Depression and the revolving door of HIV care: using multistate approaches to understand complex longitudinal exposure/outcome relationships among a cohort of people entering HIV care in Cameroon from 2019 to 2020

Abstract

Introduction

Addressing the mental health needs of people with HIV is critical to ending the HIV epidemic. Yet, empirical evidence on the downstream consequences of poor mental health at entry into HIV care is scant, limiting our ability to deploy appropriate interventions. Multistate methods can provide a nuanced picture of how longitudinal care engagement differs between those with and without mental ill-health.

Methods

From June 2019 to March 2020, we enrolled people with HIV aged 21+ entering HIV care at three clinics in Cameroon. We conducted structured interviews to ascertain demographics and depression status (Patient Health Questionnaire-9 scores>9=heightened depressive symptoms) at enrolment and extracted participant clinical record data through 1 January 2022. We estimated the proportion of individuals and time spent in six mutually exclusive and exhaustive care states: linked to clinic; engaged at clinic, prescribed antiretroviral therapy; disengaged from clinic; re-engaged at clinic; known death and known transfer out across the first 18 months following entry. Further, we explored re-engagement patterns among those who disengaged. Estimates were compared for those with vs without heightened depressive symptoms.

Results

420 people contributed 630.1 person-years of follow-up; 20% (n=84) had probable depression. A similar proportion of individuals with and without heightened depressive symptoms failed to return to the clinic after their first visit (~10%). However, those with heightened depressive symptoms were less likely to be continuously engaged in care at month 18 (prevalence difference=−14.3; 95% CI –27.0 to –2.3) and spent an average of ~40 fewer days (95% CI –78.4 to –1.2) engaged in care at their original clinic across follow-up compared with their counterparts. Of the 141 who disengaged from care, those with heightened depressive symptoms were less likely to be re-engaged 6 months later (prevalence difference=−10.2%; 95% CI −25.7 to 5.9).

Conclusions

Routine depression screening and sustained support for people with HIV with heightened depressive symptoms are warranted to uphold these individuals’ right to health and improve downstream outcomes.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Poor mental health is more common among people with HIV compared with the general population and is also associated with poor HIV treatment outcomes.

WHAT THIS STUDY ADDS

• Early HIV care continuum outcomes are similar among those entering HIV care with and without heightened depressive symptoms. However, people with HIV and comorbid depression are at heightened susceptibility to downstream care outcomes (eg, engagement in care beyond the first year following care entry and re-engagement in care) compared with their mentally well counterparts.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Routine depression screening and treatment referral that is integrated into existing HIV care service delivery systems is warranted to improve public health.

Introduction

Addressing the mental health needs of people with HIV has gained mounting traction as a strategy critical to ending the HIV epidemic.^{1 2} In the Global South, the prevalence of depression among people with HIV ranges from ~10% to 60%.³ In a 2018 meta-analysis, the pooled prevalence of depression among people with HIV in low-income and middle-income countries was 24%³ compared with just 5% in the general population. This illuminates a stark disparity in mental health outcomes among those with and without HIV.

It is well known that poor mental health can contribute to suboptimal chronic health outcomes among those with comorbid conditions.⁴ For those with HIV in the Global South, heightened depressive symptoms are associated with delayed HIV testing, delayed entry into HIV care, delayed antiretroviral treatment (ART) initiation,⁵ ART non-adherence,⁶ loss to care,⁵ viral non-suppression⁷ and early mortality.⁸ However, much of our existing evidence on the relationship between mental health and HIV care outcomes fails to disentangle the temporal relationship between depressive symptoms and downstream HIV care outcomes. Moreover, evidence exploring the more nuanced relationship between depressive symptoms and longitudinal HIV care continuum outcomes, including the relationship between depression and the timing of lapses and re-engagement in HIV care, is scant. For example, it is possible that individuals with depressive symptoms are more likely to experience early lapses in HIV care (eg, immediately after diagnosis or care initiation), but return to care sooner than their mentally well counterparts if they need other medical services. Alternatively, those with heightened depressive symptoms may be more likely to remain engaged in care early in their care-seeking journey if they experience additional psychosocial benefits from the additional support they may receive at the clinic during the HIV care initiation process. Additional research is needed to explore these differences in engagement patterns.

For the past several years, Cameroon has experienced increased conflict and insecurity due to a range of factors, including the Anglophone Crisis, an ongoing armed conflict between the Cameroonian government and Ambazonia separatist groups in the northwest and southwest regions of the country, and unrest in neighbouring regions.9,11 This has led to the displacement or death of tens of thousands of Cameroonians.9,11 The ongoing conflict has led to healthcare service delivery interruptions in much of the country, inevitably impacting both the mental and physical well-being of countless individuals, particularly those with chronic conditions such as HIV.¹² The lack of mental health legislation and specific budgeting for the provision of mental healthcare in Cameroon has further exacerbated the consequences of this violence, as a majority of mental health conditions in the country go untreated.¹³

Multistate approaches are a robust analytic tool that can account for transitions into and out of mutually exclusive and exhaustive care states across time.¹⁴ These and related approaches are being increasingly employed to provide a holistic picture of the patient experience (eg, how individuals engage and disengage in HIV care across their care-seeking journeys),15,17 understand complex longitudinal exposure/outcome relationships that may vary over time and better inform the design and deployment of HIV prevention and care programmes and policies.18,20 In this work, we aimed to understand differences in longitudinal HIV care outcomes—including time to treatment initiation and total time engaged in care—among those with and without heightened depressive symptoms at entry into HIV care in Cameroon. To establish these differences, we used a robust multistate approach. Our overarching goal was to generate evidence to better inform depression treatment interventions for the improvement of longitudinal HIV care outcomes among people with HIV in under-resourced settings in Cameroon in a heightened period of regional conflict.

Methods

Setting

From June 2019 to March 2020, we enrolled people with HIV aged 21+ entering HIV care at three urban HIV clinics—one in the Southwest Region, one in the Centre Region and one in the Northwest Region—in Cameroon. These facilities were selected due to their participation in the International epidemiology Databases to Evaluate AIDS Consortium.²¹ Notably, the clinics in the Northwest Region and Southwest Region are in areas of ongoing unrest.

Data show the prevalence of HIV in Cameroon has declined over the past two decades.²² However, recent studies suggest virologic suppression is well below The Joint United Nations Programme on HIV/AIDS (UNAIDS)’ 95-95-95 targets across the country, including the regions in which these clinics are located.²³

Procedures

Study procedures have been previously described.24,26 Briefly, we screened individuals newly entering HIV care at each of the three participating clinics for study eligibility. To participate, individuals had to (1) be 21 years of age or older, (2) be newly initiating HIV care and (3) provide written informed consent to participate. For eligible and consenting individuals, we conducted structured interviews to ascertain demographics and depression status at enrolment. We extracted participant clinical record data from physical patient files to obtain clinic visit dates, ART initiation dates, clinic transfer dates and death dates for each participant between the time of entry into care through 1 January 2022.

Measures

Depression

Depression at enrolment into care was characterised using results from the Patient Health Questionnaire-9 (PHQ-9). Potential scores range from 0 to 27, with scores >9 indicative of moderate to severe depressive symptoms, henceforth referred to as ‘heightened depressive symptoms’. This measure has been previously used in Cameroon and validated in similar settings in the Global South.27,29

Clinical outcomes

We identified six mutually exclusive and exhaustive HIV care continuum states—four non-absorbing states (ie, states that individuals can transfer both into and out of) and two absorbing states (ie, states that individuals can transfer into but not out of)—and 12 possible transitions between these states (online supplemental figure 1). Individuals were considered (1) linked to care on their first documented and attended appointment date. Individuals were considered (2) engaged at the clinic and on ART the first date they had a documented prescription for ART (prescription pick-up data were unavailable). Individuals were considered (3) disengaged from the clinic on the first date they experienced a 183-or-more-day (ie, 6 months) gap in attended clinic visits.³⁰ This definition of disengagement from care is consistent with findings that show this results in the fewest instances of outcome misclassification. Individuals were considered (4) re-engaged in care at the clinic on their first documented and attended appointment date following a 183-or-more-day gap in clinic visits (ie, an instance of disengagement from clinic). Individuals were considered a (5) known death if a date of death was documented in their physical patient file and a (6) known transfer out of clinic if a clinic transfer date was documented in their physical patient file.

Sociodemographic characteristics

Sociodemographic characteristics, including gender, age, education, relationship status, past year travel, employment status and food insecurity, were ascertained via self-report.

Analysis

In the population overall, and among those with and without heightened depressive symptoms at enrolment into HIV care, we estimated the proportion of individuals in each of the six mutually exclusive and exhaustive clinical states defined above (online supplemental figure 1) for each time point in the 18 months following enrolment into HIV care. We employed a non-parametric multistate analytic approach that uses the Aalen–Johansen estimator to allow for transitions into and out of distinct care states over time while accounting for competing events and unequal follow-up time across patients. In addition, we calculated the restricted mean time spent (ie, average time to event given some individuals were administratively censored) in these clinical states during the first 12 and 18 months following enrolment into HIV care by estimating the area under the probability-in-state curves generated using the Aalen–Johansen estimator.³¹ For each of these analyses, the date of enrolment into care served as time zero; individuals were administratively censored on either day 548 of follow-up or on the date of database closure on 1 January 2022.

In a separate analysis including only those who disengaged from care during study follow-up, we estimated the prevalence of individuals in each of the four clinical states individuals could be in following their first clinical lapse: (1) disengaged from the clinic, (2) re-engaged at the clinic, (3) known death and (4) known transfer out of clinic. We quantified outcomes across the first 6 months following an individual’s first instance of disengagement from care. For this analysis, the first date of disengagement from the clinic served as time 0, and individuals were censored on either day 183 following their first instance of disengagement from the clinic or on the date of database closure. Finally, we estimated the instantaneous hazard of disengagement from the clinic, transfer outs and re-engagement at the clinic using Kaplan-Meier methods. Each of our estimates was compared for those with vs without (referent) heightened depressive symptoms at enrolment into HIV care using differences. 95% CIs were obtained via bootstrapping with 5000 non-parametric resamples of the data.

Missing data

For individuals missing data on <10% of items in the PHQ-9 (n=6), the mean of the individual’s non-missing scale responses was imputed for the missing item. Individuals who were missing data on their first clinic date or ART initiation date were excluded (n=6).

Patient and public involvement

In-person meetings were held with members of the local community during the development of the study aims and methods, protocol development and study implementation to solicit input. Data collection protocols were iteratively developed with members of the local community and refined as needed to ensure cultural and contextual appropriateness.

Results

A total of 426 individuals entering HIV care were enrolled into the cohort; 420 were included and contributed 630.1 person-years of follow-up. A majority of included individuals were women (n=246; 58.6%), ages 21–39 (n=246; 58.6%) and in a relationship (n=245; 58.3%; table 1). Approximately one-third of participants were unemployed at the time of enrolment into HIV care (n=149; 35.5%), 28% reported moderate to severe household hunger (n=119) and 20% (n=84) had probable depression at enrolment.

Table 1. Characteristics of people living with HIV newly initiating care in Cameroon by depressive symptoms at enrolment into care.

	Total sample n=420 N (%)	Heightened depressive symptoms* n=84 N (%)	No heightened depressive symptoms* n=336 N (%)
Gender
Men	174 (41.4)	29 (34.5)	145 (43.2)
Women	246 (58.6)	55 (65.5)	191 (56.8)
Age in years
21–39	246 (58.6)	53 (63.1)	193 (57.4)
40+	174 (41.4)	31 (36.9)	143 (42.6)
Education
None	30 (7.1)	13 (15.5)	17 (5.1)
Primary	214 (51.0)	40 (47.6)	174 (51.8)
Secondary	176 (41.9)	31 (36.9)	145 (43.2)
Relationship status
Single	175 (41.7)	47 (56.0)	128 (38.1)
Partnered	245 (58.3)	37 (44.0)	208 (61.9)
Away from home>1 month in the past year
Yes	163 (38.8)	37 (44.0)	126 (37.5)
No	257 (61.2)	47 (56.0)	210 (62.5)
Employment status
Working for pay	271 (64.5)	51 (60.7)	220 (65.5)
Not working for pay	149 (35.5)	33 (39.3)	116 (34.5)
Household hunger†
None/mild	298 (71.5)	52 (62.7)	246 (73.7)
Moderate/severe	119 (28.5)	31 (37.3)	88 (26.3)

^*Heightened depressive symptoms=Patient Health Questionnaire-9 scores>9; no heightened depressive symptoms=Patient Health Questionnaire-9 scores<10.

^†Missing: household hunger n=3.

Clinical outcomes in the population overall

Under 90% of individuals initiated ART the first day they enrolled into HIV care and approximately 11% of participants did not return to the clinic after their initial enrolment appointment (online supplemental table 1; figure 1A). 12 and 18 months following enrolment into care 66.0% (95% CI 61.3 to 70.6) of individuals and 57.9% (95% CI 53.0 to 62.7) were continuously engaged in care (ie, had no prior clinical lapse) and on ART; 21.2% (95% CI 17.4 to 25.2) and 20.7% (95% CI 16.9 to 24.7) were disengaged from the clinic; 7.9% (95% CI 5.4 to 10.5) and 8.8% (95% CI 6.2 to 11.6) had a documented transfer out; and 4.5% (95% CI 2.6 to 6.6) and 8.6% (95% CI 6.0 to 11.3) were re-engaged at the clinic, respectively (online supplemental table 1). Of the 141 individuals who disengaged from the clinic at least once in the 18 months following enrolment into care, 20.6% (95% CI 14.2 to 27.7) were re-engaged at the clinic 3 months after their first instance of disengagement and 27.6% (95% CI 20.6 to 35.5) were re-engaged 6 months later (online supplemental table 2; figure 2A). Overall, individuals spent a restricted mean time of 33.1 days (95% CI 27.1 to 39.4), or approximately 1 month, disengaged from the clinic during the first 12 months following entry into care and 71.7 days (95% CI 60.1 to 84.0), or just over 2 months, disengaged during the first 18 months following entry into care (online supplemental table 3).

Figure 1. HIV care outcomes following cohort enrolment in (A) the total population, (B) those with heightened depressive symptoms at entry into care and (C) those without heightened depressive symptoms at entry into care. ART, antiretroviral treatment.

Figure 2. HIV care outcomes following first instance of disengagement from care in (A) the total population, (B) those with heightened depressive symptoms at entry into care and (C) those without heightened depressive symptoms at entry into care.

Clinical outcomes, stratified by depression status at enrolment

On day 365 following enrolment into HIV care, 57.1 (95% CI 46.1 to 67.5) of those who had heightened depressive symptoms at enrolment were continuously engaged in care at their original clinic, compared with 68.2 (95% CI 63.1 to 73.4) of individuals without heightened depressive symptoms at enrolment (figure 1B, C), yielding a prevalence difference of −11.0% (95% CI −23.4 to 0.7; table 2). On day 548 following enrolment into HIV care, 46.4% (95% CI 35.6 to 56.8) of individuals with heightened depressive symptoms at enrolment were continuously engaged in care at their original clinic compared with 60.7% (95% CI 55.3 to 66.2) of their counterparts, yielding a prevalence difference of −14.3% (95% CI –27.0 to –2.3). Of the 35 individuals with heightened depressive symptoms at enrolment into care who disengaged from their original clinic at least once during the follow-up, 17.1% (95% CI 5.6 to 30.0) and 20.0% (95% CI 7.7 to 34.3) were re-engaged at their original clinic 3 and 6 months following the first instance of disengagement, respectively (table 3, figure 2B, C). The difference in the prevalence of re-engagement among those with vs without heightened symptoms of depression at enrolment into care was −4.6% (95% CI −18.9 to 10.6) 3 months following the first instance of disengagement and −10.2% (95% CI −25.7 to 5.9) 6 months following the first instance of disengagement (table 3). The instantaneous hazard of re-engagement among those who disengaged from their original clinic at least once was higher and occurred earlier among those without heightened depressive symptoms compared with those with heightened depressive symptoms at enrolment (online supplemental figure 2). Across the first 18 months following entry into HIV care, those with heightened depressive symptoms at enrolment spent, on average, 38.5 fewer days (ie, just over 1 month) engaged in HIV clinical care compared with those without heightened depressive symptoms (95% CI –78.4 to –1.2; online supplemental table 4).

Table 2. Proportion of individuals in each stage of the care continuum in months 12 and 18 following enrolment into care, stratified by depression status at baseline.

	Month 12			Month 18
	Depressed	Non-depressed		Depressed	Non-depressed
	Prevalence (95% CI)	Prevalence (95% CI)	PD (95% CI)	Prevalence (95% CI)	Prevalence (95% CI)	PD (95% CI)
Primary states
Linked to clinic, ART naïve	0.0 (0.0 to 0.0)	0.0 (0.0 to 0.0)	0.0 (0.0 to 0.0)	0.0 (0.0 to 0.0)	0.0 (0.0 to 0.0)	0.0 (0.0 to 0.0)
Engaged at clinic, prescribed ART	57.1 (46.1 to 67.5)	68.2 (63.1 to 73.4)	−11.0 (−23.4 to 0.7)	46.4 (35.6 to 56.8)	60.7 (55.3 to 66.2)	−14.3 (−27.0 to –2.3)
Disengaged from clinic	28.6 (19.2 to 38.7)	19.3 (15.2 to 23.6)	9.2 (−1.2 to 20.4)	28.6 (19.3 to 38.6)	18.7 (14.6 to 23.1)	9.8 (−0.4 to 20.9)
Re-engaged in clinic	3.6 (0.0 to 7.9)	4.8 (2.6 to 7.3)	−1.2 (−5.5 to 3.7)	8.3 (2.8 to 14.7)	8.6 (5.7 to 11.8)	−0.3 (−6.5 to 6.8)
Absorbing states
Known transfer out	10.7 (4.4 to 17.6)	7.1 (4.5 to 10.0)	3.6 (−3.2, 11.3)	11.9 (5.4 to 19.4)	8.0 (5.2 to 11.0)	3.9 (−3.3 to 12.0)
Known death	0.0 (0.0 to 0.0)	0.6 (0.0 to 1.5)	−0.6 (−1.5 to 0.0)	4.8 (1.1 to 10.1)	3.9 (2.0 to 6.0)	0.9 (−3.6 to 6.5)
Cumulative states
Total engaged	60.7 (50.0 to 70.7)	72.9 (68.2 to 77.7)	−12.2 (−24.4 to –1.4)	54.8 (43.8 to 65.1)	69.3 (64.3 to 74.3)	−14.6 (−27.1 to –3.2)

ART, antiretroviral treatment; PD, prevalence difference.

Table 3. Proportion of individuals in individuals in four clinical states at months 1 and 6 following the first instance of disengagement from care (n=141), stratified by depression status at baseline.

	Month 3			Month 6
	Depressed	Non-depressed		Depressed	Non-depressed
	Prevalence (95% CI)	Prevalence (95% CI)	PD (95% CI)	Prevalence (95% CI)	Prevalence (95% CI)	PD (95% CI)
Primary states
Disengaged from clinic	82.9 (70.0 to 94.4)	76.4 (67.7 to 84.3)	6.4 (−8.8 to 20.7)	71.4 (55.9 to 85.4)	64.2 (54.8 to 73.1)	7.3 (−10.5 to 24.1)
Re-engaged in clinic	17.1 (5.6 to 30.0)	21.7 (14.2 to 30.1)	−4.6 (−18.9 to 10.6)	20.0 (7.7 to 34.3)	30.2 (21.7 to 39.3)	−10.2 (−25.7 to 5.9)
Absorbing states
Known transfer out	0.0 (0.0 to 0.0)	0.0 (0.0 to 0.0)	0.0 (0.0 to 0.0)	0.0 (0.0 to 0.0)	1.0 (0.0 to 3.2)	−1.0 (−3.2 to 0.0)
Known death	0.0 (0.0 to 0.0)	1.9 (0.0 to 4.9)	−1.9 (−4.9 to 0.0)	8.6 (0.0 to 18.6)	4.7 (1.0 to 9.0)	3.9 (−5.2 to 14.7)

PD, Prevalence difference.

Discussion

We used longitudinal multistate methods to characterise longitudinal HIV care continuum outcomes in a cohort of people newly initiating HIV care in Cameroon and explore differences in these outcomes by depressive symptom status at entry into HIV care. We demonstrate that care outcomes in this cohort of people newly entering HIV care remain well below UNAIDS’ 95-95-95 targets, with particularly poor outcomes among those entering care with heightened depressive symptoms. While numerous studies have documented the relationship between depression and suboptimal HIV care outcomes, the evidence across settings has varied widely and has largely been limited to analyses that rely on cross-sectional data. In this work, we highlight that people living with HIV who enter care with heightened depressive symptoms are more susceptible to both disengagement from care and delayed re-entry into care following disengagement in the first 18 months of care.

In the population overall, 90% of individuals initiated treatment the same day they enrolled into care and over 10% failed to return to the clinic after their initial enrolment visit, indicating additional resources are needed to achieve UNAIDS’ 95-95-95 targets. These proportions were similar among those with and without heightened depressive symptoms at entry into care, indicating that early retention in HIV care remains a challenge, irrespective of mental health status. Evidence from similar settings supports the idea that individuals are particularly susceptible to disengagement from care in the periods immediately following diagnosis with HIV or entry into care, even in the era of immediate access to ART (ie, Universal Test and Treat, Test and Start).¹⁸ As such, identifying and scaling interventions that specifically aim to improve the client experience early in the HIV care-seeking journey, including during testing, ART initiation, and in the period immediately after care initiation (eg, between an individual’s first and second clinic visit), is urgent.32,34 Examples of such interventions that have been shown to be effective at improving early HIV care outcomes in similar settings include small financial incentives for adults entering HIV care in Tanzania (86% retention at month 6 following entry in the intervention group vs 73% in the control group)³⁵ and structured peer support for pregnant women living with HIV in rural Nigeria (adjusted OR exploring retention among women in the structured peer support group vs unstructured peer support=5.9; 95% CI 3.0 to 11.6)³⁶ among others.³⁷

Despite similarities in early engagement in care (ie, across the first 6 months) among those with and without heightened depressive symptoms at entry into care, those with heightened depressive symptoms were less likely to be engaged in care at both months 12 and 18 compared with others. Moreover, among those who disengaged from care at least once during follow-up, individuals with heightened depressive symptoms at enrolment into care were less likely to be re-engaged in care 6 months following their first instance of disengagement, suggesting mental ill-health may impede retention and re-engagement in care beyond 6 months after entry. Ultimately, these outcomes contributed to a difference of ~40 days of time spent in care at one’s original clinic across the first 18 months following entry among those with vs without heightened depressive symptoms at entry. These findings, supported by a growing body of literature,38,40 point to the importance of routine screening for heightened depressive symptoms at the time of HIV testing and enrolment into care, as this can improve linkage to mental healthcare and ultimately improve downstream HIV care outcomes (e.g. long-term retention and re-engagement in care). Moreover, integrated depression screening and treatment for people with HIV can ensure adequate access to comprehensive healthcare, thereby upholding these individuals’ right to health and improving their quality of life.41,43 Implementation strategies that have been shown to successfully improve integration of depression screening and referral in primary care clinics in Africa include approaches such as task-shifting, train-the-trainer, learning collaboratives, audit and feedback, and provider supervision.^{44 45} These integration implementation strategies should also be employed and tested in Cameroon. Additional research is needed to understand the importance of depression screening across an individual’s HIV care-seeking journey and the potential changes in patients’ depressive symptoms (and the downstream effects of any observed changes) as those with comorbid depression and HIV are established in care and on treatment.

There are numerous interventions that have been shown to yield improved depressive symptom outcomes among individuals living with HIV in the Global South and similarly resource-constrained settings, including group-based and peer-led or ‘task-shifted’ psychotherapies.^{46 47} In addition, at least one study in South Africa demonstrated the benefits of a mental health intervention, which consisted of a task-shared cognitive-behavioural therapy for adherence and depression, for improving both depressive symptoms and HIV care outcomes (eg, ART adherence, viral suppression) in a population of individuals with depression and uncontrolled viremia.⁴⁸ Efficacious mental health interventions that have been shown to improve outcomes among those living with HIV or other chronic conditions in the Global South should be scaled up in Cameroon, as contextually feasible and appropriate, to decrease the downstream consequences of poor mental health.

Limitations

This analysis has limitations worth noting. As in other under-resourced healthcare settings across the Global South, viral load monitoring was inconsistent and poorly captured in the clinics included in this study.⁴⁹ As a result, we did not have data to reliably estimate the proportion of the population and time spent virally suppressed in this study. Viral suppression is arguably the most relevant measure of successful HIV treatment and care; thus, additional research is warranted to characterise the longitudinal relationship between depression at entry into care and downstream viral suppression outcomes in Cameroon. Second, this study spanned the height of the COVID-19 pandemic during which clinical care services and clinical care seeking have been shown to vary, including among people living with HIV in the Global South and in Cameroon.50,52 Consequently, our findings may not be representative of outcomes in more stable care provision conditions. Third, we enrolled a population of individuals newly entering HIV care and did not have data regarding these individuals’ date of diagnosis with HIV. Moreover, we did not have repeated measures of depressive symptoms at routine intervals. Because of these limitations, we were unable to explore subgroup differences in those who may have reported heightened depressive symptoms at entry into care largely because of the acute stress associated with HIV diagnosis and those with chronic depressive symptoms. Additional longitudinal research is needed to characterise how depression changes across the course of individuals’ HIV care-seeking journeys and how this impacts outcomes across the HIV care continuum. Finally, clinical outcomes were not cross-validated with other data sources (eg, death registries, national treatment registers), which could have yielded misclassification bias. Double sampling and tracking and tracing activities that have been employed in other settings to improve the reliability of HIV care continuum estimates53,58 should be conducted in future work to minimise potential misclassification.

Conclusions

Depressive symptoms at entry into HIV care are associated with downstream suboptimal HIV care continuum outcomes, particularly in the period between 6 and 18 months following entry into care. Individuals who screen positive for heightened depressive symptoms at entry into HIV care may benefit from more personalised linkage to and retention in care services such as peer navigation and HIV treatment supporters. Moreover, brief depressive symptom screening may be warranted at subsequent HIV care-seeking visits, though additional research is needed to understand how depressive symptoms change over time among people entering HIV care, and how changes in symptomatology are related to longitudinal HIV care outcomes. Given the significant impact of depressive symptoms on long-term HIV care outcomes, it is essential to integrate targeted, ongoing mental health support and symptom monitoring into HIV care protocols to optimise patient retention and improve overall health outcomes.

Data availability statement

Data are available upon reasonable request.