Abstract
This study identified prevalence and correlates of HIV-associated dementia (HAD) among people living with HIV (PLWHA) in Ho Chi Minh City, Vietnam. Four hundred PLWHA completed a self-report questionnaire and were interviewed by a trained researcher to assess HAD using the International HIV Dementia Scale (IHDS). Clinical information concerning HIV treatment was also extracted from medical records. The results indicate the prevalence of probable HAD based on IHDS score <10.5 was 39.8% (95% confidence interval [CI]: 35.0%-44.5%). Probable HAD was significantly higher among female, older PLWHA and among those with low education level (≤ primary school), moderate level of adherence to HIV medication and HIV stage 3. Those PLWHA with depressive symptoms also had higher odds of having probable HAD (odds ratio = 3.23, 95% CI: 2.05-5.11). These findings underscore the importance of early HAD screening and appropriate referral for further assessment and management of PLWHA especially those with higher risk of HAD.
Introduction
One of the many causes of dementia is human immunodeficiency virus (HIV). HIV-associated dementia (HAD) is more prevalent in people living with HIV (PLWHA) who have severe immunosuppression, high viral loads in the cerebrospinal fluid, and advanced HIV as indicated by anemia and hypoalbuminemia. 1,2 While treatment advances have increased the life expectancy of PLWHA, this increased longevity has contributed to increased rates of HAD in both developed and developing countries. 3 -5 HIV-associated dementia represents a personal, social, economic, and medical burden and is an important risk factor for death among PLWHA worldwide. 3,6 -8 People living with HIV having HAD have lower capacity for self-care, including poorer regimen adherence, compared to those without HAD. 9 The median survival time for PLWHA after HAD diagnosis is about 6 months for those with no HAD progression and about 3 months for those with rapid HAD progression. 10 Early HAD identification and treatment with medication that has high central nervous system penetration effectiveness (CNSPE) is associated with better neurocognitive outcomes and higher survival rates for PLWHA. 2,11
Research evidence worldwide has consistently reported a high prevalence of HAD among PLWHA. 12 A review of 26 studies of PLWHA across 8 countries revealed HAD prevalence estimates that ranged from 2.2% up to 52% based on a psychiatrist’s diagnosis, where different reference criteria were employed. 13 Screening tests provide an alternative, cost-effective means of identifying symptoms of HAD so that those identified as at risk can be referred for further assessment. Such scales have been widely used in resource-limited countries where widespread access to HAD diagnosis established through psychiatrist interview is not possible. The International HIV Dementia Scale (IHDS) is among the most commonly and internationally used screening scales for HAD due to its acceptable to good reliability, validity, its simplicity, and limited requirement of language proficiency and training. 14,15 Although there is a possibility of HAD overestimation using the IHDS, this is the only available scale that allows cross-country comparisons and may therefore be recommended in settings where data on HAD screening and diagnosis are not available, and international comparisons are therefore very important. 16,17 Using the IHDS, estimates of probable HAD from African countries have ranged from 21.1% to 85%. 17 -22 Atashili et al 18 suggested that high prevalence estimates may be due to the inclusion of participants who are receiving antiretroviral therapy (ART) with CD4 count less than 200 or who have an AIDS-defining illness. For example, all participants of Atashili et al’s study with HAD prevalence estimate of 85% were receiving ART, while the prevalence of 21.1% reported by Njamnshi et al was from a mixed sample, some of whom were not receiving ART. 17 The range of HAD prevalence estimates in studies conducted in the United States, Canada, and India is lower, from 30% to 40%. 12,14,23,24 Unfortunately, due to a lack of research information, the prevalence of probable HAD in Vietnamese PLWHA using the IHDS is not available.
Despite the high prevalence and negative impact of HAD, and the knowledge that appropriate HIV/AIDS management and treatment that can slow cognitive decline and extend life is available for PLWHA with HAD, this disorder may be underdiagnosed and undertreated worldwide, especially in developing countries. 15 For example, in Vietnam there is low awareness about HAD among PLWHA and medical staff, limited outpatient psychiatry services for PLWHA, 25 and mental health issues among PLWHA are given relatively low priority in health policy and planning. 26 Moreover, definitive HAD diagnosis requires expensive and time-consuming medical procedures, such as neurological examination and history, brain scanning, and lumbar puncture. 27 Resource-limited, high HIV burden countries like Vietnam are not able to routinely employ this approach for large-scale screening and diagnosis of HAD at HIV outpatient clinics (OPCs). The lack of early screening to detect HAD in resource-limited countries such as Vietnam is particularly concerning since effective HAD treatment and management requires early identification and referral so that harm reduction interventions and support from family, friends, and health professionals can be put in place to assist PLWHA maintain a good quality of physical health and well-being. 9,20 Therefore, the use of alternative less costly methods to screen for HAD in PLWHA in countries like Vietnam should be considered. This includes training nonpsychiatric OPC health staff to screen for probable HAD using a screening tool with acceptable reliability and validity, such as the IHDS. 16,17 Such screening for probable HAD among PLWHA has not been routinely conducted at HIV OPCs in Vietnam.
If large-scale screening is not feasible, it may also be useful to target certain populations for HAD screening based on factors that have been found to be associated with the development of HAD. Worldwide research has demonstrated that HAD is associated with being female, older age, and having lower education level, 18,28 while employment status, religious affiliation, smoking, drinking and current substance use have not been found to be good predictors of HAD. 29,30 It has also been found that once a person develops HAD certain factors reliably predict the course of HAD. For example, HAD has been reported as being more prevalent and rapidly progressing among PLWHA who acquired HIV through intravenous drug use compared to transmission through sexual contact. 7,10 Other factors relevant to HAD in PLWHA include the presence of depressive symptoms or anxiety symptoms since these may be the initial symptoms of HAD. 7,18 There are also a number of HIV/AIDS-related clinical indicators that have been identified in previous research as risk factors for the development and/or rapid progression of HAD. These include low CD4 count, delay in ART initiation, and low CNSPE regimen. 7,11,20 However, duration of HIV infection has not been found to differ between PLWHA with and without HAD. 8,31
Although research investigating the prevalence of HAD and the predictors of HAD has been conducted in numerous countries around the world, to date no research has investigated the prevalence or predictors of HAD in Vietnam and findings from previous studies in other countries may not be relevant to Vietnam due to low awareness about HAD, limited psychiatry services for PLWHA, 25 low priority in health policy, 26 and complexity of HAD screening and diagnosis. 27 The availability of such information can assist in targeting most at risk of HAD among PLWHA, decision-making and planning, and the introduction of strategies to maximize the well-being of PLWHA with HAD, in Vietnam. Thus, the aim of this study was to estimate the prevalence of probable HAD using the IHDS in a large sample of outpatients attending HIV clinics in Ho Chi Minh City (HCMC) and to examine the correlates of probable HAD so as to identify potential risk factors. Since research has not been conducted in Vietnam, the relationship between a range of factors will be investigated, including those that have not previously been reported as significant correlates.
Methods
Setting and Participants
Participants were recruited from 2 HIV OPCs located in HCMC. The HCMC is the largest city in Vietnam and has a high prevalence of HIV/AIDS compared to the Vietnamese average. 32 In 2012, there were about 25 000 PLWHA under care and treatment at 30 HIV OPCs, mostly located at district preventative medicine centers in HCMC. 33
The current study presents findings obtained as part of a larger project investigating mental health and risky behaviors among PLWHA, in Vietnam, which was conducted by the current authors. 34 That project was conducted in 2013 and PLWHA were recruited over a 4-week period at 2 randomly selected OPCs. The sample size was originally calculated to estimate the prevalence of PLWHA who have symptoms of a range of mental disorders including depression, anxiety, alcohol use disorder, drug use disorder as well as HAD, which is the focus of the current article. With estimated prevalence of 50%, type I error rate of α = 5% and estimation error of 5%, at least 385 participants were required. Since the reported prevalence of probable HAD ranged from 21.1% to 85%, 17 -22 the sample size was sufficiently large to estimate the prevalence of HAD. There were approximately 1000 patients at each OPC, and 1 in every 5 patients was consecutively invited to participate when they attended the OPC for their regular monthly visit. The recruitment criteria were ≥18 years old, currently under care at the study OPCs, and able to read and write. Of the 410 patients approached, 400 PLWHA consented to participate in the study (200 from each OPC), a participation rate of about 97.6%. Patients who agreed to participate were provided with a participant information statement and signed a consent form. Those who did not choose to participate in the study continued with their usual visit.
Measurements
Self-report questionnaire
Background information
The questionnaire included questions about demographic characteristics of age, sex, sexual orientation, work status, highest level of education, marital status, parental status, religious affiliation, economic status, source of HIV infection, self-rated general health status, diagnosed mental disorder, time since mental disorder diagnosis, and type of mental disorder. Nine questions concerned practical, emotional, and spiritual support received from family and from social and healthcare professionals during the last 3 months, and 12 questions concerned stressful life experiences over the past 30 days. Questions about stressful life experiences were adapted from the HIV Stressor Scale by Thompson et al. 35 Six further questions about number of children, income range, current residence, household composition, living with drug or alcohol abuser, living with a mentally ill person were not included in the current analysis.
Center for Epidemiologic Studies–Depression Scale
Center for Epidemiologic Studies–Depression Scale (CES-D) 36 is a 20-item questionnaire used to assess depressive symptoms over the last 7 days. It has excellent psychometric properties in both HIV and non-HIV populations 36 -38 and has been translated into many languages for use in high HIV burden settings. 37,38 A total CES-D score of ≥16 has been shown to discriminate those with depressive symptoms and those without depressive symptoms. 36
Phan Vietnamese Psychiatric Scale–Anxiety Scale
Phan Vietnamese Psychiatric Scale–Anxiety Scale (PVPS-AS) 39 is a 13-item questionnaire used to assess symptoms of generalized anxiety disorder. It was developed originally for Vietnamese people and takes into account Vietnamese culture and religion. It has excellent psychometric properties with internal consistency Cronbach’s α at .91 and test–retest reliability coefficient at 0.81. 39 The scale also has good criterion validity with high level of agreement with psychiatrist’s interview. A mean score of PVPS ≥1.60 has been shown to discriminate those with and those without symptoms of anxiety. 39
World Health Organization Alcohol Use Disorder Identification Test
World Health Organization Alcohol Use Disorder Identification Test (WHO-AUDIT) 40 is a 10-item questionnaire used to identify alcohol use disorder through assessing alcohol consumption, alcohol dependence, and alcohol-related health problems. It has excellent psychometric properties with high internal consistency, test–retest reliability, and concurrent validity. 41 This scale has been used worldwide both in the general population and in HIV-positive patient samples. 42,43 A total WHO-AUDIT ≥8 has been shown to discriminate those with and those without problematic drinking conditions. 40
Drug Abuse Screen Test
Drug Abuse Screen Test (DAST) 44 is a 10-item questionnaire used to assess drug abuse problems. This scale demonstrates very good psychometric properties with high levels of internal consistency, test–retest reliability, concurrent validity, criterion validity, and discriminant validity. 45,46 The DAST has also been used in many studies among PLWHA worldwide and a total score of DAST ≥3 has been shown to discriminate those with and those without drug abuse problems. 47 -49
The questionnaire was originally constructed in English and was translated into Vietnamese independently by an accredited translator and the first author. All differences between the 2 versions were compared and discussed in order to develop the final version.
Interview-based questionnaire
International HIV Dementia Scale
International HIV Dementia Scale14,22 is an interviewer-administered performance measure that assesses 3 domains: motor speed (timed finger tapping), psychomotor speed (timed alternating sequence test), and memory recall (recall of 4 items). The IHDS does not require respondents to be proficient in English and takes about 2 to 3 minutes to complete. The first domain is measured by the number of accurate taps of the first 2 fingers of the nondominant hand completed in 5 seconds, with scores ranging from 0 (0-2 taps/5 seconds) to 4 (
Clinical information extraction
After receipt of signed consent from participants, clinical information about HIV treatment was extracted from OPC records. These included number of family members with an HIV-positive diagnosis, body mass index, time since HIV diagnosis, duration of ART, CD4 count, current HIV stage, HIV regimen, previous diagnosis with an opportunistic disease, history of delaying ART, chronic conditions other than HIV, and ART adherence level (good/average/poor) as recorded by the clinic physician. Based on the drugs used for each regimen, a score for CNSPE was calculated as recommended by Letendre, where a score ≤7 indicated low drug penetration effectiveness. 11
Clinical interview
All participants were assessed individually by 1 of 4 experienced psychiatrists from HCMC University of Medicine and Pharmacy for symptoms of depression, anxiety, alcohol use disorder, substance use disorder, and HAD. The criteria for these conditions were based on the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition, Text Revision). 50 The interview took approximately 30 minutes. The psychiatrists referred participants who had symptoms of mental disorders including HAD to mental health services for further assessment and treatment. The interview was also used as a criterion to estimate sensitivity and specificity of the IHDS. Findings from this clinical interview have been reported elsewhere. 34
Procedure
Ethics approval for the study was granted by the human ethics committee at HCMC Provincial AIDS Committee, Vietnam, and by the University of Sydney, Australia. The participants were first directed to a private room at the clinic to complete the self-report questionnaire. A research assistant was in the room to answer any questions raised. Participants then placed the completed questionnaire in an envelope, sealed the envelope, and placed the sealed envelope in a secure box. After the self-report questionnaire was completed the research assistant who was trained in IHDS assessment administered the IHDS. Participants were then directed to another private clinic room for the psychiatric interview. Neither the psychiatrist nor the research assistant had access to the self-report questionnaire data. All participants received feedback directly from the psychiatrists about their current mental health status and those who had symptoms of mental disorders requiring further assessment and treatment were referred to a mental health service and advised that these services would be provided free of charge. Clinical information was extracted using the patient identification number provided by the participant on the consent form. All information for each participant was matched using a predefined study number so that no identifying information about the participant was stored.
Data Analysis
Continuous data for age, time since HIV diagnosis, most recent CD4 count, and time since ART initiation were summarized as means (M) and standard deviations (SD). Counts and percentages were used for categorical variables. For univariate analysis, the association between HAD and demographic characteristics, clinical factors, and psychosocial factors was assessed using the chi-square test for categorical variables and the t test for continuous variables. Odds ratio (OR) and 95% confidence intervals (CIs) for the cross-tabulations were calculated using logistic regression. Variables with significance levels < 0.20 in univariate analysis were used for multiple logistic regression based on procedures recommended by Hosmer et al. 51 Nonsignificant variables at α = .05 from the initial multivariable model were dropped. The likelihood ratio test was used to compare the final model to the model with each nonsignificant variable included. The change in estimate and the interaction were also investigated before fitting the final model. Model fit indices included Pearson chi-square test of goodness of fit, Hosmer-Lemeshow chi-square and Cragg and Uhler R2. The distribution of residuals was also examined. All analyses were conducted using SPSS version 21. 52
Results
Nearly two-thirds of participants (63.5%) were male and ages ranged from 20 to 67 years, with a mean of 34.8 years (SD = 6.8). The majority of participants reported acquiring HIV through sexual transmission (56.0%) followed by injecting drug use (29.6%). About 14.4% of participants did not specify how they had acquired HIV, said they could not remember, or said they had acquired HIV through traffic or work accidents, presumably through contact between exposed wounds and body fluids or blood transfusion with contaminated blood. The cutoff point of IHDS used in previous studies (<10.5) and the psychiatrist’s interview as reference result in sensitivity at 76.7% and specificity at 64.7%. The percentage of agreement between the IHDS and psychiatrist’s interview was 66.0%. Based on the IHDS, probable HAD was identified in 159 PLWHA (39.8%, 95% CI: 35.0%-44.5%). In bivariate analysis, HAD was associated with older age, lower levels of education and HIV infection reported through sources other than sexual transmission or injecting drug use. Sex, sexual orientation, work status, marital status, parental status, religion affiliation and economic status were not associated with HAD at a significance level of .05 (see Table 1).
Association between Sociodemographic and HAD in Vietnamese HIV Outpatients.a
Abbreviations: CI, confidence interval; HAD, HIV-associated dementia; OR, odds ratio; SD, standard deviation.
aN = 400.
bIncluding: not reported, not remembered, and stepping on a needle, traffic, or working accidents.
HIV-associated dementia was associated with the adherence level for HIV medication, where PLWHA with good adherence were less likely to have probable HAD than those with moderate adherence. However, self-reported general health status, body mass index, time since HIV diagnosis, time since ART initiation, most recent CD4 count, current HIV stage, penetration effectiveness score and level, physical comorbidity, history of being diagnosed with an opportunistic disease, and history of delaying ART were not associated with probable HAD (see Table 2).
Association between Clinical Factors and HAD in Vietnamese HIV Outpatients.a
Abbreviations: ART, antiretroviral therapy; CI, confidence interval; CNSPE, central nervous system penetration effectiveness; HAD, HIV-associated dementia; IQR, interquartile range; OR, odds ratio.
aN = 400.
bFor every 100 units change.
There was a statistically significant association between HAD and family support, anxiety symptoms, and depressive symptoms. People living with HIV who had probable HAD were less likely to report family support and more likely to report symptoms of anxiety and depression than those without HAD. However, HAD was not associated with family members being diagnosed HIV positive, HIV status disclosure, social support, professional support, number of stressful life experiences, and problematic alcohol use disorder or drug abuse (see Table 3).
Association between Psychosocial Factors and HAD in Vietnamese HIV Outpatients.a
Abbreviations: CES-D, Center for Epidemiologic Studies–Depression Scale; CI, confidence interval; DAST, Drug Abuse Screen Test; HAD, HIV-associated dementia; OR, odds ratio; PVPS-AS, Phan Vietnamese Psychiatric Scale–Anxiety Scale; WHO-AUDIT, World Health Organization Alcohol Use Disorder Identification Test.
aN = 400.
bYes = occasionally, frequently, and very frequently received/no = never and rarely received.
Twelve variables from Tables 1 to 3 with P < .20 were used in a multiple logistic regression with IHDS category as the dependent variable. After investigating confounders and interactors, the final model for HAD indicated a good fit with 7 predictors: sex, age, education level, source of HIV infection, HIV stage, adherence to HIV medication, and depressive symptoms (see Table 4). The Cragg and Uhler R2 was .230, while Hosmer-Lemeshow chi-square (df = 8) was 6.05, P = .641 and Pearson chi-square (df = 299) was 316.05, P = .238, indicating no difference between the observed data and the model specified. Residuals checking revealed no problems with the data and the model fitted. Female PLWHA were 1.77 times as likely as males to have probable HAD (95% CI: 1.06-2.96, P = .029). Every increase of one year of age resulted in an increase of about 5% in the odds of having probable HAD (P = .003). There was no difference in HAD between PLWHA with primary school or secondary school education (OR = 0.83, 95% CI: 0.49-1.40, P = .484), but those with high school education were less likely to have HAD than those with primary school only (OR = 0.55, 95% CI: 0.30-0.99, P = .046). People living with HIV who reported “other” as source of HIV infection were more likely to have HAD than both those who reported infection through sexual transmission (OR = 3.57, 95% CI: = 1.82-7.01) and those who reported infection from injected drug use (OR = 2.33, 95% CI: 1.13-4.81). Compared to PLWHA at the first stage of HIV, PLWHA at the third stage were 3.60 times more likely to have HAD (95% CI: 1.16-11.19). There was a difference between those with average and those with good adherence to HIV medication, where PLWHA with moderate adherence were 3.57 times more likely to have probable HAD. People living with HIV with depressive symptoms were 3.23 times more likely to have probable HAD than those without depressive symptoms (95% CI: 2.05-5.11).
Multivariate Model for Correlates of HAD in Vietnamese HIV Outpatients.a
Abbreviations: CI, confidence interval; HAD, HIV-associated dementia.
aN = 400.
Discussion
This study was the first to be conducted in HCMC, Vietnam, to investigate the prevalence and correlates of probable HAD among PLWHA at OPCs. The results show that nearly 40% of PLWHA had probable HAD based on the cutoff of <10.5 on the IHDS. This prevalence is much higher than the average estimate (ie, 12%) of symptoms of all neurocognitive disorders in the Asia Pacific region. 53 This difference may be explained by the difference in measurement, since this study employed the IHDS, while Wright et al used a combination of several psychological test batteries. 53 However, the prevalence reported here was similar to that reported in other countries using the IHDS and the same cutoff scores for classification of those with and without probable HAD. These include studies conducted in the United States (37.9%), 14 Canada (39.4%), 23 and India (35% in 2006 and 32.5% in 2014). 12,24 However, the prevalence of probable HAD in this study is lower than findings from studies conducted using the IHDS in Africa such as in Cameroon (85%), 18 Nigeria (54.3% in 2012 and 66.2% in 2013), 19,20 and South Africa (80.0% in 2008 and 81.3% in 2011). 21,22
The high prevalence found in African countries could potentially be due to more advanced HIV infection among study participants where only PLWHA with CD4 counts less than 200 or HIV stage 4 were on ART. 18,20 In the current study, the mean CD4 count was approximately 400 and most of the participants (92%) were in HIV stage 1. Alternatively, the high prevalence might also be due to the psychometric properties of the IHDS in different settings such as South Africa where sensitivity and specificity were 45% and 79% at the cutoff at ≤10. 21 The sensitivity of 45% indicates a high level of false positives. Given our findings of a high prevalence of probable HAD, we suggest there is a crucial need for screening PLWHA for HAD at all HIV/AIDS OPCs in Vietnam.
In the Vietnamese health system, there are free national mental health programs conducted by departments of mental health known as district preventative medicine centers that are located near, or at the same place, as HIV/AIDS OPCs. However, official, practical, and routine links between these 2 services have not been established. Therefore, screening activities will only be beneficial if there is an official link between these services so that PLWHA who have HAD can be referred for further diagnosis and treatment. Such screening and referral activities at HIV clinics have been shown to enhance PLWHA’s health during the course of HIV treatment. 54
In regard to correlates of HAD, HAD prevalence was higher among females and increased with age, lower education level, higher HIV stage, lower adherence to HIV treatment, and presence of depressive symptoms. As such our findings were consistent with those from previous studies. 14,16,18,28,55 The sex difference in HAD might be explained by the greater psychiatric morbidity among females including depression, the difference in chemistry and brain structure as well as hormone levels. 55 In addition, PLWHA with a lower level of education level might be less likely to seek early medical care which can result in the progression of HIV and thus a higher probability of HAD. 28 Higher HIV stage and lower treatment adherence may be associated with HAD directly, 2 while depression may be a symptom of HAD. 7 No relationship between HAD and alcohol or drug use was found in the current study. This is inconsistent with the findings of Nath et al who reported that alcohol or drug use was associated with HAD. 7 However, our findings are consistent with those reported by Dang et al in China where HAD identified using the IHDS was also not found to be associated with alcohol use. 30 Rosenbloom et al revealed that alcohol only affects the brain after a certain HIV stage such as AIDS. 56 In the current study, most PLWHA (92%) were in HIV stage 1 and thus the effect of alcohol use on HAD was not found.
Our findings show that while across the entire sample PLWHA were more likely to report sexual transmission or drug use as sources of HIV. However, PLWHA with probable HAD were less likely to report these sources of HIV infection and were more likely to indicate “other” undefined sources of infection compared to those without probable HAD. It is likely that those with probable HAD coded “other” (n = 57, 14.3%) because of poor memory due to HAD. The accuracy of self-report data for those with probable HAD is questionable, and this is a limitation of the present study.
No associations were found between HAD and HIV-related clinical factors such as CD4 count or drug penetration effectiveness. Previously, low CD4 count has been proposed as a significant risk factor for neurocognitive disorder, although the lowest CD4 during the course of the HIV illness was a better indicator than current CD4 level. 7,11 In our study, we included current CD4 count and this may explain the lack of association. Similarly, in contrast to Letendre, 11 we found no significant difference between the CNSPE score of PLWHA with and without probable HAD. It is possible that PLWHA in our sample have not used the high penetration drugs long enough to have a discernible effect in reducing the neurocognitive impairment including HAD. Further, as neurocognitive impairment might be irreversible, the use of a high-penetration effective drug at ART initiation might be more important than the usual use of CNSPE. 2,11 Our findings highlight the importance of HAD screening in the early stage of HIV/AIDS treatment.
Several limitations of the study need to be recognized. The study used the IHDS to identify probable HAD. While this had the advantage of allowing comparison with previous studies worldwide that had used this measure, the prevalence might be overestimated. In the present sample, the IHDS was more likely to identify cases of probable HAD than via the psychiatrist’s assessment. However, the IHDS is a much more cost-effective and accessible screening method compared to assessment by a psychiatrist, making it suitable for use in a resource-limited setting such as Vietnam. Additionally, overinclusion of potential cases ensures that the likelihood that someone with probable HAD is not identified using the IHDS is low. A high rate of false positives is preferable to failing to identify those who need further assessment and support. Clearly, however, the screening value of the IHDS among Vietnamese HIV patients needs to be validated. Further, the study was conducted at two randomly selected OPCs in a large city, and participants were generally in a stable HIV/AIDS condition. The findings might not be generalizable to others HIV populations such as those in hospital settings or those in rural areas. Further studies are needed to examine the psychometric properties of the IHDS in Vietnamese PLWHA and to investigate HAD among other Vietnamese samples of PLWHA, from a range of different settings including hospital and community or different backgrounds and HIV severity.
Conclusion
This is the first study conducted in Vietnam examining the prevalence of HAD and the relationship between HAD and a range of factors including demographics, clinical, and psychosocial. The prevalence of probable HAD among Vietnamese PLWHA was 40% and was associated with being female, being older, having a low education level (≤ primary school), reporting the source of HIV infection as other, having an HIV stage 3, having a moderate level of adherence to HIV medication, and experiencing depressive symptoms. Screening tests such as the IHDS are essential for directing limited resources for the diagnosis of HAD. It is likely that early screening for HAD, especially in those who are most at risk of HAD would lead to appropriate referral for further assessment and management among PLWHA in Vietnam, particularly given the established link between HIV/AIDS OPCs and the national mental health program. However, more research to validate the IHDS among Vietnamese PLWHA is needed.
Footnotes
Authors’ Note
Informed consent was obtained from all individual participants included in the study.
Acknowledgments
The authors would like to thank the people living with HIV who participated in this study as well as the individuals and institutions that made this research possible: Dr Tran Thinh, Dr Nguyen Hoang Tam, Dr Van Hung from Ho Chi Minh City Provincial AIDS Committee; Dr Bui Thi Thu Phuong and Dr Pham Van Hieu from HIV outpatient clinics; Associate Professors Do Van Dung, Mr Tran Nhat Quang, Mr Hua Thanh Liem, Ms Van Thi Thuy Duong, and Ms Kim Xuan Loan from Ho Chi Minh City University of Medicine and Pharmacy; Ms Bui Thi Hy Han from Ho Chi Minh City Institute of Public Health and Dr John Nguyen; and Dr Jeffrey Mendel from University of California at San Francisco.
Ethics Approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the human ethics committee at Ho Chi Minh City Provincial AIDS Committee, Vietnam (Approval number: IRB-03-2013, dated 17/10/2013) and the University of Sydney, Australia (Approval number: 2013/859, dated 15/11/2013) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
This work was supported by The Representative Office of Abbott Laboratories SA in Vietnam in collaboration with Ho Chi Minh city University of Medicine and Pharmacy in Hand Awareness program. No Abbott products were used or recommended to be used during the study and no trademark of Abbott appeared in the study. The travel for field site work during this study was covered by the Australia Award Scholarship. Protocol development and manuscript writing was supported by CDC-PEPFAR Vietnam, the University of California, San Francisco’s International Traineeships in AIDS Prevention Studies (ITAPS), US NIMH, R25MH064712 and the Starr Foundation.