Health-related quality of life of HIV-infected intravenous drug users

Abstract

To investigate health-related quality of life in HIV-infected intravenous drug users registered but not engaged in HIV outpatient care (missing ≥2 outpatient appointments over 1 year or non-attendance for ≥6 months) we conducted a cross-sectional study to examine health-related quality of life of HIV-infected intravenous drug users registered for care at an inner city HIV unit. EQ-5D, SF-36, SF-6D, mood disorder, clinical and substance misuse data were collected. Mean scores and preference derived utility scores were calculated. Statistical relationships between health-related quality of life and other variables were explored using univariate and multivariate analysis. Fifty-five patients were recruited, 64% were males. The mean anxiety value was 11.44 (anxious) and mean depression score was 9.3 (borderline depressed). The mean EQ-5D utility was 0.45 (95% CI 0.35, 0.55) and mean SF-6D utility was 0.52 (95% CI 0.48, 0.55). There was no statistical relationship between HIV indices, substance misuse and EQ-5D and SF-6D utility. Anxiety and depression were significantly correlated with EQ-5D and SF-6D utility values on univariate and multivariate analysis. Health-related quality of life was reduced in this HIV-infected intravenous drug user population. Whilst hepatitis C co-infection and substance misuse did not affect health-related quality of life, anxiety and depression had a significant impact on it.

Keywords

AIDS viral disease HIV health-related quality of life injecting drug use antiretroviral therapy anxiety depression

Background

The advent of effective combination therapy for HIV has transformed this once terminal condition into a chronic disease.¹ Despite a reduction in mortality, morbidity and hospitalisation rates, patients infected with HIV still have higher rates of depression and a lower health-related quality of life (HRQOL).^2-4 Part of this is multifactorial and due to co-infections such as hepatitis C (HCV), substance misuse and psychiatric co-morbidities.^2,4-6 Numerous reports have shown that patients infected with HIV have reduced HRQOL when compared with the general population and with patients with other chronic conditions such as epilepsy and diabetes.^2,3,7

Many studies assessing the HRQOL of HIV-infected individuals have been conducted.^2,3,7 Antiretroviral therapy (ART) has been shown to improve HRQOL.⁸ However, the literature has been contradictory on the relationship between CD4 count, HIV viral load (HIV VL) and HRQOL.^3,5,7,9–11 Other factors that have been found to negatively impact the HRQOL of HIV-infected individuals include lower socioeconomic status, older age, drug use, psychiatric co-morbidities, substance misuse and hospitalisation.^4,5,11,12

Whilst HRQOL has been extensively investigated in many HIV cohorts, there are currently few data available on HIV-infected intravenous drug users (IDUs), especially chaotic individuals not engaged with outpatient HIV services.^2,3,7 HIV-infected drug users are a marginalised patient population with multiple medical and psychiatric co-morbidities and high rates of illiteracy,^13,14 who are often excluded from studies. This research is focused on investigating HRQOL in this heterogeneous cohort.

Aim

The aim of this research is to examine HRQOL in non-engaging (missing two or more HIV outpatient appointments over the preceding year or non-attendance for 6 months or longer) HIV-infected IDUs, using different HRQOL measures.

Methods

Study population

This was a cross-sectional study to examine HRQOL of non-engaging HIV-infected IDUs who were registered for care at an inner city HIV specialist clinic. This is a sub-study of an ongoing study to improve clinical outcomes in HIV-infected IDUs, that involves the establishment of a dedicated HIV clinic on site with methadone services. For this sub-study, baseline HRQOL and clinical data were analysed. Eligible individuals were at least 18 years old, HIV infected, on methadone replacement and not engaged with outpatient HIV services (missed two or more HIV outpatient appointments over the preceding year or non-attendance for 6 months or longer).

Patients were recruited from HIV respite services, HIV inpatient services or the newly established HIV clinic. All patients who attended the new HIV service were provided written and verbal information on the new HIV service by staff at the methadone service. They were then approached by the HIV team to attend the new HIV clinic.

HRQOL, clinical data, anxiety, depression and illicit drug use screening were performed on all patients. Ethical approval was granted for this study from the ethics committees covering both centres

Data collection

Questionnaires and clinical data were collected at recruitment to the study. Clinical data such as CD4 and HIV VL were collected from the electronic results management system. Other clinical data such as co-morbidities were collected verbally from patients and verified by clinical note review, and or taken from clinical records. Verification of non-attendance was performed through clinical note analysis and review of non-attended appointments on the electronic appointment management system.

High rates of illiteracy have been previously recorded in this cohort, so the investigators were expected to assist most participants with physically completing the questionnaires.¹⁴ For illiterate participants, interviewers were permitted to read out questions and record answers without prompting. The data were entered onto a Microsoft Excel database. The SF-36 data were entered as raw data and composite scores for each health domain scale.

Clinical and demographic data collected

Demographic data collected included age, ethnicity, gender, mode of acquisition of HIV and current employment status.

HIV specific data collected included CD4 counts (cells/mm³), HIV VL (copies/mL), hepatitis C antibody (HCV), HCV PCR/RNA status, antiretroviral status (ART), and medical and psychiatric co-morbidities. Cirrhosis was predominantly diagnosed by radiological or endoscopic findings rather than liver biopsy as this non-engaging cohort have a high rate of non-attendance for liver biopsy.

Survey materials and instruments used

The HRQOL instruments used are the paper versions of the EQ-5D (3-level) and SF-36 (version 2).^15,16 The hospital anxiety depression scale (HADs) was used to determine levels of mood disorders.^17,18 To quantify current substance misuse, the Treatment Outcomes Profile (TOP) was also performed.¹⁹

The EQ-5D is a two-part measure that comprises a single utility score based on societal preferences, and a visual analogue scale EQ5D_VA.³ The first part consists of five domains that record health status in terms of mobility, self-care, usual activities, pain/discomfort and anxiety/depression. There are three levels of severity for each domain, giving 243 possible health states.³ These are converted into utilities, which are anchored between 0 (death) and 1 (best possible health state) using preferences derived from the general UK population. Individuals can obtain negative utilities for a given health state, which are interpreted as ‘states worse than death.’²⁰

The SF-36 was designed as a profile measure with eight multi-item scales.¹⁵ It also has two composite scales; the physical (PCS) and mental component summary (MCS) scores. These are further divided into eight health domain scales. The physical component score comprises physical functioning (PF), role limitation/physical (RP), bodily pain (BP) and general health (GH). The mental component score consists of vitality (VT), social functioning (SF), role limitations/emotional (RE) and mental health (21).

In order to compare directly with the EQ-5D, utilities were calculated by transforming the SF-36 into the SF-6D using a method described by Brazier et al.¹⁵ The population preferences used in this article are derived from a UK population,²¹ and the analysis was carried out in SPSS version 16 (the SPSS syntax is available from the EuroQol group).¹⁶ There are currently no population values available for Ireland.

The HADs was designed to screen and assess for anxiety and depression in medically ill patients.^17,18 It contains fourteen, four-point items. Seven elements assess anxiety (HADS-A), and the other seven evaluate depression (HADS-D). It is easy to administer and has good internal consistency, test-retest reliability and convergent validity.²² Scores from 0 to 7 are normal, 8–10 indicates a borderline abnormality and ≥11 are abnormal.

The TOP was developed by the English National Treatment Agency (NTA) to monitor substance misuse treatment. It comprises a 38-measure tool, divided into four sections: substance use, health risk behaviour, offending and health and SF. It includes three VAS scores on physical health, mental health and quality of life. The TOP also assesses substance misuse (including alcohol, opiates, crack, cocaine, amphetamines and cannabis), injecting drug use, forensic and employment status over the preceding month. It has been established as a reliable instrument for treatment outcomes monitoring.¹⁹

Statistical analysis

Descriptive statistics were used to describe baseline demographics; mean values, ranges and standard deviations are presented. Utility scores were derived from both the EQ-5D and SF-6D. Univariate analysis was performed using Spearman’s correlation and Wilcoxon rank. A multiple linear regression was fit for both the EQ-5D and SF-36 to include variables that were statistically significant in the univariate analysis.

Results

Patient demographics and clinical parameters

Fifty-five patients were included in this study. Sixty-four percent of patients were males (Table 1). The overall mean age of the cohort was 37 years. A total of 56% of the patients were HCV PCR positive and 23% were cirrhotic. Ninety-five percent fulfilled international criteria for ART and sixty-two percent of these were on ART. Forty-seven percent of the patients in receipt of ART were virologically suppressed (HIV VL <40). Another 33% of patients met international criteria for ART but were not taking it (CD4 count ≤350 cells/mm³ or HIV/HCV co-infection).^23,24 Alcohol and cannabis were the most frequently misused substances, and 42% of patients drank over the recommended weekly alcohol limits (14 units a week for women and 21 for men).²⁵

Table 1.

Clinical and substance misuse demographics (n = 55).

Number (%)
Characteristic
Male sex	35 (64%)
Age at time of questionnaires, in years (mean ± SD [range])	37 ± 6.4 (27–51)
Time since HIV diagnosis, years (mean ± SD [range])	10 ± 6.4 (1–26)
Ethnicity
Caucasian	55 (100%)
Mode of acquisition of HIV
IDU	48 (87%)
Heterosexual	6 (11%)
MSM (men who have sex with men)	1 (2%)
Current employment status
Employed	1 (2%)
Unemployed	51 (93%)
Attending college/course	3 (5%)
CD4 count (cells/mm³)
At questionnaire (mean ± SD [range])	257 ± 205 (3–1015)
≤200 cells/mm³	24 (44%)
≤350 cells/mm³	41 (75%)
HIV viral load (VL)/RNA at time of questionnaire, (copies/mL) (mean ± SD [range])	99,536 ± 342,838 (<40–1,696,195)
HCV PCR positive	31 (56%)
On ART at time of questionnaire	32 (58%)
Virologically suppressed	15/32 (47%)
Cirrhotic	13 (23%)
Injecting drug use over preceding month	7 (13%)
Monthly Substance misuse (mean ± SD [range])
Alcohol consumption (units)	202.65 ± 327.13 (0–1120)
Heroin consumption (gram)	0.8 ± 3.07 (0–17)
Crack cocaine consumption (gram)	2.88 ± 16.52 (0–120)
Cocaine consumption (gram)	0.11 ± 0.76 (0–6)
Cannabis consumption (joints)	645.24 ± 142.67 (0–840)

ART: Antiretroviral therapy; HCV: hepatitis C; IDU: intravenous drug user.

HRQOL characteristics

Anxiety/depression scores

The mean anxiety score was 11.44 ± 4.39 (Table 2). A value of ≥11 indicates anxiety. Sixty percent of patients had high levels of anxiety. The mean HADs-D score was 9.49 ± 4.85, which is borderline depressed. Forty-four percent of patients were depressed (value ≥11). No patients were prescribed an efavirenz-based regimen.

Table 2.

Health-related quality of life (HRQOL) characteristics.

Characteristic	Number (%)
HADs – Anxiety
Mean ± SD (range)	11.44 ± 4.39 (0–19)
Borderline anxiety (HADs 8–10)	7 (13%)
Anxiety (HADs ≥ 11)	33 (60%)
HADs – Depression
Mean ± SD (range)	9.3 ± 4.78 (1–21)
Borderline depressed	7 (13%)
Depressed	24 (44%)
Generic measures
EQ5D utility score (mean ± SD [range])	0.45 ± 0.36 (–0.38 to 1)
EQ5D utility < 0 ‘worse than death state’	6 (11%)
EQ5D < 0 and anxiety (HADS-A ≥ 11)	5 (83%)
EQ5D <0 and depressed (HADs-D ≥ 11)	4 (7%)
EQ5D_VAS scale (mean ± SD [range])	48.11 ± 25.8 (0–100)
SF6D utility value (mean ± SD [range])	0.52 ± 0.13 (0.24–0.87)
SF36 Physical health (mean ± SD [95% CI])
Physical component summary (PCS)	37.72 ± 10.19 (34.97, 40.47)
Physical functioning (PF)	61.36 ± 30.9 (53.00, 69.73)
Role limitations/physical (RP)	25.00 ± 32.62 (16.18, 33.82)
Bodily pain (BP)	51.85 ± 31.81 (43.25, 60.45)
General health (GH)	30.51 ± 22.04 (24.55, 36.47)
SF36 Mental health (mean ± SD [CI])
Mental component summary (MCS)	32.53 ± 12.31 (29.20, 35.85)
Vitality (VT)	28.36 ± 21.56 (22.53, 34.19)
Social functioning (SF)	41.36 ± 27.94 (33.81, 48.92)
Role limitations/emotional (RE)	27.88 ± 38.90 (17.36, 38.39)
Mental health (MH)	47.13 ± 24.30 (40.56, 53.70)

HADs: Hospital anxiety depression scale.

EQ5D and SF-6D values

The mean EQ5D utility score was low at 0.45, with a range of −0.36 to +1 (Table 2). Six patients (11%) had an EQ-5D utility score that was worse than death. Five (83%) of these were scored as anxious by the HADs-A. Four (67%) were also depressed. The mean EQ-5D_VAS was 48.11. The mean SF-6D utility score was 0.52, with a range of 0.24 to 0.87.

SF-36 health domain scores

The mean physical component score of this cohort was 37.72 and mental component score was 32.53 (Table 2). Role limitations, both physically and emotionally, were consistently reduced. The mean role emotional score was 27.88 ± 38.90 and role physical value was 25.00 ± 32.62.

Univariate and multivariate analysis

To investigate the impact of clinical indices, HCV co-infection, cirrhosis, mood disorder and substance misuse upon HRQOL, univariate analysis was performed with Spearman's correlation and Wilcoxon rank test for binary data (Table 3). Only anxiety and depression significantly impacted upon both EQ-5D and SF-6D (Table 3). A multiple linear regression model was then fit for both EQ-5D and SF-6D to include anxiety, depression, CD4 count and HIV VL (Table 4). CD4 count and HIV VL were included in the multivariate model as they were close to significance. Anxiety and depression remained statistically significant in both multivariate models.

Table 3.

Univariate analysis of health-related quality of life (HRQOL).

EQ-5D utility	SF-6D utility
CD4 count (Spearman’s correlation, p value)	0.166, p = 0.232	0.315, p = 0.02
HIV VL (Spearman’s correlation, p value)	−0.134, p = 0.34	−0.26, p = 0.06
HCV co-infection (Wilcoxon’s rank, p value)	2285, p = 0.110	1951, p = 0.85
Cirrhosis (Wilcoxon’s rank, p value)	1627, p = 0.329	1373, p = 0.705
Anxiety (Spearman’s correlation, p value)	−0.455, p = 0.001	−0.391, p = 0.003
Depression (Spearman’s correlation, p value)	−0.424, p = 0.001	−0.574, p = <0.01
Intravenous drug use over preceding month (Wilcoxon’s rank, p value)	1253.5, p = 0.224	1096, p = 0.543
Monthly alcohol consumption (Spearman’s correlation, p value)	0.53, p = 0.706	0.206, p = 0.138
Monthly opiate consumption (Spearman’s correlation, p value)	0.027, p = 0.844	−0.42, p = 0.766
Monthly cocaine consumption (Spearman’s correlation, p value)	0.31, p = 0.826	0.1, p = 0.477
Monthly crack cocaine consumption (Spearman’s correlation, p value)	0.121, p = 0.385	0.132, p = 0.345
Monthly cannabis consumption (Spearman’s correlation, p value)	0.54, p = 0.706	−0.199, p = 0.153

Table 4.

Multivariate analysis of factors affecting health-related quality of life (HRQOL).

Domain/Item	Β coefficient	SE (standard error)	p Value
Model 1: EQ-5D	0.4185
Intercept	0.8931	0.217	p < 0.001
Anxiety	−0.024	0.007	p < 0.001
Depression	−0.0318	0.007	p < 0.001
CD4 Count	0.0377	0.030	p = 0.205
HIV VL	−0.0014	0.010	p = 0.897
Model 2: SF-6D	0.4819
Intercept	0.7018	0.074	p < 0.001
Anxiety	−0.0098	0.002	p < 0.001
Depression	−0.0117	0.002	p < 0.001
CD4 count	0.0121	0.010	p = 0.231
HIV VL	−0.0031	0.003	p = 0.391

VL: viral load.

Discussion

This study examined HRQOL in non-engaging HIV-infected IDUs. The results of this research clearly demonstrated that this cohort have poor HRQOL as defined by generic and disease-specific measures and clinical indices.

HIV infection has been extensively found to reduce HRQOL.^2,3,7 To date, little research has focused on HIV-infected IDUs, especially the non-engaging cohort. In this discussion, HRQOL will be compared with other equitable chronic conditions and HIV populations from the literature. Factors specific to the HRQOL in this cohort are explored and compared with results from other studies.

EQ-5D utility in this HIV IDU cohort was compared to other chronic disease cohorts. The mean EQ-5D score of 0.45 obtained in this HIV IDU cohort is lower than that reported by Longworth and Bryan for patients awaiting liver transplantation (0.517).²⁶ However, rheumatology cohorts have reported similar values to this non-engaging HIV IDU population; 0.43 in rheumatoid and 0.49 in psoriatic arthritis.²⁰ This was attributed in particular to the poor pain scores present in the arthritis cohort.²⁰

The mean HRQOL scores calculated from the SF-6D were also compared to the liver transplant, arthritis and HCV mono-infected cohorts. The mean SF-6D utility for the HIV IDU cohort was 0.52. In contrast, other studies in patients with liver co-morbidities such as HCV mono-infection had a mean utility of 0.67,²⁷ whilst those awaiting liver transplant had a mean of 0.606.²⁶

Other studies have published on EQ-5D utilities in other HIV populations.^3,17 Whilst most of these have been focused on non-drug-using populations, comparison between cohorts is still useful. Miners et al.³ studied the EQ-5D in a UK-based population where 98% of patients had acquired their HIV sexually.³ They had a mean utility value of 0.65 ( ± 0.28).³ Stavem et al.²⁸ had a more heterogeneous cohort, with 20% IVDUs and their mean EQ-5D utility was 0.77 (±0.26).²⁸ Both of these mean utility values were higher than this HIV IDU cohort (0.51).

The results of the SF-36 health domain scales in this HIV IDU population are consistently lower than what has been presented in the literature in other HIV cohorts.^2,29 The HIV IDU population results were contrasted with the Fleming HIV/HCV and Heslin non-HIV IDU cohorts.^2,30 The demographics of the cohort described by Fleming et al.² is most similar to this HIV IDU population, with 89% IDUs and all patients HIV/HCV co-infected.² The non-HIV IDU population described by Heslin et al. were patients enrolled in a trial of buprenorphine, who required medical management for opioid withdrawal.³⁰ All SF-36 scores of the HIV IDU were lower than the HIV/HCV cohort.² The HIV IDUs also scored consistently lower than the non-HIV IDU cohort except for BP and SF.³⁰ Whilst these populations are not matched, it appears that non-engaging HIV-infected IDUs perform worse than their non-drug-user HIV counterparts.

For the HIV IDU cohort, both the physical and emotional components of the role limitations had the lowest of all the health domain scores. The mental health components of the health domain scores also tended to be lower than their physical health counterparts. The rates of anxiety and depression were high in this cohort. Forty-four percent of individuals were depressed, which is higher than the 22−32% previously quoted in the literature.³¹

Many factors have been found to deleteriously affect HRQOL. Some of these include: HCV infection, substance misuse, depression and mood disorder.^2,4,6,32 Over half of the patients included in this study were HCV co-infected. Individually, HCV and HIV mono-infection have been shown to reduce HRQOL.^27,33 However, co-infection with both viruses does not appear to further worsen HRQOL; Fleming et al.² found HRQOL in co-infected patients to be statistically similar to patients mono-infected with either virus. Therefore, it is unlikely that the poorer HRQOL in this HIV IDU cohort can be attributed to co-infection alone.

Previous literature has found current drug use to be strongly associated with reduced mental and physical HRQOL, where mental health was more severely affected.^6,32 Further investigation of the relationship between drug use, mood disorder and HRQOL was investigated by Sherbourne et al.³² The negative impact of substance misuse on HRQOL disappeared once they corrected for mood disorder.³² In this HIV IDU cohort, there was not a meaningful relationship between substance misuse, injecting drug use and HRQOL. However, anxiety and depression appeared to have a deleterious effect on HRQOL. There was a high baseline rate of anxiety and depression in this cohort, and both were significantly associated with the EQ-5D and SF-6D utility values. The patients who had an EQ-5D utility ‘worse than death’ were also predominantly depressed or anxious. The integration of mental health and HIV services as undertaken in this study may help to manage these challenging patients and improve their physical and mental health and HRQOL.

There are limitations to this research. Patient numbers included in this study were small. This is in part due to the challenges posed and time required to recruit and retain chaotic IDUs with high rates of illiteracy in studies. In the absence of Irish societal preferences, UK values were used to calculate utilities. Other authors have questioned the validity of using preferences from other populations.³⁴ Levels of substance misuse, injecting drug use and forensic history may also have been under-reported due to the high proportion of illiteracy in this cohort necessitating assistance with questionnaire completion. Whilst this was a predominantly Caucasian heterosexual population, anxiety and depression had a more deleterious effect on HRQOL than substance misuse, which enables these results to be extrapolated to other cohorts such as HIV-infected men who have sex with men (MSM) IDUs and non-Irish HIV-infected IDUs.

To summarise, HRQOL was reduced in this non-engaging HIV-infected IDU population. The HRQOL of this population appears to be worse than previously published for other HIV cohorts in the literature and many other comparable chronic conditions such as HCV and patients awaiting liver transplantation. Whilst HCV co-infection and substance misuse do not affect the HRQOL in this population, anxiety and depression appear to have a significant impact upon it. The impact of integrating HIV and addiction services on HRQOL, clinical and cost outcomes is awaited.

Footnotes

Conflict of interest

The authors declare no conflict of interest.

Funding

This research was funded by an education grant from Tibotec and Merck, Sharp & Dohme.

References

Palella

Delaney

Moorman

. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. New Eng J Med 1998; 338: 853–860.

Fleming

Christiansen

Nunes

. Health-related quality of life of patients with HIV disease: impact of hepatitis C coinfection. Clin Infect Dis 2004; 38: 572–578.

Miners

Sabin

Mocroft

. Health-related quality of life in individuals infected with HIV in the era of HAART. HIV Clin Trial 2001; 2: 484–492.

Braitstein

Montessori

Chan

. Quality of life, depression and fatigue among persons co-infected with HIV and hepatitis C: outcomes from a population-based cohort. AIDS Care 2005; 17: 505–515.

Briongos Figuero

Bachiller Luque

Palacios Martin

. Assessment of factors influencing health-related quality of life in HIV-infected patients. HIV Med 2010; 12: 22–30.

Korthius

Zephyrin

Feleishman

. Health-related quality of life in HIV-infected patients: the role of substance use. AIDS Pat Care STD 2008; 22: 859–867.

Hays

Cunningham

Sherbourne

. Health-related quality of life in patients with human immunodeficiency virus infection in the United States: results from the HIV Cost and Services Utilization Study. Am J Med 2000; 108: 714–722.

Jelsma

Maclean

Hughes

. An investigation into the health-related quality of life of individuals living with HIV who are receiving HAART. AIDS Care 2005; 17: 579–588.

Delate

Coons

. The use of 2 health-related quality-of-life measures in a sample of persons infected with human immunodeficiency virus. Clin Infect Dis 2001; 32: E47–E52.

10.

Gill

Griffith

Jacobson

. Relationship of HIV viral loads, CD4 counts, and HAART use to health-related quality of life. J Acquir Immune Defic Syndr 2002; 30: 485–492.

11.

Kauf

Roskell

Shearer

. A predictive model of health state utilities for HIV patients in the modern era of highly active antiretroviral therapy. Value Health 2008; 11: 1144–1153.

12.

Mathews

May

. EuroQol (EQ-5D) measure of quality of life predicts mortality, emergency department utilization, and hospital discharge rates in HIV-infected adults under care. Health Qual Life Outcomes 2007; 5: 5–5.

13.

Altice

Kamarulzaman

Soriano

. Treatment of medical, psychiatric, and substance-use comorbidities in people infected with HIV who use drugs. Lancet 2010; 376: 367–87.

14.

Clarke

Delamere

McCullough

. Assessing limiting factors to the acceptance of antiretroviral therapy in a large cohort of injecting drug users. HIV Med 2003; 4: 33–37.

15.

Brazier

Roberts

Deverill

. The estimation of a preference-based measure of health from the SF-36. J Health Econ 2002; 21: 271–292.

16.

EQ-5D: a standardised instrument for use as a measure of health outcome [online], http://www.euroqol.org (accessed 2012).

17.

Zigmond

Snaith

. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983; 67: 361–370.

18.

Bjelland

Dahl

Haug

. The validity of the Hospital Anxiety and Depression Scale. An updated literature review. J Psychosom Res 2002; 52: 69–77.

19.

Marsden

Farrell

Bradbury

. Development of the treatment outcomes profile. Addiction 2008; 103: 1450–1460.

20.

Adams

Walsh

Veale

. Understanding the relationship between the EQ-5D, SF-6D, HAQ and disease activity in inflammatory arthritis. Pharmacoeconomics 2010; 28: 477–487.

21.

Dolan

Gudex

Kind

. The time trade-off method: results from a general population study. Health Econ 1996; 5: 141–154.

22.

Savard

Laberge

Gauthier

. Evaluating anxiety and depression in HIV-infected patients. J Pers Assess 1998; 71: 349–367.

23.

Williams I, Churchill D, Anderson J, et al. BHIVA guidelines for the treatment of HIV-1positive adults with antiretroviral therapy, http://www.bhiva.org/TreatmentofHIV1_2012.aspx (2012, accessed 2012).

24.

Reiss P, Battegay M, Clumeck N, et al. European Guidelines for the Treatment of HIV Infected adults in Europe, version 6, http://www.europeanaidsclinicalsociety.org/images/stories/EACS-Pdf/EACSGuidelines-v6.0-English.pdf (2011, accessed 20 December 2012).

25.

HSE. Public information: Your Health: Alcohol, http://www.hse.ie/eng/services/healthpromotion/alcohol/ (2011, accessed 2011).

26.

Longworth

Bryan

. An empirical comparison of EQ-5D and SF-6D in liver transplant patients. Health Econ 2003; 12: 1061–1067.

27.

Dan

Kallman

Srivastava

. Impact of chronic liver disease and cirrhosis on health utilities using SF-6D and the health utility index. Liver Transpl 2008; 14: 321–326.

28.

Stavem

Froland

Hellum

. Comparison of preference-based utilities of the 15D, EQ-5D and SF-6D in patients with HIV/AIDS. Qual Life Res 2005; 14: 971–980.

29.

Arpinelli

Visona

Bruno

. Health-related quality of life in asymptomatic patients with HIV. Evaluation of the SF-36 health survey in Italian patients. Pharmacoeconomics 2000; 18: 63–72.

30.

Heslin

Stein

Heinzerling

. Clinical correlates of health-related quality of life among opioid-dependent patients. Qual Life Res 2011; 20: 1205–1213.

31.

Bing

Burnam

Longshore

. Psychiatric disorders and drug use among human immunodeficiency virus-infected adults in the United States. Arch Gen Psychiatry 2001; 58: 721–728.

32.

Sherbourne

Hays

Fleishman

. Impact of psychiatric conditions on health-related quality of life in persons with HIV infection. Am J Psychiatry 2000; 157: 248–254.

33.

Mrus

Sherman

Leonard

. Health values of patients coinfected with HIV/hepatitis C: are two viruses worse than one? Med Care 2006; 44: 158–166.

34.

Johnson

Luo

Shaw

. Valuations of EQ-5D health states: are the United States and United Kingdom different? Med Care 2005; 43: 221–228.