The integration of HIV and family planning services is supported by many donors and governments [1–3]. This endorsement is based on the assumption that both family planning and HIV services target individuals with sexual risk behavior. Combining these services has a potential to improve programmatic effectiveness and efficiency [4–10]. Additionally, family planning may reduce mother-to-child transmission of HIV by reducing unplanned pregnancies among individuals who are at risk for HIV infection [11–15]. Last, better access to family planning services may help to slow the HIV epidemic by increasing contraceptive use and reducing sexual risk behavior among HIV-positive patients [8,16–18] (J. Kagaayi, G. Kigozi, T. Lutalo, F. Makumbi, F. Nalugoda, Z. Namukwaya, et al., in preparation).
Studies [10,19] have shown that clients attending voluntary HIV counseling and testing (VCT) services have an unmet contraceptive need for family planning and are receptive to receiving contraceptive services. Additionally, cost-effectiveness analyses indicate that, if offered as a part of VCT programs, family planning would be more cost-effective than antiretroviral treatment for preventing mother-to-child transmission of HIV [13,14,20]. However, no study to date has shown that integrated service delivery increases contraceptive uptake among VCT clients.
Similar to most countries with a serious HIV/AIDS problem, Ethiopia considers VCT an essential part of its effort to prevent and treat AIDS . The value that the government attaches to VCT is reflected in the phenomenal increase in counseling and testing facilities, from 658 in 2004 to 1469 in 2008. Paralleling the increase in facilities is the number of people tested, 448 000 in 2004 and 4.6 million in 2008 . Ethiopia also has one of Africa's most ambitious reproductive health policies, including a target of increasing contraceptive prevalence from 14% in 2005 to 60% by 2010 . Considering the priority placed on VCT and family planning, it is not surprising that the government also has a policy to include family planning as a standard component of VCT [24,25].
Although the policy environment is quite favorable for integrating family planning into VCT, there are major impediments to integration. Ethiopia is a large, sparsely populated country, but its population of 80 million is increasing rapidly at an annual rate of 2.5% . The adult HIV prevalence is an estimated 2.1%, with over 1.1 million people infected . The most difficult hurdle for effective integration is a lack of resources. Donors have been very generous in supporting the country's HIV/AIDS program, especially the United States, which provided $852 million in the last 4 years under the U.S. President's Emergency Plan for AIDS Relief (PEPFAR) . External funding for family planning has not approached this level and, indeed, PEPFAR, the principal funder of VCT services, did not fund family planning during our study period.
The Voluntary HIV Counseling and Testing Integrated with Contraceptive Services (VICS) study is a proof-of-concept study designed to evaluate what happens when quality family planning services are introduced into VCT facilities. We aimed to assess whether counselors could feasibly offer family planning during VCT and whether clients would accept such services. In order to better understand the intervention's impact, we also examined what kinds of clients received contraceptive counseling and methods. Last, we aimed to assess the potential effect of the family planning intervention on clients' intent to use condoms consistently.
VICS is a pre and postintervention survey that was conducted in eight public sector VCT facilities in Oromia Region, Ethiopia, from 2006 to 2008. The intervention was implemented by Pathfinder International, Ethiopia, and included developing family planning messages for VCT clients, training counselors, ensuring contraceptive supplies in VCT facilities and monitoring services. Unlike typical family planning clients, VCT clients may not have an interest in or knowledge of contraceptives. As a result, family planning counseling messages were developed specifically for young, single and premarital clients and included basic information on family planning benefits and methods. Counselors were authorized to provide family planning counseling, condoms and pills during VCT sessions. Referrals were made to on-site family planning nurses for clinical methods, except when VCT counselors were also trained as nurses and could provide injectables. Service sites were chosen on the basis of available human resources, interest in participating and geographic proximity to Pathfinder's local implementing partners. The intervention and facilities were previously described in more detail elsewhere (D. Gillespie, H. Bradley, M. Woldegiorgis, K. Kidanu, S. Karklins, in preparation).
At the time of our study, VCT services were delivered in two parts. First, a counseling professional assessed the client's level of HIV risk and administered an HIV test. After the testing algorithm was completed by an on-site laboratory technician, the counselor provided the client with the HIV test result and further counseling depending on the client's serostatus. At the time of our study, clients waited up to several hours between pre and posttest counseling. This service delivery mode, however, is currently changing, as rapid tests are now widely available.
This study utilizes two cross-sectional rounds of data collection. Preintervention surveys were conducted between July and October 2006, during which time 4019 men and women attending standard VCT services in study clinics were interviewed. In 2008, approximately 18 months after the introduction of family planning services, 4027 additional clients were interviewed. Sample size requirements were calculated on the basis of estimation of a 35% difference in sex-stratified contraceptive uptake between HIV-negative and HIV-positive clients, assuming 15% HIV prevalence in study facilities at baseline.
Each client was interviewed twice: after pretest counseling and after posttest counseling when they knew their HIV status. Clients were eligible for the study if they were 15–49 years of age and consented to release their HIV results to the study. Clients were sampled consecutively, and of those who were approached, only 145 of 4138 (3.5%) men and 186 of 4239 (4.4%) women refused participation. We did not collect demographic information on these refusals. Due to the limited number of interviewers and the time required to complete each interview, 2 h on an average, it was logistically impossible to interview many clients. A review of the facilities' service statistics indicated no statistically significant differences between the demographic characteristics of participating and nonparticipating clients who accepted family planning services. Interviewers were sex-matched to VCT clients and blinded to clients' HIV test results, which were collected from facility records.
Counselors were interviewed about their socio-demographic characteristics, family planning knowledge and attitudes towards work, before and after their training in family planning. Information on facility characteristics, such as number of counseling rooms, was collected from facility supervisors at the time of both surveys, and the information provided was verified by the research staff. We obtained written informed consent from all study participants, and both the Johns Hopkins Bloomberg School of Public Health Institutional Review Board and the Ethiopian Science and Technology Commission gave ethical approval for this study.
This analysis has three main outcome variables: whether the client received contraceptive counseling during VCT, whether the client obtained a contraceptive method during VCT and whether, after VCT, the client intended to use condoms consistently during the next 2 months for disease prevention or transmission. All outcomes are dichotomous and based on self-report.
Our regression analysis of the first two outcomes, receipt of contraceptive counseling and obtaining a contraceptive method, was limited to those attending VCT after intervention. Preintervention study participants were not included as family planning services in VCT were essentially nonexistent at that time. For the consistent condom use outcome, clients from both pre and postintervention samples were used. However, we excluded individuals who intended to be abstinent for at least the next 2 months.
Clients were linked to VCT providers and facilities at the time of data collection. Data from client, provider and facility questionnaires were merged. Prefamily planning training provider surveys were linked to preintervention clients and postfamily planning training provider surveys linked to postintervention clients. All clients from one VCT facility were excluded because the facility did not implement the intervention, for example, only 1% of clients received family planning counseling. Thus, 456 of 4019 (11%) clients were excluded at preintervention, and 500 of 4027 (12%) clients were excluded at postintervention.
Some VCT counselors did not complete provider questionnaires because they left their positions before the study interviews were conducted. Clients whose providers were not interviewed were excluded from the present analyses. Before the intervention, eight of 23 VCT providers were missing information, resulting in 1127 of 3563 (31.6%) clients excluded. At postintervention, one of 17 providers did not complete the questionnaire, resulting in 99 of 3527 (2.8%) clients excluded. We compared modeled outcomes including clients with and without provider-level information. None of the associations between client characteristics and study outcomes were different in terms of direction or statistical significance, and only small differences in the magnitude of associations were observed. Because many variables were measured during the second interview, after VCT, we then excluded 57 (2%) clients at preintervention and 54 (<2%) clients at postintervention who did not complete this interview, resulting in a final sample size of 2379 preintervention (27% of facilities' eligible clients) and 3374 postintervention clients (28%).
Bivariate analyses were conducted separately for men and women to examine associations between individual, provider and facility-level variables and each study outcome. Associations that were statistically significant at α equal to 0.10 were examined further in multivariate analyses. We found significant differences in intervention impact by facility, with the percentage of clients receiving contraceptive counseling varying widely from 19 to 61%. Therefore, we utilized multilevel models with provider and facility-level random intercepts to account for clustering of observations within providers and facilities.
Our data have a three-level hierarchical structure, with clients nested within VCT providers, who are nested within VCT facilities. We used generalized linear and latent mixed models (GLLAMM) in STATA version 10.0 (STATA Inc., College Station, Texas, USA) to create multilevel logistic regression models for the three outcomes, and each model was run separately by sex.
We began by fitting a null model with two random intercepts for each outcome. We then added separate blocks of related covariates such as client demographic characteristics or counselor attitudes. Variables significantly related (P < 0.10) to the outcome within the block of variables were added into the sex-stratified full models.
In order to determine whether excluding clients without provider information made a difference to the associations found in our three-level models, we ran two-level models with only facility-level random intercepts on the full dataset for each outcome. This sensitivity analysis indicated that clients' missing provider-level information were not substantially different from those included in the three-level models.
Table 1 shows the characteristics of VCT clients before and after introduction of family planning in VCT. Clients are young, mostly single, highly educated and urban, and these characteristics tend to be more pronounced among the postintervention sample. For example, the mean age of women was 24 years in the preintervention survey and 22 years at postintervention (P < 0.01). In the postintervention survey, 78% of men had never been married compared with 69% in the earlier survey. Both men and women were well educated, with 71% of men and 61% of women having attained secondary or higher education.
Notably, many clients in our study were not sexually active. In the postintervention survey, 39% of women and 29% of men had never had sex, whereas only 29% of women and 22% of men had sex during the 30 days preceding testing. Of currently sexually active clients, significantly more were contracepting after the intervention, with 74% of women and 20% of men already using contraception at the time of their HIV test. Contraceptive use was much lower among men, who were only asked to report their own contraceptive use and not their partner's. HIV prevalence was 7.6% among women and 4.3% among men in the postintervention sample, which was lower than in the preintervention sample most likely due to demographic changes in the client profile.
Change over time
Table 2 shows the change over time in main study outcomes. We observed major improvements in contraceptive counseling. After integration, nearly 41% of women and 29% of men received contraceptive counseling compared with only 2 and 3%, respectively, before intervention. Despite improved counseling, contraceptive uptake was modest. Only 6.5% of women and 6% of men accepted contraceptives, still a significant improvement. Contraceptive provision was limited to condoms for men. Very few women received hormonal methods.
There were no significant pre and postintervention differences in condom use intent. However, in the adjusted analyses (see Table 3), men were more than twice as likely to intend to use condoms consistently after the introduction of family planning services.
Tables 3–5 show bivariate and multivariate findings for three client-level outcomes: receiving contraceptive counseling, receiving contraceptive methods and intending to use condoms consistently after counseling. Adjusted odd ratios come from multivariate models controlling for other covariates in the model and provider and facility-level random intercepts, so the results should be interpreted as representing a given provider and facility.
The adjusted analyses revealed very few differences in study outcomes by the demographic variables in Table 1. Among female clients, there were no demographic differences between those who did and did not receive contraceptive counseling or methods. However, married men were 2.6 times more likely than single men to receive contraceptive counseling (results not shown). Seeking VCT with a sexual partner was more important than marital status in terms of our study outcomes. Clients attending VCT with a partner were 2.5 and 3.4 times more likely (men and women, respectively) to receive contraceptive counseling.
Other client-level variables were differentially associated with receiving contraceptive counseling by sex. Contracepting women were 1.8 times more likely to receive contraceptive information than those not contracepting. HIV status was important for men, with HIV-positive men three times more likely to receive family planning information than those testing negative.
Although there were few differences in the types of clients receiving contraceptive information, clients with more perceived or actual sexual risk were more likely to obtain contraceptive methods. For example, women and men testing simply ‘to know’ their status were approximately 50% less likely to receive methods than those testing due to perceived HIV risk. Currently sexually active women were four times more likely to obtain contraceptive methods than those who had never had sex, whereas sexually active men were nearly 12 times more likely to receive contraceptive methods. Notably, men testing with female partners were 74% less likely to obtain methods than those testing alone, although they were more likely to receive contraceptive counseling.
Not surprisingly, receiving contraceptive counseling increased the odds that a client would obtain a method. Women and men who received contraceptive counseling were six and 3.9 times more likely, respectively, to obtain methods than those not receiving counseling.
Clients with more HIV risk were also more likely to express condom use intentions. Among women not intending to abstain from sex for the next 2 months, those testing ‘to know’ their HIV status or testing before new sexual relationships were 39 and 54% less likely, respectively, to intend consistent condom use than those testing because of perceived HIV risk. Similarly, men testing ‘to know’ their status or before entering a new sexual relationship were significantly less likely to intend consistent condom use than those who suspected they were HIV positive. Sexually active women were more than twice as likely as those who had never had sex to intend consistent condom use. Last, HIV-positive women and men were 3.3 and nine times more likely, respectively, to hold consistent condom use intentions after counseling compared with HIV-negative clients.
Study outcomes varied most consistently by two provider-level variables: the counselor's level of knowledge about contraceptive side effects and perceived adequacy of contraceptive supplies. For example, men receiving services from providers who were more knowledgeable about contraceptive side effects were 2.5 times more likely to receive contraceptive counseling than men counseled by providers with less knowledge. Women were 1.6 times more likely to intend consistent condom use if their VCT provider was more knowledgeable about contraceptive side effects.
The VCT provider's perception about adequacy of contraceptive supplies was important for both contraceptive uptake and condom use intent outcomes. Women and men whose counselors felt that contraceptive supplies were adequate were two and four times more likely, respectively, to receive methods than those whose providers felt that supplies were inadequate. Male and female clients who were counseled by providers who felt that contraceptive supplies were adequate were also significantly more likely to intend to use condoms consistently.
None of the facility-level variables were significantly associated with study outcomes for women. However, male clients attending facilities with more rooms available for simultaneous counseling sessions and smaller client loads were much more likely to receive contraceptive counseling and methods. For each increase in the number of counseling rooms, men were 9.7 times more likely to receive contraceptive counseling and 3.9 times more likely to receive a method. Men attending facilities serving more than 15 clients daily were 88% less likely than those attending facilities serving fewer clients to receive contraceptive counseling and 93% less likely to receive a method.
For most of the outcomes, the variance of both the provider and facility-level random intercepts declined substantially between the null and final models (see Tables' footnotes). With a few exceptions, the variances of the provider and facility-level random intercepts in the final models were essentially zero, and the standard errors were similar in size or larger than the variance. Thus, much of the initial variability in outcomes between providers and facilities captured by the random intercepts in the null models was explained by the covariates in the final models.
The integration of family planning into VCT programs is based on the reasonable premise that persons testing for a sexually transmitted disease have a need for family planning services. However, we found a smaller than expected immediate contraceptive need among study clients. At the time of the postintervention survey, only 29% of women had been sexually active in the last 30 days, and 39% had never had sex. Additionally, 74% of sexually active women were already using contraceptives, and more than 60% of contracepting women were using modern, female-controlled methods (not shown).
Why were people at a very low or no risk testing for HIV? Although this question will be examined in more detail in the future, a preliminary analysis suggests that these clients feel they are at a risk of being infected with HIV through nonsexual behavior. Although over 95% of all clients knew how HIV is transmitted, 95% also believed that ‘sometimes people are infected with HIV for no apparent reason’, and only 21% believed people who abstain from sex are completely protected from HIV.
Considering the low risk this young and largely sexually inactive population had for unwanted pregnancy, the family planning outcomes achieved are noteworthy. The number of women and men receiving counseling increased, by 20-fold and 10-fold, respectively. Contraceptive uptake was modest after the integration of family planning, 6.5% for women and 6% for men, yet this was a significant improvement from zero, making this study the first to show significantly increased contraceptive uptake as a result of an integration intervention.
Nearly 60% of women and 70% of men did not receive contraceptive counseling, a factor strongly associated with obtaining contraceptive methods. Our findings suggest that, in addition to client characteristics, counselor and facility-level factors should be considered carefully when implementing integration programs. Men attending facilities with lower client loads and more counseling rooms were much more likely to receive contraceptive counseling and methods. Adequacy of contraceptive supplies also emerged as an important variable. Although concrete steps were taken to ensure a dependable supply of contraceptives, the perception of inadequate supplies persisted among counselors, and this perception was highly associated with their clients' contraceptive uptake and condom use intentions.
However, and perhaps most importantly, it is clear that there are subgroups of clients that are more receptive than others to receiving contraceptives in VCT, and providers should prioritize these clients in terms of contraceptive service delivery. Sexually active men and women and those with more perceived HIV risk were more likely to obtain contraceptive methods. For example, currently sexually active women and men were four and 12 times more likely, respectively, to accept contraceptives than clients who never had sex. The benefits of integrating family planning and VCT services may be more pronounced among higher risk populations.
The study's major limitation is its dependency on two cross-sectional surveys. Because clients are not followed up, little can be said about attitudinal and behavior changes over time. We measured contraceptive uptake rather than continuation, and condom use intent rather than subsequent use. Assessing the true effectiveness of an integrated approach requires a prospective study that would measure contraceptive use and unintended pregnancy.
We relied on clients' self-reported sexual activity level and receipt of contraceptive counseling and methods, which may be subject to bias. Because clients were only asked whether or not they discussed contraceptive methods in VCT, we are unable to distinguish between clients who refused family planning information vs. those who were not offered information. Additionally, we excluded clients who were missing provider-level information. The characteristics of these providers may have been different from those who stayed in their positions, introducing potential selection bias. However, the associations between client and facility-level variables and our study outcomes did not differ significantly by whether or not provider-level information was missing.
Last, the study's client population and service sites may not be representative of other parts of Ethiopia. Because our facilities were not sampled using probability methods, they may have performed better or been different in terms of client catchment populations than other VCT facilities. Although such representation is not necessary for a proof-of-concept study, it does mean that extrapolating from the surprising findings of high contraceptive use among the sexually active clients and the very high percentage of nonsexually active clients is difficult.
Our findings call to mind the wise, but often ignored, maxim ‘know your epidemic, know your response’ . Although this analysis suggests that clients with more sexual risk are receptive to receiving family planning services in VCT, testing for a sexually transmitted disease and providing contraceptives to a population in which those who have never had sex outnumber the sexually active is not likely to be the optimal use of resources. An education and mobilization campaign that focuses on what constitutes risk behavior and who should be tested could well result in a client profile more in need and desirous of HIV testing and family planning services.
We appreciate the support of Pathfinder International, Ethiopia, and Chandrakant Ruparelia, Johns Hopkins Program for International Education in Gynecology and Obstetrics (JHPIEGO), Baltimore, Maryland, USA.
This study was funded by The David and Lucile Packard Foundation, The William and Flora Hewlett Foundation and The Bill and Melinda Gates Institute for Population and Reproductive Health. The funders had no role in the design or conduct of the study.
Conflicts of interest: None.
1. Strachan M, Kwateng-Addo A, Hardee K, Subramaniam S, Judice N, Agarwal K. An analysis of family planning content in HIV/AIDS, VCT, and PMTCT policies in 16 countries
. 2004. Policy Working Paper Series No. 9.
2. World Health Organization. The Glion call to action on family planning and HIV/AIDS in women and children
. Geneva, Switzerland: WHO; 2004.
3. United Nations Population Fund (UNFPA). The New York call to commitment: linking HIV/AIDS and sexual and reproductive health
4. Maggwa BN, Askew I. Integrating STI/HIV management strategies into existing MCH/FP programs: lessons from case studies in East and Southern Africa
. Nairobi, Kenya: The Population Council; 1997.
5. O'Reilly KR, Dehne KL, Snow R. Should management of sexually transmitted infections be integrated into family planning services: evidence and challenges. Reprod Health Matters 1999; 7:49–59.
6. Dehne KL, Snow R, O'Reilly KR. Integration of prevention and care of sexually transmitted infections with family planning services: what is the evidence for public health benefits? Bull World Health Organ 2000; 78:628–639.
7. Berer M. Integration of sexual and reproductive health services: a health sector priority. Reprod Health Matters 2003; 11:6–15.
8. 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.
9. Maharaj P, Cleland J. Integration of sexual and reproductive health services in KwaZulu-Natal, South Africa. Health Policy Plan 2005; 20:310–318.
10. Reynolds HW, Beaston-Blaakman A, Burke H, Janowitz B, Nasution M, Reuben E, et al
. Integrating family planning services into voluntary counseling and testing centers in Kenya: operations research results
. Arlington, Virginia: Family Health International; 2006.
11. Allen S, Serufilira A, Gruber V, Kegeles S, Van de Perre P, Carael M, Coates TJ. Pregnancy and contraception use among urban Rwandan women after HIV testing and counseling. Am J Public Health 1993; 83:705–710.
12. King R, Estey J, Allen S, Kegeles S, Wolf W, Valentine C, Serufilira A. A family planning intervention to reduce vertical transmission of HIV in Rwanda. AIDS 1995; 9(Suppl 1):S45–S51.
13. Stover J, Fuchs N, Halperin D, Gibbons A, Gillespie D. Adding family planning to PMTCT sites increases the benefits of PMTCT. USAID issue brief
. Washington, DC: USAID, Population and Reproductive Health; 2003.
14. Sweat MD, O'Reilly KR, Schmid GP, Denison J, de Zoysa I. Cost-effectiveness of nevirapine to prevent mother-to-child HIV transmission in eight African countries. AIDS 2004; 18:1661–1671.
15. Hoffman IF, Martinson FE, Powers KA, Chilongozi DA, Msiska ED, Kachipapa EI, et al
. The year-long effect of HIV-positive test results on pregnancy intentions, contraceptive use, and pregnancy incidence among Malawian women. J Acquir Immune Defic Syndr 2008; 47:477–483.
16. Askew I, Berer M. The contribution of sexual and reproductive health services to the fight against HIV/AIDS: a review. Reprod Health Matters 2003; 11:51–73.
17. Best K. Averting HIV-infected births. Network 2004; 233:21.
18. Kane MM, Colton TC. Integrating SRH and HIV/AIDS services: Pathfinder International's experience synergizing health initiatives. Watertown, Massachusetts: Pathfinder International; 2005.
19. Hoke TH, Reuben EB. Assessment of how family planning and HIV services can be integrated in Zimbabwe. Durham, North Carolina: Family Health International; 2006.
20. Reynolds HW, Janowitz B, Homan R, Johnson L. The value of contraception to prevent perinatal HIV transmission. Sex Transm Dis 2006; 33:350–356.
22. Ethiopian News Agency. ‘Ministry Says Number of People Undergoing VCT on Increase’. 2009. http://www.etharc.org
. [Accessed 24 April 2009]
Keywords:© 2009 Lippincott Williams & Wilkins, Inc.
condom use; family planning; integration; multilevel; voluntary HIV counseling and testing