history. PLWH were asked “Over your whole life, how much time have you spent in jail, prison, or a correctional facility?” with response options including: none, less than a week, 7 days to less than a month, 1 month to a year, or more than a year. Participants were positive for this syndemic indicator if they endorsed spending any time in jail.

5. Non-partnered relationship status. PLWH were asked to identify their relationship status by selecting one of the following options: married or living with someone as if married, non-cohabiting relationship, single, divorced or separated, or loss of long-term partner/widowed. Participants were considered positive for this syndemic indicator if they endorsed being single, divorced or separated, or widowed/having lost a long-term partner.

6. Immigration History. Participants were asked to identify their country of birth. Participants born outside of the US were considered positive for this syndemic indicator should they have moved to the US after the age of 18 years.

Psychosocial life instability.

An additive index of psychosocial life instability (0–5) was created by taking the sum of 5 observed psychosocial syndemic conditions. All 5 psychosocial syndemic conditions were dichotomized so that participants were positive (1) or negative (0) for that syndemic indicator.

1. Depression. Depression was assessed via the 9-item Patient Health Questionnaire ⁵¹. Responses were on a 4-point scale from 0 (not at all) to 3 (nearly every day) with greater scores indicating greater number of symptoms. PLWH were considered positive for this syndemic indicator if their total score was 10 or higher.

2. Anxiety. Participants’ anxiety was assessed using a symptom thermometer adapted from the distress thermometer and problem list for patients as created by the National Comprehensive Cancer Network ⁵². The anxiety thermometer is a single item self-report measure that assesses anxiety in the past week using a visual analog scale from 0 (no anxiety) to 10 (extreme anxiety). PLWH were considered positive for this syndemic indicator if they reported an anxiety score of 4 or greater.

3. Problematic substance use. Participants were asked to report if they used crack, cocaine, heroin, other opioids, amphetamines, hallucinogens, ecstasy/MDMA, sedatives/tranquilizers, marijuana/cannabis, and or any other drugs in the past 30 days. Score options ranged from 0 (no use) to 4 (about every day). Additionally, participants reported their average alcohol consumption on days when they drank. PLWH who reported consuming 4 or more drinks on average during days of drinking and or using an illicit substance at least once a week in the past 30 days were positive for this syndemic condition. Items were adapted from the ASI-Lite ⁵³.

4. Interpersonal violence. PLWH endorsed whether they had ever experienced childhood sexual abuse, relational abuse, or abuse as an adult to evaluate whether participants were victims of interpersonal violence. If participants responded yes to any of these forms of violence, they were positive for this syndemic condition.

5. Trauma history. Using the Brief Trauma Questionnaire ⁵⁴, participants were asked to endorse if they had (a) experienced several different types of traumatic events and (b) either seriously injured/hurt or believed that they would be seriously injured/killed. PLWH were considered positive for this syndemic condition if they endorsed being seriously injured/hurt or perceiving that they would be seriously injured/hurt for any of the ten traumatic events on the list.

Data analytic plan

Analyses were performed in IBM SPSS Statistics-Version 29.0 ⁵⁵. Authors examined descriptive statistics to characterize the sample and to ensure that assumptions for linear regression were met. Linear regression models assessed the relationships between (1) baseline index of structural life instability and follow-up index of psychosocial syndemics, (2) follow-up index of psychosocial syndemics and ART adherence, and (3), ART adherence and HIV log RNA viral load. We included all prior predictors as covariates in the subsequent linear regressions. Finally, authors ran a sequential mediation model using the PROCESS macro (model 6)⁵⁶ to examine the indirect effects of the baseline life instability index on log RNA viral load at follow up through the follow-up psychosocial syndemic index and ART adherence as intermediary variables. Recorded results include unstandardized coefficients, standard errors, p values, and 95% confidence intervals (CI). For the indirect effect, we determined significance by evaluating the Bootstrapped 95% confidence intervals ^57–59.

Results

The parent study enrolled 1360 PLWH, who completed a baseline survey between April 2017 and July 2023. All enrolled participants were eligible for follow-up assessments, which could happen after one year during their clinic visits. However, the follow-up was added some time after the study’s inception, was not required or consistently offered to all participants who completed a baseline and was further impeded by COVID-restrictions at the clinic. Therefore, of the participants who completed baseline assessments, 241 also completed a follow-up assessment. Further details regarding reasons for loss -to follow-up were not available to our study team.

Approximately half of the participants identified as women at baseline (50.6%, N = 122) and were assigned female at birth (51.0%, N = 123). Participants were on average 51.1 years old (SD = 11.0 years) and most self-identified as straight or heterosexual (78.4%, N = 189). Participants predominately self-identified as racially African American/Black (76.3%, N = 184) and ethnically non-Hispanic/Latino (77.2%, N = 186). Eighty-eight percent (N= 212) of the participants completed the survey in English, 10.0% (N = 24) in Spanish, 1.7% (N = 4) in English with some Spanish, and 0.4% (N = 1) in Haitian Creole.

The average number of structural life instability factors endorsed at baseline was 2.23 (SD = 1.05) while the mean number of psychosocial syndemic indicators experienced by participants at follow-up was 3.05 (SD = 1.30). The two most endorsed life instability factors were under-employment (73.4%) and incarceration history (53.5%) while the two most prevalent psychosocial syndemic conditions were trauma history (87.6%) and depression (72.6%). Participants’ average ART adherence in the past 30 days was 90.35% and their mean log viral load at follow-up was 1.47. A more comprehensive review of descriptive statistics can be found in Table 1.

Table 1:

Demographics for Participants Receiving Care at a Public HIV Clinic (N = 241)

Variable	N	%
Racial identity
Black	184	76.3
White	53	22.0
Other	3	1.2
Ethnicity
Hispanic/Latino/a/e	54	22.4
Not Hispanic/Latino/a/e	186	77.2
Sex assigned at birth
Male	118	49.0
Female	123	51.0
Current gender identity
Male	116	48.1
Female	122	50.6
Trans woman	2	0.80
Other	1	0.40
Sexual orientation
Heterosexual	189	78.4
Gay, lesbian, or homosexual	32	13.3
Bisexual	15	6.2
Other	3	1.2
Structural life instability *
Education attainment (less than high school or GED equivalent)	80	33.2
Housing instability	33	13.7
Under-employment (not working full or part-time)	177	73.4
Incarceration history	129	53.5
Non-partnered relationship status	72	29.9
Immigration history (immigrated to the US after age 18 years)	64	26.6
Psychosocial life instability *
Depression	175	72.6
Anxiety	113	46.9
Problematic substance use	74	30.7
Interpersonal violence	161	66.8
Trauma history	211	87.6
Other Variables	M	SD
Age	51.13	10.99
Structural life instability at baseline	2.23	1.05
Psychosocial syndemic indicator at follow-up	3.05	1.30
% ART adherence in the past 30 days	90.35	19.38
Log viral load at follow-up	1.47	1.21

^*Number and percentage of participants who met criteria for this structural life instability or psychosocial syndemic condition.

Structural life instability at baseline predicted psychosocial syndemic problems at follow-up, such that for every additional structural life instability factor experienced, there was an average increase of 0.23 endorsed syndemic problems (b = 0.23, SE = 0.08, p < 0.01, 95% CI [0.06, 0.39]). For every additional psychosocial syndemic problem at follow-up there was a 1.37 percentage point decrease in participants’ past-month ART adherence (b = −1.37, SE = 0.66, p = 0.04, 95% CI [−2.68, −0.07]) when controlling for baseline structural life instability. ART adherence was, in turn, significantly inversely associated with participants’ continuous log RNA viral load (b = −0.04, SE = 0.01, p < 0.001, 95% CI [−0.05, −0.03]) when holding baseline life instability and follow-up psychosocial syndemic index scores constant. Complete details on the effects of covariates for each outcome can be found in Table 2.

Table 2:

Direct and Indirect Effects for Structural Life Instability, Psychosocial Syndemic Index, ART Adherence, and Log RNA Viral Load among PLWH (N = 241)

						95% Confidence Interval
Outcome	Predictor/covariate	b	SE	t	p	Lower	Upper
Direct effects
FU Psychosocial Syndemic Problems	BL Structural Life Instability	0.23	0.08	2.68	<0.001 **	0.06	0.39
FU Past Month ART Adherence	BL Structural Life Instability	−1.06	0.86	1.23	0.22	−2.76	0.64
	FU Psychosocial Syndemic Problems	−1.37	0.66	−2.07	0.04 *	−2.68	−0.07
FU Log RNA Viral Load	BL Structural Life Instability	0.02	0.07	0.27	0.79	−0.12	0.15
	FU Psychosocial Syndemic Problems	0.01	0.05	0.13	0.90	−0.10	0.11
	FU Past Month ART Adherence	−0.04	0.01	−7.28	<0.001**	−0.05	−0.03
Indirect Effects ***
FU Log RNA Viral Load	BL Structural Life Instability −> FU Past Month ART Adherence	0.04	0.03	-	n.s.	−0.01	0.10
	BL Structural Life Instability −> FU Psychosocial Syndemic Problems	0.00	0.01	-	n.s.	−0.03	0.03
	BL Structural Life Instability −> FU Psychosocial Syndemic Problems −> FU Past Month ART Adherence	0.01	0.01	-	significant	0.00	0.03

^***Significance of the indirect effect was assessed with bootstrapping in lieu of generating a p value.

Key: BL = Baseline, FU = Follow-up

^*p < 0.05

^**p < 0.001

There was a significant marginal indirect effect of baseline structural life instability on log RNA viral load at follow-up through both the psychosocial syndemic index and ART adherence variables (b = 0.0114, SE = 0.0071, 95% Boostrap CI [0.0005, 0.0274]). Thus, for every additional structural life instability factor endorsed at baseline, participants’ RNA viral load increased on average by 1 log unit at follow-up, sequentially mediated by psychosocial syndemic burden and ART adherence. The model explained 18.95% of the variance in log RNA viral load (F (3, 237) =18.47, p< 0.001, R² = 0.19). A more thorough review of model results can be found in Table 2.

Discussion

This study sought to examine how structural life instability and psychosocial syndemic burden impacts HIV clinical-care cascade outcomes among multiply marginalized PLWH receiving care from a large catchall public clinic in South Florida. Overall, PLWH enrolled faced high levels of structural and psychosocial barriers with baseline structural life instability significantly predicting subsequent psychosocial syndemic burden at follow-up. Participants’ psychosocial syndemic burden was significantly negatively associated with self-reported ART adherence at follow-up when controlling for baseline structural life instability. Furthermore, self-reported ART adherence was inversely significantly associated with continuous log RNA viral load when holding both baseline structural life instability and psychosocial syndemic burden at follow-up constant. Finally, a significant indirect effect of baseline structural life instability on follow-up log RNA viral load was observed, sequentially mediated by both the psychosocial syndemic index and ART adherence. This relationship accounted for nearly 19% of the variance in log RNA viral load.

PLWH with more baseline structural life instability were likely to experience greater psychosocial syndemic burden at follow up, demonstrating a link between structural barriers and subsequent psychosocial functioning. This finding is consistent with published literature examining the relationship between structural barriers and psychosocial challenges among PLWH receiving care in community-based settings both within the U.S. and abroad ^30,44,60,61. Thus, structural barriers that facilitate life instability such as a history of incarceration or low educational attainment, may heighten psychosocial distress in PLWH by undermining mental health, disrupting personal safety, and impacting the ease by which coping mechanisms can be employed to manage destabilizing day-to-day contexts. Although the temporality tested within this study was that baseline life instability predicted poor psychosocial functioning at follow-up, it is probable that this relationship is both bidirectional and cyclical such that experiencing a greater psychosocial burden may also result in PLWH experiencing more life destabilizing factors too. As such, more comprehensive longitudinal research, with additional time points, is needed to truly tease out the causal pathway between these two constructs.

Proximal psychosocial syndemic burden resulted in greater impact on HIV care outcomes than distal structural conditions of life instability highlighting an important area for intervention. When controlling for life instability at baseline, PLWH’s ART adherence decreased by 1.37% for every additional psychosocial syndemic indicator experienced, and in turn, signaled a decrease in participants’ continuous log RNA viral load. The presence of significant downstream effects in the absence of direct effects from baseline life instability to follow-up HIV-related health outcomes centers the importance of considering both socioenvironmental and psychosocial factors within the same study/model. Poor psychosocial functioning can disrupt engagement in the HIV-clinical care cascade at several stages including affecting a person’s ability to adhere to their daily ARTs, attend regularly scheduled medical appointments, and maximize behaviors that prevent HIV transmission. For example, an individual who utilizes substances to cope with depression or anxiety may inadvertently experience inconsistent ART adherence and subsequent viral non-suppression. This cascade of negative health consequences may then further increase the likelihood for HIV transmission through condomless sex or intravenous substance use alongside poorer general health outcomes due to undermanaged HIV. Selective attention to only one set of factors may result in suppressed or confounded effects, yielding an incomplete understanding of how these factors may maintain HIV inequities. Of note, we observed a significant indirect effect of structural life instability on follow-up RNA viral load only in the presence of the full mediation pathway, accounting for the sequential effects of psychosocial syndemics and adherence. Given the pharmacologic forgiveness and improved efficacy of modern ART regimens, assessment of and intervention on all relevant modifiable factors along the causal cascade may be necessary to see effects on relevant HIV health outcomes. Furthermore, because of the interconnected nature of many of these psychosocial syndemic indicators, multipronged or transdiagnostic interventions⁶² that address several psychological challenges may be a more efficient way to improve HIV-related care outcomes among PLWH facing both structural and psychological barriers to care.

The mean age of participants in this study was 51.1 years old highlighting the increasing importance of studying HIV within an aging context ⁶³. More than one-half of PLWH in the U.S. are over the age of 50 years ⁶⁴ with models projecting that this number will almost double by 2045 ⁶⁵. Additionally, certain subpopulations of older adults, such as those living with marginalized identities (e.g., sexual, gender, racial, ethnic minorities) or HIV, are increasingly burdened by structural and psychosocial factors of life instability compared with their non-marginalized peers that negatively impact their effectively manage their HIV care ^66–69. As such, furthering examining these findings with a specific lens towards aging may illuminate areas for tailored intervention to help older PLWH cope with experiences of past and ongoing factors of life instability and psychosocial stress while also drawing attention to the need for policy-related changes at local, state, and national levels to improve the lives of under-resourced older PLWH across the U.S.

This study possesses limitations. Even though data was collected from participants at two distinct time points, authors are unable to truly assess causal relationships between structural life instability and both psychosocial syndemic burden and HIV-related care outcomes due to the potential conflation of concurrent syndemic burden at baseline and due to only having two timepoints and multiple predictors and potential mediators. Similarly, unmeasured confounders may better explain the relationship between predictors and outcomes in this study highlighting a focus for future longer term cohort studies. Further, some of the structural life instability indicators evaluated may not carry the same influence over time (e.g., destabilizing nature of immigration may become less important the more years someone lives in the U.S.) which may have impacted the magnitude of observed findings. Bias, particularly as it relates to the subjective measurement of certain sensitive constructs like medication adherence or psychosocial syndemic burden, may have indirectly impacted findings. However, the use of more objective metrics such as log RNA viral load and education level may have attenuated the potential challenges associated with self-report data. Finally, enrolled PLWH appear to be particularly resilient as evidenced by their lower rates of structural life instability and psychosocial syndemic burden and high ART adherence and HIV viral suppression. And, as such, it is possible that we were unable to identify meaningful associations between our predictors and measure of HIV disease severity (i.e., continuous log RNA viral load) due to most participants being undetectable with low log RNA viral load at follow-up. Although this is a strength of the sample, it is possible that experiences of participants enrolled in this study may not be indicative of other subgroups of PLWH receiving care at community-based clinics within other HIV “hot-spots” across the country; thus, generalizability is cautioned.

Conclusion

PLWH receiving care from community-based clinics in regions with high poverty, immigration, and inconsistent access to healthcare face significant HIV-related health inequities due to a disproportionate burden of structural and psychosocial challenges. Findings from this study emphasize the importance of addressing upstream determinants such as structural life instability and psychosocial barriers when developing and optimizing interventions along the HIV clinical-care cascade, especially among underserved PLWH receiving medical and psychiatric care in real-world settings where compounded barriers to care are highly prevalent. As such, the need for multilevel intervention that concurrently targets both structural components of life instability and comorbid psychosocial challenges is particularly important during this moment due to significant cuts to global funding for HIV prevention and treatment programming and general uncertainty regarding the future of HIV-care services worldwide. By capitalizing and expanding upon already established transdiagnostic approaches that address syndemic problems at the individual level, underserved PLWH may be provided with strategies to more adeptly cope with factors of life instability even in societal systems that are keen on maintaining structural barriers to care (e.g., Medicaid restrictions, de-prioritization of scientific research, anti-LGBTQ+ and immigrant legislation, etc.).

Data Availability Statement:

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.