Secondary Logo

Share this article on:

Integration of family planning services into HIV care and treatment in Kenya: a cluster-randomized trial

Grossman, Daniela,b; Onono, Maricianahc; Newmann, Sara J.b; Blat, Cinthiab; Bukusi, Elizabeth A.c; Shade, Starley B.d; Steinfeld, Rachel L.b; Cohen, Craig R.b

doi: 10.1097/QAD.0000000000000035
Supplement Articles

Objective: To determine whether integrating family planning services into HIV care is associated with increased use of more effective contraceptive methods (sterilization, intrauterine device, implant, injectable or oral contraceptives).

Design: Cluster-randomized trial.

Setting: Eighteen public HIV clinics in Nyanza Province, Kenya.

Participants: Women aged 18–45 years receiving care at participating HIV clinics; 5682 clinical encounters from baseline period (December 2009–February 2010) and 12 531 encounters from end-line period (July 2011–September 2011, 1 year after site training).

Intervention: Twelve sites were randomized to integrate family planning services into the HIV clinic, whereas six clinics were controls where clients desiring contraception were referred to family planning clinics at the same facility.

Main outcome measures: Increase in use of more effective contraceptive methods between baseline and end-line periods. Pregnancy rates during the follow-up year (October 2010–September 2011) were also compared.

Results: Women seen at integrated sites were significantly more likely to use more effective contraceptive methods at the end of the study [increased from 16.7 to 36.6% at integrated sites, compared to increase from 21.1 to 29.8% at controls; odds ratio (OR) 1.81, 95% confidence interval (CI) 1.24–2.63]. Condom use decreased non-significantly at intervention sites compared to controls (OR 0.64, 95% CI 0.35–1.19). No difference was observed in incident pregnancy in the first year after integration comparing intervention to control sites (incidence rate ratio 0.90; 95% CI 0.68–1.20).

Conclusions: Integration of family planning services into HIV care clinics increased use of more effective contraceptive methods with a non-significant reduction in condom use. Although no significant reduction in pregnancy incidence was observed during the study, 1 year may be too short a period of observation for this outcome.

Supplemental Digital Content is available in the text

aIbis Reproductive Health, Oakland, California

bBixby Center for Global Reproductive Health, Department of Obstetrics, Gynecology, and Reproductive Sciences, University of California, San Francisco, USA

cKenya Medical Research Institute, Kisumu, Kenya

dCenter for AIDS Prevention, Department of Medicine, University of California, San Francisco, USA.

Correspondence to Daniel Grossman, MD, Ibis Reproductive Health, 1330 Broadway, Suite 1100, Oakland, CA 94612, USA. Tel: +1 510 986 8941; e-mail: dgrossman@ibisreproductivehealth.org

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (http://www.AIDSonline.com).

Back to Top | Article Outline

Introduction

An estimated 13 million HIV-infected women live in sub-Saharan Africa [1]. Among the HIV-infected women in this region, studies indicate that 62–93% of pregnancies are unintended [2–4]. The WHO has included the prevention of unintended pregnancy as one of four pillars of a comprehensive prevention of mother-to-child transmission (PMTCT) strategy [5]. Improved access to family planning among HIV-infected individuals is also expected to decrease maternal morbidity and mortality, as well as poor neonatal outcomes [6,7].

In many settings in sub-Saharan Africa, contraceptive services are provided in family planning clinics separate from clinics providing antiretroviral therapy (ART) and related care for HIV-infected individuals [8]. Recognizing the structural barriers associated with this model of care, at least six international statements have recommended integrating family planning and HIV services [9]. Advocates of integration hypothesize that it will increase contraceptive uptake, prevent pregnancy and empower women and men to determine their family size. However, there is little rigorous research evidence to support this hypothesis, and there have been no randomized trials on integrating family planning into HIV services [10]. A cohort study in Nigeria found no difference in contraceptive use between women at ART clinics providing enhanced family planning services compared to those at ART clinics providing more basic family planning counselling [11]. A recent cohort study in Kenya found that integrated services were associated with a significant increase in condom use, but there was no effect on pregnancy incidence; use of non-condom family planning methods decreased at the integrated site [12].

We aimed to determine whether integration of family planning services into HIV care and treatment improved the uptake of more effective contraceptive methods [sterilization, intrauterine device (IUD), sub-dermal implants, injectables and oral contraceptives]. A secondary aim of the study was to determine whether pregnancy incidence was affected by integration of family planning services. Because the intervention focused on changing the service-delivery model at the healthcare facility and to minimize contamination, the unit of randomization was the facility.

Back to Top | Article Outline

Methods

We performed a cluster-randomized trial comparing contraceptive prevalence and pregnancy incidence among women at HIV care and treatment clinics that offered integrated family planning services to HIV clinics that referred clients seeking family planning services to a separate maternal–child health and family planning (MCH-FP) clinic at the same facility. Data from clinical encounters were treated as serial cross-sections to assess change in study outcomes over time. More details of the study design are available in the Supplemental Digital Content (http://links.lww.com/QAD/A405).

Eligible sites were public sector HIV clinics at dispensaries, health centres, and sub-district and district hospitals in Kisumu East, Nyatike, Rongo and Suba Districts of Nyanza Province, Kenya. Prior to initiating the study, we performed site assessments at all eligible sites, and sites were excluded if they had already integrated family planning services into HIV care or if they were not offering ART on site. All sites were supported by Family AIDS Care and Education Services (FACES), a collaboration between University of California, San Francisco (UCSF) and the Kenyan Medical Research Institute (KEMRI) [13]. Women were eligible to contribute data to the study if they had a visit at one of the HIV clinic sites during the study period and were between the age of 18 and 45 years.

The study was approved by the Committee on Human Research at UCSF and the Ethical Review Committee at KEMRI. We originally planned to obtain consent from individual women at study sites; however, shortly after beginning the study, we obtained approval from both UCSF and KEMRI for an evaluation protocol that allowed analysis of de-identified data from FACES patients without individual consent.

After assessing sites for eligibility, facilities were stratified based on whether they were large (>700 patients enrolled in HIV care) or small (<700 enrolled HIV patients) and randomly allocated within these strata to receive the intervention or control in a 2 : 1 ratio. The 2 : 1 ratio was selected to respond to the Ministry of Health's interest in moving forward with integration. The allocation sequence was generated by the study's biostatistician who was not involved in fieldwork for the study. Sites were identified and recruited, and written permission to perform the study at these sites was obtained from the Provincial Medical Officer and District Medical Officers of Health prior to randomization.

Clinics, healthcare providers, patients and researchers involved in implementing the study were not blinded to the allocation. Data abstraction from medical records occurred at clinic sites, and it was not possible to blind staff who abstracted data. All outcome data for the trial were extracted from an electronic medical record system (EMRS). Although all study investigators were not blinded to allocation during the conduct of the trial, investigators did not view data by study arm until after the statistical analysis plan was finalized, data cleaning was completed and the database was locked.

In preparation for the study, starting in March 2010, peer educators at all sites were trained to conduct group educational health talks about family planning to clients waiting to be seen at the HIV clinics. These health talks focused on why some HIV-infected individuals choose to use family planning and reviewed all available contraceptive methods, including their effectiveness and common side effects.

Staff at study sites underwent training between June and August 2010 (see the Supplemental Digital Content, http://links.lww.com/QAD/A405 for more information on training). HIV clinic staff were trained to ask all clients about their current use of contraception, condom use and their interest in using family planning. Condoms were generally provided at the HIV clinic except during a period of nationwide shortage. At HIV clinics assigned to be control sites, staff continued the standard practice of referring clients interested in receiving non-condom family planning to a separate MCH-FP clinic at the same facility.

HIV clinics assigned to the intervention integrated family planning counselling and provision into the HIV clinic according to guidelines established by the Kenyan Government [14,15]. In addition to asking about interest in using family planning, HIV clinic staff at integrated sites also provided all reversible family planning methods within the HIV clinic. At control sites, all reversible family planning methods were available at the MCH-FP clinic, but family planning services were generally available at different times than the HIV clinic, often requiring women to wait or return at another time. At all sites, patients interested in female sterilization were kept on a list to receive services when a roving team came to the facility to perform the procedures, generally once every 1–3 months. Patients interested in vasectomy were referred to an outside provider.

At all sites, clinical encounters were recorded on paper forms that were then entered into an EMRS database (OpenMRS versions 1.6.1–2) [16]. More information on EMRS data entry is given in the Supplemental Digital Content (http://links.lww.com/QAD/A405).

The primary outcome for the study was reported use of more effective contraception (sterilization, IUD, implant, injectable or oral contraceptives) compared to no contraception or use of a less effective method (barrier methods, including condoms used alone, or natural family planning). We decided to focus on contraceptive prevalence, rather than just uptake of new methods, since we hypothesized that the intervention would affect both contraceptive uptake and continuation. Secondary outcomes included reported use of any family planning method, use of condoms, either alone or with another method (‘dual method use’ [17]) and incident pregnancy. We attempted to collect data on whether the pregnancy was intended or not, but this variable was missing from a large proportion of pregnant women's medical records. Prior to performing any analyses, we examined baseline contraceptive use among all pooled study sites (December 2009–February 2010) and observed a high prevalence of reported condom use (reported in approximately 57% of women's encounters). We therefore decided to focus on the use of more effective contraception as the primary outcome of the study.

Contraceptive method use and pregnancy status on a visit were ascertained from patient medical records. More information on study outcome ascertainment is given in the Supplemental Digital Content (http://links.lww.com/QAD/A405).

We present descriptive statistics on study variables collected during the baseline period. Categorical variables are presented using frequencies and proportions and continuous variables using median and inter-quartile range (IQR). Contraceptive prevalence estimates and odds ratios (ORs) comparing contraceptive use across intervention and control arms were derived from a series of generalized estimating equation (GEE) models fitted with robust standard errors to account for clustering of patients within sites. Each model regressed a contraceptive use outcome on integration arm, study period (end line compared to baseline), and a study period by intervention arm interaction term. The OR for the effect of the intervention was obtained from the coefficient for the interaction term, which reflects the increase in the odds of the outcome in the intervention arm compared to the control arm between baseline and follow-up. Contraceptive use prevalence estimates reflect model projections of the predicted probability of the outcome by intervention arm and study period.

Negative binomial regression was used to quantify pregnancy rates during the first year after integration (October 2009–September 2010). In addition, we present an approximate pregnancy rate over 100 person-years of follow-up for each study arm. More information on the statistical analysis of the pregnancy rate is given in the Supplemental Digital Content (http://links.lww.com/QAD/A405).

The primary outcome upon which we based our sample size calculation was prevalence of use of modern contraception. According to the 2003 Kenya Demographic and Health Survey [18], which were the most recent data available at the time the study was planned, approximately 22% of women in a union in Nyanza Province used a modern contraceptive method. We estimated that an 8% increase in contraceptive prevalence would represent a clinically significant increase. Therefore, our sample size estimate was based on the ability to detect an 8% difference in contraceptive prevalence of 30% in intervention sites compared to 22% in the control sites. Using this information, the necessary sample size was 140 women at each of the 18 sites, for a total sample of 2520 women. This estimate assumed an intra-cluster correlation coefficient (ICC) of 0.01, a two-sided alpha of 0.05 and 80% power.

SAS version 9 (Cary, North Carolina, USA) was used for data management and analysis. This study is registered at ClinicalTrials.gov, number NCT01001507.

Back to Top | Article Outline

Results

Twenty-nine health facilities in Nyanza Province, Kenya, were assessed for eligibility. Eleven sites were excluded: six sites had already integrated family planning services into the HIV clinic and five were not providing ART. Eighteen sites were randomized. Among the intervention sites, six were large and six small sites, whereas three of the control sites were large and three small (Fig. 1). Table 1 shows characteristics of the study sites in December 2009 prior to intervention launch. All sites were rural except for two located in towns with populations around 7000. The sites where only a sample of encounters were entered into OpenMRS were sites A, F, I, J, K, Q and R. In general, family planning services were free at integrated sites and control sites M, N and P; HIV clinic patients referred to the MCH-FP clinic at sites O, Q and R were intermittently charged USD $0.20–1.20 depending on the method.

Fig. 1

Fig. 1

Table 1

Table 1

During the baseline period (December 2009–February 2010), contraception data were available for 3584 eligible women who had at least one encounter at the study sites. Data on the size of the clusters at baseline are presented in Fig. 1 and Table 1. Table 2 shows demographic and clinical information for participants with contraception data during the baseline period. Information on education and marital status was only collected at the time of enrolment into HIV care and was available for 61.6% of women at intervention sites and 76.7% of women at control sites.

Table 2

Table 2

The study flow diagram (Fig. 1) gives information on the number of encounters analyzed at baseline and end line (July–September 2011), as well as the size of the clusters. The size of the clusters increased during the study period because more women entered HIV care over time.

At integrated sites, the prevalence of use of more effective family planning methods increased from 16.7 to 36.6%, whereas at control sites, it increased from 21.1 to 29.8% (Table 3). Most of the differential increase in contraceptive use at integrated sites was due to implants and injectables. The prevalence of dual method use increased from 10.1% at baseline to 20.9% at end line at integrated sites; at control sites, dual use increased from 11.5 to 19.1%. The use of less effective family planning methods (almost exclusively condoms used alone) decreased at intervention sites from 50.7% at baseline to 36.9% at end line, and at control sites increased slightly from 39.6 to 39.9%. The use of any family planning method increased at integrated sites from 67.5 to 73.4%, and at control sites from 60.8 to 69.7%.

Table 3

Table 3

Figure 2 presents the prevalence of contraceptive method use (use of more effective and less effective family planning and non-use of family planning) for women with encounters at integrated sites and control sites over the entire study period. The 12 months of the study following completion of training at all sites is divided into 3-month periods. By the final period of observation, the prevalence of more effective family planning use increased at both integrated and control sites, with a greater increase at integrated sites.

Fig. 2

Fig. 2

The odds of more effective family planning use during the final 3 months of the study relative to baseline was 1.81 [95% confidence interval (CI) 1.24–2.63] for women at intervention sites compared to control sites (Table 3). The odds of condom use for this period was 0.64 (95% CI 0.35–1.19) for women at intervention sites compared to control sites, and the odds of dual method use during this period was 1.30 (95% CI 0.77–2.17) for women at intervention sites compared to control sites. The odds of use of any family planning method during this period was 0.90 (95% CI 0.50–1.60) for women at intervention sites compared to control sites. Imputation of data for long-acting and permanent methods resulted in a change to 842 of 66 650 records (1.3%) over the course of the study. Sensitivity analyses with and without imputation resulted in similar estimates (data not shown).

During the final year of the study, the pregnancy rate was 1.5 and 1.7 new pregnancies per 100 clinic visits at integrated and control sites, respectively [incidence rate ratio (IRR) 0.90, 95% CI 0.68–1.20] (Table 3). This corresponds to a rate of 5.5 (95% CI 4.8–6.4) and 6.1 (95% CI 4.6–8.2) pregnancies per 100 person-years of follow-up at integrated and control sites, respectively. A sensitivity analysis adjusting for baseline pregnancy rates revealed a similar result (data not shown).

Back to Top | Article Outline

Discussion

The trial showed that integration of family planning services into HIV care and treatment is associated with significantly higher use of more effective contraceptive methods. We did not see a change in use of any family planning method, probably due to the high prevalence of reported condom use at baseline. Although we did not see a significant reduction in pregnancy incidence during the study, it is likely that 1 year is too short a period of observation for this outcome. In addition, the study was not powered to detect a reduction in pregnancy. Our findings build on the results of a non-randomized trial in Kenya that found that non-condom contraceptive use increased and pregnancy incidence decreased among HIV-infected women participating in a clinical trial after launching a multipronged intervention that promoted family planning use and provision in the study clinic [19]. Additional analyses from our trial will examine the cost, acceptability and feasibility of integrated services, as well as the facility characteristics at integrated sites associated with higher levels of contraceptive use.

To the best of our knowledge, this is the first randomized trial to evaluate the effectiveness of integrating family planning services into HIV care and treatment, and our results provide encouraging evidence in support of the programmatic push toward integration. Whereas cluster-randomized trials are prone to certain bias [20], we believe that this study helps to demonstrate the value of this study design to answer questions regarding models of healthcare delivery. Specifically, we found significant increases in the use of more effective family planning at control sites during follow-up, probably as a result of the system strengthening provided by the study team, including efforts to ensure that commodities were always available at all sites.

We observed a small reduction in condom use at integrated sites, although the overall change in condom use at end line between integrated and control sites was not significantly different. Dual method use was stressed in patient counselling, and we observed that reported dual method use increased as women used more effective family planning methods, although this change was not significantly different between study arms. Studies from other settings have found a similar prevalence of dual method use [8,21], and clearly additional research is needed to identify interventions to increase condom use along with more effective family planning methods. Dual method use among HIV-infected women is critical for the prevention of heterosexual HIV transmission. Among people living with HIV in sub-Saharan Africa who are in stable relationships, up to half have an uninfected partner [22]. In addition, one study has suggested that women who use depomedroxyprogesterone acetate might have an increased risk of HIV transmission [23].

The study has several limitations. We have limited information about the study population from the EMRS and have more missing data from the time of enrolment in HIV care at integrated compared to control sites. Therefore, it is possible that unobserved differences between the two populations affected study outcomes, although the randomized design should eliminate selection bias and confounding. In addition, condom use was based on self-report, and social desirability bias may have increased reported consistent use. However, this bias is likely to be similar in both study arms. In addition, the small fee intermittently charged to patients seeking contraception at three of the control sites might have reduced use of more effective methods at these sites. There are additional limitations related to our pregnancy data, including the fact that not all incident pregnancies (especially early pregnancies) were likely detected during the follow-up period, and we did not have information on pregnancy intendedness. Finally, our findings are specific to the model of family planning–HIV integration studied here.

In addition to the cluster-randomized design, strengths of this study include the system-strengthening training that was performed at both intervention and control sites to isolate the effectiveness of integration. We also included facilities of different levels of care.

The rollout of integrated family planning and HIV services will require leadership at the community, regional, national and international levels. Advocates will need to focus their efforts on including integration in national plans and budgets. Country leaders and ministries of health will need to take ownership and establish a ‘minimum package’ of integrated services, as has been done in Kenya [14,15]. In addition, the international donor community will need to work closely with stakeholders to support family planning–HIV integration. Whereas this study demonstrates that the ‘one-stop shop’ improves the use of more effective contraception among HIV-infected women, only a concerted effort by governments, donors and other key stakeholders will move family planning–HIV integration forward to help meet the unmet need for family planning services in sub-Saharan Africa.

Back to Top | Article Outline

Acknowledgements

D.G. originated and supervised study, participated in study site training, led study design, data interpretation and writing. M.O. coordinated all aspects of study implementation and monitoring, contributed to data interpretation and writing. S.J.N. contributed to study design, training, data interpretation and writing. C.B. led data management and analysis, and contributed to writing. S.B.S. participated in study design, supervised data management and analysis, and contributed to writing. R.L.S. performed literature search, developed study and training materials, monitored study and contributed to writing. E.A.B. supervised training and data collection, contributed to data interpretation and writing. C.R.C. supervised all aspects of study, including development of the study design, analysis and writing of the manuscript.

We thank the Kenyan men and women who participated in the study. We acknowledge the important logistical support of the KEMRI-UCSF Collaborative Group and especially Family AIDS Care and Education Services (FACES). We gratefully acknowledge the Director of KEMRI, the Director of KEMRI's Centre for Microbiology, and the Nyanza Provincial Ministries of Health for their support in conducting this research. We also thank Norah Bett, Catherine Amayo, Christine Auma, Beatrice Hongo, Rose Odeny, Elizabeth Onyango, Gorrety Akinyi Ong’udi, Shiko Kingori, Katie D. Schwartz, Harriette Sande, Pauline Wekesa, Janet Opala, Christine Alindi, Violet Nasenya, Micahel Onyalo, Mercy Njuguna, Kevin Owuor, Benard Otieno, Charles Kibaara, Jane Ivory Ernstthal, Meghan Flanagan, the data clerks, the Community and Clinic Health Assistants, and the Community Advisory Group members for their important contributions to this research.

Back to Top | Article Outline

Conflicts of interest

None of the authors declared a conflict of interest. This work was supported by The Bill and Melinda Gates Foundation (grant number OPPGH5361) and the Tides Africa Fund (grant numbers TFR10-01364 and TFR08-02986). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Back to Top | Article Outline

References

1. WHO, UNICEF and UNAIDS. Global HIV/AIDS response. Epidemic update and health sector progress towards Universal Access, Progress report 2011. http://www.unaids.org/en/media/unaids/contentassets/documents/unaidspublication/2011/20111130_UA_Report_en.pdf]. [Accessed 6 November 2012]
2. Schwartz SR, Rees H, Mehta S, Venter WD, Taha TE, Black V. High incidence of unplanned pregnancy after antiretroviral therapy initiation: findings from a prospective cohort study in South Africa. PLoS One 2012; 7:e36039.
3. Rochat TJ, Richter LM, Doll HA, Buthelezi NP, Tomkins A, Stein A. Depression among pregnant rural South African women undergoing HIV testing. J Am Med Assoc 2006; 295:1376–1378.
4. Homsy J, Bunnell R, Moore D, King R, Malamba S, Nakityo R, et al. Reproductive intentions and outcomes among women on antiretroviral therapy in rural Uganda: a prospective cohort study. PLoS One 2009; 4:e4149.
5. United Nations Population Fund (UNFPA). The Glion call to action on family planning and HIV/AIDS in women and children, 2004. http://www.unfpa.org/public/publications/pid/1435. [Accessed 1 March 2013]
6. Duerr A, Hurst S, Kourtis AP, Rutenberg N, Jamieson DJ. Integrating family planning and prevention of mother-to-child HIV transmission in resource-limited settings. Lancet 2005; 366:261–263.
7. Brocklehurst P, French R. The association between maternal HIV infection and perinatal outcome: a systematic review of the literature and meta-analysis. BJOG 1998; 105:836–848.
8. Mutiso SM, Kinuthia J, Qureshi Z. Contraceptive use among HIV infected women attending Comprehensive Care Centre. East Afr Med J 2008; 85:171–177.
9. Wilcher R, Petruney T, Reynolds HW, Cates W. From effectiveness to impact: contraception as an HIV prevention intervention. Sex Transm Infect 2008; 84 (Suppl 2):ii54–ii60.
10. Lindegren ML, Kennedy CE, Bain-Brickley D, Azman H, Creanga AA, Butler LM, et al. Integration of HIV/AIDS services with maternal, neonatal and child health, nutrition, and family planning services. Cochrane Database Syst Rev 2012; 9:CD010119.
11. McCarraher DR, Vance G, Gwarzo U, Taylor D, Chabikuli ON. Changes in contraceptive use following integration of family planning into ART Services in Cross River State, Nigeria. Stud Fam Plann 2011; 42:283–290.
12. Kosgei RJ, Lubano KM, Shen C, Wools-Kaloustian KK, Musick BS, Siika AM, et al. Impact of integrated family planning and HIV care services on contraceptive use and pregnancy outcomes: a retrospective cohort study. J Acquir Immune Defic Syndr 2011; 58:e121–e126.
13. Lewis Kulzer J, Penner JA, Marima R, Oyaro P, Oyanga AO, Shade SB, et al. Family model of HIV care and treatment: a retrospective study in Kenya. J Int AIDS Soc 2012; 15:8.
14. Ministry of Medical Services and Ministry of Public Health and Sanitation. Family planning orientation package for HIV service providers. Nairobi: Government of Kenya, 2009. http://integrationforimpact.org/wp-content/uploads/2012/05/National_FP-HIV_Integration_Package.pdf. [Accessed 26 November 2012]
15. Ministry of Medical Services and Ministry of Public Health and Sanitation. Minimum package for reproductive health and HIV integrated services. Nairobi: Government of Kenya, 2012. http://iframe.k4health.org/sites/default/files/Minimum%20Package%20for%20RH%20HIV%20Integrated%20Services.pdf. [Accessed 26 November 2012]
16. OpenMRS home page. http://openmrs.org/. [Accessed 6 November 2012]
17. Eisenberg DL, Allsworth JE, Zhao Q, Peipert JF. Correlates of dual-method contraceptive use: an analysis of the National Survey of Family Growth (2006–2008). Infect Dis Obstet Gynecol 2012; 2012:717163.
18. Central Bureau of Statistics (CBS) [Kenya], Ministry of Health (MOH) [Kenya], and ORC Macro. 2004. Kenya. Demographic and Health Survey 2003. Calverton, Maryland: CBS, MOH, and ORC Macro.
19. Ngure K, Heffron R, Mugo N, Irungu E, Celum C, Baeten JM. Successful increase in contraceptive uptake among Kenyan HIV-1-serodiscordant couples enrolled in an HIV-1 prevention trial. AIDS 2009; 23 (Suppl 1):S89–S95.
20. Puffer S, Torgerson D, Watson J. Evidence for risk of bias in cluster randomized trials: review of recent trials published in three general medical journals. Br Med J 2003; 327:785–789.
21. Seutlwadi L, Peltzer K. The use of dual or two methods for pregnancy and HIV prevention amongst 18-24-year-olds in a cross-sectional study conducted in South Africa. Contraception 2013; 87:782–789.
22. Chemaitelly H, Cremin I, Shelton J, Hallett TB, Abu-Raddad LJ. Distinct HIV discordancy patterns by epidemic size in stable sexual partnerships in sub-Saharan Africa. Sex Transm Infect 2012; 88:51–57.
23. Heffron R, Donnell D, Rees H, Celum C, Mugo N, Were E, et al. Use of hormonal contraceptives and risk of HIV-1 transmission: a prospective cohort study. Lancet Infect Dis 2012; 12:19–26.
Keywords:

family planning; HIV treatment; integration; Kenya

Supplemental Digital Content

Back to Top | Article Outline
© 2013 Lippincott Williams & Wilkins, Inc.