Background
Worldwide, HIV counseling and testing (HCT) is a key intervention for HIV prevention and a critical entry point into life-sustaining treatment and care programmes [
1‐
4]. As the scale-up to anti-retroviral treatment access continues to accelerate in Sub-Saharan Africa, innovative strategies to increase access to HCT services should be simultaneously implemented [
5]. In the past decade, a variety of HIV counseling and testing approaches have been used, including provider-initiated testing and counseling as part of medical care, and client-initiated or voluntary testing and counseling (VCT) [
6]. Empirical evidence has demonstrated the public health benefits of HIV testing such as reduction in risky sexual behaviour [
7,
8] and linking HIV infected individuals to HIV care, treatment and support [
9,
10]. In Africa, however, low access and limited reach of facility-based HIV testing services have been an impediments to global attempts to prevent HIV transmission and scale-up of HIV care and treatment at population level [
11].
Home-Based HIV Counseling and Testing (HBHCT) has the potential to address the challenges of limited access to testing. It involves the use of mostly community health workers to provide door-to-door counseling and testing services to consenting members of the community [
12]. HBHCT has been shown to increase uptake of HIV testing, improve access to testing, while simultaneously limiting the costs for travelling to health facilities and reducing the potential stigma [
8,
12,
13]. Uganda relies heavily on the conventional facility-based HIV testing models in line with the national policy recommendations [
14], but these approaches mostly reach patients visiting health facilities or individuals who seek voluntary HIV testing following known exposure to HIV infection [
15,
16].
Although HBHCT may be effective in increasing access of HIV testing, a recent systematic review suggested that there is limited evidence and experience to support widespread scale-up of implementation of this model in Sub-Saharan Africa [
11]. Few published studies have evaluated the feasibility, acceptance and uptake of HBHCT in rural [
8,
9,
17] and in mixed rural-urban populations in Uganda [
18]. Urban community residents represent a unique risk subgroup for HIV as evidenced by patterns of higher HIV sero-prevalence in urban compared to rural settings [
19,
20]. This pilot study was done to evaluate further the acceptance of HBHCT, to estimate the HIV sero-prevalence among individuals tested at home, and to determine the factors associated with acceptance of HBHCT in an urban setting. The information generated will add to the body of knowledge towards community-based interventions for HIV testing in urban settings.
Methods
Study Setting and Population
A cross-sectional house-to-house survey was nested in an ongoing community cough survey conducted in Rubaga division of Kampala city from January-June 2009. The division has a population of about 250,000 people, with 50% adults 15 years or older. It is composed of 13 administrative units called parishes which are further subdivided into 127 smaller units called local council villages or zones [
21]. The division is mainly served by two private not-for-profit hospitals, two public health care clinics and is also proximal to the national referral hospital. All these health facilities offer provider-initiated HIV testing as part of routine medical care routine. However, HIV testing services are also available for a minimal cost at other stand-alone voluntary HIV counseling and testing centers in Kampala. Almost all participants resided no more than approximately 5-10 kilometers from at least one of the health facilities.
Eligibility Criteria
Adults aged 15 years and older who were residents of Rubaga division were eligible to enrol in the study. We excluded persons who already knew their HIV status to be positive, those who were not able to communicate in English or Luganda, who did not give written consent to participate or who were absent from home at the time of survey. The study was approved by the Makerere University School of Public Health Research and Ethics committee and the Uganda National Council for Science and Technology.
Sampling
Rubaga division was conveniently selected for the study because of its proximal location (~10 Km) to the Mulago National Referral Hospital which offers free comprehensive HIV/AIDS testing and treatment services. A computer-based random number generator was used randomly select five villages for the study. The total numbers of subjects to enroll per village were estimated using the proportion-to-population size approach. The specific houses within the selected villages were also selected randomly using an approach that was previously employed in a similar setting and described in detail elsewhere [
22]. Eligible participants who were found at home were consecutively enrolled from the selected households at the time of the study visit. The five villages included in the study were similar in their population characteristics.
Study Outcomes
The main study outcome was acceptance of HIV test which was operationalized as consenting, actually taking the test and receiving test results offered in the participants' home. The secondary outcome was the HIV sero-prevalence defined as participants who had a positive rapid HIV test result out of those who had a test performed.
Data Collection
During the home visits a trained study team of six; three nurse counselors and three interviewers worked in pairs to obtain written informed consent, administered face-to-face interviews and, performed HIV testing. The information gathered included socio-demographics, occupation, weekly income, previous HIV test history, knowledge if current HIV status and, willingness to accept an HIV test. Quality assurance principles of counseling were observed before, during and after the HIV test in accordance with the national HV policy guidelines [
14]. Eligible members of the same household were offered the options of individual or group pre-test counseling. For household members who preferred group counseling, the study counselors provided the standard HIV information about testing in one seating after which they were requested to give written consent to be tested. Post-test counseling and delivery of results was offered according to the participants' preference.
Home-Based HIV Testing Procedures
Rapid HIV testing was performed according to the testing algorithm recommended by the Uganda national HIV testing policy [
14]. Determine HIV-1/2 assay (Abbott Laboratories, Illinois, United States of America) was used for screening, the HIV-1/2 STAT-PAK Dipstick assay (Chembio Diagnostic System Inc, New York, USA) was used as the confirmatory test and the Uni-Gold test (Trinity Biotech, Wicklow, Ireland) as the tie-breaker. The results were reported as positive if the Uni-gold test results were positive. The average test result turnaround was 15-20 minutes.
Referral for Care and Medical Evaluation
Participants who were identified to be HIV sero-positive during the survey were advised to seek further evaluation and referred to HIV care centres of their preference. Individuals who tested HIV negative received HIV prevention counseling. The interviewers also encouraged participants to disclose their HIV status to their sexual partners and/or family members as they felt comfortable.
Statistical Analysis
The main outcome was coded as a binary variable (HIV test acceptance, 'Yes' or 'No'). Acceptance was expressed as the proportion (with 95% confidence intervals) of participants who voluntarily accepted and took the HIV test among those who were offered a test. The HIV sero-prevalence was estimated among those who accepted to be tested from the HBHCT as the proportion who were HIV sero-positive. Differences in proportions among groups for the unordered categorical variables were tested with Pearson's chi-square; trends in proportions of ordered categorical variables were tested for using the Cochran-Armitage test. Multivariable logistic regression analysis was used to determine factors that were associated with acceptance of HBHCT. The best prediction model was selected based on a measure of model goodness of fit, the Akaike Information Criterion (AIC) [
23]. Factors included in the analysis were sex, age, religion, education, marital status and previous HIV testing. All chi-square P-values ≤ 0.05 were considered statistically significant. Data management and analysis was done using Statistical Analysis Software (SAS) and cross-validated using STATA (version 11.0; StataCorp, LLC, College Station, TX, USA).
Discussion
We found that home-based HIV counseling and testing was feasible and acceptable (69%) in this urban setting in Kampala. The factors associated with acceptance of HIV counseling and testing in the home were being 25 years or older, being male, previously married and previous HIV testing history. The HIV sero-prevalence in the study population was similar to the most recent national HIV sero-prevalence estimate of 6.4% [
24]. These findings were consistent with population-based studies done elsewhere that showed high acceptance for home-based HIV counseling and testing [
3,
25]. However, we did not find any published quantitative urban population-based studies for Uganda.
Our study findings support the use of a home-based HIV counseling and testing strategy in an urban setting. Urban population-based studies done in Zambia and Kenya reported similar acceptance levels; 71% and 78% of participants agreed to HIV testing and counseling in the home, respectively [
3,
25]. In contrast, rural population-based studies done in Africa consistently showed higher levels of acceptance for home-based HIV counseling and testing ranging from 84% to 98% [
12,
17,
26]. The urban-rural differences in acceptance for home-based HIV counseling and testing could be due to underlying disparities in availability and access to testing sites. Regardless of the contextual factors, the high demand for home-based HIV testing suggests that this strategy could potentially address known barriers associated with facility-based HIV testing including stigma, discrimination, fear to receive results, lack of confidentiality, privacy, long distances and cost of transportation to testing sites [
5,
9,
13,
16]. Based on the reasons given by those individuals for not accepting to take the home-based test in our study such as not being emotionally prepared and having to consult spouses or parents, it is evident that HIV programs would still need to be aware of existing barriers and seek to address them.
Like other studies [
25], we found that men were more likely to accept the test than women during this home-based survey. Differences in gender r roles may explain some of the results in this study. Women of reproductive age are more likely than men to come into frequent contact with the health system and thus to access HIV testing services. For example, pregnant women have an opportunity to learn their HIV sero-status through the routine testing during Antenatal Care and Prevention of Mother-to-Child HIV transmission programs [
19]. On the other hand, the higher acceptance of HBHCT by men presents an opportunity to reach entire families through community-based approaches. From a social-cultural perspective, targeting married men with home-based HIV counseling and testing could be beneficial for promoting couple counseling and testing [
10]. In Uganda, men often play a dominant role in the decision making process related to couple and family health issues.
Consistent with a similar study done in Malawi [
27], socio-demographic factors also influence acceptance of home-based HIV testing, older individuals were less likely to accept a home-based HIV test compared to younger individuals. A possible explanation is that older individuals are usually married or in stable relationships and may have a lower HIV risk perception. However, previously married (widowed or separated) individuals were more likely to accept the HIV test; perhaps motivated by a higher risk perception for HIV infection in this subgroup. In Uganda, being widowed often results from death of spouse from HIV/AIDS [
27] and has been shown to be a risk factor for HIV infection [
20]. Previous testing especially within 12 months prior to the HBHCT was negatively associated with acceptance compared with individuals never previously tested. The Uganda HIV testing policy recommends repeating an HIV testing within six months of a previous test [
14] but there is need to be a more practical strategy for its implementation.
The HIV sero-prevalence in the overall study population was 6.5% while the sero-prevalence among those who accepted to test was 7.4%. These findings are very similar to the national sero-prevalence of 6.4% and provides external validity to our study [
28]. A lower proportion of individuals self-reported as previously "never tested" (39%) compared to 55%-65% national estimates for 15-49 year old urban residents [
19]. Furthermore, a high proportion of previously "never tested" individuals were HIV sero-positive. Similarly, a study done in rural western Uganda was effective in identifying previously unknown HIV positive individuals as well as discordant couples among those who had "never tested" [
17]. Since Uganda's HIV epidemic has reached a stable and generalized state, it is critical to identify the hidden drivers for effective infection prevention and control. This study and others [
1,
9,
17] point to the potential value of home-based HIV counseling and testing as a way to reach untested individuals.
The WHO/UNAIDS policy guidelines recommends four key strategies for HIV counseling and testing that are mainly facility-based [
16]. However, it is apparent that complementary strategies such as home based HIV counseling and testing that expand access to HIV testing services in underserved or hard to reach communities are needed. Such strategies could enhance early identification of HIV positive individuals, prompt linkage to treatment and care, and may curtail risky sexual behaviours that lead to HIV transmission [
29]. The successful Implementation of HBHCT could be hampered by shortage of trained health workers and high operational costs in resource limited settings. In order to circumvent such challenges, delivery models that leverage community resources such as trained community lay providers who can perform rapid HIV testing under supervision by medical personnel, have been used elsewhere [
12] and are also endorsed by the Uganda HIV policy guidelines [
14]. Integration of HBHCT with other community screening programs such as those for tuberculosis and/or malaria may also increase cost-effectiveness [
13].
Study Limitations
Our study findings should be interpreted in light of some limitations. First, the sero-prevalence of the study could be an underestimate because we are missing information about the individuals who did not accept the HIV test. If we apply the estimated 6.4% sero-prevalence to our study sample then we would have expected a slightly higher HIV prevalence. We also excluded children younger than 15 years; therefore our findings can only be generalized to the adult population in similar urban settings. Second, there is selection bias as the study population comprises mostly women; this may lead to an overestimation of the HIV sero-prevalence because women are at a greater risk for HIV than men [
20]. The strength of this study is that it was conducted in the general urban population which gives a picture of the operational experiences and challenges that could be encountered when HBHCT is undertaken in similar settings. Previous community HIV testing has been mainly targeted towards high risk household members of already known HIV/AIDS infected or tuberculosis index cases [
10].
Acknowledgements
The authors would like to acknowledge the funding support of the Doris Duke Charitable Foundation. We thank the staff of Uganda-Case Western Reserve University Research Collaboration, particularly Yusuf Mulumba and LaShonda Malone for their skilful data management of the project, and Dr. Sarah Zalwango for her technical support to the Doris Duke community study. We acknowledge the study home health visitors and interviewers; Josylene Nabisere, Joyce Nalubowa, Joan Nassuna, Esther Nakayenga and Godfrey Kijjambu who worked very hard to enrol community participants. We would like to thank the Rubaga community local leaders for their administrative support and the residents who volunteered their time to participate in the study.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JNS: Designed the study, coordinated recruitment of participants, collected data, analyzed data and drafted the manuscript. CCW: Designed the study, participated in data analysis and reviewed all drafts of the manuscript. HS: Designed the study, participated in data collection, analysis, and reviewed all drafts of the manuscript. JL: Designed the study, coordinated recruitment of participants, collected data, analyzed data and reviewed drafts of the manuscript. AS: Participated in design, data analysis and reviewed all draft manuscript. MA: Participated in design, data collection, analysis, reviewed drafts of the manuscript. XY: Guided and help with data analysis, reviewed drafts of the manuscript. All authors read and approved the final manuscript.