Final Report for the
AIDSCAP Program in Kenya: Lessons Learned and Recommendations

This report comprehensively summarizes the FHI/AIDSCAP program in Kenya (1992-1997). The report lists program accomplishments, constraints and outcomes, as well as supplying information on lessons learned and recommendations.

1. Behavior Change Communication

• Programming in HIV/AIDS BCC will be improved and facilitated when the national strategy is in place. BCC activities in Kenya are still highly fragmented, with no clear sense of a national IEC strategy within which individual donors and programs play a role. A new national strategy, targeted primarily at youth (among other target groups) has been developed by NASCOP in connection with the STI Project. At the same time, NCPD and UNFPA have jointly developed a national population IEC strategy. It could multiply the impact and commitment of IAs if they perceived themselves to be part of a national strategy, particularly if sharing of materials and experiences among different IEC implementers was improved.
• Rural areas are still relatively neglected in AIDS prevention interventions, despite the recent epidemiological evidence that they are as hard hit, in some cases more so, than urban areas. In the next five years, we need to find creative, low-cost but reliable ways to get messages out to the villages. The requests for information made to worksite peer educators when they go home to the rural areas are an example of this need to extend programs. It would be valuable to map the intensity of AIDS prevention programming against the level of HIV seroprevalence to see if interventions are focused in the most needy areas.
• There remains a gap between attitude change and behavior change, noted by the worksite, university, and medical student interventions. However, anecdotal evidence suggests that changes are taking place: men report that they frequent bars less often and report having fewer casual partners due to fear of AIDS and sex workers report declines in business. Condom use in Kenya continues to increase, suggesting a change in attitudes toward condoms, at least among some groups, although some groups are still hostile or mistrustful about condoms. Knowledge of HIV transmission has been high in Kenya for a long time, but quantitative documentation of significant behavior change among key target groups participating in direct interventions has not yet been reliably and validly accomplished.
• Many people still deny their personal risk. Future interventions need to address the issue of realistic personal risk assessment. The tendency to see one's own high behavior as safe, even if similar behavior by another person would be labeled high-risk remains. Effective messages dealing with this issue remain in short supply.
• AIDS prevention is still vulnerable to negative media publicity. In particular, the wide attention given to Pearl Omega, advertised as a cure for AIDS, was a real, if temporary setback, to peer education work, since it was an excuse to deny the need to change behavior. Should there be repeats of such media stories in the future, concerted efforts should to be made to mobilize partners to refute such stories and to gain media attention for rebuttals.
• IA's should not be intimidated by the "vocal minority". In HIV/AIDS prevention work, there is always likely to be some opposition and it should be recalled that persons and groups who approve or who are indifferent to messages or programs rarely express their opinions in an organized or public way. The experience of PATH should serve as an example: some of their materials were burned by opponents in a public ceremony in a Nairobi park. The result of this publicity was increased demand for their materials by NGO Consortium members and by the general public. Increased demand for a manual produced by the Scouts Association (UNFPA-funded) also resulted from negative publicity. This case is particularly interesting since the manuals had to be purchased for about $ 6.00.
• Working through an umbrella organization to provide training in materials development to member NGOs is a cost-effective approach. Working with PATH/Kenya and the NGO Consortium, AIDSCAP funded an on-the-job training course for representatives for 20 NGOs, which took them through all stages of materials development, from identification of target audiences and formative research to working with printers for materials production and developing a materials distribution plan. Among the essential features of this course were clear criteria for selection of participants, the formation of groups to work together on target audiences, selection of those target audiences by the participants, field work, and supervision and feedback at every stage. The only change suggested by trainees was to shorten the two-year course to six months, to minimize drop-outs. It is recommended that this course be formalized by PATH and AIDSCAP in published form as a model for other countries.
• Interpersonal education methods need to respond to the schedules and work and life-styles of target groups. For example, the peer education method did not work for security guards, a large workforce who work as individuals at isolated sites. This constraint was overcome by changing the strategy to build in an AIDS prevention component to the ongoing training courses given by security companies. Training for the hard-to-reach group of matatu operators was also designed to reach them at times and places convenient to them.
• The curricula and training schedules of peer education and other HIV/AIDS prevention education projects need to be flexible so they can evolve as peer educators become more experienced. In addition, a method of assessing needed changes to the content of training and scope of the activities should be built into project design. Focus groups with the worksite peer educators showed that, having been well informed about the facts of AIDS transmission and prevention, they wanted more focus on interpersonal skills, including crisis counseling, on home-care, and on how to reach their families (especially youth) and communities.
• Future interventions should consider setting AIDS prevention in a broader context that relates to wider concerns of the target audience. Several university anti-AIDS clubs broadened their scope to become reproductive health clubs, and found this attracted more members and made their messages more relevant to students. The matatu intervention was designed using adult education principles to address a range of pressing concerns for this audience including relationships with the police, traffic laws, courtesy and gender relations, dealing with stress, and job-seeking skills, as well as STDs and HIV/AIDS. This increased the sense of commitment and ownership among the matatu operators.
• Radio soap-opera in local languages appeared to be an effective way to stimulate awareness and discussion among rural audiences. Even after one year, focus groups in the end of project evaluation still vividly remembered the content and messages in the Maajabu programs, and over 27,000 letters were received by the producers.
• An attempt to assess the impact of mass media and drama interventions on audiences should be made. Since national surveys are expensive, USAID should consider working with partners involved in IEC mass media to fund a joint regular audience survey to assess audience rates, audience profiles, recall rates, and if possible, impacts on attitude and behavior change. Program managers reported that drama and video heightened audiences' sense of the epidemic's relevance to their lives, and increased their sense of personal risk, while also maximizing the number of people reached by the intervention. Evidence of this would encourage IA's and would provide important information to program funders and planners.
• Peer educators suggested that individual counseling, and women-only discussions, provided women with safer environments to discuss sexual behavior and risk reduction options than joint counseling with partners and mixed-sex groups.
• Program managers and project staff reported significant obstacles to maintaining staff commitment to prevention programs. Obstacles included: the personal impact of HIV/AIDS prevention work on volunteer project staff, increased job demands, and lack of recognition. Periodic innovations in programming, and tokens demonstrating appreciation and recognition of efforts, were cited as contributing positively to morale and motivation.
• Linking BCC activities with service provision increases program effectiveness and provides a measure of BCC impact. The worksite and university peer education interventions were strengthened by the existing links with on-site health clinics. The implementing agency, FPPS, had already been working at many sites to support the clinics with family planning and STD training. The end of project evaluation indicated that in both interventions, health clinic staff stated that more clients were reporting STDs to the clinic.
• Effective materials distribution at field level is still a problem. Peer educators at the worksite needed more materials, but field visits showed that when these were provided, there was no regular system for distribution. Training on how to use materials should be included in peer education projects.
• AIDSCAP interventions only reached a small proportion of those who still need AIDS prevention education. We still need to identify low-cost, fast, effective models for replicating effective pilot interventions (such as that with matatu operators) on a national scale.

2. STDs

• Syndromic management of STDs is acceptable in the private sector, both to clinicians working in large worksite clinics as well as to individual private practitioners. Perceived advantages and disadvantages of the syndromic approach are significantly different for clinicians working in different types of practice, and these differences are important to understand when attempting to motivate health workers to adopt a new approach. For example, support from management is essential for changes in treatment protocols and drug availability in private industry, while issues of profitability and perceived competitive advantage appear to be potent motivators for individuals in private practice.
• Post-training supervision is necessary, though on-site supervisory visits are time-consuming. More efficient models of health worker monitoring and supervision should be explored. Feedback and continued motivation of health workers can be linked to monitoring and supervision, and refresher courses and newsletters are potential tools.
• Improvements in the proportion of patients receiving prevention advice on condom use and partner referral appear to lag behind those gained in correct diagnosis and treatment. This may be due to the relatively insensitive methods used for capturing this information. Operations research into improved methods of counseling, recognizing the time constraints faced by a busy health worker, would be useful.
• Notifying and treating sexual partners is perceived to be problematic by most health workers. Several direct and indirect approaches are used, though with inconsistent results. Use of partner referral cards seems to be helpful, though most patients, when they do refer, notify only one partner, almost invariably the regular partner. Operations research into new approaches to improved partner referral, whether through better clinician/patient communication or messages targeted to the community about the dangers and asymptomatic nature of STDs, should be a priority.
• Risk assessment (as recommended by WHO) was taught as a supplementary tool for detection of women at increased risk for cervical infection. Emphasis was placed, however, on the application of the Kenya national guidelines which do not include use of risk assessment. While some health workers were comfortable applying risk assessment, others were apparently confused by the multiple approaches. If risk assessment is used in the future, more time should be allotted in training to ensure that health workers are comfortable in using it.
• Prescribing habits for STDs vary widely, especially among individual private practitioners. Cost is an important consideration, although efficacy may be even more important. Many health workers expressed willingness to employ more expensive syndromic regimens if patients responded well and were satisfied. Others reverted to "pet drugs", some effective, others not. Operations research in to the range of drug choices and factors influencing their use would be useful.

3. Condoms

Future condom distribution strategies could be enhanced if the following lessons learned are considered:

• If diversification of condom access is an important strategy goal, this needs to be clearly and prominently articulated by the donor and understood and accepted by all parties whose cooperation is required to implement it.
• Attitudes and practices in Kenya regarding condom use appear to be in a period of transition. While certain groups remain opposed to condom use for "moral or ethical" reasons, because they continue to hold misconceptions about condom effectiveness in preventing disease, or due to continuing (but lessening) image problems, condom use continues to rise rapidly. The number of condoms distributed free in Kenya has risen from 8 million in 1988 to over 90 million in 1996. In addition, 500,000 socially-marketed condoms per month were sold in 1995. Much work remains to be done and the potential market is by no means saturated.
• To ensure that action is taken to ensure that condom distribution forms an integral part of IA's activities, "collaboration" on developing effective condom distribution schemes among separately-funded IAs needs to be a contractual obligation which is subject to reporting requirements and which forms part of the activities which are regularly monitored by the IA and by FHI/AIDSCAP. Merely suggesting that IA's collaborate with condom distributors (free or socially-marketed), even if this is in writing in a sub-agreement or LOA, has not proven to be effective.
• It is not realistic for IAs institutionally-wedded to one condom distribution mechanism to temporarily, for a specific contract, embrace a radically different one. This was, for example, evident in the interventions managed by Family Planning Private Sector. FPPS was selected to manage peer education programs at institutions because of its history of supporting clinics at those institutions. The clinics had, however, provided free Ministry of Health condoms to their clients, using an additional mechanism of free condom dispensers. Combining this mechanism with the introduction of socially marketed condoms proved unsuccessful. In selecting IA's, realistic choices therefore must be made on which elements of a program might need to be sacrificed.
• Institutions without much/any experience with market-driven condom distribution mechanisms may require external TA to develop a "menu" of effective options to propose to management of worksites and institutions where "structural" obstacles may appear to preclude distribution alternatives.
• We still do not have a very good picture of the profile of the typical condom user or purchaser. It would be useful to undertake operational research to determine this profile as it would help in better designing not only social-marketing messages and strategies, but messages and strategies for free distribution where applicable.
• The high acceptance rate by both the women and their male partners in the female condom study indicates that the device could have a valuable role in HIV/STD prevention in Kenya. Research into an acceptable price and appropriate distribution mechanisms and promotion for a socially marketed product should be carried out as soon as possible.

4. Policy

• Information and its presentation must be placed within a context of on-going contact and interaction with policy makers and should avoid relying on special or one-time events. A series of AIDS Impact Model (AIM) presentations did occur around the country. Some of the presentations were financially and technically supported by AIDSCAP and surveillance data collection was also supported by AIDSCAP. The presentations did bring important impact information to district and provincial officials, but offered little specific direction for subsequent action by those officials. The interventions carried out by MAP and the NGO Consortium, on the other hand, responded to identified needs, were designed to be cumulative (done in a series of steps), worked within structures known and acceptable to Kenyan participants and authorities, included follow-up to training activities, and channeled positive responses to the process into extensive networks.
• The staff of KANCO and MAP and many of the participants in the policy development process were empowered by the realization that they can do "policy." It is important to provide further support to policy development as a component of HIV/AIDS prevention interventions. Policy is often perceived as something coming from "higher levels", not a process which takes place at all levels and into which interested groups and parties can feed. Many Kenyans at all levels and across most sectors are anxious to know more and are prepared to identify policy and programmatic issues and seek to frame appropriate responses. The majority of workers in HIV/AIDS prevention and care or in affected communities had not realized that they had a role to play in policy development and advocacy. Many - in government, secular and religious organizations - had been waiting for policies to come down to them from those in authority. The desire to discuss issues and learn from common experiences was confirmed and provided a foundation for the creation of at least three national HIV/AIDS networks: Kenya Christian AIDS Network, the district network of KANCO, and the Legal and Ethical Network on AIDS.
• Policy development is not only a labor intensive and time consuming process, but promoters and facilitators of the process must not rush the process of consensus building. While many activists feel that progress in addressing issues and gaining resources has been too slow, in fact, the speed at which policy issues have emerged and been debated in the country compares very favorably with comparable efforts in the family planning field.
• In many cases, policy makers do not have sufficient information to ask the right questions and are waiting for guidance, suggestions, and workable solutions from specialists (although few policy makers will admit to their need).
• Policy does not just happen, the process has to be guided and facilitated. There is a clear role in policy development for facilitators--people who have a good overview of the process and strategy, people who can make important contacts, people who can provide information, people who can advocate with policy makers. HIV/AIDS policy issues arise from many sources: health system clients and providers, NGO and government program staff, financial and planning analysts, the media, lawyers' offices, and many other points where people are affected by the epidemic.
• While there was an initial concern that government, political and religious authorities would oppose KANCO and MAP involvement in policy development, it became increasingly clear, as the process worked upwards, that provincial administrations and denominational leaders welcomed and supported the policy initiatives.
• In Kenyan society, stated intent and public profile carry a great deal of influence. The greater openness throughout society to confronting HIV/AIDS is a major achievement, both for the solid policy foundation that has been laid and the increased flexibility offered to activists. Given the prominence of public profile, the few legal and legislative changes during the AIDSCAP period may be less importance than the changes in attitudes.
• Solid data is an important tool for engaging leaders and policy-makers in discussion. A baseline survey on sexual behavior of church youth became an important tool for engaging church leaders in discussions about HIV/AIDS. Sentinel surveillance data can play a similar role, and the AIM model also used data as such a tool.

5. Evaluation and Behavioral Research

• There is no easily accessible central database of quantitative and qualitative research studies. Much valuable information on target groups may therefore not be consulted in future project design. Besides learning from each other's experience and minimizing duplication, the opportunity to develop standardized approaches to measuring sexual behavior is also missed because of this lack of coordination. For example, the STI Project has developed quantitative and qualitative instruments to measure sexual attitudes and practices among youth for the Nyanza district. If other organizations used the same instruments, a wider baseline could be established to assess the need for behavior change in Kenya nationally.
• Few practical instruments exist to measure interim stages of behavior change, those which do exist are not consistent across donors and programs. AIDSCAP's KABPs, like the WHO/GPA questionnaires and the KDHS, measure knowledge of prevention, number of partners, and condom use at last sexual intercourse. However, there are no reliable indicators to measure interim stages of change, such as discussion with partners or friends. Project managers and staff often used observation in place of more formal research to inform project design, and to monitor and evaluate activities throughout the life of the project. As a result, although many projects are convinced change has occurred, it is hard to prove, which can be discouraging.
• The methodology for segmenting target audiences should be reviewed and refined to assess whether it results in the most effectively targeted programs. "Youth" can, for example, be interpreted as any age from 13-35. Sexually active adults tend to be segmented by profession/socio-economic level, but not by how they correspond to a behavior change continuum.
• Behavioral research results were completed too late to be disseminated to field interventions. Future programs should monitor the timing of behavioral research more closely and build in an improved system for disseminating results to NGOs and collaborating with them to review and revise interventions based on those results. Design of the follow-on USAID/Kenya AIDS prevention project could also include opportunities to revisit the information, data, and findings of the AIDSCAP program, for example through adding data presentations to intervention design sessions.
• The effectiveness of any condom distribution system cannot be measured without pre-project and end of project measures of distribution and without reference to the number of persons targeted.

6. Implementation and Management

• AIDSCAP's combination of roles as a funding agency and as an implement of well-defined programs was at times perceived as "restrictive" to implementing agencies. More flexibility in terms of program design would allow a more process-oriented approach with partner agencies, allowing more collaboration in the program design phase.
• A mechanism for inter-implementing agency communication should be established such as regular quarterly meetings in addition to the one-on-one interactions between AIDSCAP/Kenya and implementing agencies. This could improve program impact through peer pressure and better information exchange and enhance social-communicative learning.
• AIDSCAP's experience is that technical assistance provided to already-strong institutions was more efficiently used and will most likely be better sustained. For example, institutions such as MAP International had pre-existing human, institutional and financial resources that could be mobilized for HIV/AIDS prevention and care interventions after an initial project with AIDSCAP support.
• The "formative" type of technical assistance provided to agencies by a highly-skilled specialist resulted in measurably improved quality of HIV/AIDS interventions by implementing agencies. The long-term relationships between consultants and agencies allowed for trust-building and better understanding of an agencies' particular needs. The two projects with the Kenya AIDS NGOs Consortium exemplified this long-term technical assistance.