HIV prevalence in Kenya in 2007 was estimated at 7.4% in the general population and 8.7% among women . Despite the proportion of Kenyans testing for HIV doubling over the previous 4 years, due in large part to a strong programme of client-initiated voluntary counselling and testing (VCT), in 2007 only one-third of Kenyans reported having ever tested . To date, relatively little attention has been paid, in Kenya and elsewhere, to systematically testing the opportunities for offering HIV testing and counselling (HTC) within reproductive health services.
To expand access to and use of HIV testing, especially in generalized epidemics, WHO and UNAIDS [2,3] have developed guidance for including ‘provider-initiated testing and counselling’ (PITC) in various health facility settings and to various patient groups, including those attending for family planning. Concerns have been raised about potential rights infringements, quality of counselling and stigmatization , but little evidence exists of the feasibility, acceptability or effectiveness of introducing PITC, beyond experiences with antenatal care as an entry point for prevention of mother-to-child transmission (PMTCT). PITC is supported within the newly revised HTC policy guidelines in Kenya .
The 2003 Kenya Demographic and Health Survey (DHS) showed that 38% of all sexually active women (i.e. had intercourse in the previous month) and 44% of unmarried sexually active women reported currently using a modern contraceptive method . As the majority of Kenyan women use the injectable and pills, the majority of these women regularly attend clinics for family planning services. There are, therefore, a sizeable proportion of sexually active women accessing family planning services who could benefit from being offered HTC. Whether this would be feasible, acceptable and what effect it might have on use of HTC and other HIV prevention service was the objective of this study, which was undertaken from 2005 to 2007. Given concerns that integrating HIV services may adversely affect the quality of the family planning services, this study also assessed the effect of the intervention on the family planning services themselves; the results of this assessment are reported elsewhere .
The policy environment in Kenya has been highly conducive to linking HIV/AIDS services with reproductive health services for several years [8,9] and a Task Force jointly convened by the Division of Reproductive Health (DRH) and the National AIDS and STI Control Programme (NASCOP) coordinates services integration within the ministries of health (MOH). A Reproductive Health and HIV/AIDS strategy has recently been finalized , which recommends several bi-directional linkages, including HIV services into family planning and family planning services within HIV services.
This policy framework guided the operations research study described here. An intervention was developed and tested that linked HTC (and other HIV prevention services) with existing family planning services and was evaluated using several HIV-related outcome measures. Two versions of the model were compared in response to the MOH's expressed interest in increasing PITC in different settings.
The intervention strengthened the quality of the family planning consultation itself using the ‘Balanced Counselling Strategy Plus’ tool (BCS+) . The original BCS was demonstrated to be effective in improving the quality of family planning counselling  and consists of an algorithm to guide the provider through a semi-structured consultation to identify a contraceptive method appropriate for the client's needs; a set of method-specific cards with information for counselling the client; and a method-specific brochure to be taken home by the client. The BCS was adapted for use in high sexually transmitted infection/HIV (STI/HIV) settings and renamed the BCS Plus, by adding steps to the algorithm that enable the provider and client to discuss STI/HIV transmission and prevention, conduct a risk assessment, discuss dual protection and discuss and offer opportunities for HTC [13,14].
Two models for integrating HIV testing into the family planning service were compared. For the ‘referral’ model, those family planning clients expressing an interest in having an HIV test after the HIV counselling were referred by the family planning provider to a VCT clinic, either on-site (if available) or elsewhere, and given a referral voucher to take with them. The ‘testing’ model was implemented in facilities that did not have on-site testing available by building the capacity of the family planning providers and family planning unit in these facilities to provide HTC (using a rapid test) in accordance with national guidelines; this testing facility was also made available to anyone seeking VCT at these facilities.
Staff in the intervention clinics were updated on contraceptive methods, STIs and HIV, reproductive rights, informed choice and consent, safe sex and dual protection, values clarification, risk assessment and reduction, use of the BCS+, record keeping, logistics management and client referral. The residential training was 5 days for the 47 providers implementing the referral version and 9 days for the 28 providers requiring training in PITC for the testing version. Trainees' competency was tested before and after the training, and there was a marked improvement for both groups. District health management teams (DHMTs) were supported to plan for and procure key commodities and supplies. Existing family planning registers were revised to record services relevant for the study so as to allow monitoring and evaluation.
A prospective prepost intervention design without a control was used. The pre and post design allowed for a comparison between samples of clients not exposed and exposed to the intervention. A control was not included because the key outcome of the intervention, the offer of HIV testing during family planning consultations, was known not to be happening in any other facilities. Moreover, the short period of evaluation (10 months, May 2006 to February 2007) reduced possible exposure of the postintervention sample to exogenous factors that may have influenced their behaviour during the family planning consultation differently than those in the preintervention sample.
Two districts in Central Province were purposively selected by the MOH. Districts were the primary sampling level because the DHMT plans, budgets and manages health services and so programmatic lessons from the study would be relevant for guiding future implementation by these and other DHMTs. One district (Nyeri) represented a setting that had only a few (3) HIV testing facilities, so that the MOH could address situations with limited access to testing; the ‘testing’ model was introduced in nine facilities, none of which had on-site testing. Another district (Thika) was selected to represent settings in which testing facilities were relatively widely available (16 in 2005), so that the MOH could address situations in which clients could be referred; the ‘referral’ model was introduced into 14 facilities (Box 1).
Criteria for selecting those health facilities in which the intervention would be introduced were determined through consultations between the DRH, NASCOP, the two DHMTs and the Provincial Health Management Team. To implement the intervention as designed and to recruit sufficient clients into the study, only facilities that met the criteria of a minimum of 100 family planning clients per month (new and revisit) and that had a minimum of two providers qualified in, and currently providing, family planning services were eligible for inclusion.
For the referral model in Thika district, two subsets were selected from among the 14 eligible facilities. One set, comprising two hospitals and eight health centres, had on-site testing but not within the family planning unit; referral here could be either to the on-site testing or to another facility. The other set, comprising the four dispensaries, had no on-site testing and so all referrals were to other facilities.
In Nyeri district, the DHMT decided to select one facility from each of the seven administrative divisions within the district to increase access across the whole district. First, all eligible facilities were identified and then the facilities with the highest family planning client load for each division (two facilities with high family planning client loads were selected in one division) were selected. The provincial general hospital was also selected because of its high client load, bringing the total to nine study facilities.
To evaluate the intervention's effect on the quality of counselling and on the key outcome variables, observations of samples of family planning consultations were undertaken using a ‘health facility assessment’ methodology that has been validated previously . Consultations were randomly selected by sequentially inviting family planning clients to participate in the study immediately prior to their consultation until the desired number was achieved. After obtaining the provider's consent to observe their consultations, clients were recruited through a standardized verbal invitation to participate that obtained their informed consent and ensured confidentiality through not recording their (or the provider's) names. A trained nurse observed the family planning client–provider interactions and used a structured checklist with a series of standard items to record the quality of care received. On exit, clients were then invited to participate in an interview by a trained research assistant using a structured questionnaire about the acceptability and quality of services received, their fertility intentions and experiences with HIV prevention behaviours.
Due to resource limitations and the nature of the data collection procedures, it was not possible to prospectively stratify sampling by new and revisit clients. Approximately, 24 observations of family planning consultations were undertaken at each facility, with the expectation (based on record reviews) that approximately five of these would be new clients; this would provide aggregate sample sizes of 45 new clients for the testing model and 70 new clients for the referral model. Using a one-tailed test, 95% confidence interval, 80% power and a design factor of 2, these samples would be sufficiently large to detect a statistically significant difference between the pre and post intervention measures for each model of 25% points, a magnitude of change programme managers considered sufficient for them to use the finding for decision-making. The sample sizes achieved are given in Table 1. The smallest sample sizes for new clients were in the testing model (27–29), which decreased the study's ability to detect a statistically significant difference to 35% points for this subsample.
To describe the acceptability of the intervention and learn about the feasibility of implementation, focus group discussions (FGDs) were held before and again after its introduction. Two FGDs were held with providers, one FGD in each district, and four FGDs with family planning clients, one at a health centre and one at a dispensary in each district.
Data entry screens for all quantitative data were set up in Epidata and double entered. Analysis used Stata 9.2 to generate descriptive statistics together with the relevant tests of significance (tests of proportions for percentages and the Student's t-test for means). The indicator scores were obtained as the additive sum of items representing specific items observed during consultations. Qualitative data were analysed manually using content analysis techniques.
All data were collected following internationally accepted ethical principles, by obtaining the participants' informed consent, ensuring anonymity by not requesting or recording their name or other identifiers, informing participants of protections such as the confidentiality arrangements for data storage and analysis, and offering the opportunity to not participate in the study without jeopardy. The study was considered by the Kenyan Ministry of Health to be an evaluation of already approved services and so did not require review by a national ethical board. This study also complied with the ethical review procedures of the Population Council's International Review Board (IRB).
Before the intervention, the majority of facilities had the minimum basic equipment, supplies, infrastructure and staff required to integrate the new services; for those missing basic equipment and supplies at baseline, NASCOP and DRH provided sufficient amounts to bring all facilities to the minimum requirements needed to implement the intervention.
All clients were female. Most of them were married, around 30 years of age, and had two or three living children. Most of the clients had completed 5–8 years of schooling. Between 13 and 19% of clients were ‘new’, that is, clients obtaining family planning services from the clinic for the first time; the remainder were ‘revisits’, that is, clients returning for either further commodities or for other services that include check up, information or advice. The majority of clients were using hormonal contraception, mostly injectables (70–78%).
Quality of counselling was assessed using a composite index that consisted of 26 indicators of actions observed to be taken by the provider, or issues discussed with the client, during the consultation. These indicators have been validated in previous studies that have assessed quality of integrated services . Overall, the quality of care increased significantly, from a mean score of 9.65 to 15.84 per client (within a range of 0–26); full details of this analysis are reported elsewhere [7,16].
There were changes in the amount of time spent with clients. Before the intervention, the median consultation time was 10 min for all family planning clients (15 min for new and 10 min for revisit clients). Implementing the intervention added, on average, 2–3 min per consultation, across both models. For clients participating in the testing model who decided to have an HIV test with the family planning provider, the median time increased by 7 min, from 10 to 17 min; this included both carrying out the rapid test and counselling the client on the result. Although the result was not recorded by the observer and so it is not possible to compare the median duration of counselling for positive or negative outcomes, very few clients tested positive.
Improvements in counselling for STI/HIV issues were observed in both versions across all five items assessed, but the low postintervention scores indicate more attention is needed (Table 2). Providers seem to be most comfortable having general discussions about STIs and HIV and to a lesser extent discussing risk factors, but family planning providers remain unwilling to engage their clients in a discussion of their personal sexual behaviour.
All items relating to counselling on condom use improved significantly (Table 3), though this aspect had been particularly weak before the intervention. Much improvement is still needed, but these changes indicate that providers are willing and able to discuss condom use in a dual protection perspective with family planning clients.
Condom use at last sex remained low across all groups, at 0–3%. When asked about condom use in the previous month, there was no change among new clients (around 5–8%), but among revisit clients, there was a significant increase, from 4 to 13%, which may indicate an effect from exposure to the intervention. Moreover, much of this increase reflects initiation of dual method use, with the proportion of revisit clients currently using a condom with another family planning method increased significantly from 1 to 6%.
As Table 4 indicates, all indicators concerning HTC during family planning consultations increased significantly, suggesting that including discussions about HIV testing during family planning consultations is both feasible and acceptable.
Substantial proportions of clients were offered an HIV test during the consultation (Table 5a and b). Providers implementing the testing version were more likely to offer their clients an HIV test than those in the referral model; FGDs with providers suggest that this may be because they were able to offer the test themselves. The lower proportion of revisit than new clients being offered testing may also be because most revisit clients make regular visits to the clinic for resupply of injectables or pills and so the provider may assume (or know) that the client had been offered testing recently and so not make an offer. In addition, counselling for revisit family planning clients is usually less thorough than for new clients and this may also influence their decision about offering HIV testing.
Among those clients offered the opportunity of having an HIV test, 57% of new clients and 66% of revisit clients indicated that they would like to have a test. These proportions were higher among clients attending clinics using the referral model, which may substantiate statements by clients during FGDs that they preferred to be tested by someone else or at another time; as they are not being tested immediately, a referral gives them the opportunity to think it through further and/or discuss with their partner.
Table 5 indicates that, overall, approximately one quarter of both new and revisit family planning clients chose to have an HIV test. Overall, the testing model appears to be more effective, with 35% of all clients being tested compared with 20% of clients in the referral model receiving a referral voucher for a test; there were no differences between new or revisit clients for either model.
For the referral model, although there was little difference in the proportions of clients being offered a test at facilities with on-site referral compared with referral to another facility, there were significant differences between these two versions of the referral model in the proportions of clients choosing to have a test and receiving a referral voucher – overall, 17% of clients at the facilities referring on-site took a voucher compared with 29% at facilities referring off-site, which perhaps reflects a preference to have the opportunity to reflect further on the decision before having a test.
Overall, approximately, one-third of the family planning clients (both new and revisit) choosing to have an HIV test had never had a test before; this masks substantial differences between the two models, however, with only 15% of those in the testing model being first-time testers, compared with 51% of those in the referral model. Thus, although the testing model proved more effective in terms of the proportion of family planning clients being tested, the vast majority of these had previously had a test, suggesting that this model is attractive to those who are familiar with the procedure and would appreciate the opportunity to have a test within their family planning consultation. The referral model would seem to appeal to those for whom the test would be their first as it gives them the opportunity to think and discuss before proceeding with taking a test.
Approximately, two-thirds of new clients reported previously having had an HIV test and this did not change over time, nor was there any difference between the models. This suggests that during the duration of the study, there were no external influences that may have increased likelihood of being tested among new clients. For revisit clients, there were significant increases in the proportions reporting ever having had a test, particularly among those exposed to the testing version. As most revisit clients will have been exposed to the intervention during the preceding 10-month period, this would suggest that the intervention may have increased the likelihood of a revisit family planning client ever being tested.
The study had limitations. The number and profiles of participating health facilities were different for each version of the model, as sampling reflected the need to generate results that would be both statistically valid and serve the planning needs of the DHMTs. National HIV surveillance data estimate HIV prevalence at 10% in Thika compared with 2.4% in Nyeri, which may influence client awareness, behaviours and attitude towards HIV testing in each district. Thika is primarily an urban and periurban district, whereas Nyeri is largely rural, which influences the mobility of health facility clients. The inclusion criteria meant that only facilities with high family planning client loads participated, which may not be representative of other family planning clinics in Kenya. The inability to confirm whether family planning clients given referral vouchers were actually tested limits the validity of conclusions about the effectiveness of this model.
The study demonstrated that both versions were feasible to implement and acceptable to providers and clients and that the majority of facilities, at all levels, had the capacity to integrate HIV prevention counselling and HTC within existing family planning services. The additional time required for an integrated consultation (3 min) and for testing those wanting on-site rapid testing (7 min) was reasonable. The incremental cost per family planning client ranges from $5.60 (hospital) to $9.53 (dispensary)  and compares favourably with an estimated cost of $27 per client at stand-alone VCT . The BCS+ proved to be an effective algorithm for improving the quality of family planning counselling and for integrating PITC services. Both versions of the intervention are equally effective in increasing the number of women accessing HTC through offering the service during family planning consultations. The testing version may be more attractive to clients looking for a repeat test as they are familiar with the procedure and are comfortable being tested immediately.
The present study was made possible by the generous support of the American people through the President's Emergency Plan for AIDS Relief (PEPFAR) and the United States Agency for International Development's U.S. Agency for International Development (USAID) under the terms of Cooperative Agreement No. HRN-A-00-98-00012-00 and In-house Project No. 8816.53101. The contents are the responsibility of the authors and do not necessarily reflect the views of USAID or the United States Government. We wish to acknowledge the following persons and officials who played critical roles in the process of implementing the integration of voluntary counselling and testing within family planning services: Margaret Gitau, Mary Gathitu, Andrew Mboche, Dr Anthony Wanyoro, Peter Mohammed Njuguna, Dr John Njoroge and Judith Maua. Erick Oweya is specifically acknowledged for performing data analysis. We wish to thank also Dr Marsden Solomon and Dr Rick Homan of FHI, members of the district health management teams (DHMTs) in Nyeri and Thika districts, the Central Province PHMT, the Hospital Management Teams in Thika and Gatundu hospitals.
Conflicts of interest: None.
1. National AIDS and STI Control Programme. Kenya AIDS Indicator Survey 2007: Preliminary Report.
Nairobi: Ministry of Health, Kenya; 2008.
2. World Health Organization (WHO)/Joint United Nations Programme on HIV/AIDS (UNAIDS). WHO and UNAIDS Secretariat Statement on HIV Testing and Counselling
. Geneva: WHO/UNAIDS; 2006.
3. World Health Organization (WHO)/Joint United Nations Programme on HIV/AIDS (UNAIDS). Guidance on Provider Initiated HIV Testing and Counselling in Health Facilities
. Geneva: WHO/UNAIDS; 2007.
4. UNAIDS Reference Group on HIV and Human Rights. Seventh Meeting Issue Paper: provider-initiated testing and counselling in health facilities; 12–14 February 2007.
5. National AIDS and STI Control Programme, Ministry of Public Health and Sanitation, Kenya. Guidelines for HIV Testing and Counselling in Kenya
. Nairobi: NASCOP; 2008.
6. Central Bureau of Statistics, Ministry of Health, and ORC Macro. Kenya Demographic and Health Survey 2003. Calverton, Maryland: CBS, MOH, and ORC Macro; 2004.
7. Liambila W, Warren C, Mullick S, Askew I, Homan R, Mohammed I, et al.
Feasibility, acceptability, effect, and cost of integrating counseling and testing for HIV within family planning services in Kenya. Washington, DC: Population Council.
8. Ministry of Health. The National Health Sector Strategic Plan 2005–2010
. Nairobi: Ministry of Health; 2004.
9. Ministry of Health. National Reproductive Health Policy: enhancing reproductive health status for all Kenyans. Nairobi: Ministry of Health; 2007.
10. Ministry of Health. National Reproductive Health and HIV/AIDS Integration Strategy. National AIDS and STI Control Program (NASCOP) and the Division of Reproductive Health (DRH); Nairobi, Kenya; August 2009.
11. Mullick S, Raney L, Liambila W, Menziwa M, Khoza D, Askew I. The Balanced Counseling Strategy Plus: a toolkit for family planning service providers working in high STI/HIV prevalence settings. Washington, DC: Population Council; 2008.
12. León F, Vernon R, Martin A, Bruce L. The Balanced Counseling Strategy: a toolkit for family planning service providers. Washington, DC: Population Council; 2008.
13. Ministry of Health. Integrating counseling and testing for HIV into family planning services, participants manual. 1st ed. Nairobi, Kenya; 2008.
14. Ministry of Health. Integrating counseling and testing for HIV into family planning services, national training manual. 1st ed. Nairobi, Kenya; 2008.
15. Rivero-Fuentes E, Ramarao S, Estrada R, Warren C, Mullick S, Birungi H, et al
. ‘Assessing integration methodology (AIM): a handbook for measuring and assessing the integration of family planning and other reproductive health services. Washington, DC: Population Council; 2008.
16. Peterman A, Mullick S, Liambila W. ‘Use of the Balanced Counseling Strategy-plus to improve family planning and HIV integration in Central Kenya’, unpublished paper. Nairobi: Population Council; 2009.
17. Sweat M, Gregorich S, Sangiwa G, Furlonge C, Balmer D, Kamenga C, et al
. Cost-effectiveness of voluntary HIV-1 counselling and testing in reducing sexual transmission of HIV-1 in Kenya and Tanzania. Lancet 2000; 356:113–121.