Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Jun 1.
Published in final edited form as: AIDS Behav. 2013 Jun;17(5):1713–1723. doi: 10.1007/s10461-012-0282-2

Alcohol consumption as a barrier to prior HIV testing in a population-based study in rural Uganda

Robin Fatch a, Ben Bellows b, Fred Bagenda c, Edgar Mulogo c, Sheri Weiser a, Judith A Hahn a,*
PMCID: PMC3769173  NIHMSID: NIHMS400008  PMID: 22878790

Abstract

Early receipt of HIV care and ART is essential for improving treatment outcomes, but is dependent first upon HIV testing. Heavy alcohol consumption is common in sub-Saharan Africa, a barrier to ART adherence, and a potential barrier to HIV care. We conducted a population-based study of 2,516 adults in southwestern Uganda from November-December 2007, and estimated the relative risk of having never been tested for HIV using sex-stratified Poisson models. More men (63.9%) than women (56.9%) had never been tested. In multivariable analysis, compared to women who had not consumed alcohol for at least 5 years, women who were current heavy drinkers and women who last drank alcohol 1-5 years prior, were more likely to have never been tested. Alcohol use was not associated with prior HIV testing among men. HIV testing strategies may thus need to specifically target women who drink alcohol.

Keywords: alcohol, HIV testing, Uganda, barriers, sub-Saharan Africa

Introduction

Sub-Saharan Africa accounts for nearly 70% of the global HIV burden, with an estimated 22.5 million people living with HIV/AIDS in 2009 (1). In Uganda, the prevalence of HIV has stabilized between 6.5-7% since 2001, with 1.2 million people living with HIV/AIDS in 2009 (1). Increased access to HIV treatment has led to a decline of AIDS-related deaths in sub-Saharan Africa by 20% from 2004 to 2009 (2). Early HIV treatment has been associated with better response to treatment, including decreased progression to AIDS and death (3-7). Early presentation for HIV care and treatment has also been shown to have additional benefits, including decreased health care costs (8), as well as potentially decreasing the risk of transmission to others (9, 10).

Timely receipt of follow-up HIV care and antiretroviral therapy (ART), however, is dependent upon knowledge of HIV status via testing (11). In Uganda, HIV counseling and testing (HCT) services have been expanded to include voluntary counseling and testing (VCT), home-based counseling and testing services, as well as opt-out routine testing and counseling, with the goal of universal access (12). However, while the percentage of Ugandans who have previously tested for HIV has increased over time, it remains low. According to the 2009 World Health Organization (WHO) Progress Report, only 24% of Ugandans know their HIV status (13). As “seek, test, and treat” garners more attention as a potential strategy for preventing HIV at both the individual and community level (14, 15), identification and reduction of barriers to HIV testing remains a critical step in improving HIV care and prevention in sub-Saharan Africa.

Alcohol is currently the most widely distributed and commonly used recreational drug in Africa; even the most rural areas have reliable production and distribution systems (16-22). The total per capita adult (age 15 and over) consumption of pure alcohol is 11.9 liters per year in Uganda (23). In Uganda, 48.6% of the men and 39.0% of the women drank in the past year; 33.7% of these men and 11.2% of these women drank heavily (60g or more of pure alcohol on at least one occasion, weekly) (23).

In Africa, older age, less education and household wealth, being single, and no history of pregnancy have often been found to be barriers to HIV testing (24-28). Additional barriers to testing have also included possessing less knowledge about HIV/AIDS and testing site locations, having less access to media, holding stigmatizing beliefs, fears, or negative attitudes, lacking willingness to discuss HIV/AIDS with others, and lacking willingness to disclose one's own HIV status (26, 28-33). On the other hand, having poor health has been associated with readiness for VCT among older age groups (25-49 years old, versus 15-24 years) (34). Sexual behaviors such as never using a condom and having a high-risk sex partner have also been associated with participation in VCT (35, 36); however, self-perceived risk for HIV infection has been associated with VCT acceptance in some but not all of the studies that examined this issue (34-36).

Substance abuse has been found to be a barrier to utilization of general health care services (37), while alcohol consumption and drinking venue attendance have been associated with increased HIV infection and risk behaviors, such as number of sexual partners, unprotected sex, and commercial sex in sub-Saharan Africa (38-42). However, despite consistent associations of alcohol consumption with HIV infection in sub-Saharan Africa (43), few studies have examined the association between alcohol use and utilization of HIV testing. The small number of studies that have examined the relationship between alcohol consumption and HIV testing in sub-Saharan Africa have had mixed findings. Among pregnant women in South Africa, alcohol use was associated with a decreased odds of participation in post-test counseling and receipt of testing results (44). Alternatively, women reporting alcohol abuse within the past year were more likely to accept HIV testing offered in a study of high risk South African women, as compared to those without alcohol abuse (45). In a cross-sectional survey of households in South Africa assessing factors associated with HIV testing, having ever taken alcohol was not significantly associated with testing, among either men or women (28). Alcohol consumption by sexual partners may also be associated with decreased receipt of HIV test results. Among pregnant women in Tanzania, those who consumed alcohol or who had a partner who consumed alcohol, were less likely to return to receive HIV results within one week of testing (46). The conflicting results from these few studies indicate the need for further research to assess the relationship between alcohol use and HIV testing. Therefore, our main goal was to examine whether alcohol use was associated with decreased HIV testing in a large population-based cross-sectional study conducted in southwestern Uganda.

Methods

Study population

This study is an analysis of data collected in a cross-sectional population-based survey that was conducted in November and December 2007 as part of a study examining the effect of a voucher program on sexually transmitted infection (STI) treatment (47). The STI voucher program subsidized non-HIV STI treatment, with the intent of improving the uptake of STI treatment. Survey data were collected from four districts (Mbarara, Ibanda, Isingiro and Kiruhura) where the STI voucher pilot program was active, and a fifth control district (Bushenyi) with no voucher program. The survey had a four-stage design using the 2002 Uganda census. In the first stage, 41 parishes (administrative units of subcounties within the districts) were sampled with probability proportional to population size from the five districts. In the next stage, a total of 82 villages were selected from the 41 parishes with probability proportional to population size. The third stage consisted of a random sample of 2,952 enumerated households in the selected villages. In the fourth stage, one resident aged 15-49 years was randomly selected from each household for interview. Information on the number of individuals who were not at home and thus not interviewed, and those refusing participation, was not collected. There were 2,516 respondents in 2007 with sufficient data to include in the analysis.

Dependent variable

The dependent variable was never having been tested for HIV, as reported by the participant.

Independent variables

The primary predictor variable was alcohol use. Participants were asked when they last took alcohol (if ever). Past drinkers were split into two groups to differentiate those who may have quit drinking recently due to poor health: distant past drinkers were those who had consumed alcohol more than 5 years ago, and recent past drinkers were those who had consumed alcohol 1-5 years ago. “Current drinkers” were defined as participants who had taken any alcohol within the past year. We split current drinkers into heavy and not heavy drinkers to examine whether the level of drinking was associated with having ever been tested for HIV. Current drinkers were classified as current heavy drinkers if they reported any of the following (based on the 75th percentile for these questions): ever taking six or more drinks on one occasion in the past year; spending a total of at least 15,000 Ugandan shillings (approximately $10 at the time of the survey) on alcohol taken in the past 30 days; being intoxicated on 3 or more of the past 30 days. We used these measures rather than the AUDIT-C, a standard measure of self-reported hazardous alcohol consumption (48), because the AUDIT-C presented difficulties in quantifying the number of drinks consumed in an area where non-standard drink sizes and alcohol concentration are common (49).

We also examined alcohol consumption by sexual partners within the past year as a potential barrier to HIV testing, categorized as follows: no sex partners took alcohol in the past year; sex partner(s) took alcohol but didn't become intoxicated at least monthly; sex partner(s) took alcohol and became intoxicated at least monthly.

Covariates

Covariates thought to be potential barriers to health care and HIV testing were also examined, including participant sex, age, marital status, religion, education, self-reported health status, and utilization of general health care services. We created a household wealth index using principal components analysis (50) to group households based on ownership of durable goods, housing quality, and available energy sources. This index score was split into three categories: the bottom 40% was classified as “poor”, the middle 40% as “middle”, and the top 20% as “rich” (51). The Household Food Insecurity Access Scale (HFIAS) was used to classify household food insecurity as follows: “food secure”, “mildly food insecure access”, “moderately food insecure access”, and “severely food insecure access” (52). We also examined two risky sexual behaviors as potential confounders of an association between alcohol consumption and prior HIV testing.

These variables were the number of sexual partners in the prior 6 months (0 or 1 partners versus more than 1 partner), and whether or not the participant reported receiving money, goods or favors in the prior 6 months in exchange for sex.

Statistical analysis

We calculated frequency distributions for categorical variables and medians and interquartile ranges for continuous variables. As the overall proportion of participants who had never been tested for HIV was quite high (60.5%), we estimated the relative risk (RR) of having never tested for HIV (rather than the odds ratio) (53). Healthcare utilization and HIV testing are consistently found to be higher among women compared to men (24, 28, 32, 54, 55), thus we stratified the analyses by gender in order to allow the correlates of HIV testing to vary by sex. We conducted stratified bivariate and multivariable analyses to estimate the RR and adjusted risk ratio (ARR) for never being tested for HIV. We used Poisson regression models to estimate the RRs, a roughly equivalent approach (56) to using log-binomial models that sometimes failed to converge. Multivariable analysis included all covariates of interest.

Most variables were missing some data points; the range of missingness for each variable was 0-12%. However, in multivariable models using listwise deletion, 30% of all observations were missing. We examined the relationships between missing covariates and HIV testing history as well as alcohol use; there were no clear patterns of missingness. The only associations we found were that missing religion was associated with not having been tested for HIV, and that missing marital status was associated with decreased alcohol use. As a certain number of statistically significant associations could be expected due to the multiple testing nature of this exploration, and we did not expect that the observed associations would bias our overall results, we proceeded with multiple imputation, which assumes that the data are missing at random. Multiple imputation allowed us to retain observations with missing values and thereby minimize the loss of power that we would have otherwise encountered had we used listwise deletion (57). We imputed the missing values using multiple (five) imputations by chained equations in Stata and conducted multivariable analysis of the imputed dataset, using sex-stratified Poisson models to estimate the ARR for never having been tested for HIV as described above. The multivariable results using the imputed dataset did not substantially differ from the results obtained using listwise deletion. Thus, the multivariable results presented below are those from the imputed data.

Results

Sample characteristics

Approximately half the 2,516 participants were female. Sixty percent of participants had never before been tested for HIV, and more men than women reported never testing for HIV (63.9% compared to 56.9%, χ2(1) = 12.78, p<0.01). Among the participants who had ever been tested for HIV, 96% reported receiving their test results, with 12% receiving HIV-positive test results.

Nearly two thirds of all participants were lifetime abstainers or distant past drinkers, and 15.0% of participants were recent past drinkers. Thirteen percent of all participants were current heavy drinkers, and 8.0% were current but not heavy drinkers (Table 1). A much higher proportion of men (21.3%) were current heavy drinkers, as compared to women (5.4%) (χ2(1) = 129.83; p<0.01).

Table 1. Demographic and behavioral characteristics of participants of the STI Prevalence and Treatment Seeking Behavior in Mbarara Study, November – December 2007.

Total n (%) Female n (%) Male n (%) χ2 (p-value)
All 2516 (100.0) 1242 (100.0) 1274 (100.0)
Demographics
Age 8.65 (0.01)
15-24 yrs old 781 (31.1) 416 (33.6) 365 (28.7)
25-34 yrs old 993 (39.5) 484 (39.1) 509 (40.0)
35-49 yrs old 738 (29.4) 338 (27.3) 400 (31.4)
Marital status 126.91 (<0.01)
Married/cohabitating 1636 (66.8) 829 (69.0) 807 (64.6)
Previously married/separated 256 (10.5) 190 (15.8) 66 (5.3)
Never married 558 (22.8) 182 (15.2) 376 (30.1)
Religion 2.00 (0.57)
Protestant 1379 (55.4) 672 (55.1) 707 (55.7)
Catholic 828 (33.3) 411 (33.7) 417 (32.9)
Muslim 230 (9.2) 107 (8.8) 123 (9.7)
Other 51 (2.1) 29 (2.4) 22 (1.7)
Highest level of education 14.31 (<0.01)
Secondary school or more 858 (34.4) 387 (31.6) 471 (37.2)
Primary school 1359 (54.5) 680 (55.5) 679 (53.6)
No school/preschool only 275 (11.0) 159 (13.0) 116 (9.2)
Household wealth 3.59 (0.17)
Rich 449 (20.1) 228 (20.8) 221 (19.4)
Middle 893 (39.9) 416 (37.9) 477 (41.8)
Poor 897 (40.1) 454 (41.4) 443 (38.8)
Household food insecurity access (HFIAS) 15.24 (<0.01)
Food secure 293 (12.0) 126 (10.5) 167 (13.5)
Mildly food insecure 175 (7.2) 89 (7.4) 86 (6.9)
Moderately food insecure 1150 (47.0) 610 (50.6) 540 (43.5)
Severely food insecure 829 (33.9) 380 (31.5) 449 (36.2)
Alcohol use
Alcohol use 187.78 (<0.01)
Never or distant past 1506 (63.7) 891 (75.7) 615 (51.8)
drinker (>5 years prior)
Past drinker (1-5 years prior) 355 (15.0) 159 (13.5) 196 (16.5)
Current not heavy drinker 188 (8.0) 64 (5.4) 124 (10.4)
Current heavy drinker 316 (13.4) 63 (5.4) 253 (21.3)
Did any sex partners drink alcohol (past year)? 127.86 (<0.01)
No sex partners drank alcohol 1720 (70.1) 738 (60.7) 982 (79.2)
Yes, but did not become intoxicated once a month or more 139 (5.7) 62 (5.1) 77 (6.2)
Yes, and became intoxicated at least once a month 596 (24.3) 415 (34.2) 181 (14.6)
Sexual risk behaviors
Number of sexual partners, past 6 months 231.86 (<0.01)
0 or 1 partner 1998 (80.5) 1138 (92.8) 860 (68.5)
More than 1 partner 484 (19.5) 89 (7.3) 395 (31.5)
Received money, goods or favors, in exchange for sex, past 6 months? 0.50 (0.48)
No 2156 (87.8) 1056 (87.3) 1100 (88.2)
Yes 301 (12.3) 154 (12.7) 147 (11.8)
Health and health care utilization
History of HIV testing 12.78 (<0.01)
Never been tested for HIV 1521 (60.5) 707 (56.9) 814 (63.9)
Ever been tested for HIV 995 (39.6) 535 (43.1) 460 (36.1)
Health status (self-rated) 6.25 (0.04)
Excellent/very good 547 (22.1) 251 (20.5) 296 (23.7)
Good 869 (35.1) 420 (34.4) 449 (35.9)
Fair/poor 1057 (42.7) 551 (45.1) 506 (40.5)
Sought healthcare services, past 6 months? 12.31 (<0.01)
Not ill, past 6 months 1143 (48.3) 520 (44.7) 623 (51.7)
Ill, but did not seek care 243 (10.3) 123 (10.6) 120 (10.0)
Ill, and sought care 982 (41.5) 521 (44.8) 461 (38.3)
Open in a new tab

Bivariate and Multivariable results

Alcohol use

In bivariate analysis among women, compared to lifetime abstainers and distant past drinkers, current heavy drinkers and recent past drinkers were significantly more likely to have never been tested for HIV. This association persisted in multivariable analysis among the women (Table 2); compared to lifetime abstainers and distant past drinkers, current heavy drinkers (ARR: 1.32, 95%CI: 1.10-1.58) and recent past drinkers (ARR: 1.20, 95%CI: 1.05-1.37) were more likely to have never been tested for HIV. Current not heavy drinkers had increased risk of having never been tested but this did not reach statistical significance (ARR: 1.20, 95%CI: 0.98-1.47). Alcohol use by sexual partners was not associated with HIV testing among women.

Table 2. Factors associated with having never been tested for HIV among women. STI Prevalence and Treatment Seeking Behavior in Mbarara Study, November – December 2007.
Never tested for HIV Bivariate Relative Risk (95% Confidence Interval) Bivariate Poisson Regression z- statistic (p-value) Adjusted Relative Risk (95% Confidence Interval) Multivariable Poisson Regression t- statistic (p-value)
All 56.9%
Demographics
Age
15-24 yrs old 61.3% 1.00 1.00
25-34 yrs old 55.0% 0.90 (0.80, 1.00) -1.93 (0.05) 0.93 (0.82, 1.05) -1.16 (0.25)
35-49 yrs old 53.9% 0.88 (0.78, 1.00) -2.04 (0.04) 0.89 (0.77, 1.03) -1.56 (0.12)
Marital status
Married/cohabitating 54.8% 1.00 1.00
Previously married/separated 51.1% 0.93 (0.80, 1.09) -0.90 (0.37) 0.89 (0.76, 1.05) -1.39 (0.17)
Never married 70.9% 1.29 (1.16, 1.45) 4.52 (<0.01) 1.19 (1.04, 1.37) 2.54 (0.01)
Religion
Protestant 57.4% 1.00 1.00
Catholic 57.7% 1.00 (0.90, 1.12) 0.07 (0.94) 0.98 (0.88, 1.08) -0.41 (0.68)
Muslim 47.7% 0.83 (0.67, 1.02) -1.75 (0.08) 0.90 (0.73, 1.12) -0.96 (0.34)
Other 48.3% 0.84 (0.57, 1.23) -0.89 (0.37) 0.79 (0.55, 1.15) -1.23 (0.22)
Highest level of education
Secondary school or more 53.0% 1.00 1.00
Primary school 56.5% 1.07 (0.95, 1.20) 1.09 (0.28) 1.06 (0.94, 1.19) 0.90 (0.37)
No school/preschool only 67.3% 1.27 (1.10, 1.47) 3.27 (<0.01) 1.30 (1.10, 1.53) 3.14 (<0.01)
Household wealth
Rich 44.3% 1.00 1.00
Middle 54.1% 1.22 (1.03, 1.45) 2.30 (0.02) 1.22 (1.04, 1.44) 2.44 (0.02)
Poor 65.4% 1.48 (1.26, 1.73) 4.77 (<0.01) 1.49 (1.26, 1.76) 4.67 (<0.01)
Household food insecurity access (HFIAS)
Food secure 65.9% 1.00 1.00
Mildly food insecure 51.7% 0.78 (0.62, 0.99) -2.01 (0.05) 0.82 (0.65, 1.03) -1.69 (0.09)
Moderately food insecure 53.6% 0.81 (0.70, 0.94) -2.77 (<0.01) 0.79 (0.68, 0.91) -3.15 (<0.01)
Severely food insecure 60.0% 0.91 (0.78, 1.06) -1.22 (0.22) 0.84 (0.72, 0.97) -2.29 (0.02)
Alcohol use
Alcohol use
Never or distant past drinker (>5 years prior) 54.7% 1.00 1.00
Past drinker (1-5 years prior) 65.4% 1.20 (1.05, 1.36) 2.75 (<0.01) 1.20 (1.05, 1.37) 2.69 (<0.01)
Current not heavy drinker 60.9% 1.11 (0.91, 1.37) 1.04 (0.30) 1.20 (0.98, 1.47) 1.80 (0.07)
Current heavy drinker 73.0% 1.34 (1.14, 1.57) 3.51 (<0.01) 1.32 (1.10, 1.58) 3.00 (<0.01)
Did any sex partners drink alcohol (past year)?
No sex partners drank alcohol 57.2% 1.00 1.00
Yes, but did not become intoxicated once a month or more 61.3% 1.07 (0.87, 1.32) 0.66 (0.51) 1.00 (0.81, 1.23) -0.04 (0.97)
Yes, and became intoxicated at least once a month 55.7% 0.97 (0.88, 1.08) -0.50 (0.62) 0.92 (0.82, 1.03) -1.46 (0.15)
Sexual risk behaviors
Number of sexual partners, past 6 months
0 or 1 partner 56.9% 1.00 1.00
More than 1 partner 59.6% 1.05 (0.88, 1.25) 0.51 (0.61) 0.91 (0.75, 1.12) -0.88 (0.38)
Received money, goods or favors, in exchange for sex, past 6 months?
No 55.8% 1.00 1.00
Yes 65.6% 1.18 (1.04, 1.33) 2.51 (0.01) 1.09 (0.96, 1.25) 1.31 (0.19)
Health and health care utilization
Health status (self-rated)
Excellent/very good 57.0% 1.00 1.00
Good 57.1% 1.00 (0.88, 1.15) 0.04 (0.97) 0.97 (0.85, 1.11) -0.38 (0.70)
Fair/poor 57.5% 1.01 (0.89, 1.15) 0.15 (0.88) 0.95 (0.83, 1.09) -0.70 (0.48)
Sought healthcare services, past 6 months?
Not ill, past 6 months 57.7% 1.00 1.00
Ill, but did not seek care 67.5% 1.17 (1.01, 1.35) 2.15 (0.03) 1.14 (0.98, 1.32) 1.64 (0.10)
Ill, and sought care 53.9% 0.93 (0.84, 1.04) -1.22 (0.22) 0.96 (0.86, 1.07) -0.73 (0.47)
Open in a new tab

Among men, neither alcohol use by the participant nor by their sexual partner(s) was associated with HIV testing in bivariate or multivariable analyses (Table 3).

Table 3. Factors associated with having never been tested for HIV among men. STI Prevalence and Treatment Seeking Behavior in Mbarara Study, November – December 2007.
Never tested for HIV Bivariate Relative Risk (95% Confidence Interval) Bivariate Poisson Regression z- statistic (p-value) Adjusted Relative Risk (95% Confidence Interval) Multivariable Poisson Regression t-statistic (p-value)
All 63.9%
Demographics
Age
15-24 yrs old 67.1% 1.00 1.00
25-34 yrs old 63.3% 0.94 (0.85, 1.04) -1.19 (0.24) 0.94 (0.84, 1.05) -1.07 (0.28)
35-49 yrs old 61.8% 0.92 (0.83, 1.02) -1.55 (0.12) 0.89 (0.78, 1.02) -1.70 (0.09)
Marital status
Married/cohabitating 62.7% 1.00 1.00
Previously married/separated 69.7% 1.11 (0.94, 1.31) 1.24 (0.22) 1.10 (0.93, 1.30) 1.10 (0.27)
Never married 65.4% 1.04 (0.95, 1.14) 0.92 (0.36) 1.04 (0.93, 1.16) 0.63 (0.53)
Religion
Protestant 61.8% 1.00 1.00
Catholic 68.6% 1.11 (1.02, 1.21) 2.34 (0.02) 1.11 (1.02, 1.21) 2.38 (0.02)
Muslim 57.7% 0.93 (0.79, 1.10) -0.83 (0.41) 0.99 (0.84, 1.17) -0.07 (0.95)
Other 68.2% 1.10 (0.82, 1.48) 0.66 (0.51) 1.13 (0.85, 1.49) 0.82 (0.41)
Highest level of education
Secondary school or more 55.2% 1.00 1.00
Primary school 68.6% 1.24 (1.13, 1.37) 4.45 (<0.01) 1.20 (1.09, 1.32) 3.60 (<0.01)
No school/preschool only 70.7% 1.28 (1.11, 1.48) 3.40 (<0.01) 1.24 (1.07, 1.44) 2.90 (<0.01)
Household wealth
Rich 56.6% 1.00 1.00
Middle 61.2% 1.08 (0.94, 1.24) 1.14 (0.25) 1.06 (0.92, 1.23) 0.85 (0.40)
Poor 71.1% 1.26 (1.10, 1.43) 3.45 (<0.01) 1.19 (1.02, 1.37) 2.31 (0.02)
Household food insecurity access (HFIAS)
Food secure 64.7% 1.00 1.00
Mildly food insecure 55.8% 0.86 (0.69, 1.07) -1.32 (0.19) 0.87 (0.70, 1.08) -1.26 (0.21)
Moderately food insecure 64.6% 1.00 (0.88, 1.14) -0.01 (0.99) 0.96 (0.84, 1.09) -0.64 (0.52)
Severely food insecure 63.0% 0.97 (0.85, 1.11) -0.38 (0.70) 0.94 (0.82, 1.08) -0.85 (0.40)
Alcohol use
Alcohol use
Never or distant past drinker (>5 years prior) 62.1% 1.00 1.00
Past drinker (1-5 years prior) 66.3% 1.07 (0.95, 1.20) 1.10 (0.27) 1.07 (0.94, 1.21) 1.01 (0.32)
Current not heavy drinker 66.9% 1.08 (0.94, 1.24) 1.06 (0.29) 1.08 (0.93, 1.25) 1.05 (0.30)
Current heavy drinker 66.0% 1.06 (0.95, 1.18) 1.10 (0.27) 1.08 (0.96, 1.22) 1.35 (0.18)
Did any sex partners take alcohol (past year)?
No sex partners drank alcohol 64.8% 1.00 1.00
Yes, but did not become intoxicated once a month or more 59.7% 0.92 (0.76, 1.11) -0.84 (0.40) 0.94 (0.78, 1.14) -0.61 (0.54)
Yes, and became intoxicated at least once a month 62.4% 0.96 (0.85, 1.09) -0.59 (0.56) 0.98 (0.86, 1.11) -0.35 (0.73)
Sexual risk behaviors
Number of sexual partners, past 6 months
0 or 1 partner 65.4% 1.00 1.00
More than 1 partner 60.3% 0.92 (0.84, 1.01) -1.70 (0.09) 0.93 (0.85, 1.03) -1.37 (0.17)
Received money, goods or favors, in exchange for sex, past 6 months?
No 64.6% 1.00 1.00
Yes 57.8% 0.90 (0.78, 1.04) -1.49 (0.14) 0.90 (0.78, 1.05) -1.37 (0.17)
Health and health care utilization
Health status (self-rated)
Excellent/very good 61.2% 1.00 1.00
Good 64.8% 1.06 (0.95, 1.19) 1.00 (0.32) 1.03 (0.92, 1.16) 0.59 (0.59)
Fair/poor 65.2% 1.07 (0.95, 1.19) 1.14 (0.26) 1.02 (0.91, 1.14) 0.27 (0.79)
Sought healthcare services, past 6 months?
Not ill, past 6 months 61.8% 1.00 1.00
Ill, but did not seek care 70.8% 1.15 (1.01, 1.31) 2.05 (0.04) 1.09 (0.95, 1.25) 1.24 (0.22)
Ill, and sought care 64.0% 1.04 (0.94, 1.14) 0.74 (0.46) 1.02 (0.93, 1.12) 0.39 (0.69)
Open in a new tab

Covariates

Among women, in bivariate analysis, the likelihood of having never been tested for HIV decreased with age, and increased for those who reported having received money or goods in exchange for sex in the past 6 months and those who had been ill in the past 6 months but had not sought care, but these associations were no longer statistically significant after adjusting for other covariates in the multivariable analysis (Table 2). Lower level of education, lower household wealth, and single marital status were associated with decreased HIV testing in bivariate and multivariable analyses. In bivariate analysis, women from mildly and moderately food insecure households were significantly less likely to have never been tested for HIV compared to those from food secure households, while in multivariable analysis, women from moderately and severely food insecure households were significantly less likely to have never been tested for HIV. Among men, those who had been ill in the past 6 months but who had not sought care, compared to men who had not been ill, were less likely to have ever been tested for HIV in bivariate analysis; however, this association was no longer significant after adjusting for other covariates in the multivariable model (Table 3). Lower education level, lower household wealth, and being Catholic versus Protestant were associated with decreased HIV testing among men in both bivariate and multivariable analyses.

Discussion

While Uganda has increased efforts to expand access to HIV testing and care, we found that 60% of participants in this study of the general population near Mbarara, Uganda, had never before tested for HIV in 2007. This figure is comparable to current estimates by the WHO, which reports that only 24% of Ugandans know their current HIV status (13). Similar to other studies (28, 29), more women than men in our study (43.1% versus 36.1%) reported having ever been tested for HIV. This difference may be due in part to HIV testing during pregnancy via participation in prevention of mother to child transmission (PMTCT) programs, or may be due to the higher rates of healthcare utilization often seen among women (55).

Alcohol use was associated with decreased HIV testing among women. Women who were current heavy drinkers, and those who had last consumed alcohol 1-5 years ago, were more likely to have never been tested for HIV compared to distant past drinkers and abstainers. There was a similar association of lack of HIV testing among women reporting current but not heavy drinking, but this association did not reach statistical significance. These results suggest that alcohol use by women may be associated with decreased utilization of HIV testing services. As health behaviors often tend to cluster together, it may be that women who participate in activities with potentially adverse health effects, such as alcohol consumption, may also be less likely to engage in positive health behaviors, including HIV testing. Alcohol use among women could also be a barrier to HIV testing if these women are spending limited resources on alcohol, rather than using this money to access health services such as HIV testing. It is also possible that women who had recently consumed alcohol could have been hesitant to seek HIV testing if they feared their healthcare provider might hold stigmatizing beliefs or negative attitudes towards alcohol consumption by women. Transactional sex has been associated with gender-based violence as well alcohol abuse (42); it may be that women who drink alcohol have little relationship power, and for this reason are less able to access health services. Alcohol use is consistently associated with risky sexual behavior in sub-Saharan Africa (38-42) (and was associated with an increased number of sexual partners, and receipt of money or goods for sex in our study, data not shown), and thus may be expected to be associated with increased likelihood to test for HIV due to higher perceived risk for HIV infection. However, sexual risk behavior was not significantly associated with HIV testing in multivariable analysis, nor did it alter the association between testing and alcohol use, indicating that it was not a confounder of this association among women. Alcohol use among men was not associated with HIV testing, nor was drinking by the participants' sexual partners.

Consistent with prior studies, less education was associated with a decreased likelihood of HIV testing (25, 27, 28), as was lower household wealth (25-27). These results were consistent in men and women, suggesting that less educated and economically disadvantaged individuals may be less aware of and/or less able to access HIV testing services, and may thus be key targets for education and outreach related to these services. Surprisingly, however, moderate and severe household food insecurity was associated with a significantly increased likelihood of ever having been tested among women, indicating that the association with socioeconomic factors and HIV testing may not be clear cut. Severe food insecurity was associated with increased hospitalizations and opportunistic infections in rural Uganda (58); thus, it is possible that the association seen in this study was due to poorer health and increased exposure to healthcare services, including HIV testing, among these groups. While we did adjust for self-rated health status and recent healthcare utilization in the multivariable model, these questions may not have been detailed enough to capture the full relationship in our data. Additionally, as many non-governmental organizations such as The AIDS Support Organization, which has a service center in Mbarara (59), embrace comprehensive approaches to HIV patient care, their services may be particularly attractive to those in need of other basic services in addition to HIV testing and care, including food for their households. For this reason, people from food insecure households may be more likely to seek out and participate in such programs, and thus be more likely to have previously been tested for HIV.

A limitation of this study is the cross-sectional study design, preventing us from making any conclusions regarding the temporal relationships between HIV testing and these factors. Additionally, as participants were simply asked whether or not they had ever been tested in their lifetime, there may have been misclassification if participants forgot past testing. We also have no information on reasons for HIV testing (for example, if participants voluntarily sought testing, if they were tested because of illness, or as part of a PMTCT program), which may potentially further elucidate some of the differences observed between men and women. Only approximately 20% of all participants reported taking alcohol in the past year, which is much lower than the WHO estimates for Uganda (44%), but closer to the overall estimates for Africa (29.2%) (23). This suggests the possibility that alcohol consumption was under-reported in this study, which could cause bias to the null. Another possibility may be that people who drink heavily were less willing to participate in the survey; however, information on participation refusals was not collected and therefore cannot be examined. As expected, however, more men than women reported heavy drinking in the past year, with 21.3% of men and only 5.4% of women classified as current, heavy drinkers. Finally, we had a large number of observations with missing data, which would have led us to exclude ∼30% of our observations in multivariable analysis; therefore, we conducted multiple imputation. While it is not possible to know whether the assumptions of multiple imputation were met (57), we were reassured that the results from the multivariable analysis using imputed data were similar to those from the multivariable analysis deleting all cases with any missing value.

The strength of this study is that it was conducted among a large random sample of persons aged 15-49 in rural Uganda, and thus provides a preliminary assessment of the factors associated with having never been tested for HIV. The high overall percentage of participants who had never before been tested for HIV demonstrates that a key step in achieving universal access to HIV care and services will be further identification and reduction of barriers to HIV testing. We found that alcohol use among women was associated with decreased HIV testing. Given that alcohol consumption is also a risk factor for HIV infection, this is problematic: female drinkers may be infected without their knowledge, and at high risk of transmitting HIV to others. HIV testing programs should target women who drink alcohol, and measures to prevent the transmission of HIV among drinkers should also be increased. These findings add to the growing body of literature suggesting that attention to heavy alcohol consumption as a critical part of HIV prevention, diagnosis and treatment efforts is greatly needed (60, 61).

Acknowledgments

Ben Bellows received funding from the German Development Bank (KFW) and University of California Berkeley for this project. Judith A. Hahn received support from the Liver Center (P30 DK026743). The authors have no known conflicts of interest.

References

  • 1.Global Report: UNAIDS report on the global AIDS epidemic 2010: UNAIDS2010
  • 2.Global Report Fact Sheet: Sub-Saharan Africa: UNAIDS2009
  • 3.Sterne JA, May M, Costagliola D, et al. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: a collaborative analysis of 18 HIV cohort studies. Lancet. 2009 Apr 18;373(9672):1352–63. doi: 10.1016/S0140-6736(09)60612-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Lawn SD, Harries AD, Anglaret X, Myer L, Wood R. Early mortality among adults accessing antiretroviral treatment programmes in sub-Saharan Africa. AIDS. 2008 Oct 1;22(15):1897–908. doi: 10.1097/QAD.0b013e32830007cd. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Lanoy E, Mary-Krause M, Tattevin P, et al. Frequency, determinants and consequences of delayed access to care for HIV infection in France. Antivir Ther. 2007;12(1):89–96. doi: 10.1177/135965350701200111. [DOI] [PubMed] [Google Scholar]
  • 6.Egger M, May M, Chene G, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet. 2002 Jul 13;360(9327):119–29. doi: 10.1016/s0140-6736(02)09411-4. [DOI] [PubMed] [Google Scholar]
  • 7.Sabin CA, Smith CJ, Gumley H, et al. Late presenters in the era of highly active antiretroviral therapy: uptake of and responses to antiretroviral therapy. AIDS. 2004 Nov 5;18(16):2145–51. doi: 10.1097/00002030-200411050-00006. [DOI] [PubMed] [Google Scholar]
  • 8.Krentz HB, Auld MC, Gill MJ. The high cost of medical care for patients who present late (CD4 <200 cells/microL) with HIV infection. HIV Med. 2004 Mar;5(2):93–8. doi: 10.1111/j.1468-1293.2004.00193.x. [DOI] [PubMed] [Google Scholar]
  • 9.Attia S, Egger M, Muller M, Zwahlen M, Low N. Sexual transmission of HIV according to viral load and antiretroviral therapy: systematic review and meta-analysis. AIDS. 2009 Jul 17;23(11):1397–404. doi: 10.1097/QAD.0b013e32832b7dca. [DOI] [PubMed] [Google Scholar]
  • 10.Donnell D, Baeten JM, Kiarie J, et al. Heterosexual HIV-1 transmission after initiation of antiretroviral therapy: a prospective cohort analysis. Lancet. 2010 Jun 12;375(9731):2092–8. doi: 10.1016/S0140-6736(10)60705-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Giordano TP, Suarez-Almazor ME, Grimes RM. The population effectiveness of highly active antiretroviral therapy: are good drugs good enough? Curr HIV/AIDS Rep. 2005 Nov;2(4):177–83. doi: 10.1007/s11904-005-0013-7. [DOI] [PubMed] [Google Scholar]
  • 12.Uganda National Policy Guidelines for HIV Counselling and Testing: Uganda Ministry of Health2005
  • 13.progress report 2009. World Health Organization; 2009. Towards universal access: scaling up priority HIV/AIDS interventions in the health sector. [Google Scholar]
  • 14.Powers KA, Ghani AC, Miller WC, et al. The role of acute and early HIV infection in the spread of HIV and implications for transmission prevention strategies in Lilongwe, Malawi: a modelling study. Lancet. 2011 Jul 16;378(9787):256–68. doi: 10.1016/S0140-6736(11)60842-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Gardner EM, McLees MP, Steiner JF, Del Rio C, Burman WJ. The spectrum of engagement in HIV care and its relevance to test-and-treat strategies for prevention of HIV infection. Clin Infect Dis. 2011 Mar 15;52(6):793–800. doi: 10.1093/cid/ciq243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Parry CD. South Africa: alcohol today. Addiction. 2005 Apr;100(4):426–9. doi: 10.1111/j.1360-0443.2005.01015.x. [DOI] [PubMed] [Google Scholar]
  • 17.Parry CD, Bhana A, Myers B, et al. Alcohol use in South Africa: findings from the South African Community Epidemiology Network on Drug use (SACENDU) Project. J Stud Alcohol. 2002 Jul;63(4):430–5. doi: 10.15288/jsa.2002.63.430. [DOI] [PubMed] [Google Scholar]
  • 18.Obot IS. Substance abuse, health and social welfare in Africa: an analysis of the Nigerian experience. Soc Sci Med. 1990;31(6):699–704. doi: 10.1016/0277-9536(90)90252-n. [DOI] [PubMed] [Google Scholar]
  • 19.Gillis LS, Stone GL. The fate of drinkers: a six-year follow-up study of a community survey. S Afr Med J. 1977 May 28;51(22):789–91. [PubMed] [Google Scholar]
  • 20.Adomakoh CC. Alcoholism: the African Scene. Ann N Y Acad Sci. 1976;273:39–46. doi: 10.1111/j.1749-6632.1976.tb52861.x. [DOI] [PubMed] [Google Scholar]
  • 21.Omoluabi PF. A review of the incidence of nonprescription psychoactive substance use/misuse in Nigeria. Int J Addict. 1995 Mar;30(4):445–58. doi: 10.3109/10826089509048736. [DOI] [PubMed] [Google Scholar]
  • 22.Acuda SW, Sebit MB. Prevalence of psychoactive substance use among psychiatric in-patients in Harare, Zimbabwe. Cent Afr J Med. 1997 Aug;43(8):226–9. [PubMed] [Google Scholar]
  • 23.Global Status Report on Alcohol and Health: World Health Organization2011
  • 24.Obermeyer CM, Osborn M. The utilization of testing and counseling for HIV: a review of the social and behavioral evidence. Am J Public Health. 2007 Oct;97(10):1762–74. doi: 10.2105/AJPH.2006.096263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Gage AJ, Ali D. Factors associated with self-reported HIV testing among men in Uganda. AIDS Care. 2005 Feb;17(2):153–65. doi: 10.1080/09540120512331325635. [DOI] [PubMed] [Google Scholar]
  • 26.Siziya S, Muula AS, Rudatsikira E, Mataya RH. Correlates of HIV testing among women in Malawi: results from the 2006 Multiple Indicator Cluster Survey. Trop Med Int Health. 2008 Nov;13(11):1351–6. doi: 10.1111/j.1365-3156.2008.02155.x. [DOI] [PubMed] [Google Scholar]
  • 27.Sambisa W, Curtis S, Mishra V. AIDS stigma as an obstacle to uptake of HIV testing: evidence from a Zimbabwean national population-based survey. AIDS Care. 2010 Feb;22(2):170–86. doi: 10.1080/09540120903038374. [DOI] [PubMed] [Google Scholar]
  • 28.Venkatesh KK, Madiba P, De Bruyn G, Lurie MN, Coates TJ, Gray GE. Who gets tested for HIV in a South African urban township? Implications for test and treat and gender-based prevention interventions. J Acquir Immune Defic Syndr. 2011 Feb 1;56(2):151–65. doi: 10.1097/QAI.0b013e318202c82c. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Mitchell S, Cockcroft A, Lamothe G, Andersson N. Equity in HIV testing: evidence from a cross-sectional study in ten Southern African countries. BMC Int Health Hum Rights. 2010;10(23):23. doi: 10.1186/1472-698X-10-23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Bwambale FM, Ssali SN, Byaruhanga S, Kalyango JN, Karamagi CA. Voluntary HIV counselling and testing among men in rural western Uganda: implications for HIV prevention. BMC Public Health. 2008;8(263):263. doi: 10.1186/1471-2458-8-263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.MacPhail C, Pettifor A, Moyo W, Rees H. Factors associated with HIV testing among sexually active South African youth aged 15-24 years. AIDS Care. 2009 Apr;21(4):456–67. doi: 10.1080/09540120802282586. [DOI] [PubMed] [Google Scholar]
  • 32.Peltzer K, Matseke G, Mzolo T, Majaja M. Determinants of knowledge of HIV status in South Africa: results from a population-based HIV survey. BMC Public Health. 2009;9(174):174. doi: 10.1186/1471-2458-9-174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Sherr L, Lopman B, Kakowa M, et al. Voluntary counselling and testing: uptake, impact on sexual behaviour, and HIV incidence in a rural Zimbabwean cohort. AIDS. 2007 Apr 23;21(7):851–60. doi: 10.1097/QAD.0b013e32805e8711. [DOI] [PubMed] [Google Scholar]
  • 34.Fylkesnes K, Siziya S. A randomized trial on acceptability of voluntary HIV counselling and testing. Trop Med Int Health. 2004 May;9(5):566–72. doi: 10.1111/j.1365-3156.2004.01231.x. [DOI] [PubMed] [Google Scholar]
  • 35.Matovu JK, Gray RH, Makumbi F, et al. Voluntary HIV counseling and testing acceptance, sexual risk behavior and HIV incidence in Rakai, Uganda. AIDS. 2005 Mar 25;19(5):503–11. doi: 10.1097/01.aids.0000162339.43310.33. [DOI] [PubMed] [Google Scholar]
  • 36.Wringe A, Isingo R, Urassa M, et al. Uptake of HIV voluntary counselling and testing services in rural Tanzania: implications for effective HIV prevention and equitable access to treatment. Trop Med Int Health. 2008 Mar;13(3):319–27. doi: 10.1111/j.1365-3156.2008.02005.x. [DOI] [PubMed] [Google Scholar]
  • 37.Neale J, Tompkins C, Sheard L. Barriers to accessing generic health and social care services: a qualitative study of injecting drug users. Health Soc Care Community. 2008 Mar;16(2):147–54. doi: 10.1111/j.1365-2524.2007.00739.x. [DOI] [PubMed] [Google Scholar]
  • 38.Kalichman SC, Simbayi LC, Kaufman M, Cain D, Jooste S. Alcohol use and sexual risks for HIV/AIDS in sub-Saharan Africa: systematic review of empirical findings. Prev Sci. 2007 Jun;8(2):141–51. doi: 10.1007/s11121-006-0061-2. [DOI] [PubMed] [Google Scholar]
  • 39.Woolf-King SE, Maisto SA. Alcohol use and high-risk sexual behavior in Sub-Saharan Africa: a narrative review. Arch Sex Behav. 2011 Feb;40(1):17–42. doi: 10.1007/s10508-009-9516-4. [DOI] [PubMed] [Google Scholar]
  • 40.Lewis JJ, Garnett GP, Mhlanga S, Nyamukapa CA, Donnelly CA, Gregson S. Beer halls as a focus for HIV prevention activities in rural Zimbabwe. Sex Transm Dis. 2005 Jun;32(6):364–9. doi: 10.1097/01.olq.0000154506.84492.61. [DOI] [PubMed] [Google Scholar]
  • 41.Fritz K, Woelk G, Bassett M, et al. The association between alcohol use, sexual risk behavior, and HIV infection among men attending beerhalls in Harare, Zimbabwe. AIDS and Behavior. 2002;6(3) [Google Scholar]
  • 42.Dunkle KL, Jewkes RK, Brown HC, Gray GE, McIntryre JA, Harlow SD. Transactional sex among women in Soweto, South Africa: prevalence, risk factors and association with HIV infection. Soc Sci Med. 2004 Oct;59(8):1581–92. doi: 10.1016/j.socscimed.2004.02.003. [DOI] [PubMed] [Google Scholar]
  • 43.Fisher JC, Bang H, Kapiga SH. The association between HIV infection and alcohol use: a systematic review and meta-analysis of African studies. Sex Transm Dis. 2007 Nov;34(11):856–63. doi: 10.1097/OLQ.0b013e318067b4fd. [DOI] [PubMed] [Google Scholar]
  • 44.Peltzer K, Mlambo G, Phaweni K. Factors determining prenatal HIV testing for prevention of mother to child transmission of HIV in Mpumalanga, South Africa. AIDS Behav. 2010 Oct;14(5):1115–23. doi: 10.1007/s10461-009-9662-7. [DOI] [PubMed] [Google Scholar]
  • 45.Luseno WK, Wechsberg WM. Correlates of HIV testing among South African women with high sexual and substance-use risk behaviours. AIDS Care. 2009 Feb;21(2):178–84. doi: 10.1080/09540120802017594. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Msuya SE, Mbizvo E, Uriyo J, Stray-Pedersen B, Sam NE, Hussain A. Predictors of failure to return for HIV test results among pregnant women in Moshi, Tanzania. J Acquir Immune Defic Syndr. 2006 Sep;43(1):85–90. doi: 10.1097/01.qai.0000225016.50890.7e. [DOI] [PubMed] [Google Scholar]
  • 47.Bellows B, Hamilton M. Vouchers for health: Increasing utilization of facility-based STI and safe motherhood services in Uganda Health Systems 20/20 project. Bethesda, MD: Abt Associates Inc; 2009. [Google Scholar]
  • 48.Bush K, Kivlahan DR, McDonell MB, Fihn SD, Bradley KA. The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test. Arch Intern Med. 1998 Sep 14;158(16):1789–95. doi: 10.1001/archinte.158.16.1789. [DOI] [PubMed] [Google Scholar]
  • 49.Papas RK, Sidle JE, Wamalwa ES, et al. Estimating alcohol content of traditional brew in Western Kenya using culturally relevant methods: the case for cost over volume. AIDS Behav. 2010 Aug;14(4):836–44. doi: 10.1007/s10461-008-9492-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Vyas S, Kumaranayake L. Constructing socio-economic status indices: how to use principal components analysis. Health Policy Plan. 2006 Nov;21(6):459–68. doi: 10.1093/heapol/czl029. [DOI] [PubMed] [Google Scholar]
  • 51.Filmer D, Pritchett LH. Estimating wealth effects without expenditure data--or tears: an application to educational enrollments in states of India. Demography. 2001 Feb;38(1):115–32. doi: 10.1353/dem.2001.0003. [DOI] [PubMed] [Google Scholar]
  • 52.Coates J, Swindale A, Bilinsky P. Household Food Insecurity Access Scale (HFIAS) for Measurement of Household Food Access: Indicator Guide. Washington D.C.: Food and Nutrition Technical Assistance Project, Academy for Educational Development; 2006. [Google Scholar]
  • 53.Deddens JA, Petersen MR. Approaches for estimating prevalence ratios. Occup Environ Med. 2008 Jul;65(7):481, 501–6. doi: 10.1136/oem.2007.034777. [DOI] [PubMed] [Google Scholar]
  • 54.Ostermann J, Kumar V, Pence BW, Whetten K. Trends in HIV testing and differences between planned and actual testing in the United States, 2000-2005. Arch Intern Med. 2007 Oct 22;167(19):2128–35. doi: 10.1001/archinte.167.19.2128. [DOI] [PubMed] [Google Scholar]
  • 55.Nabyonga J, Desmet M, Karamagi H, Kadama PY, Omaswa FG, Walker O. Abolition of cost-sharing is pro-poor: evidence from Uganda. Health Policy Plan. 2005 Mar;20(2):100–8. doi: 10.1093/heapol/czi012. [DOI] [PubMed] [Google Scholar]
  • 56.Barros AJ, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol. 2003 Oct 20;3(21):21. doi: 10.1186/1471-2288-3-21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009;338:b2393. doi: 10.1136/bmj.b2393. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Weiser SD, Tsai AC, Gupta R, et al. Food insecurity is associated with morbidity and patterns of healthcare utilization among HIV-infected individuals in rural Uganda. AIDS. 2011 Sep 6; doi: 10.1097/QAD.0b013e32834cad37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59. [cited 2011];The AIDS Support Organisation (TASO) Available from: http://www.tasouganda.org/
  • 60.Hahn JA, Woolf-King SE, Muyindike W. Adding fuel to the fire: alcohol's effect on the HIV epidemic in Sub-Saharan Africa. Curr HIV/AIDS Rep. 2011 Sep;8(3):172–80. doi: 10.1007/s11904-011-0088-2. [DOI] [PubMed] [Google Scholar]
  • 61.Fritz K, Morojele N, Kalichman S. Alcohol: the forgotten drug in HIV/AIDS. Lancet. 2010 Aug 7;376(9739):398–400. doi: 10.1016/S0140-6736(10)60884-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES