Uncovering gaps in HIV care in belgium through patient reported outcome measures and neurocognitive screening

Abstract

Background

Patient-reported outcome measures (PROMs) are critical for evaluating health-related quality of life (HRQoL) among people living with HIV (PLHIV). This study explores HRQoL using PROMs and assesses the prevalence and implications of HIV-associated neurocognitive disorders (HAND) within the HIV Reference Centre (HRC) of Ghent University Hospital, a tertiary outpatient clinic providing routine multidisciplinary HIV care.

Methods

Between November 2021 and June 2023, 836 of 1,473 PLHIV followed completed a PROMs questionnaire assessing physical and mental health, sleep, social support, sexual health, stigma, and substance use. Screening for HAND was conducted using the European AIDS Clinical Society (EACS) guidelines, Montreal Cognitive Assessment (MoCA), and International HIV Dementia Scale (IHDS).

Results

Seventy-seven percent of the participants were male and mean age was 52 years (SD = 11.68). Satisfaction rates were high for physical (77%) and mental health (71%). Notably, 47% reported significant sleep disturbances, and 18.5% regularly used substances. Forty-four percent reported one or more neurocognitive complaints on the EACS screening questions, of whom 14% had impairment across all three cognitive domains. Of those reporting complaints, 36% underwent formal cognitive screening; within this screened subgroup, impaired scores were identified in 45% on the MoCA and 63% on the IHDS. Regression analyses indicated that negative feelings, sleep difficulties, low social support, and neurocognitive functioning significantly predicted overall HRQoL.

Conclusions

Sleep, substance use, and neurocognitive functioning emerge as points of interest among Belgian PLHIV. Systematic identification of cognitive difficulties—when integrated with appropriate referral pathways, cardiovascular risk management, and psychosocial support rather than as isolated screening—may contribute to preserving HRQoL. The potential benefits of screening must be weighed against the risk of increasing anxiety among PLHIV.

Introduction

Combination antiretroviral therapy (cART) has transformed HIV infection into a manageable chronic condition, with life expectancies of people living with HIV (PLHIV) now approaching those of HIV-negative individuals. With this success, attention in the HIV field has shifted beyond viral suppression toward optimising HRQoL in PLHIV [1]​. The WHO proposed adding a “fourth 95” target to the UNAIDS 95-95-95 goals—i.e. that 95% of those with virologically controlled HIV should achieve a good HRQoL [2]​. Reaching this goal requires understanding and addressing the factors that influence HRQoL in PLHIV [3, 4]. In this context, it is essential to identify the key determinants of HRQoL in contemporary HIV care.

HRQoL in PLHIV is known to be influenced by various psychosocial and health factors [5]. Previous studies have demonstrated that PLHIV experience higher levels of anxiety and depressive feelings compared to the general population [6, 7]. Indeed, depression is highly prevalent among PLHIV and has consistently been identified as a critical predictor of poor HRQoL. Depressive symptoms often exert a stronger impact on HRQoL than clinical HIV markers, and might overshadow neurocognitive symptoms in some cases, although the relationship may be bidirectional, as neurocognitive complaints can also be a manifestation of depression [8, 9]. Additionally, a UK survey of gay and heterosexual men living with HIV found that approximately 35% reported concerns about HIV-related sexual rejection [10], and more than half experience sleep disturbances [11, 12] highlighting the role of stigma-related psychosocial stressors that, while documented primarily in male populations, may affect PLHIV more broadly. Internalization of stigma contributes to reduced self-worth and feelings of shame among PLHIV [13–15].

Beyond individual psychological factors, social determinants also play a critical role in shaping HRQoL. Previous research has highlighted a moderate overall quality of life (QoL) among PLHIV, with particular concerns in the domains of social relationships and gender-specific differences in perceived well-being [16, 17]. In Belgium, a survey involving 505 PLHIV revealed that 43% experienced poor social support, and 26% had a diagnosis of depression. Moreover, 41% of respondents identified 'less stigma in society' as a primary need for improved living with HIV, and 65% had experienced discrimination due to their HIV status [18]. Consequently, social determinants such as employment status, social support, and experiences of stigma play critical roles in the overall well-being of PLHIV, underscoring the importance of considering multiple domains when evaluating HRQoL.

Neurocognitive impairment (NCI) is another important comorbidity that may compromise HRQoL in this population [19]. HIV can affect the central nervous system, and a spectrum of HAND persists even in the cART era. Estimates of the prevalence of NCI in treated PLHIV vary widely, from roughly 20% in some cohorts to almost 50% in others​ [20–24]. For instance, the large CHARTER study reported that nearly half of participants had measurable neurocognitive deficits despite effective therapy​ [20, 24]. Most cases are mild or asymptomatic impairments rather than overt dementia​. Nonetheless, even milder forms of NCI can impact everyday functioning—they have been associated with difficulties in employment, medication adherence and other activities​ [2]. NCI is also linked to lower self-reported HRQoL in PLHIV​ [2]. As the HIV-positive population ages, neurocognitive issues are expected to become increasingly prominent, potentially impeding the goal of maximising HRQoL [25]. However, the utility of the Frascati-based HAND criteria has been increasingly questioned. Recent consensus recommendations by the International HIV-Cognition Working Group advocate for moving away from solely quantitative methods of diagnosing cognitive impairment in PLHIV and emphasize the importance of clinical context [26]. The Frascati criteria may yield high false-positive rates, potentially overestimating the burden of cognitive impairment in the cART era [27, 28]. Furthermore, the conceptual separation of HIV-associated brain injury from other causes of brain injury in PLHIV has been proposed, as many comorbidities including aging, cardiovascular risk, substance use, and depression can independently contribute to low cognitive performance [29]. We acknowledge these limitations and note that the HAND terminology is considered by some as outdated; throughout this paper we therefore use the term “low cognitive performance” (LCP) to describe below-cutoff performance on screening tools, without implying that these findings represent confirmed cognitive impairment or are necessarily HIV-attributable.

To capture these multidimensional aspects of health, PROMs have emerged as valuable tools to address HRQoL and comorbidities in PLHIV. PROMs allow to capture a persons’ perception of their own health through questionnaires. They enable persons to report on their physical, social, mental, sexual and emotional well-being. Their use in clinical practice allows to detect “red flags” (i.e. difficulties that affect HRQoL) among PLHIV, resulting in appropriate counseling and encouraging them to self-care [30]. Using PROMs in everyday practice helps close the gap between how clinicians perceive clinical reality and how patients experience it, and it allows treatment plans to be personalized according to the patient’s preferences and specific needs. Research proves that incorporating information from PROMs results in better communication and decision-making between healthcare providers and PLHIV [31]. Interestingly, many of the factors assessed by PROMs—such as substance use, mental health, and sleep quality—may also affect neurocognitive functioning and offer insight into the underlying contributors to NCI. PROMs help improve the quality of care and maximizing involvement of PLHIV, thus help attaining the UNAIDS fourth 90 goal of providing comprehensive and patient-centered care [32].

In parallel, patient satisfaction represents a complementary dimension of patient-centred care that may interact with HRQoL. Satisfaction may influence QoL both directly—by making PLHIV feel heard, respected, and supported—and indirectly, by encouraging sustained engagement in care and adherence to treatment, ultimately leading to better clinical outcomes and improved QoL [33]. However, this relationship is not always linear; for instance, individuals with low HRQoL may still report high satisfaction with their care, and vice versa. Assessing satisfaction alongside HRQoL can therefore provide a more comprehensive view of patient-centered outcomes in HIV care. This assessment can be done through patient-reported experience measures (PREMs), questionnaires that focus on the care process and how patients experience it. Despite its importance, research on satisfaction among people living with HIV remains limited, and it is unclear whether factors such as NCI or poor HRQoL negatively influence how patients perceive their care.

These domains can be integrated within a broader conceptual framework linking clinical factors to PROMs. Our study is conceptually grounded in the biopsychosocial model of health, which posits that biological, psychological, and social factors interact to determine health outcomes, and in the Wilson-Cleary conceptual framework linking clinical variables to PROMs [34]. HRQoL, as distinct from the broader concept of QoL, specifically captures the impact of health status and healthcare interventions on an individual’s perceived physical, mental, and social well-being. Whereas QoL encompasses general life satisfaction including non-health domains such as financial security and environment, HRQoL focuses on health-related dimensions and is thus more directly amenable to clinical intervention. In this study, we primarily assess HRQoL and use this term consistently throughout, reserving “QoL” for discussions of broader life satisfaction where appropriate.

In this study, we assess HRQoL and patient satisfaction with healthcare services in a cohort of PLHIV on cART, and we examine how these relate to neurocognitive functioning and other psychosocial factors. Specifically, our objectives were: (1, descriptive) describe patient-reported HRQoL and satisfaction with care, (2, analytical) identify psychosocial and clinical predictors of HRQoL, and (3, descriptive) estimate the prevalence of HAND and explore its correlates, among PLHIV on cART followed at a Belgian reference centre. Together, these descriptive estimates provide the necessary context for the subsequent identification of psychosocial and clinical predictors of HRQoL. By addressing these, we hope to inform interventions to improve long-term HRQoL and patient experience for PLHIV, aligning with the goal of truly patient-centered HIV care.

Materials and methods

Study design and participants

We conducted a cross-sectional observational study among PLHIV on cART receiving routine care at the HRC of Ghent University Hospital in Belgium. Participants ≥ 18 years were approached during clinic visits during the study period November 2021 and June 2023 and provided informed consent to complete a survey and/or brief assessment as part of this study. The research protocol was approved by the Ethical Committee of Ghent University Hospital (BC-11836-AM01). Clinical trial number: not applicable.

Data collection

PROMs and PREMs

Participants underwent a tablet-based assessment including 24 items on mental and physical health, sleep quality, social support, sexual health, stigma, alcohol- and substance use (before or during sexual activities) and neurocognitive complaints. Questionnaires were available in Dutch, English and French. These PROMs were examined using a five-point scale ranging from "very unsatisfied" to "unsatisfied," "neither satisfied nor unsatisfied," "satisfied," and "very satisfied.1" For the frequency-related questions, participants were asked to indicate how often they engage in certain behaviors or experience certain phenomena. The response options were "always," "very often," "often," "rarely," and "never." Furthermore, participants were asked to rate their overall satisfaction with the (para)medical care and services they receive for their HIV (such as communication with healthcare providers, ease of access to the clinic, and perceived quality of services). These PREMs were examined using a 5-point Likert scale from 1 (“very dissatisfied”) to 5 (“very satisfied”). Demographic variables (age, sex at birth, educational level, current employment status) and clinical information (latest CD4 + T-lymphocyte count (CD4), time since diagnosis, antiretroviral regimen) was obtained from medical records.

The domains assessed (mental health, physical health, sleep, social support, sexual health, stigma, substance use) were selected on the basis of the EACS guidelines and informed by the domains reported in a prior exploratory Belgian cross-sectional survey among PLHIV [18]. These domains correspond conceptually to domains assessed in validated HIV-specific HRQoL instruments such as the WHOQOL-HIV-BREF and MOS-HIV [35, 36], although our single-item Likert-scale questionnaire has not itself undergone formal psychometric validation and should therefore not be considered equivalent to these validated tools. In the remainder of this manuscript, 'patient-reported health status' refers to the descriptive distribution of Likert responses on the individual PROM domains, as shown in Figs. 1, 2 and 3. It is distinct from HRQoL, which is the composite sum score of physical and mental health satisfaction items used as the primary outcome in Table 3.

Fig. 1.

Distribution of self-reported satisfaction across five health domains among 836 PLHIV. Bars depict the proportion of responses from ‘very unsatisfied’ (red) to ‘very satisfied’ (dark green) and ‘not applicable’ (grey)

Fig. 2.

Self-reported patterns of substance use and perceived impact among respondents. Left panel: frequency of alcohol consumption and other (non-prescription) substance use. Right panel: frequency of chemsex engagement and the perceived negative impact of substance use on daily life. Bars show the percentage of respondents in each frequency category. Color shading transitions from dark green to red with increasing frequency

Fig. 3.

Frequency of reported experiences of stigma and discrimination among respondents. Categorized by type of stigma and frequency of occurrence. Responses are presented as percentages of total respondents per category

Table 3.

Regression analysis predicting HRQoL based on gender, frequency of negative feelings, sleep satisfaction, social support, sexual health satisfaction and neurocognitive functioning. Variables retained in the multivariable model based on omnibus likelihood-ratio test (p < 0.05); p-values shown are for individual category levels versus reference

	Beta	95% CI	p-value
Gender
Male		-	-
Female	0.31	0.03, 0.58	0.03
Negative feelings
Never	-	-
Rarely	-0.22	-0.48, 0.05	0.11
Often	-1.0	-1.3, -0.67	< 0.001
Very often	-2.0	-2.4, -1.5	< 0.001
Always	-1.9	-2.8, -1.1	< 0.001
Sleep
Very unsatisfied	-	-
Unsatisfied	0.46	-0.02, 0.95	0.061
Neither satisfied nor unsatisfied	0.87	0.39, 1.3	< 0.001
Satisfied	1.0	0.54, 1.5	< 0.001
Very satisfied	1.3	0.78, 1.8	< 0.001
Social support
Not applicable	-	-
Very unsatisfied	-0.29	-1.0, 0.42	0.4
Unsatisfied	-0.90	-1.6, -0.15	0.02
Neither satisfied nor unsatisfied	-0.10	-0.53, 0.33	0.7
Satisfied	-0.02	-0.39, 0.36	> 0.9
Very satisfied	0.20	-0.17, 0.58	0.3
Sexual health
Very unsatisfied	-	-
Unsatisfied	0.90	0.44, 1.4	< 0.001
Neither satisfied nor unsatisfied	0.90	0.48, 1.3	< 0.001
Satisfied	1.1	0.64, 1.5	< 0.001
Very satisfied	1.5	0.98, 2.0	< 0.001
Neurocognitive functioning	-0.17	-0.27, -0.08	< 0.001

The p < 0.05 threshold refers to the pooled likelihood-ratio test comparing the full model to a reduced model omitting each variable (i.e., in the multivariable context). Variables not reaching significance at p < 0.05 were excluded from the displayed table for parsimony but were retained in the full model. Multiple imputation (50 datasets, predictive mean matching) was applied to the complete sample of 836 respondents

Neurocognitive assessment

To screen for NCI, participants were initially assessed using the standard EACS screening questions [37]. Individuals who reported cognitive complaints, defined as answering positively to all of the three questions, were subsequently contacted by phone and invited for further neurocognitive evaluation at our clinic. Those consenting underwent additional screening, employing both the MoCA and the IHDS [38, 39]. The IHDS is a brief screening tool specifically recommended for identifying signs of HIV-associated NCI in clinical settings, assessing motor speed, psychomotor speed, and memory-recall. Scores range from 0 to 12, with lower scores indicating worse performance. The MoCa is validated as a highly sensitive tool for early detection of mild cognitive impairment (MCI). It includes the following 8 cognitive domains: executive functions, visuospatial skills, attention, concentration and work rate, language, short-term memory and orientation, with a total sum score of 30 points. Following published guidelines, NCI was defined as an IHDS score ≤ 10, which suggests at least mild impairment. Additionally, cognitive impairment was indicated by a MoCA score below 26 (out of 30) [40]. Participants were administered the IHDS and MoCA in their preferred language. Those with results below established cutoff values on one or both screening instruments (i.e. MoCA score < 26 and/or IHDS score ≤ 10) — suggestive of cognitive difficulties warranting further evaluation — were offered referral for further neuropsychological evaluation at the department of Neurology of Ghent University hospital.

Statistical analysis

Descriptive statistics were calculated as medians and interquartile ranges for continuous variables and counts (percentages) for categorical variables. Missing data were handled under the Missing At Random assumption by performing 50 imputations using predictive mean matching (mice, m = 50, maxit = 5, seed = 1), with convergence assessed by inspection of trace plots. The primary outcome variable “HRQoL” was operationalized as a composite sum score of the individual responses on physical health satisfaction and mental health satisfaction items (each scored 1–5, total range 2–10). Higher scores indicate better self-reported HRQoL. This composite approach was chosen because physical and mental health are consistently identified as the core components of HRQoL in the literature. The individual domain scores (sleep, social support, sexual health, stigma, substance use, neurocognitive functioning) were included as predictor variables in the regression models. For the neurocognitive functioning predictor, a composite score was created by summing responses on three items assessing self-reported difficulties in reasoning, memory, and attention (each scored 0–1, total range 0–3).

For each of the five outcomes—patient‐reported HRQoL, satisfaction with medical care, satisfaction with paramedical care, MoCA score, and IHDS score—a full linear regression model was fit to each imputed dataset including demographic factors (sex, age, region of origin), clinical covariates (time since diagnosis, current CD4 count, antiretroviral regimen), psychosocial measures (negative affect, sleep quality, social support), substance‐use indicators (alcohol and drug use) and composite neurocognitive or stigma scores where applicable. To test the significance of each variable, a reduced model omitting that variable was compared to the full model via likelihood‐ratio tests: the Meng & Rubin (1992) pooling method was applied for the HRQoL and satisfaction outcomes, and the Li et al. (1991) approach for the MoCA and IHDS scores [40, 41]. Pooled χ² statistics, degrees of freedom, p-values and relative increase in variance (RIV) were reported, with p < 0.05 deemed significant and p < 0.10 noted as trends. Finally, model assumptions of linearity, homoscedasticity and normality of residuals were verified graphically using residuals‐versus‐fitted and Q–Q plots for each imputed dataset. All analyses were performed using the R package (R Core Team (2024)), for multiple imputation the mice package was used [42]. Throughout, we use 'primary predictors of interest' to refer to the six psychosocial and clinical variables specified a priori on the basis of the Wilson-Cleary framework and EACS guidance, and 'covariates' to refer to the sociodemographic and clinical variables included for adjustment. Each variable is used in only one role per model.

HRQoL was assessed using regression analysis, with beta coefficients (β) representing the change in HRQoL per unit increase in each predictor variable; positive β indicates improved HRQoL, and negative β indicates reduced HRQoL. For each outcome, we fitted multivariable linear regression models (ordinary least squares) to each of the multiply imputed datasets. Full models included all prespecified covariates (Table 1) based on clinical relevance and prior literature, rather than data-driven stepwise selection, thus avoiding the pitfalls of arbitrary variable selection. Model fit was assessed through inspection of residuals-versus-fitted and Q-Q plots for each imputed dataset, confirming adequate linearity, homoscedasticity, and approximate normality of residuals.

Table 1.

Overview of variables included for each analysis

	HRQoL (N = 836)	Satisfaction medical care (N = 836)		Satisfaction paramedical care (N = 836)	MoCA (n = 43)	IHDS (n = 43)
Gender	x	x	x		x	x
Age	x	x	x		x	x
Negative feelings	x	x	x		x	x
Sleep	x	x	x		x	x
Social support	x	x	x		x	x
Sexual health	x	x	x		x	x
Alcohol use	x	x	x		x	x
Substance use	x	x	x		x	x
Chemsex	x
Neurocognitive functioninga	x	x	x
Stigmaa	x	x	x
Current CD4 count	x
CD4 nadir					x	x
Date diagnosis	x	x	x		x	x
Sexual orientation	x	x	x
ART regimen	x	x	x		x	x
Origin	x	x	x		x	x
Mental healtha		x	x
Physical healtha		x	x
Language of questionnaire		x	x
Adherence		x	x
Transmission route		x	x
Satisfaction with general information		x	x
Satisfaction with information about therapy options		x	x

^aFor variable ‘neurocognitive functioning’ a sum score was made of the components “reasoning”, “memory” and “attention”. Similar for “stigma” for which a sum score was made of “professional stigma”, “discrimination” and “self-stigma”. The response variable “HRQoL” was based on the sum score of “mental health” and “physical health”. ‘Chemsex’ was defined as the intentional use of psychoactive substances (typically crystal methamphetamine, mephedrone, GHB/GBL, or ketamine) immediately before or during sexual activity to initiate, sustain, or enhance sexual engagement

All 836 respondents completed the full PROMs questionnaire. The regression analyses for HRQoL, satisfaction with medical care, and satisfaction with paramedical care were based on the full sample (N = 836), with missing data handled through multiple imputation. The MoCA and IHDS analyses were restricted to the 43 participants who underwent cognitive screening. The sample size for each analysis is specified in Table 1.

Table 1 gives an overview of which variables were included for each analysis.

Results

Participants response and inclusion

Out of the 1473 PLHIV followed at the HRC Ghent during the study period, 836 (57%) completed the questionnaire (Dutch: n = 657; English: n = 93; French: n = 86). Among these respondents, 118 (14%) individuals provided three positive answers on the EACS neurocognitive screening questions. Of these 118 individuals, 35 (30%) could not be reached despite multiple contact attempts, and 40 (34%) declined participation for various reasons (e.g., lack of interest, insufficient time). Ultimately, 43 individuals (36%) agreed to participate and were included in the neurocognitive assessment analysis.

Socio-demographic characteristics

Out of the 836 respondents, the majority were male (77%, N = 646). The mean age was 52 (SD = 11.68) years. Regarding sexual orientation, 42% (N = 349) of participants identified as homosexual, 31% (N = 257) as heterosexual, and 2% (N = 22) as bisexual; however, sexual orientation data was missing or unknown for approximately one quarter of the respondents (25%, N = 208). The majority of the participants (73%; N = 609) were of European origin. All respondents were on cART, of which the majority on a integrase based regimen (84%, N = 704). A detailed summary of socio-demographic variables is presented in Table 2.

Table 2.

Socio-demographic profile and clinical characteristics of the 836 respondents. Data presented as number (%) or median (Q1, Q3). CD4 count expressed in cells/µl

	N = 836
Gender
Male	646 (77%)
Female	190 (23%)
Male-to-female ratio	3.4
Mean Age (SD)	52 (11.68)
Language
Dutch	657 (79%)
French	93 (11%)
English	86 (10%)
Current CD4 count (in cells/µl)	740 (548,935)
Transmission route
Homo/bisexual contact	234 (28%)
Injecting drug user	3 (0.4%)
Homo/bisexual + injecting drug user (IDU)	4 (0.5%)
Heterosexual contact	95 (11%)
Heterosexual contact + IDU	2 (0.2%)
Transfusion	2 (0.2%)
Perinatal	6 (0.7%)
Unknown	481 (58%)a
Sexual orientation
Homosexual	349 (42%)
Heterosexual	257 (31%)
Bisexual	22 (2.6%)
Unknown	208 (25%)a
Origin
European	609 (73.1%)
Asiatic	42 (5%)
Latin-American	24 (2.9%)
Sub-Saharan African	155 (19%)
Other	6 (0.7%)
ART regimen
PI-based	52 (6.2%)
Integrase based	704 (84%)
NRTI/NNRTI combination	80 (9.6%)

^aMissing data were present for the following variables: transmission route (58% unknown, due to electronic medical record registration issues), sexual orientation (25% unknown), and scattered missingness (< 5%) for individual PROMs items. The Missing At Random assumption was considered appropriate given that the dominant source of missingness (transmission route) was attributable to administrative registration practices in the electronic medical record rather than to patient characteristics or outcomes. Convergence of the imputation model was confirmed by inspection of trace plots.

Patient-reported health status

The majority of respondents indicated being satisfied (57%, n = 480) or very satisfied (20%, n = 164) with their physical health. Similar results were reported for mental health, with 52% (n = 437) expressing satisfaction and 19% (n = 157) reporting being very satisfied. Approximately one in five respondents (17.6%, n = 147) expressed dissatisfaction with their sexual life. Nearly half of the respondents (47.3%, n = 393) reported experiencing sleep difficulties. Most respondents (76%, n = 638) indicated satisfaction with the social support they received from friends and/or family (see Fig. 1).

Eighty-eight respondents (11%) reported daily alcohol consumption, while 34% (N = 284) drink alcohol weekly. 18.5% indicated monthly to daily use of substances. Additionally, 14.2% of respondents reported frequent (often to always) use of narcotics or stimulants in relation to sexual activities (chemsex). Among those engaging in chemsex, 8.1% reported experiencing a negative impact on their HRQoL (see Fig. 2).

One hundred and ten respondents (13.2%) frequently (often to always) encountered discrimination, stereotypes, and/or negative attitudes in their personal lives. A majority (60%, n = 500) had not disclosed their HIV status at work, and 31% (n = 261) reported having ever experienced discrimination or negative attitudes in a professional context. Additionally, approximately one-third of respondents (n = 274) reported experiencing HIV-related self-stigma, such as feelings of guilt or shame (see Fig. 3).

Most respondents expressed satisfaction or high satisfaction with the medical and paramedical care received at the hospital. Specifically, 25% (n = 212) were satisfied and 69% (n = 577) were very satisfied with medical care, whereas 20% (n = 167) were satisfied and 40% (n = 333) were very satisfied with paramedical care (see Fig. 4).

Fig. 4.

Distribution of respondents' satisfaction levels for paramedical and medical care (PREMs). The stacked bar chart illustrates the proportion (%) of respondents reporting varying degrees of satisfaction, highlighting notable differences between the two categories. "No answer" was relevant only for paramedical care

Factors associated with HRQoL, satisfaction with care, and neurocognitive performance

HRQoL analysis

Women reported slightly higher HRQoL compared to men (β = 0.31, 95% CI [0.03, 0.58], p = 0.03). Additionally, frequent experience of negative feelings (p < 0.001), poor sleep quality (p < 0.001), lack of social support (p = 0.02), and sexual health (p < 0.001) and impaired neurocognitive functioning (p < 0.001) were correlated with poor HRQol (see Table 3).

Satisfaction of medical care

Mental health, negative feelings, social support, and transmission route emerged as significant variables of satisfaction of medical health (see Table 4). Individuals reporting being "very satisfied" with their mental health expressed significantly higher satisfaction with medical care compared to those who were "very unsatisfied" (β = 0.61, 95% CI [0.10, 1.1], p = 0.019).

Table 4.

Regression analysis predicting satisfaction with medical care (PREMs). Beta coefficients (β), 95% confidence intervals (CI), and p-values are reported for negative feelings, mental health, social support, transmission route. Reference categories for each predictor are indicated by a dash (–). Variables retained in the multivariable model based on omnibus likelihood-ratio test (p < 0.05); p-values shown are for individual category levels versus reference

	Beta	95% CI	p-value
Negative feelings
Never	-	-
Rarely	0.08	-0.10, 0.47	0.4
Often	0.28	0.04, 0.51	0.02
Very often	0.55	0.21, 0.90	0.002
Always	0.62	0.05, 1.2	0.034
Mental health
Very unsatisfied	-	-
Unsatisfied	0.02	-0.47, 0.50	> 0.9
Neither satisfied nor unsatisfied	0.39	-0.10, 0.88	0.12
Satisfied	0.49	0.00, 0.97	0.052
Very satisfied	0.61	0.10, 1.1	0.019
Social support
Not applicable	-	-
Very unsatisfied	-0.72	-1.2, -0.25	0.003
Unsatisfied	-0.71	-1.2, -0.22	0.005
Neither satisfied nor unsatisfied	0.24	-0.05, 0.52	0.10
Satisfied	0.10	-0.15, 0.35	0.4
Very satisfied	0.28	0.03, 0.53	0.029
Transmission route
Homo/bisexual	-	-
Injecting drug user	-1.3	-2.3, -0.37	0.007
Homo/bisexual + injecting drug user	0.11	-0.76, 0.98	0.8
Transfusion	0.06	-1.2, 1.3	> 0.9
Heterosexual contact	0.12	-0.12, 0.36	0.3
Heterosexual contact + IDU	-1.9	-3.1, -0.68	0.002
Perinatal	0.07	-0.68, 0.83	0.9
Unknown	0.01	-0.14, 0.17	0.9

Interestingly, negative feelings also significantly predicted satisfaction. Compared to individuals who never experienced negative feelings, those reporting negative feelings "often" (β = 0.28, 95% CI [0.04, 0.51], p = 0.02), "very often" (β = 0.55, 95% CI [0.21, 0.90], p = 0.002), and "always" (β = 0.62, 95% CI [0.05, 1.2], p = 0.034) reported significantly higher satisfaction with medical care.

Social support was another significant variable. Compared to participants for whom social support was "not applicable," those who were "very unsatisfied" (β = -0.72, 95% CI [-1.2, -0.25], p = 0.003) or "unsatisfied" (β = -0.71, 95% CI [-1.2, -0.22], p = 0.005) reported significantly lower satisfaction, whereas those "very satisfied" reported significantly higher satisfaction (β = 0.28, 95% CI [0.03, 0.53], p = 0.029).

Furthermore, transmission route also significantly was associated with satisfaction with medical care. Injecting drug users (IDU) (β = -1.3, 95% CI [-2.3, -0.37], p = 0.007) and those reporting heterosexual contact combined with IDU (β = -1.9, 95% CI [-3.1, -0.68], p = 0.002) reported significantly lower satisfaction with medical care compared to individuals with homo/bisexual transmission routes.

Satisfaction of paramedical care

Similar to satisfaction of medical care, mental health, negative feelings, social support, and transmission route emerged as significant variables (see Table 5).

Table 5.

Regression analysis predicting satisfaction with paramedical care (PREMs). Beta coefficients (β), 95% confidence intervals (CI), and p-values are reported for negative feelings, mental health, social support, transmission route. Reference categories for each predictor are indicated by a dash (–). Variables retained in the multivariable model based on omnibus likelihood-ratio test (p < 0.05); p-values shown are for individual category levels versus reference

	Beta	95% CI	p-value
Negative feelings
Never	-	-
Rarely	0.08	-0.09, 0.26	0.4
Often	0.28	0.05, 0.51	0.019
Very often	0.55	0.21, 0.90	0.002
Always	0.63	0.07, 1.2	0.029
Mental health
Very unsatisfied	-	-
Unsatisfied	0.01	-0.47, 0.48	> 0.9
Neither satisfied nor unsatisfied	0.38	-0.10, 0.86	0.12
Satisfied	0.47	-0.01, 0.95	0.05
Very satisfied	0.60	0.10, 1.1	0.02
Social support
Not applicable	-	-
Very unsatisfied	-0.71	1.2, -0.24	0.003
Unsatisfied	-0.71	1.2, -0.23	0.004
Neither satisfied nor unsatisfied	0.24	-0.05, 0.52	0.10
Satisfied	0.10	-0.15, 0.35	0.4
Very satisfied	0.28	0.03, 0.53	0.028
Transmission route
Homo/bisexual	-	-
Injecting drug user	-1.3	2.3, -0.37	0.007
Homo/bisexual + injecting drug user	0.13	-0.74, 1.0	0.8
Transfusion	0.05	-1.2, 1.3	> 0.9
Heterosexual contact	0.13	-0.12, 0.37	0.3
Heterosexual contact + IDU	-1.9	-3.1, -0.68	0.002
Perinatal	0.06	-0.69, 0.81	0.9
Unknown	0.01	-0.15, 0.17	0.9

Neurocognitive screening

Cognitive difficulties were equally divided among the three domains (memory (29%, n = 237), reasoning (28%, n = 235), and attention (29%, n = 240)). Eighteen percent (n = 154) of participants reported difficulties in just one cognitive domain, while 12% (n = 101) reported difficulties in two domains. Fourteen percent (n = 118) of respondents experienced difficulties across all three cognitive domains. A total of 55% (n = 463) of participants indicated no cognitive difficulties. No significant differences were observed in demographic variables between groups of participants reporting no difficulties, difficulties in one, two or all domains. Among those who experienced difficulties across all three cognitive domains, 36% (n = 43) were subsequently screened for NCI using the MoCA and IHDS assessments. There were no significant demographic differences between participants who completed the assessments and those who either could not be reached or chose not to participate.

The average MoCA score among participants was 25.1 out of 30 (range: 16–30), while the average IHDS score was 9.2 out of 12 (range: 3–12). Cross-tabulation of MoCA and IHDS scores indicated varying levels of agreement between the tests. Among participants, 17 individuals (40%) demonstrated cognitive impairment according to both MoCA (score < 26) and IHDS (score ≤ 10). Conversely, two individuals (4.7%) who showed impairment according to the MoCA had normal IHDS scores (> 10). Additionally, 10 individuals (23.3%) scored as impaired on the IHDS but normal (≥ 26) on the MoCA. Fourteen individuals (32%) showed normal cognitive function on both tests. These incongruencies were further confirmed statistically, revealing a significant difference between the average scores of the two screening tools (p < 0.001).

Montreal cognitive assessment

Several factors significantly influenced the MoCA scores. Specifically, participants who were satisfied with their social support showed marginally significant better MoCA scores compared to those without applicable social support (β = 5.8, 95% CI [0.05, 12], p = 0.048). Similarly, participants who were very satisfied with their social support also demonstrated marginally significant better cognitive scores (β = 4.9, 95% CI [0.01, 9.8], p = 0.05) (see Table 6).

Table 6.

Factors associated with neurocognitive functioning scores. Beta coefficients (β), 95% confidence intervals (CI), and p-values from regression analysis assessing the relationship between neurocognitive functioning scores and variables: sleep quality, social support, origin, age and alcohol use.

	Beta	95% CI	p-value
MOCA
Sleep
Unsatisfied	1.0	-4.3, 6.4	0.7
Neither satisfied nor unsatisfied	-3.4	-9.4, 2.6	0.2
Satisfied	-1.8	-7.7, 4.1	0.5
Very satisfied	-7.2	-15, 0.48	0.064
Social support
Not applicable	-	-
Very unsatisfied	-0.28	-7.8, 7.3	>0.9
Unsatisfied	-1.0	-7.0, 4.9	0.7
Neither satisfied nor unsatisfied	-0.41	-5.8, 5.0	0.9
Satisfied	5.8	0.05, 12	0.048
Very satisfied	4.9	0.01, 9.8	0.05
Origin
Asiatic	-	-
European	-7.6	-19, 4.0	0.2
Latin-American	-4.4	-17, 7.7	0.4
Sub-Saharan	-8.3	-20, 3.1	0.14
IHDS
Age	-0.10	-0.19, 0.00	0.044
Alcohol use
Never	-	-
Monthly	0.76	-2.3, 3.8	0.6
Weekly	1.6	-0.82, 4.0	0.2
Daily	2.8	-0.11, 5.8	0.058
Origin
Asiatic	-	-
European	-7.6	-16, 1.2	0.085
Latin-American	-8.9	-18, 0.36	0.058
Sub-Saharan	-10	19, -1.3	0.028

Exploratory analysis; individual category-level p-values should be interpreted with caution given the limited sample size (N = 43. Multiple imputation (m = 50, predictive mean matching) was used to handle item-level missingness within this subgroup only; unscreened participants were not imputed.

International HIV dementia scale

Age, alcohol use, and ethnic origin were significantly associated with IHDS scores. Older age was marginally significant associated with worse IHDS scores (β = -0.10, 95% CI [-0.19, 0.00], p = 0.044). Participants of Sub-Saharan origin scored significantly lower on the IHDS compared to Asiatic participants (β = -10, 95% CI [-19, -1.3], p = 0.028) (see Table 6).

Discussion

In this cross-sectional study among 836 PLHIV on cART followed at the HRC of Ghent University Hospital, we aimed to (1) describe patient-reported HRQoL and satisfaction with care, (2) identify psychosocial and clinical predictors of HRQoL, and (3) estimate the prevalence of HAND and explore its correlates. Three findings emerge as the principal targets for future intervention. First, sleep disturbances were reported by 47% of respondents and were an independent predictor of lower HRQoL, consistent with a growing literature on the under-recognition of sleep problems in virologically suppressed PLHIV. Second, substance use was common, with 18.5% reporting monthly-to-daily use and 14.2% engaging in chemsex; in the latter subgroup, 8.1% reported a direct negative impact on HRQoL, situating chemsex at the intersection of sexual, mental, and infectious-disease care. Third, 44% of respondents reported one or more neurocognitive complaints on the EACS screening questions, and, among the subgroup who underwent formal screening, below-cutoff performance was observed in 45% on the MoCA and 63% on the IHDS, confirming that cognitive concerns remain prevalent in the cART era while also highlighting the well-known limitations of screening tools in this population.

The similarity in gender ratio, age, and ethnicity between the study participants and the overall population of the HRC in Ghent (N = 1473; man/woman ratio 2.6, mean age 48.2; 66% European, 23% Sub-Saharan African, 2% Latin-American, 3% Asiatic and 5% other) indicate that the study sample is broadly representative of the overall HRC population of Ghent. Compared to the 2018–2023 national HIV surveillance data of PLHIV [43], the study sample was representative of most sociodemographic characteristics. Nonetheless, among the study participants there were proportionally slightly more MSM and individuals of Belgian nationality, and fewer persons with Sub-Saharan African nationalities. Though, this corresponds with surveillance data showing that non-Belgian nationalities among heterosexuals were proportionally more present in the Brussels region than in the other regions [43].

HRQoL was most strongly associated with mental health status: reporting negative feelings had a profound negative impact on HRQoL. Numerous studies have identified depression as the single most important predictor of poor HRQoL among PLHIV [8]​. For example, a study of older HIV-positive adults in a European cohort found that depressive symptoms and social role satisfaction were key determinants of HRQoL, with cognitive function also playing a role [8]​. Our multivariable analysis similarly found negative feelings to have the largest effect size on HRQoL. An important interpretive caveat is that the association between negative feelings and reduced HRQoL may partly reflect mood-congruent reporting bias rather than a direct causal pathway. Depressed or distressed individuals tend to evaluate their own health and functioning more negatively than objective measures might suggest [8]. Since both our HRQoL measure and negative feelings assessment rely on self-report, shared-method variance may inflate the observed association. This bidirectional relationship between mood and self-reported HRQoL has been previously documented in PLHIV, where depressive symptoms were found to be the strongest predictor of perceived poor HRQoL, potentially overshadowing other objective functional limitations [8]. Nevertheless, regardless of whether mood directly impairs functioning or merely colours self-perception, the clinical implication remains the same: addressing negative affect through screening and intervention is likely to improve both experienced and reported HRQoL.

Even mild negative feelings had an appreciable impact, so early identification and treatment (whether through counseling, peer support, or pharmacotherapy) could significantly elevate PLHIV’ HRQoL. Interventions such as antidepressant therapy or cognitive-behavioural therapy might yield dual benefits for mood and cognitive function, thereby improving overall HRQoL.

The observed prevalence of substance use among our respondents — 11% reporting daily alcohol consumption, 18.5% indicating monthly to daily use of substances, and 14.2% (of which 83% MSM) engaging in chemsex—underscores the multifaceted challenges faced by PLHIV. Notably, 8.1% of those participating in chemsex reported a negative impact on their HRQoL. Importantly, previous research showed that substance use has been associated with decreased adherence to antiretroviral therapy [44] and has been strongly associated with sexual behaviors linked to increased risk of acquiring sexually transmitted infections (STIs) [45, 46]. This highlight the need for targeted prevention and intervention strategies, which requires bridging the gap between drug addiction services and sexual health expertise to ensure more comprehensive and accessible care [47, 48].

Sleep disturbances emerged as a substantial concern in our study, affecting 47% of participants, aligning with findings from recent literature indicating a prevalence exceeding 50% among PLHIV [12]. Such disturbances are linked to various biological, psychological, and therapeutic factors, including neuroinflammation and side effects of certain cART [12, 49]. Sleep deficiencies in PLHIV are not only common but also underdiagnosed, potentially impacting adherence to ART and overall health outcomes [50]. Moreover, sleep disturbances have been identified as independent risk factors for increased suicide risk in this population [51]. Importantly, sleep difficulties independently contributed to lower HRQoL in our cohort, underscoring the clinical relevance of addressing sleep issues and offering sleep hygiene education.

Even in a clinic population with high viral suppression and good overall care engagement, almost half of PLHIV (44%) have reported neurocognitive complaints based on EACS guidelines. Of those further assessed with MoCA and IHDS, 68% had lower neurocognitive scores on one or both tests and were referred for a comprehensive neurocognitive evaluation. Previous studies have quoted prevalence estimates from 20% to as high as 50% depending on the population and assessment methods [19]. The CHARTER study in the USA, which performed comprehensive neuropsychological testing, found that 47% of participants had some degree of impairment (mostly asymptomatic) [20,24]. Our measurement of NCI was based on screening tools (MoCA and IHDS) rather than a comprehensive neuropsychological battery. We acknowledge that the MoCA cutoff of < 26, while widely used, has been associated with high false-positive rates across multiple populations [39]. Rosca et al. [39] specifically noted that a lower threshold is probably more useful for optimal screening, as it lowers false-positive rates and improves diagnostic accuracy. A meta-analysis has proposed a cutoff of 23 as optimal for detecting mild cognitive impairment [52]. Had we applied this lower cutoff, the proportion of screen-positive participants would have been substantially lower. Our findings should therefore be interpreted as reflecting “low cognitive performance on screening tools” rather than confirmed cognitive impairment. Furthermore, these cutoffs have not been specifically validated in a Belgian HIV population, and factors such as language barriers (21% completed the questionnaire in a non-native language), education level, and cultural background are known to influence MoCA performance [29]. We recommend that future studies explore population-specific cutoff validation and consider analyzing cognitive scores as continuous variables to avoid the limitations of arbitrary dichotomization.

The IHDS appeared more sensitive than the MoCA, identifying impairment in 10 cases where MoCA did not, but it may have lower specificity. Conversely, MoCA showed higher specificity but may miss subtle cases (lower sensitivity). Notably, 32% of individuals had normal results on both tests while keeping in mind that our screening already filtered out the most severe cases by focusing only on persons with 3 affirmative answers. These findings support using both MoCA and IHDS to improve diagnostic accuracy, as each test complements the other's strengths and limitations [53]. To provide a more conservative estimate: of the total 836 respondents, 118 (14%) reported cognitive complaints on all three EACS screening questions. Of the 43 who subsequently underwent MoCA/IHDS screening, 29 (67%) were screen-positive on at least one test. Extrapolating cautiously, if we restrict our interpretation to those with concordant cognitive screening abnormalities — i.e. both subjective complaints and below-cutoff screening performance — the estimated proportion of the total cohort would be approximately 5% (43/836 × 67%), a figure considerably lower than crude prevalence estimates often cited in the literature and more aligned with conservative estimates of clinically meaningful cognitive impairment in virologically suppressed PLHIV [26].

Our findings indicate that low cognitive performance on screening tools remains prevalent among PLHIV in the cART era, although its etiology is likely multifactorial. Below-cutoff screening performance in this aging population may reflect not only potential HIV-associated brain injury but also age-related neurodegenerative processes, cardiovascular comorbidities, substance use, depression, and sociocultural factors influencing test performance [26, 29]. Importantly, attributing cognitive difficulties solely to HIV without comprehensive neuropsychological evaluation and exclusion of confounders is not warranted.

The significant association between ethnic origin and both MoCA and IHDS scores is noteworthy and likely reflects cultural, linguistic, and educational influences on cognitive test performance rather than true differences in cognitive ability [29, 54]. Almeida et al. [54] demonstrated that migrant status was an independent predictor of low MoCA scores in an HIV population, with the MoCA performing worse in migrant populations. This finding reinforces the importance of interpreting screening tool results within their clinical and sociocultural context, as advocated by the International HIV-Cognition Working Group [26].

NCI independently contributed to lower HRQoL, although its effect was smaller than that of symptoms of reporting negative feelings. There has been some debate in the literature about the extent to which mild NCI affects self-reported QoL. Some earlier studies, especially those that controlled for depression, did not find a significant independent effect of HAND on QoL [55]. For instance, one report noted that after accounting for depressive disorders, HAND status was not associated with SF-36 QoL scores in their cohort, implying mood was the primary driver of QoL [44]​. In contrast, other studies have reported that cognitive impairment does influence certain QoL domains—for example, researchers have documented worse mental health and role functioning scores in PLHIV with cognitive deficits [55]. Our results support an intermediate view: NCI has a measurable adverse impact on HRQoL, but a substantial portion of its apparent effect may be mediated through or confounded by symptoms of negative feelings and related factors.

The significant association found between NCI and reduced HRQoL underscores the importance of cognitive health assessments in HIV care, given the robust negative impact of neurocognitive issues on HRQoL is clear and clinically relevant. This suggests that cognitive difficulties—such as forgetfulness, mental slowing, or executive dysfunction—directly diminish one’s perceived HRQoL, perhaps by limiting daily activities, independence, or self-efficacy ​[56]. Therefore, while treating depression is crucial, it is also important to address cognitive health. Preventing neurocognitive decline through early intervention, regular monitoring, and risk factor management is equally critical, as it may help preserve HRQoL before significant impairments occur. Cognitive rehabilitation strategies, mental exercises, or occupational therapy could be beneficial. Additionally, ensuring that clinical services are cognitively accessible (e.g. use memory aids, simplified dosing regimens) can help those with NCI manage their health better and potentially improve their QoL.

One striking result was the uniformly high patient satisfaction with care in this cohort. Over 85% of participants, including those with significant reporting of negative feelings or NCI, reported being very satisfied with their HIV care. High satisfaction is encouraging because it likely facilitates trust and retention in care. As noted in prior research, patient satisfaction can directly influence adherence to treatment and clinic attendance [33]. Our findings imply that even individuals facing challenges (cognitive or emotional) feel that their healthcare needs are being met—a testament to patient-centered care. Interestingly, the lack of difference in satisfaction by NCI status may indicate that our clinicians effectively communicate and tailor care for cognitively impaired individuals (for example, using clear explanations or involving caregivers when needed). Another interpretation could be that individuals with NCI may have somewhat reduced expectations or awareness regarding care; however, given their similar ratings on specific items, it appears they genuinely felt well cared for. This result highlights that maintaining a high standard of patient experience is feasible and beneficial, even as the HIV population ages or accumulates comorbidities. Investing in staff training, reducing wait times, and nursing compassionate care may yield high satisfaction levels, which in turn can reinforce the care continuum.

This study has some limitations to consider. First, its cross-sectional design limits our ability to infer causality. We identified associations between factors and HRQoL, but we cannot be certain of the directionality. Longitudinal studies would help clarify temporal relationships, such as whether baseline NCI predicts future declines in QoL or vice versa. Second, missing data required the use of multiple imputation. We assume data were missing at random given observed variables, but if there were unmeasured factors associated with missingness, biases could persist. We attempted to include a wide range of variables in the imputation models to mitigate this. Third, we relied on self-reported questionnaires for QoL, one should be aware of socially desirability in answering. Especially for the EACS screening questions, this may have introduced the risk of under-diagnosing persons who do not complain and over-diagnosing those who exaggerate their difficulties. Furthermore, the EACS algorithm suspects neurocognitive problems starting from one affirmative answer on the EACS screening questions, whereas, to manage the workload, this project only focused on the respondents with 3 affirmative answers. Future research should would be valuable to ascertain whether further neurocognitive testing is necessary with one or two affirmative responses to the EACS screening questions. Fourth, anxious feelings were not questioned in this study, but we must be aware that anxiety and depression are often interlinked, i.e. individuals reporting higher levels of anxiety were significantly more likely to also suffer from depression [56, 57]. It is therefore highly recommended to explore anxious feelings in future PROMs. Next, selection bias is likely present due to the voluntary nature of questionnaire participation. The most vulnerable subgroups may have been underrepresented, particularly individuals facing language barriers, illiteracy, or other factors limiting their ability to complete the survey. This underrepresentation could have influenced the reported satisfaction rates regarding mental and physical well-being, potentially leading to an overestimation of these outcomes. We further acknowledge that the 57% response rate, while comparable to other survey-based studies in routine HIV care settings [58, 59], introduces a risk of nonresponse bias, particularly among individuals with cognitive difficulties, language barriers, or lower engagement in care who may have been underrepresented. Similarly, the small sample size for cognitive screening (n = 43 of 118 eligible, 36%) limits the statistical power and generalizability of the MoCA/IHDS analyses. These findings should therefore be interpreted as hypothesis-generating rather than definitive prevalence estimates. Finally, in terms of generalizability, our cohort was drawn from a single urban HIV clinic with a well-established multidisciplinary care program. Consequently, the results may not be directly generalizable to other settings or to populations with lower engagement in care. HRQoL and satisfaction might be lower in settings where PLHIV face more barriers to care or where health systems are overstretched. Likewise, the prevalence of NCI might differ in other contexts—for example, our cohort had relatively longstanding HIV (many infected in pre-cART era), whereas younger cohorts might have different cognitive outcomes.

Recommendations

Our study underlines a few actionable points for HIV care providers and programs aiming to improve long-term outcomes for PLHIV.

• Routine screening for depressive symptoms, integrating mental health services and prompt mental health interventions should be standard in HIV care.

• Cognitive screening (with tools like IHDS or MoCA) can be beneficial, especially for older PLHIV or those reporting cognitive complaints. In addition, preventive measures—such as managing cardiovascular risk factors, promoting mental health, and encouraging healthy lifestyle choices—are also key to preserving cognitive function over time.

• Referral to vocational programs, peer support groups, or social services for those unemployed or isolated could enhance participants’ sense of purpose and support, thereby improving mental health and HRQoL.

• Systematically collecting PROMs is vital as it may facilitate early identification of potential difficulties.

• Clinics should continue to solicit patient feedback. Promoting a tailored approach that considers individual patient needs, alongside the principles of value-based healthcare, can further enhance care quality and outcomes.

Conclusion

This study addressed three objectives in a cohort of Belgian PLHIV: (1) evaluating HRQoL and care satisfaction, (2) identifying predictors of HRQoL, and (3) estimating the prevalence and correlates of low cognitive performance. We found that mental health status — particularly reporting negative feelings—is the strongest predictor of HRQoL, alongside sleep satisfaction, social support, and sexual health satisfaction. Low cognitive performance on screening tools was prevalent but likely multifactorial in origin, and was independently associated with reduced HRQoL even after accounting for mood. Patient satisfaction with HIV care was high, including among those with cognitive or emotional challenges, suggesting current care practices effectively meet patients’ expectations [60].

Author contributions

CVB Conceptualization, methodology, data collection, formal analysis, writing—original draft, writing—review & editing, and project administration LD Conceptualization, data collection, writing—review & editing SD Conceptualization, methodology, review & editing EVC Conceptualization, data collection, writing—review & editing, and project administration SC Conceptualization, writing—review & editing MADS Conceptualization, methodology, writing—review & editing, supervision. funding acquisition.

Funding

The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Data availability

The data presented in this article is all provided within the manuscript.

Declarations

This article includes sections revised with the assistance of an AI language model to improve clarity and coherence. All content was reviewed and approved by the authors.

Ethics approval and consent to participate

The research protocol was approved by the Ethical Committee of Ghent University Hospital (BC-11836-AM01) in accordance with the principles of the Declaration of Helsinki. Written informed consent for participation was obtained from all individual participants prior to participation and is part of the general ICF used at the department (see supplementary data). Clinical trial number: not applicable.

Consent to publish

Not applicable.

Competing interests

The authors declare no competing interests.