Assisted partner services (APS) are a longstanding component of public health efforts to control sexually transmitted infections, including human immunodeficiency virus (HIV), in many high income nations.1–4 These services typically involve public health professionals interviewing persons with sexually transmitted infections and actively attempting to ensure their sex partners' notification, medical evaluation and treatment. However, APS are not widely employed in sub-Saharan Africa. In the absence of such assistance, fewer than half of potentially exposed partners are tested,5–11 though in some instances disclosure of HIV status is common.12,13
Assisted partner services is efficacious and can be effective in sub-Saharan Africa. Two randomized controlled trials conducted in Malawi found that APS increased testing among the sex partners of persons with HIV, and that the intervention is cost-effective.14–16 Data from Cameroon, where a faith-based organization has provided APS to over 15,000 persons with HIV since 2007, strongly suggest that the intervention is effective and can be brought to scale.17
Despite mounting evidence, APS programs remain unusual in the majority of countries most affected by HIV. We implemented a pilot APS project at the Machava II Health Center in Maputo Province, Mozambique. This paper presents data on the acceptability, effectiveness and safety of APS in a large, urban clinic in Maputo.
Community health workers (CHWs) initially interviewed individuals newly diagnosed with HIV infection (index patients [IPs]) on their day of diagnosis and counseled them to disclose their HIV status to their partners and to refer partners and children age ≤5 for HIV testing (Fig. 1). Community health workers initiated efforts to re-interview IPs four weeks after the IP's initial diagnosis (follow-up interview 1) to assess partner notification testing outcomes and adverse events (AEs). At the time of reinterview, CHWs offered IPs APS and requested permission to contact their sex partners with the goal of assuring their notification and testing, and verifying IP reported outcomes. Community health workers initiated efforts to re-interview IPs a second time at 8 weeks (follow-up interview 2) to ascertain final notification and testing outcomes and AEs. We compared outcomes at first and second follow-up interviews to estimate the value of APS over passive services in improving partner notification and testing. The project was designed to last 4 months and to enroll 150 persons. This sample size would provide 80% power to detect a difference between 4- and 8-week outcomes if 15% of IPs had a partner test HIV positive by 4 weeks and APS increased this percentage to 30% (α = 0.05, 2-sided z test).
Partner Notification Training
The pilot was modeled on the Cameroonian Baptist Convention Health Services APS program.17 Members of the Mozambican team attended an APS training in Cameroon before initiation of the Maputo program, and the program's trainings and data collection instruments were adapted from those used in Cameroon.
Mozambican CHWs and supervisors attended a 2-week training before the program's initiation. This included both classroom and practice-based work focusing on program objectives and procedures, interviewing, disclosure counselling, ethical dilemmas, intimate partner violence (IPV), and data collection.
The program enrolled IPs aged ≥18 years who newly tested HIV-positive at Machava II, June to September, 2014. Machava II is an urban clinic operated by the Ministry of Health on the outskirts of Maputo. Testing staff referred potential IPs to a CHW, who explained the program and obtained IP consent to participate. The program initially employed 5 CHWs, though this decreased to a staff of 3 over the course of the project. Community health workers attempted to enroll all persons who tested HIV positive during the approximately 20 hours per week that they were available. Given the small staff, we could not provide APS to all eligible persons, and the enrolled population was a convenience sample.
Data Collection and Partner Notification Procedures
Community health workers interviewed IPs using structured interview forms that included fields for information about children age ≤5 years and information on up to 4 persons with whom the IP had sex in the prior 3 years. Community health worker asked IPs questions about IPV, including whether named partners had ever physically hurt them and whether they feared that a partner would hurt them if they told the partner they tested HIV positive.
Community health workers counseled patients to disclose their HIV status to partners and to ask partners to come to the clinic for HIV testing. Counseling included reinforcing the benefits of disclosure, exploration of potential partner reactions, and discussion of how and when to disclose. Community health workers also created a plan for testing children.
Community health workers reinterviewed IPs at the time they brought a partner to the clinic for testing or, if the IP did not bring a partner for testing without assistance, attempted to conduct telephone or in-person interviews with IPs approximately 4 weeks after the IP's diagnosis. Interviews asked IPs about disclosure to sex partners, partner HIV testing, and test results of both partners and children. At the end of the interview, CHWs asked IPs' permission to contact named partners to ensure their notification and testing. Community health workers never revealed the names of IPs without the IP's consent.
When contacted, CHWs informed partners of their possible HIV exposure, offered to help arrange HIV testing, and interviewed partners about their HIV testing history. The program protocol was for CHWs to make up to 3 attempts to notify partners by telephone and, if this was unsuccessful, to notify partners via a visit to their home or workplace. However, a lack of travel funds and concerns about CHW safety delayed implementation of field notification efforts until eleven weeks after program initiation.
Community health workers completed a second IP follow-up interview approximately 8 weeks after the IP's diagnosis. This interview collected information about notification and testing status of each partner and AEs, including IPV, partnership dissolution, and loss of economic support. Community health workers asked IPs if AEs were related to their HIV diagnosis.
Before closing cases, CHWs completed a final partner disposition form indicating if each partner HIV tested and the partner's test results. The final disposition reflected knowledge derived through interviews with the IP and partner and review of medical records. Community health workers defined testing outcomes as verified if they observed the partner test or confirmed their test results through review of clinic records.
Data Management and Statistical Analysis
The program coordinator reviewed data collection forms weekly and facilitated weekly case reviews. Staff double data-entered information into CSPro (US Census).
We excluded partners with previously diagnosed HIV infection from analyses of HIV testing and case-finding. We defined partners as being previously diagnosed if the IP reported that the partner was HIV-positive at their baseline interview or if the partner themselves reported testing HIV positive and the CHW defined the partner as HIV-positive on the final disposition. We defined partners' HIV testing status at follow-up 1 based on IP report, and defined partners as both HIV tested and HIV-diagnosed at follow-up 1 if the IP reported the partner was HIV-positive or the partner self-reported being HIV-positive when interviewed. Final partner HIV testing and HIV status was based on CHW's final partner disposition, verified or unverified. In cases where the CHW final partner disposition indicated that a partner never tested but IPs reported that a partner HIV tested at follow-up 1 interview, we defined the partner as not having tested. We defined partners as being notified, tested, or diagnosed as a result of IP efforts, not APS, if they met any of the following criteria: (1) the CHW interviewed the partner <28 days after the initial IP interview; (2) a CHW interviewed the partner before the IP's 4-week interview; or (3) the partner self-reported testing HIV-positive when interviewed by the CHW. Each of these criteria suggests that notification or testing occurred before the protocol specified provision of APS 4 weeks after the initial IP interview.
We used Stata 12.0 for analyses (College Station, Tex). We calculated the number of persons CHWs needed to interview (NNTI) to notify and test a new partner, and to identify one person with undiagnosed HIV infection and one person in an ongoing serodiscordant partnership. The NNTI is the number of persons receiving APS divided by the number of outcomes attributable to APS.17–19 We used the cii command to define exact CIs for a binomial proportion for NNTI estimates.
The University of Washington Institutional Review Board defined the project as a programmatic activity and issued a nonresearch determination. The similar protocol was approved by the Mozambican Bioethical Committee.
Machava II staff newly diagnosed 1225 patients with HIV infection during the 4 months of project enrollment. Community health worker offered 223 patients APS, and 220 (99%) accepted; 206 (94%) of 220 participants completed both follow-up visits. We report outcomes on these 206 persons.
Of the 206 IPs, 162 (79%) were women, 101 (49%) of whom were diagnosed through prenatal care (Table 1). The median age of IPs was 29 years (interquartile range, 24–34 years) and 152 (74%) were married. Twelve participants (4.6%), including 11 women, reported experiencing IPV with a current partner in the prior 2 months. The 206 participants provided information on 283 sex partners (mean, 1.4 partners; range, 1–4), and 140 (50%) participants provided information on >1 partner. Among 283 named partners, 168 (59%) were spouses and 209 (74%) were men. At baseline, 30 (11%) of 206 IPs reported having a partner with previously diagnosed HIV infection; CHWs confirmed that 21 of these partners were HIV-positive and we excluded these partners from outcome analyses. Thus, our partner related outcomes include 262 partners. IPs provided CHWs with phone numbers for 191 (73%) of these partners.
Although most IPs reported disclosing their HIV status to partners prior to receiving APS, relatively few partners tested for HIV. Community health workers conducted follow-up visit 1 a median of 32 days following the initial interview (range, 8–146 days). At that time, IPs reported notifying 169 (65%) of their 262 identified partners, but that only 82 (31%) partners had HIV tested. Index patient and partner reports identified 34 partners (13%) who were newly diagnosed with HIV infection by the time of follow-up 1 (Fig. 2).
A total of 163 IPs had children. Index patients agreed to test 113 (99%) of 114 children aged <5 years, of whom 98 (87%) tested and 2 (2%) were HIV-positive.
Community health workers' interviews with partners generally confirmed follow-up visit 1 IP reported outcomes. Index patients gave CHWs permission to contact 222 (85%) of their 262 partners. Index patients were more likely to allow CHWs to contact their most recent sex partner than other partners (90% vs 72%, P < 0.0001); willingness to allow CHWs to contact partners was not associated with gender or IP age. Community health workers successfully interviewed 173 (66%) partners a median of 44 days (range, 0–154 days) after the initial IP interview. Among the 173 interviewed partners, 141 (81%) reported being notified, 24 (14%) stated they had not been notified, and 8 (5%) originally tested at the same time as the IP who named them. A total 139 (93%) of the 150 partners who IP’s reported already notifying confirmed that a partner had told them that they had tested HIV positive. Of the 173 interviewed partners, 69 (40%) reported they had never HIV tested and 104 (60%) reported previously HIV testing. Ninety-five (91%) of the previously tested partners reported testing HIV negative and 9 (9%) reported being HIV-positive.
The number of partners tested increased significantly following APS, as indicated by the final partner dispositions recorded at follow-up 2 (Fig. 2). Community health workers conducted follow-up visit 2 a median of 40 days after follow-up visit 1 (range 1–155 days). Based on final dispositions, 193 (74%) of 262 partners were notified, 174 (66%) HIV tested, and 83 (32%) were newly diagnosed with HIV infection. We were unable to confirm the time of HIV testing for 9 (5%) partners, including 6 persons who tested HIV positive, and excluded these partners from analyses assessing the impact of APS. Assisted partner services resulted in 83 partners (32% of all partners) testing for HIV and 43 new HIV diagnoses; 17% of all partners were diagnosed with HIV as a result of APS, and 52% of partners newly diagnosed with HIV infection were diagnosed as a result of APS. Among the 43 partners newly diagnosed as a result of APS, CHWs verified the HIV status of 37 (88%) and interviewed 40 (95%). Among those interviewed, 21 reported never having previously HIV tested and 19 reported previously testing HIV negative. Excluding outcomes in the 9 partners for whom we could not define the timing of HIV testing, in relative terms, APS increased partner notification by 13%, testing by 101% and new HIV diagnosis by 126%.
Ninety-one partners tested HIV negative. Seventy-two (35%) IPs reported that they anticipated having sex in the future with 77 partners who tested HIV negative (i.e. partnerships were ongoing).
Community health workers provided APS to 4.8 IPs to identify 1 new case of HIV infection, and 5.3 IPs to identify one ongoing HIV serodiscordant partnership (Table 2). Combining these outcomes, the NNTI to identify one new case of HIV or one previously unidentified, ongoing serodiscordant partnership was 2.5.
Assisted partner services primarily identified male partners with HIV and serodiscordant partnerships in which the male partner was HIV-uninfected. Among the 43 partners diagnosed with HIV infection as a result of APS, 37 (86%) were men. Among the 77 ongoing serodiscordant partnerships, 57 (74%) included an HIV-uninfected man.
None of the 173 interviewed partners reported experiencing IPV after their partners’ HIV diagnosis. One IP reported that a partner left her for reasons unrelated to HIV. Two other women reported that partners left them because of HIV, one of whom also reported loss of financial support.
We found that APS was acceptable, safe and effective when provided by CHWs in an urban clinic in Mozambique. The intervention only had a very small impact on partner notification because 65% of IPs reported disclosing their HIV status to their partner(s) before receiving APS. However, APS more than doubled partner HIV testing, the identification of HIV-infected partners, and the identification of ongoing serodiscordant partnerships. These findings build on the results of previous APS studies,14–17,20,21 further supporting the effectiveness and feasibility of APS in sub-Saharan Africa.
The model of APS used in Mozambique differs from those employed elsewhere in sub-Saharan Africa, and our findings may be particularly important as APS becomes part of routine HIV prevention programs in the region. In APS studies undertaken in Malawi and Kenya, the intervention relied on study personnel,14,16,22 whereas an ongoing Cameroonian program uses HIV educators, nurses, social workers, chaplains and support group coordinators. Our program employed part-time CHWs who earned approximately US $40 per month. Although this minimized costs, it also presented significant challenges. The program required hiring new supervisors, low salary affected staff retention, and staff sometimes struggled to understand APS forms. Also, due to funding limitations and concerns for staff safety, CHWs' seldom visited partners' homes and workplaces. This lack of outreach likely diminished our program’s effectiveness, but allowed us to evaluate a low-intensity model of partner follow-up that may be more scalable in low resource environments.
To our knowledge, this study is the first to report outcomes related to integrating child HIV testing into APS. Fewer than 20% of HIV-exposed infants in resource limited settings receive an HIV test,23 and we hypothesized that APS might identify undiagnosed HIV-infected children. We found that integrating child testing into APS was acceptable, but that only 2% of tested children were HIV-infected. The utility of testing children as part of APS requires further evaluation.
Fear of IPV and other AEs is an impediment to APS implementation. Intimate partner violence is common in many nations with generalized HIV epidemics.13,24 In some populations, IPV frequently occurs after disclosure of HIV status, though a study in Zimbabwe found it was more common before HIV status disclosure than after it.13 Evaluations in Malawi,14,16 Cameroon,17 Kenya,21 and now Mozambique suggest that IPV is very uncommon following APS. Of note, in our program, several IPs with a history of IPV elected to notify their partners and, despite the fact that our protocol was to not contact partners if IPs had an IPV history, several CHWs did so. This deviation from planned processes highlights the need for greater emphasis on IPV training. At the same time, the absence of AEs, even with these protocol deviations, is reassuring. While continued consideration of IPV is needed, we believe that this concern should not be an impediment to the scale-up of APS.
Our study has a number of limitations. First, we compared outcomes achieved approximately 4 weeks following IP HIV diagnosis (before APS) to those obtained approximately 8 weeks after diagnosis (after APS). The unvalidated assumption underlying this approach was that partners not tested within four weeks of IP diagnosis were unlikely to test in the near future. The choice of a four week timeframe for evaluating outcomes achieved through unassisted counseling was somewhat arbitrary and reflected a desire to give IPs time to promote their partners' testing while also assuring IP follow-up. It is possible that more partners would have tested if given a longer period of unassisted follow-up. Second, several factors could have led us to misclassify outcomes. We were not able to define the timing of partner testing for 5% of partners and elected to exclude those persons from analyses looking at the impact of APS. Also, follow-up visit 1 partner testing outcomes were based primarily on IP and partner report, whereas our final outcomes were based on a composite of verified outcomes and IP and partner reports. Index patients may have overreported partner notification due to social desirability bias, or underreported testing if they didn’t know if their partners tested. However, >90% of interviewed partners confirmed IP reports that they had been notified and, among the 40 interviewed partners we classified as newly diagnosed after APS, 19 said that they were HIV-negative and 21 self-reported that they had never HIV tested when interviewed by CHWs, indicating that their HIV diagnosis was, in fact, new. This high level of agreement between IP and partner reports suggests that our findings were valid. Also, CHWs verified most of our final HIV testing outcomes and, among unverified outcomes, most were based on interviews with partners, not IPs. Thus, we do not believe that misclassification led to a substantial overestimate of the effect of APS. When viewed in isolation, the generalizability of our findings is uncertain. Our study population was drawn from a single public, urban clinic in Mozambique and included a large number of IPs receiving prenatal care. However, APS has now been shown to be effective in diverse African settings and among populations excluding women pregnant women. The consistency of findings across studies strongly suggests that our findings are generalizable. Finally, because the project had limited staff, we offered APS to a minority of persons diagnosed with HIV at Machava II during the project period. We do not have data to compare APS recipients to nonrecipients, and these populations may have differed. Future studies should assess the workforce capacity needs and scalability of APS.
In conclusion, we found that a model of APS that relies on a CHW labor force is acceptable to patients, safe, and effective in promoting HIV testing among the sex partners of newly diagnosed HIV-infected persons in Maputo, Mozambique. This model of APS increased the identification of serodiscordant partnerships and was highly effective in finding partners, particularly male partners, with previously undiagnosed HIV infection. Public health and clinical programs in low resource settings should integrate APS into routine HIV testing as a means to identify both persons with undiagnosed HIV infection and ongoing serodiscordant partnerships in which the risk of transmission is high, but avertable with antiretroviral therapy.
1. Hogben M, McNally T, McPheeters M, et al. The effectiveness of HIV partner counseling and referral services in increasing identification of HIV-positive individuals: A systematic review. Am J Prev Med 2007; 33(2 suppl):S89–S100.
2. Katz DA, Hogben M, Dooley SW Jr, et al. Increasing public health partner services for human immunodeficiency virus: Results of a second national survey. Sex Transm Dis 2010; 37:469–475.
3. Golden MR, Faxelid E, Low N. Partner notification for STD and HIV. In: Holmes KK, ed. Sexually Transmitted Diseases. Fourth ed. New York: McGraw-Hill, 2009.
4. Ferreira A, Young T, Mathews C, et al. Strategies for partner notification for sexually transmitted infections, including HIV. Cochrane Database Syst Rev 2013; 10:CD002843.
5. Temmerman M, Moses S, Kiragu D, et al. Impact of single session post-partum counselling of HIV infected women on their subsequent reproductive behaviour. AIDS Care 1990; 2:247–252.
6. Lie GT, Biswalo PM. HIV-positive patient's choice of a significant other to be informed about the HIV-test result: Findings from an HIV/AIDS counselling programme in the regional hospitals of Arusha and Kilimanjaro, Tanzania. AIDS Care 1996; 8:285–296.
7. Farquhar C, Mbori-Ngacha DA, Bosire RK, et al. Partner notification by HIV-1 seropositive pregnant women: association with infant feeding decisions. Aids 2001; 15:815–817.
8. Antelman G, Smith Fawzi MC, Kaaya S, et al. Predictors of HIV-1 serostatus disclosure: A prospective study among HIV-infected pregnant women in Dar es Salaam, Tanzania. Aids 2001; 15:1865–1874.
9. Nebié Y, Meda N, Leroy V, et al. Sexual and reproductive life of women informed of their HIV seropositivity: A prospective cohort study in Burkina Faso. J Acquir Immune Defic Syndr 2001; 28:367–372.
10. Salami AK, Fadeyi A, Ogunmodede JA, et al. Status disclosure among People Living with HIV/AIDS in Ilorin, Nigeria. West Afr J Med 2011; 30:359–363.
11. Irungu E, Chersich MF, Sanon C, et al. Changes in sexual behaviour among HIV-infected women in west and east Africa in the first 24 months after delivery. Aids 2012; 26:997–1007.
12. Batte A, Katahoire AR, Chimoyi A, et al. Disclosure of HIV test results by women to their partners following antenatal HIV testing: A population-based cross-sectional survey among slum dwellers in Kampala Uganda. BMC Public Health 2015; 15:63.
13. Shamu S, Zarowsky C, Shefer T, et al. Intimate partner violence after disclosure of HIV test results among pregnant women in Harare, Zimbabwe. PLoS One 2014; 9:e109447.
14. Brown LB, Miller WC, Kamanga G, et al. HIV partner notification is effective and feasible in sub-Saharan Africa: Opportunities for HIV treatment and prevention. J Acquir Immune Defic Syndr 2011; 56:437–442.
15. Rutstein SE, Brown LB, Biddle AK, et al. Cost-effectiveness of provider-based HIV partner notification in urban Malawi. Health Policy Plan 2014; 29:115–126.
16. Rosenberg NE, Mtande TK, Saidi F, et al. Recruiting male partners for couple HIV testing and counselling in Malawi's option B+ programme: An unblinded randomised controlled trial. Lancet HIV 2015; 2:e483–e491.
17. Henley C, Forgwei G, Welty T, et al. Scale-up and case-finding effectiveness of an HIV partner services program in Cameroon: An innovative HIV prevention intervention for developing countries. Sex Transm Dis 2013; 40:909–914.
18. Golden MR, Hogben M, Potterat JJ, et al. HIV partner notification in the United States: A national survey of program coverage and outcomes. Sex Transm Dis 2004; 31:709–712.
19. Ahrens K, Kent CK, Kohn RP, et al. HIV partner notification outcomes for HIV-infected patients by duration of infection, San Francisco, 2004 to 2006. J Acquir Immune Defic Syndr 2007; 46:479–484.
20. Brown LB, Miller WC, Kamanga G, et al. Predicting partner HIV testing and counseling following a partner notification intervention. AIDS Behav 2012; 16:1148–1155.
21. Cherutich P, Golden MR, Wamuti B, et al. Effectiveness of Partner Services for HIV in Kenya: A cluster randomized trial. Conference on Retroviruses and Opportunistic Infections 2016; February 22–25, 2016, 2016; Boston, MA, USA.
22. Wamuti BM, Erdman LK, Cherutich P, et al. Assisted partner notification services to augment HIV testing and linkage to care in Kenya: Study protocol for a cluster randomized trial. Implement Sci 2015; 10:23.
23. Kellerman S, Essajee S. HIV testing for children in resource-limited settings: What are we waiting for? PLoS Med 2010; 7:e1000285.
24. Coker AL, Richter DL. Violence against women in Sierra Leone: Frequency and correlates of intimate partner violence and forced sexual intercourse. Afr J Reprod Health 1998; 2:61–72.