Mwandagalirwa, Kashamuka MHS*; Jackson, Elizabeth F. MHS, PHD†‡; McClamroch, Kristi PHD§; Bollinger, Robert MD, MPH∥; Ryder, Robert W. MD, MSC¶; Weir, Sharon S. PHD#
From the *KSPH/UNC-DRC Program, Kinshasa School of Public Health, Kinshasa University, †Department of Epidemiology, School of Public Health, University of North Carolina, Chapel Hill, North Carolina; ‡Health Services Research, Family Health International, Research Triangle Park, North Carolina; §Department of Epidemiology and Biostatistics, School of Public Health, University of Albany, Albany, New York; ∥Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland; ¶Department of Medicine, School of Medicine, University of California, San Diego, California; and #Department of Epidemiology, School of Public Health, University of North Carolina, Chapel Hill, North Carolina.
The authors thank the Fogarty teams at UNC and Johns Hopkins University for help in administrative and financial matters. The authors thank field workers and participants without whom the study would not have been a success. The authors thank Dr. Steven Callens for comments on the study design and proposal. Finally, the authors thank all administrative officials, the representative and police chiefs of the Kinshasa study communes, the PNLS staff, provincial medical authorities, and the Ministry of Health for their efforts to facilitate this study.
Supported by a grant from the Fogarty International Center, US National Institutes of Health, Program of International Training Grants in Epidemiology related to AIDS, D43 TW00010-AITRP and also by USAID (to Jackson and Weir).
Correspondence: Sharon S. Weir, PhD, Carolina Population Center, 123 West Franklin St, Suite 304, Chapel Hill, NC 27516. E-mail: Sharon_Weir@unc.edu.
Received for publication November 22, 2007, and accepted February 6, 2009.
HIV prevention programs in Kinshasa need to be based on current dynamics of the HIV epidemic in the local setting. Estimates of HIV prevalence among pregnant women in Kinshasa have remained between 2% and 5% for the last 2 decades.1–8 Estimated HIV prevalence among female sex workers and sexually transmitted infection (STI) patients, although much higher than among pregnant women, decreased from 29% to 23% among sex workers and from 12% to 9% among STI patients between 1997 and 2002.1,2 Prevalence of other STIs has been relatively stable over the past 2 decades.9 The apparent stabilization of HIV prevalence among antenatal clinic (ANC) patients and decline among sex workers and STI patients is somewhat unexpected given the political instability and conflict experienced in Kinshasa since 1991. One explanation is that stable HIV prevalence reflects high mortality among those infected with HIV coupled with high HIV incidence.10 In this resource constrained setting beset with conflict, trends reflected in sequential cross-sectional surveys must be interpreted with caution. Ongoing cohort studies are not available, and there is insufficient data to assess whether there have been important changes in the population participating in surveillance activities. Not only differential mortality, but also differences in other underlying demographic factors and survey participation could provide misleading estimates of HIV prevalence trends.
To contribute to the investigation of HIV prevalence among different subpopulations in Kinshasa, we implemented a study using the Priorities for Local AIDS Control Efforts (PLACE) methodology in Eastern Kinshasa aimed at finding undetected pockets of HIV infection. PLACE is a venue-based epidemiologic approach that interviews community informants to identify locations where people meet new sexual partners. New partnerships increase the probability of sexual contact between susceptible and infected persons and such contacts are crucial to the spread of the HIV epidemic. Information on the characteristics and behaviors of persons with high rates of new sexual partnership formation who socialize at these places can be used to inform prevention programs and identify groups at risk of acquiring and transmitting HIV.11,12 The PLACE method makes it possible to target prevention activities to persons with high rates of new partnership formation who do not self-identify as commercial sex workers.
This article presents results from the 2003 Kinshasa PLACE study. In addition, to determine whether this venue-based methodology identifies high-risk populations not identified using more traditional methodologies, demographic and behavioral characteristics and levels of HIV in this group were compared with those in clients at a major antenatal care clinic and several STI clinics in the same Eastern Kinshasa neighborhood.
MATERIALS AND METHODS
In March 2002, a team of researchers from the Congolese National AIDS Program (PNLS) and the University of North Carolina School of Public Health carried out a PLACE study in 3 contiguous, poor, densely populated sections of Kinshasa. The study team was interested in assessing whether these areas contained sexual networks with high rates of new sexual partnership formation. A total of 1350 community informants were asked to identify places where people go to meet new sexual partners and identified 916 unique sites that the study team was able to visit and map. An owner, manager, or other knowledgeable person at each venue was interviewed about the characteristics of each site such as the number of men and women who socialize there, site patron behaviors such as drinking and finding new sex partners, and on site HIV prevention activities such as the sale or distribution of condoms.
Mapping revealed 9 geographic clusters of sites. One of the largest clusters (133 sites) was selected for further investigation pending Institutional Review Board approval. Men and women socializing at 59 sites in this cluster were interviewed about their characteristics and behavior and tested for HIV between October 2003 and January 2004. Interviews were conducted at 59 venues that were either most frequently identified by community informants as places where people meet new sexual partners or located along a major road in the area. When venues identified in 2002 were no longer in business, a similar site nearby was chosen as a replacement.
HIV prevalence at PLACE sites was compared with prevalence among patients at the largest ANC clinic in the area and at 7 STI clinics within the selected geographic cluster. Patient volume at the ANC at the time of the study was estimated to be 800 persons per month, and at the 7 STI clinics, patient volume ranged between an average of 11 and an average of 104 patients per months at the various sites. Patients at the STI clinics were overwhelmingly female and poor. Typically, male STI patients come to Kinshasa STI clinics as a last resort after having unsuccessfully tried to treat their STI symptoms through over the counter medications available at pharmacies (K. Mwandagalirwa, 2008; personal communication). Treatment at STI clinics in the study is through syndromic management and data on STI prevalence is not available. HIV care was not provided at the 7 STI clinics at the time of the study.
After authorities at each PLACE site or ANC and STI clinic voluntarily agreed to host the study, private areas in which to conduct interviews and HIV testing were identified. At each STI clinic, 1 or 2 nurses were trained as interviewers and conducted informed consent, pretest counseling for HIV, interviews, and phlebotomies with each respondent individually. At the ANC and PLACE venues, patients and venue patrons were informed about the study and received counseling about voluntary HIV testing as a group. Subsequently, at the ANC, 3 study interviewers conducted the informed consent process with individual patients. If informed consent was obtained, participants were interviewed and directed to a study nurse who performed a phlebotomy. ANC and STI patients were encouraged to return to each clinic to receive their HIV test results. Approximately 1140 ANC women were informed about the study during the group HIV pretest counseling session. Of these, 24 women choose not to participate and 1116 women were interviewed and tested for HIV, representing a consent rate of approximately 98%. A valid HIV test result was available for 1106 of these women. At the STI clinic, approximately 69 men and 236 women were invited to participate in the study. Of these, 66 men and 229 women consented to participate in both the interview and blood draw, representing a response rate of 96% for men and 97% for women. Data collection at ANC and STI clinics occurred in October and November of 2003.
At PLACE sites, after all patrons received pretest counseling and study information, interviewers approached patrons individually to explain the informed consent process and ask each person if he or she would be willing to be interviewed and tested for HIV. When verbal consent was obtained, respondents proceeded to a private area and written informed consent was obtained. Interviewers administered the questionnaire and a study nurse performed a phlebotomy for HIV testing. Participants received money for transportation to obtain their HIV test results. PLACE site participants received test results at their choice of 4 health centers providing voluntary counseling and testing.
Of the approximately 1229 men and women introduced to the study at PLACE venues, approximately 98% (1199) agreed to participate in the interview and blood draw. At the first 6 sites, approximately 30 of 67 persons in the HIV pretest counseling sessions declined to participate (a response rate of 45%). The exact number and gender of those who declined is not known due to problems with record keeping which were subsequently resolved. This initial low response rate was thought to be due to a lack of appropriate communication about the study with venue staff. After an improved explanation of the purpose of the study to participants and venue owners, participation rose to nearly 100%.
A respondent’s sexual partnership level was defined as follows: “high” if he or she reported having a new sexual partner or more than one sexual partner in the past 4 weeks; “moderate” if he or she did not have new or multiple partners in the past 4 weeks but did in the past 12 months; and “low” if he or she had one long-term partner in the past 12 months or no partner. Women were identified as commercial sex workers by self-report. Transactional sex was defined as having received or given money or a gift in exchange for sexual intercourse. We were interested in the extent to which persons with high partnership rates did not report engaging in commercial or transactional sex.1,2
Blood specimens were tested for HIV antibodies using the Vironostika HIV1/2–O Elisa. Enzygnost EIA HIV1–2 and the Determine HIV1/2 rapid test were used to confirm positive results. Specimens with a negative confirmatory test were retested for HIV RNA and HIV p24 core antigen to identify newly infected samples not yet antibody positive.
Pools of 100 HIV antibody-negative blood specimens were tested for HIV RNA at the PNLS laboratory, using reverse transcriptase-polymerase chain reaction.13 Blood was collected in EDTA tubes and the Amplicor 1.5 test from Roche was used to extract RNA and amplify it for quantification. The lower level of detection for this sensitive method was 400 copies/mL. Pools of 100 specimens that tested positive were to be retested in smaller and smaller pools until the positive specimen was identified. Control tests confirmed that the PCR technique at the PNLS lab was well performed. The INNOTEST HIV Antigen was the enzyme immunoassay used for the detection and quantification of p24 core antigen of the HIV-1, group O, and HIV-2.14,15
Data were entered in Epi Info Version 6 and analyzed using STATA 9.0. The sociodemographic and behavioral characteristics of men and women interviewed and tested at PLACE sites, men and women interviewed and tested at STI clinics, and women interviewed and tested at the ANC are described and compared using the Pearson χ2 test. Whenever 5% or more of the respondents in any group were missing data for a particular variable, the percentage of respondents missing data for that variable was reported for every group. Prevalence of HIV by sociodemographic and behavioral characteristics was examined for each group. The ability of the PLACE methodology to identify HIV-positive individuals relative to STI and ANC clinics was evaluated by odds ratios comparing the odds of HIV infection among PLACE women with the odds of HIV infection among women at the ANC and STI clinics. We used multiple logistic regression analysis with interaction terms for site type and characteristic to calculate odds ratios of HIV infection within each stratum of selected sociodemographic and behavioral characteristics.
Characteristics of Women at PLACE Sites, ANC, and STI Clinics
Two hundred and forty-seven women socializing at social sites, 1116 pregnant women from the ANC, and 229 female STI patients were interviewed and tested for HIV infection (Table 1). The 3 populations were significantly different in age, education, condom use with last sex partner, partnership level, and lifetime condom use (P < 0.05 for each). About half of all women surveyed in each setting were aged 20 to 29 and the STD clinics had the highest proportion of women 35 and older.
Women at PLACE sites were the least educated and had the highest rates of sexual partnership formation. Approximately one-fourth of women at PLACE sites self-identified as sex workers. One-third of all women at PLACE sites reported engaging in transactional sex and half reported having had new or multiple sexual partnerships within the past 4 weeks. However, of those who reported having new or multiple partners in the past 4 weeks, only 8% reported engaging in both commercial and transactional sex and 30% denied engaging in commercial or transactional sex. Rates of sexual partnership formation among STI clinic and ANC patients were much lower than among women at social venues: only 20% of female STI clinic patients and 5% of ANC patients reported having new or multiple partnerships within the past year. Despite their high rate of sexual partnership formation, 44% of PLACE women said they had never used a condom.
HIV in Women
We estimated the prevalence of HIV in each of the 3 study populations by age, education, hospitalization, and sexual behavior (Table 2). No respondents were identified as being recently infected with HIV (through presence of HIV RNA or p24 core antigen before antibody seroconversion). HIV prevalence was higher among STI patients (16.5%) and PLACE patrons (12.3%) and lower among pregnant women (4.1%). The pattern of HIV prevalence by age was different for each study population. Among ANC clients, prevalence did not vary greatly by age (Table 2). Among STI clinic patients and women at PLACE sites, prevalence was lowest among women 15 to 24. HIV prevalence peaked at ages 25 to 34 among PLACE women (21.3%) and among women 35 and older at STI clinics (35.0%). The odds of HIV infection among PLACE patrons aged 25 to 34 were 2.6 times those of STI patients and 5.6 times those of ANC patients of the same age group (95% CI: 2.8, 11.1) (Table 3).
Having new or multiple sex partnerships was strongly associated with HIV infection among women at PLACE sites; having any new or multiple partnerships in the past year was associated with an over 4-fold increase in the odds of HIV infection relative to women with no new or multiple partnerships in the past year (95% CI: 1.5, 14.7). Women reporting new or multiple partnerships in the past 4 weeks had the highest prevalence (19.3%) followed by women who had new or multiple partners in the past year but not in the past month (10%). HIV prevalence among those with the highest partnership level at STI clinics was similar to that of PLACE women with a high partnership level (18.6%).
Characteristics of Men at PLACE Sites and at the STI Clinics
A total of 952 men socializing at PLACE sites and 66 male STI patients participated in the study (Table 1). Men at PLACE sites were younger and more likely to report new and multiple partnerships than men at STI clinics (P < 0.05 for each). Over one-half of men interviewed at PLACE sites reported having had new or multiple sexual partnerships within the past 4 weeks, and 20% more reported having had new or multiple partnerships within the past year. Most men at STI clinics denied having either new or multiple partnerships.
Although the age distribution of men and women at PLACE sites was similar, male patrons reported a higher level of education than females (P < 0.05). Almost three-fourths of men reported having had new or multiple partnerships in the past year compared with 58% of women (P < 0.05). Transactional sex was reported almost as frequently by men at PLACE sites as it was by women at PLACE sites (29% vs. 34%). Male respondents interviewed at PLACE sites were less likely to report condom use with their most recent new sex partner than females (P < 0.05).
HIV in Men
No cases of recent HIV infection were identified among men. HIV prevalence was an order of magnitude higher among the STI clinic population than among PLACE patrons (23% vs. 2%, P < 0.05). For male PLACE patrons, HIV prevalence increased with age, professional employment, a recent visit to an STD clinic, and having a new sexual partner in the past year but not using a condom. Among men interviewed at the STI clinic, HIV prevalence increased with age (P < 0.05). Partnership level was not associated with HIV infection in either population.
HIV prevalence among ANC attendees in our study (4%) was similar to HIV prevalence at other Kinshasa ANC in 2002 (3%).1 Since 2002, prevalence has decreased to 1.6% among women attending Kingasani ANC, where this study took place.8 HIV prevalence was higher among STI clients in our study than among other Kinshasa STI clients in 2002,1 and higher than the prevalence among sex workers attending Matonge STI clinic, also in Kinshasa.9 Matonge clinic has had HIV prevention programming since 19859 and the lower HIV prevalence at Matonge may reflect successful efforts there. HIV prevalence among sex workers interviewed at PLACE venues in our study was 28.6%, substantially higher than the HIV prevalence reported at Matonge (12.4%). In the 2002 Matonge study, HIV prevalence among sex workers varied by type of sex work and ranged from 24% among home-based workers to 6.6% among occasional or clandestine sex workers. Unfortunately, we do not have data describing the type of sex work carried out by women at PLACE venues.
The study had sufficient power to compare differences in the HIV prevalence and social and behavioral characteristics of female ANC, PLACE, and STI clinic populations. The study had 92% power to detect a difference between the HIV prevalence of ANC patients and PLACE patrons (of 3% vs. 6%), 89% power to detect a difference between the HIV prevalence of STI patients and PLACE patrons (of 12% vs. 6%), and over 99% power to detect a difference between the HIV prevalence of STI patients and ANC patients (of 12% vs. 3%). One of the original objectives of this study was to estimate the prevalence of new HIV infections among persons recruited from ANC and STI clinics and venues where people meet new sexual partners. Unfortunately, funds were insufficient to recruit the extremely large samples required to assess differences among these 3 apparently low-incidence populations. The apparent low level of new HIV infections in our study population may be reflected in subsequent decreases in HIV prevalence among Kinshasa ANC attendees.8 Unfortunately, the small sample size among male STI patients limits the usefulness of the STI data for men and limits comparison between STI and PLACE data. As discussed earlier, most STI clients in Kinshasa are female. The opposite gender imbalance occurs at PLACE venues, where the majority of patrons socializing are male.
Our study was designed to compare levels of HIV infection among PLACE patrons and ANC and STI clinic patients by age, and not as a comparison of the behavioral characteristics of respondents at these 3 site types. The PLACE survey was the only instrument to include questions about employment, exchange of money for sex, and additional details about condom use, to assess whether the characteristics and sexual behavior of people socializing at each site were associated with each site’s geographic location, size, or type (bar, hotel, etc). Because PLACE surveys are focused on formation of new sexual partnerships, questions about marital status are thought to create bias and are often not asked. Our study asked respondents about their total number of sexual partners and their total number of new sexual partners over the past 4 weeks and 12 months. We were not able to characterize the partnerships of persons who reported more than one partner during these time periods as concordant or consecutive. In addition, the 1-year delay between identification of PLACE sites and on-site interviewing of PLACE patrons may have introduced bias because sites still in operation in 2003 may have been different from 2002 sites no longer operating in 2003.
The percentage of missing data were high for survey questions concerning recent hospitalization, new sexual partnerships within the past year, and condom use at last sex with last new partner. Approximately one-fourth of STI patients did not give information about recent hospitalization and over 10% of STI patients did not indicate whether they had a new sexual partner in the past year. Over 40% of female STI patients and 27.1% of ANC patients who had a new sex partner in the past year did not indicate whether they used a condom the last time they had sex with their most recent new partner.
Our study highlights the value of HIV outreach testing among patrons of sites where people meet new sexual partners. First of all, many more sites where people meet new partners were identified than expected based on ongoing surveillance of sex workers in Kinshasa. The number and diversity of sites that we recorded suggested a more extensive system of sexual networks than anticipated. Patrons of sites where people meet new sex partners were more likely than STI clinic attendees and much more likely than ANC attendees to report new and multiple sexual partnerships. HIV prevalence among women at PLACE sites (12.3%) was much higher than expected and was strongly associated with partnership formation levels. Self-identified commercial sex workers at PLACE sites had a very high HIV prevalence (28.6%) relative to those who did not identify as sex workers (6.3%), but there was also a high prevalence of infection (16%) among women interviewed at PLACE sites who reported engaging in transactional sex but who did not identify themselves as sex workers. The number of HIV-infected women detected by outreach testing among all female patrons of social venues was double the number that would have been found if testing had been limited to women who self-identified as sex workers.
In addition, HIV screening at PLACE sites is likely to be as or more effective for identifying infected women as screening at STI clinics. A similar number of women 25 to 34 were tested at STI clinics and PLACE sites (81 and 89, respectively), yet on-site testing at PLACE sites identified 19 infected women and 8 infected women at STI clinics. Two-thirds of the infected women at STI clinics were older than 35 and did not report new or multiple partners in the past year. The infected women at PLACE venues had a higher partnership rate than infected women at the STI clinics, suggesting that prevention at PLACE sites might be a more effective strategy for limiting transmission of HIV to additional sexual partners.
Finally, this study demonstrates the feasibility of providing voluntary HIV counseling and testing to men and women socializing at sites where people meet new sex partners. This is the first Kinshasa study to collect blood for HIV testing and behavioral data regarding HIV/AIDS in the general population at social venues during their busiest night-time periods. By including all women in the outreach testing, stigma associated with outreach testing targeting specific groups such as sex workers was avoided. Forty-three percent of persons interviewed at PLACE sites collected their HIV test result. Higher rates of return for results were found among ANC attendees (88%) and STI clients (67%). The PNLS, UNICEF, and the Kinshasa School of Public Health used PLACE data from this study to target local HIV prevention programs.
1. Mulanga C, Bazepeo SE, Mwamba JK, et al. Political and socioeconomic instability: How does it affect HIV? A case study in the Democratic Republic of Congo. AIDS 2004; 18:832–834.
2. Mulanga-Kabeya C, Nzilambi N, Edidi B, et al. Evidence of stable HIV seroprevalences in selected populations in the Democratic Republic of the Congo. AIDS 1998; 12:905–910.
3. Mayala M, Minlangu M, Nzila N, et al. Prevalence and incidence of HIV-1 infection among employees of a large textile business and their wives in Kinshasa, 1991–1996 [in French]. Rev Epidemiol Sante Publique 2001; 49:117–124.
4. Batter V, Matela B, Nsuami M, et al. High HIV-1 incidence in young women masked by stable overall seroprevalence among childbearing women in Kinshasa, Zaire: Estimating incidence from serial seroprevalence data. AIDS 1994; 8:811–817.
5. Magazani K, Laleman G, Perriens JH, et al. Low and stable HIV seroprevalence in pregnant women in Shaba province, Zaire. J Acquir Immun Defic Syndr 1993; 6:419–423.
6. Democratic Republic of the Congo Ministry of Health and PNLS. Report of sentinel surveillance 2003–2004.
7. Ryder RW, Ndilu M, Hassig SE, et al. Heterosexual transmission of HIV-1 among employees and their spouses at two large businesses in Zaire. AIDS 1990; 4:725–732.
8. Kinoshita-Moleka R, Smith JS, Atibu J, et al. Low prevalence of HIV and other selected sexually transmitted infections in 2004 in pregnant women from Kinshasa, the Democratic Republic of the Congo. Epidemiol Infect 2008; 136:1290–1296.
9. Vandepitte JM, Malele F, Dieu-Merci K, et al. HIV and other sexually transmitted infections among female sex workers in Kinshasa, Democratic Republic of Congo, in 2002. Sex Transm Dis 2007; 34:203–208.
10. Denolf D, Musongela JP, Nzila N, et al. The HIV epidemic in Kinshasa, Democratic Republic of Congo. Int J STD AIDS 2001; 12:832.
11. Weir SS, Morroni C, Coetzee N, et al. A pilot study of a rapid assessment method to identify places for AIDS prevention in Cape Town, South Africa. Sex Transm Infect 2002; 78(suppl 1):i106–i113.
12. Weir SS, Tate JE, Zhusupov B, et al. Where the action is: Monitoring local trends in sexual behaviour. Sex Transm Infect 2004; 80(suppl 2):ii63–ii68.
13. Quinn TC, Brookmeyer R, Kline R, et al. Feasibility of pooling sera for HIV-1 viral RNA to diagnose acute primary HIV-1 infection and estimate HIV incidence. AIDS 2000; 14:2751–2757.
14. Coombs RW, Collier AC, Allain JP, et al. Plasma viremia in human immunodeficiency virus infection. N Engl J Med 1989; 321:1626–1631.
15. Eyster ME, Ballard JO, Gail MH, et al. Predictive markers for the acquired immunodeficiency syndrome (AIDS) in hemophiliacs: Persistence of p24 antigen and low T4 cell count. Ann Intern Med 1989; 110:963–969.