Underutilization of HIV testing services remains a principal obstacle to effective implementation of HIV prevention and treatment programs.1-5 Early HIV diagnosis facilitates timely access to care and antiretroviral therapy that can subsequently improve mortality outcomes and may reduce HIV transmission rates.1-3,6,7 In sub-Saharan Africa, as many as 80% of HIV-infected adults were estimated to be unaware of their serostatus in 2007.1,8 Additionally, a 2007 estimate suggested that 90% of the adult population has never received an HIV test, and only 2.2% of adults received such testing and counseling annually.1,8,9
Traditionally, standard HIV counseling and testing has been offered in 2 types of settings. The first is known as client-initiated opt-in voluntary counseling and HIV testing. The second is diagnostic counseling and testing, initiated by health care providers for those clients who demonstrate signs or symptoms of HIV infection or have other identifiable risk factors for HIV infection. With increasing recognition of the limitations of these standard approaches, a third approach has evolved: routine opt-out provider-initiated testing and counseling (PITC).2,10-13 In recent years, opt-out antenatal HIV testing and counseling has been successfully integrated into mother-to-child transmission prevention programs.11,14,15 Beginning May 2007, the World Health Organization (WHO) and the Joint United Nations Programme on HIV/AIDS officially recommended that PITC be offered to “all adults and adolescents seen in all health facilities” in areas with an HIV prevalence greater than 1% and where appropriate prevention, treatment, and care resources are available.1
Despite a call for its expansion, PITC remains underutilized, in part due to increased requirements for test kits, laboratory equipment, and human resources.16 An increase in total HIV cases detected in the general patient population using PITC would support its widespread adoption. However, because standard counseling and testing programs serve volunteering clients or those with identifiable risk factors, they are expected to detect a higher number of HIV cases per unit cost when compared with routine testing programs.17 Therefore, these additional costs associated with higher testing volume may be further justified if PITC programs are able to detect a proportionally equal or higher percentage of infected individuals than the standard approach. To our knowledge, no study to date has compared PITC HIV detection rates against standard non-PITC programs in a general patient population in rural sub-Saharan Africa. We conducted a randomized clinical trial among adults attending a mobile medical clinic in rural setting in southern Zambia to compare HIV detection rates by testing strategy and to assess attitudes toward PITC.
Between July 14, 2008, and July 22, 2008, we provided HIV counseling and testing services to 1221 adults aged 15-49 years who presented to a mobile clinic at any 1 of 4 rural sites in southern Zambia. The clinic was designed to provide acute medical, dental, optical, and wound care. At each site, patients with medical chief complaints were seen and treated on a first-come, first-serve basis by 1 of 22 health service providers (10 physicians, 9 registered nurses, 2 clinical officers, and 1 nurse practitioner) who were instructed at study outset to refer patients for on-site HIV counseling and testing by 1 of 2 different methods. The number of providers designated to each referral method was based on the estimates of clinic laboratory and counseling capacity. Sixteen providers (4 physicians, 9 registered nurses, 2 clinical officers, and 1 nurse practitioner) were trained to refer those patients who requested HIV testing or who demonstrated signs or symptoms of HIV/AIDS according to the WHO staging criteria or exhibited other identifiable risk factors for HIV infection (standard non-PITC). The remaining 6 providers were randomly selected among the physicians to refer every patient for HIV counseling and testing, regardless of presenting condition or complaint (PITC). PITC providers were further asked to document any patient initiation or request for testing, as well as the presence of signs and symptoms that would have prompted referral under traditional circumstances, in the same fashion as the standard non-PITC group. All patients were repeatedly informed of the availability of on-site counseling and testing services in both English (the national language) and the local language, Tonga. Patients could request counseling and testing at any time during their visit (ie, in registration, triage, examination room, or the pharmacy).
Patients, therefore, presented for HIV testing for 3 reasons: (1) client request (standard counseling and testing), (2) provider referral based on clinical presentation (also a standard form of counseling and testing), and (3) routine opt-out referral (PITC) (Fig. 1). 249 patients were referred through multiple pathways. Although patients presented to mobile clinics for different types of care (medical, dental, optical, wound), only patients seeking medical care were included, to improve the generalizability of our findings to a general adult patient population presenting for primary care in rural Zambia. The primary analysis was via intent-to-treat analysis with the PITC group including all patients referred by the routine opt-out approach and the standard non-PITC group including all remaining patients referred either by client initiation or by high-risk clinical presentation. Hereafter, we respectively refer to these 2 groups as the PITC and standard non-PITC groups.
After receiving medical care, patients with referral for HIV testing and counseling by either standard non-PITC or PITC were directed to a separate area designated for all clinic testing. Trained Zambian nurses fluent in English and Tonga provided pretest and posttest counseling according to the WHO guidelines. Patients were tested for HIV infection using Determine HIV-1/2 (Abbott Laboratories, Abbott Park, IL), with confirmation of positives using Unigold HIV (Trinity Biotech, County Wicklow, Ireland), in accordance with the Ministry of Health guidelines. Guidelines have been validated locally via the dual rapid test algorithm and a quality control procedure described elsewhere.18 Test results were linked confidentially to patient referral cards and stored securely by laboratory personnel. All patients testing positive for HIV received explicit referral to the nearest antiretroviral treatment site.
A separate group of clinic assistants, masked to patient HIV status, administered a brief sociodemographic questionnaire before pretest counseling to a convenience sample of PITC and standard non-PITC patients. The questionnaire was developed from previously tested survey instruments19,20 and was translated to Tonga and back translated to English by independent professional translators. Bilingual interviewers used a standardized protocol to conduct the survey in either language as per patient preference. The structured interview collected demographic information on age, sex, occupation, education, income, relationship status, access to health care services, and prior HIV test history. Knowledge about and stigmatizing attitudes toward HIV testing were assessed as score-based variables using validated 10-question and 4-question survey tools, respectively,19,20 with scores treated as binary for analysis. Similarly, the patient's personal and perceived partner attitude toward routine opt-out HIV testing were obtained using questions derived from previously tested instruments.21
The study evaluated 3 primary outcomes stratified by testing referral method: (1) HIV serostatus, (2) history of prior HIV test, and (3) attitude toward PITC. Univariate comparisons of patient characteristics by referral group, HIV status, and patient favorability toward PITC were conducted using rank sum and χ2 tests. We employed multivariable logistic regression models to assess the association between referral methods and HIV status. We adjusted for age, sex, and clinic site as potential predictors of HIV status. Age was a continuous predictor that was nonlinear with the log odds of HIV-positive status. To relax our linearity assumption, we modeled age using a restricted cubic spline with 3 knots. The objective of this study was to assess equivalence of HIV prevalence for 2 referral types, standard non-PITC and PITC.
Additionally, we sought to identify patient characteristics predictive of a history of HIV testing. We used multivariable logistic regression modeling, with age, sex, stigma toward HIV, knowledge of HIV, and education serving as predictors of testing history. Predictors were selected a priori. R software 2.9.2 (R Foundation for Statistical Computing, Vienna, Austria, available at: http://www.r-project.org, accessed February 22, 2010) was used for all data analyses.
Both the Vanderbilt University Institutional Review Board and the Namwianga Mission, an institutional member of Churches Health Association of Zambia, reviewed the study protocol and provided ethical approval. Depending on literacy level, written or verbal consent was obtained from all participants.
A total of 1221 adults aged 15-49 years who presented to our mobile clinic with medical complaints or requesting screening services received an HIV test. We selected 317 adults (26%) to complete study questionnaires based on interviewer and patient availability.
The PITC group consisted of 681 patients (56%). Of these, one third (n = 222; 33%) would have been referred for testing in the absence of a PITC approach, including 142 (64%) via client request, 63 (28%) via the provider's clinical suspicion of HIV infection, and 17 (8%) via both.
The standard non-PITC group consisted of 540 patients (44%), including 454 (84%) referred for testing by client request, 59 (11%) referred strictly by provider clinical suspicion of HIV infection, and 27 (5%) referred for both (Fig. 1). Altogether, PITC tested 114 adults per practitioner, whereas standard non-PITC tested 34 adults per practitioner. PITC and standard non-PITC groups were similar in all baseline characteristics except for sex (proportionately more women in PITC vs. standard non-PITC groups; P < 0.01) and clinic location (P < 0.01) (Table 1).
Most participants were women (72%) and married (66%). The mean age of the study population was 30 years (interquartile range: 22-38). Among married men, 28% reported 2 or more wives. Among married women, 30% reported that their husbands had multiple wives. Only 59% of patients had a full primary education, and 76% reported at least 1 health clinic visit in the past 12 months.
Overall, 53% of patients reported a prior HIV test. Among those previously tested, 37% had been tested in the past 12 months, primarily through regional sites providing standard non-PITC services. When adjusted for other variables, being a woman [odds ratio (OR): 2.4; 95% confidence interval (CI): 1.1 to 4.0; P < 0.001] and possessing fewer stigmatizing attitudes toward people living with HIV/AIDS (OR: 1.4; 95% CI: 1.1 to 1.9; P = 0.02) were predictive of prior HIV testing (data not shown). Older age was associated with prior HIV testing, but when adjusted for sex, knowledge, stigma, and education, was not statistically significant (P = 0.07). Prior HIV testing was not significantly associated with clinic location (P = 0.65) or additional health clinic visits in the past 12 months (P = 0.10) (data not shown). Because marital status had a significant association with HIV testing, we ran a second multivariable model of HIV testing history adjusting for marital status that revealed no effect modification.
Of the 1221 adults who underwent HIV testing, 130 were found to be seropositive, resulting in an overall prevalence of 10.6% (95% CI: 2.9% to 12.4%). HIV-positive individuals tended to be older than uninfected individuals (P < 0.01) and had a greater number of sexual partners in the past 12 months (P = 0.05) (Table 2). HIV status was not significantly associated with either patient sex (P = 0.34) or prior HIV testing (P = 0.86), although adults previously tested for HIV had a lower HIV prevalence (10.1%) than those never tested (11.3%). HIV prevalence in Kanchindu (6.4%) was significantly lower than that in the other 3 sites (P < 0.01).
HIV prevalence did not differ by testing referral type (Table 2). HIV prevalence in the PITC group was 11.2% (95% CI: 8.8% to 13.5%), whereas prevalence in the standard non-PITC group was 10.0% (95% CI: 7.5% to 12.5%). Among all patients who requested HIV testing, prevalence was 10.4%. In all patients whose clinical presentation prompted referral for HIV testing, prevalence was 16.3%. In the subgroup of PITC patients without additional indication for testing (ie, no client request or clinical presentation implicating HIV), prevalence was 9.6%. There was no significant difference in HIV prevalence rates according to testing strategy (P = 0.08). PITC detected 76 cases of HIV by 6 providers, whereas standard non-PITC detected 54 cases by 16 providers. Thus, PITC detected more than 3 times more HIV infections per practitioner than standard non-PITC (12.7 vs. 3.4).
Multivariable analysis of HIV status adjusting for selected patient characteristics did not alter our results (Table 3). The PITC group was not associated with a lower prevalence of HIV than the standard non-PITC group (OR: 1.01; 95% CI: 0.67 to 1.52; P = 0.95). Moreover, when compared with all study patients who either requested testing and/or had a clinical presentation suggesting HIV, the subgroup of PITC patients without additional indications for HIV testing did not have a statistically different HIV prevalence (OR: 0.83; 95% CI: 0.55 to 1.24; P = 0.36).
Of the patients who completed questionnaires, most (73%) reported prior knowledge of the concept of routine opt-out HIV testing and 96% reported being “in favor” of PITC. Previous HIV testing was associated with knowledge of PITC (P = 0.02) (data not shown).
Patients Declining HIV Testing
Due to clinic limitations in monitoring patient flow, it was logistically not possible to track adults who either did not consent for HIV testing or defaulted before receiving HIV testing. However, among those not receiving HIV testing, 89 completed questionnaires. This sample was compared with the 317 adults with questionnaires who were tested. They were found to be similar in all baseline demographic characteristics and attitudes toward PITC. Information regarding reasons for declining an HIV test was not formally assessed. Anecdotally, the most common reasons cited for declining an HIV test were known HIV positivity, recent prior HIV testing, and religious/cultural conflicts.
To our knowledge, this is the first randomized clinical trial in a rural African setting to compare HIV prevalence rates based on screening protocol-PITC opt-out testing vs. standard non-PITC testing per client request or clinical referral. We found HIV prevalence to be comparable between the 2 groups. Moreover, the subset of asymptomatic patients referred by PITC who did not request testing had a similar rate of HIV infection as those who had. This suggests that, at least in rural Zambia, HIV prevalence among patients from the general patient population may equal that in traditionally defined volunteering or high-risk groups who have served as targets of standard non-PITC programs. Opt-out testing increased the number of patients tested per practitioner by 3 times compared with standard non-PITC and also increased the number of HIV-infected persons detected per practitioner more than 3-fold. Of the 459 PITC patients who lacked other reasons for testing, 44 (9.6%) were found to be HIV positive and referred for ART who would not have been screened using a standard non-PITC approach.
The overall HIV prevalence in 2008 of 10.6% in our study population is slightly higher than estimates for the surrounding general rural population in 2008 (12% nationally, 5-9.9% regionally).22,23 This likely reflects the fact that our group of clinic patients may be sicker than the general public as many of them came to the mobile clinic services because they had a chief complaint. Lower HIV prevalence and higher volume of standard non-PITC testing relative to PITC in Kanchindu may in part reflect an increased number of low-risk clients requesting testing as a result of the health-promoting activities of 2 mobile voluntary counseling and HIV testing clinics in Kanchindu shortly before our study. Although, across sub-Saharan Africa, the percentage of adults who have ever been tested for HIV may be as low as 10%,1 we found a higher rate, with 53% of adults reporting having received HIV testing and counseling in the past. This may reflect earlier efforts by nongovernmental organizations and the Zambian Ministry of Health to extend testing coverage in this region. Nevertheless, given the burden of disease in this region, this proportion remains lower than the WHO recommendations, especially considering that most patients (63%) were not tested within the past year.
Whereas previous studies of opt-out HIV testing and counseling have focused on specific populations-infant,24 pediatric,25 antenatal,14,15,26-28 sexually transmitted infection,29 or patients with tuberculosis30,31-our study addresses the need for PITC evaluation in a general adult patient population.32 Additionally, this report is one of the first cross-sectional comparisons of PITC with a standard non-PITC approach, and our finding of approximately equal HIV detection rates between testing modalities parallels results of the few other direct comparisons published.29,30 Our study therefore avoids the confounding effects inherent to retrospective comparison analyses, including evolving population attitudes and disease epidemiology. Finally, by implementing both PITC and standard non-PITC approaches at the same sites, we reduce the confounding effect of intersite variability.32
Although anecdotally we observed low levels of test refusal, one limitation of this study was our inability to quantify the exact number of adults opting out. We suspect a primary driver of test refusal was the long wait time (20-40 minutes) for pre- and posttest counseling; however, because wait times were experienced uniformly by all patients, we do not suspect PITC and standard non-PITC dropout populations to differ greatly. Nevertheless, lack of data on the extent of or reasons for opting out indicate potential for sampling bias between referral groups. Additionally, long wait times resulted in 16 (1.3%) individuals defaulting before receiving test results; though, because HIV test results had already been gathered for these individuals, this did not affect data collection, analysis, or interpretation. Future high-volume PITC programs should incorporate methods to reduce wait time and premature patient departure. Additionally, because randomization of patients was not possible due to clinic logistical constraints, we instead randomized providers in an unblinded fashion. Although this technique may have introduced additional bias between referral groups, we do not suspect significant effects on study results owing to low interprovider variability in provider referral practices. Finally, an additional limitation of this study was our ability to detect a 50% difference in HIV detection rates with adequate (>80%) power; though, smaller undetected differences in detection rates may carry less clinical significance and even be permissible due to PITC's corresponding ability to significantly augment total testing volume.
Our findings may have substantial implications for HIV prevention policies. Although the WHO recommends PITC for all areas with generalized epidemics,1 standard of care testing and counseling programs in most parts of sub-Saharan Africa still selectively target individuals who either request testing or possess clinical indications for it.2 Previous arguments in support of PITC have focused on its ability to detect greater absolute numbers of HIV-positive patients.14,15,25,27,28,30,33 However, given the substantial increase in resources that would be required by PITC at every standing health facility, the persistence of standard approaches may reflect an implicit assumption that patients who request testing or exhibit risk factors for HIV infection have a higher likelihood of infection than the general adult population and therefore deserve the focus of limited screening resources.17 Although further confirmatory studies are needed, our study may offer initial evidence against this hypothesis for high-prevalence rural areas of southern Africa. Adults targeted by current standard testing modalities may not have a higher likelihood of HIV infection than the general adult patient population in the same area. This is particularly concerning because asymptomatic HIV-infected individuals without clear risk factors may be missed by standard screening programs.
PITC has been shown repeatedly to be highly acceptable in sub-Saharan Africa.14,15,21,24,26-28,34 The high percentage (76%) of adults in our study reporting at least 1 clinic visit during the past 12 months suggests the potential for national clinic-based PITC to increase HIV testing coverage.35 However, the implementation of PITC in our study required substantial human and material resources and resulted in increased referrals for life-long ART. A national PITC program would consequently incur substantial additional costs at a time when targeted President's Emergency Plan for AIDS Relief treatment funds seem to be leveling off.36 Given the potential that routine opt-out testing may identify proportionally just as many HIV-infected persons as current standard non-PITC programs, a formal cost-effectiveness comparison between the 2 approaches is needed. Although previous cost-effectiveness analyses of HIV counseling and testing have focused on net infections averted and resultant costs saved,37,38 future models should account for PITC's potential to increase population knowledge about HIV/AIDS and thereby reduce stigma as well.32 We now believe that PITC should be compared further to home-based VCT as to coverage and cost-effectiveness.39 We also acknowledge that population incidence could decrease if enough persons got tested, linked to care, and placed on ART to reduce community viral load.40-43 Finally, although expansion of PITC to all established health facilities remains an urgent need in nations like Zambia, we suggest that mobile-based PITC, which may provide equivalent testing coverage while consolidating costs by limiting services to one site at any given time, when offered in conjunction with other medical services may be a relatively inexpensive and highly effective method for expanding HIV screening to remote regions.
The authors thank the dedicated staff of Zambia Medical Mission and Namwianga Mission in Kalomo, Zambia and all of the study volunteers. They also thank the Zambian Ministry of Health and Centre for Infectious Disease Research in Zambia for their assistance in supplying HIV test materials and identifying local ART referral sites.
2. Matovu JKB, Makumbi FE. Expanding access to voluntary HIV counselling and testing in sub-Saharan Africa: alternative approaches for improving uptake, 2001-2007. Trop Med Int Health
3. Bassett IV, Walensky RP. Integrating HIV screening
into routine health care in resource-limited settings. Clin Infect Dis
. 2010;50(Suppl 3):S77-S84.
6. The Voluntary HIV-1 Counseling and Testing Efficacy Study Group. Efficacy of voluntary HIV-1 counselling and testing in individuals and couples in Kenya, Tanzania, and Trinidad: a randomised trial. Lancet
7. Granich RM, Gilks CF, Dye C, et al. Universal voluntary HIV testing
with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet
11. US Centers for Disease Control and Prevention. Introduction of routine HIV testing
in prenatal care-Botswana, 2004. MMWR Morb Mortal Wkly Rep
12. Branson B. Current HIV epidemiology and revised recommendations for HIV testing
in health-care settings. J Med Virol
. 2007;79(Suppl 1):S6-S10.
13. Manzi M, Zachariah R, Teck R, et al. High acceptability of voluntary counselling and HIV-testing but unacceptable loss to follow up in a prevention of mother-to-child HIV transmission programme in rural
Malawi: scaling-up requires a different way of acting. Trop Med Int Health
14. Moses A, Zimba C, Kamanga E, et al. Prevention of mother-to-child transmission: program changes and the effect on uptake of the HIVNET 012 regimen in Malawi. AIDS
15. Chandisarewa W, Stranix-Chibanda L, Chirapa E, et al. Routine offer of antenatal HIV testing
(“opt-out” approach) to prevent mother-to-child transmission of HIV in urban Zimbabwe. Bull World Health Organ
16. Monjok E, Smesny A, Mgbere O, et al. Routine HIV testing
in health care settings: the deterrent factors to maximal implementation in sub-Saharan Africa. J Int Assoc Physicians AIDS Care
17. Bartlett JG, Branson BM, Fenton K, et al. Opt-out testing for human immunodeficiency virus in the United States: progress and challenges. JAMA
18. Wright RJ, Stringer JSA. Rapid testing strategies for HIV-1 serodiagnosis in high-prevalence African settings. Am J Prev Med
19. Joint United Nations Programme on HIV/AIDS (UNAIDS). Stigma
and discrimination indicator 1. In: Joint United Nations Programme on HIV/AIDS (UNAIDS). National AIDS Programmes: A Guide to Monitoring and Evaluation. Geneva, Switzerland: UNAIDS; 2000. Available at: http://www.indicatorregistry.org/upload/UNAIDS
.%20National%20AIDS%20Programmes.%20A%20guide%20to%20monitoring%20and%20.pdf. Accessed December 21, 2009.
20. Central Statistical Office (Zambia
), Central Board of Health (Zambia
), ORC Macro. Zambia
Demographic and Health Survey 2001-2002. Calverton, MD: Central Statistical Office, Central Board of Health, and ORC; 2003.
21. Weiser SD, Heisler M, Leiter K, et al. Routine HIV testing
in Botswana: a population-based study on attitudes, practices, and human rights concerns. PLoS Med
24. Rollins N, Mzolo S, Moodley T, et al. Universal HIV testing
of infants at immunization clinics: an acceptable and feasible approach for early infant diagnosis in high HIV prevalence settings. AIDS
25. Kankasa C, Carter RJ, Briggs N, et al. Routine offering of HIV testing
to hospitalized pediatric patients at university teaching hospital, Lusaka, Zambia
: acceptability and feasibility. J Acquir Immune Defic Syndr
26. Perez F, Zvandaziva C, Engelsmann B, et al. Acceptability of routine HIV testing
(“opt-out”) in antenatal services in two rural
districts of Zimbabwe. J Acquir Immune Defic Syndr
27. Homsy J, Kalamya JN, Obonyo J, et al. Routine intrapartum HIV counseling and testing
for prevention of mother-to-child transmission of HIV in a rural
Ugandan hospital. J Acquir Immune Defic Syndr
28. Creek TL, Ntumy R, Seipone K, et al. Successful introduction of routine opt-out HIV testing
in antenatal care in Botswana. J Acquir Immune Defic Syndr
29. Leon N, Naidoo P, Mathews C, et al. The impact of provider-initiated (opt-out) HIV testing
and counseling of patients with sexually transmitted infection in Cape Town, South Africa: a controlled trial. Implement Sci
30. Pope DS, Deluca AN, Kali P, et al. A cluster-randomized trial of provider-initiated (opt-out) HIV counseling and testing
of tuberculosis patients in South Africa. J Acquir Immune Defic Syndr
31. US Centers for Disease Control and Prevention. Provider-initiated HIV testing
and counseling of TB patients-Livingstone District, Zambia
, September 2004-December 2006. MMWR Morb Mortal Wkly Rep
33. Bassett IV, Giddy J, Nkera J, et al. Routine voluntary HIV testing
in Durban, South Africa: the experience from an outpatient department. J Acquir Immune Defic Syndr
34. Wanyenze RK, Nawavvu C, Namale AS, et al. Acceptability of routine HIV counselling and testing, and HIV seroprevalence in Ugandan hospitals. Bull World Health Organ
35. Ivers LC, Freedberg KA, Mukherjee JS. Provider-initiated HIV testing
Haiti: low rate of missed opportunities for diagnosis of HIV in a primary care clinic. AIDS Res Ther
36. Cohen J. HIV/AIDS: Botswana's success comes at steep cost. Science
37. Creese A, Floyd K, Alban A, et al. Cost-effectiveness of HIV/AIDS interventions in Africa: a systematic review of the evidence. Lancet
38. Stover J, Bertozzi S, Gutierrez J, et al. The global impact of scaling up HIV/AIDS prevention programs in low- and middle-income countries. Science
39. Bateganya M, Abdulwadud OA, Kiene SM. Home-based HIV voluntary counselling and testing (VCT) for improving uptake of HIV testing
. Cochrane Database Syst Rev
. 2010 Jul 7;(7):CD006493.
40. Burns DN, Dieffenbach CW, Vermund SH. Rethinking prevention of HIV type 1 infection. Clin Infect Dis
41. DeGruttola V, Smith DM, Little SJ, et al. Developing and evaluating comprehensive HIV infection control strategies: issues and challenges. Clin Infect Dis
. 2010;50(Suppl 3):S102-S107.
42. Kamali A. What is new in HIV/AIDS research in developing countries? Trop Med Int Health
43. Kurth AE, Celum C, Baeten JM, et al. Combination HIV Prevention: Significance, Challenges, and Opportunities. Curr HIV/AIDS Rep
. 2010 Oct 13.