Improving Uptake of Family Planning Services in Western : Evaluation Findings and Key Learnings

Clair Null With support from the David and Lucile Packard Foundation, the “Reversing the Anitha Sivasankaran Stall in Fertility Decline in Western Kenya” project was launched in 2009 to counter Kimberly Smith stagnating and deteriorating reproductive health indicators in western Kenya. Imple- mented by a consortium of partners, this six-year initiative, which became known as the Packard Western Kenya (PWK) project, aimed to strengthen the supply of and generate demand for family planning services—and ultimately reduce unmet need for contraception and fertility rates—with a focus on long-acting and permanent methods (LAPM) and sustainability. This brief summarizes key findings from a mixed-methods evaluation of the PWK project. The PWK Project Consortium Challenged by an alarmingly high population to 4.7 by the 1990s. Community-based distribution • The African Population and growth rate at independence, Kenya became an (CBD) of contraceptives, funded primarily by donors, Health Research Center early leader on family planning (FP) by making it an is often credited with the concurrent and equally (APHRC) coordinated all essential component of the country’s development dramatic increases in the modern contraceptive project activities. strategy. Adopting its first national population policy prevalence rate (CPR). However, as donor atten- • Marie Stopes Kenya (MSK) and FP program in 1967, Kenya went on to have a tion and funding shifted to HIV/AIDS, funds for FP and Family Health Options strong and successful FP program in the 1980s and declined, compromising gains in FP service provision Kenya (FHOK) implemented project activities in Busia 1990s, resulting in a dramatic decline in the fertility and contraceptive use and leading to a prolonged and Counties, respec- rate—from 8.1 children per woman in the 1970s stall (and reversal in some areas) in fertility decline. tively. • Great Lakes University of (GLUK) conducted monitoring of project activities. PWK Project at a Glance During the expansion phase (2012-2015), the project Objective 1: Improve the supply of FP services also coordinated with other • Community Health Volunteers (CHVs) are recruited and trained to provide quality family planning Packard Foundation grantees (FP) information and counseling, and to distribute pills and condoms targeting specific subpopula- tions, including: • Facility staff are trained to deliver long acting and reversible contraceptive (LARC) methods and reduce contraceptive stock-outs through improved commodity management (the project pro- • Center for the Study of Ado- vided emergency buffer stocks) lescence (CSA) • Mobile outreaches are organized to bring a full range of FP methods, including long-acting and • Forum for African Women permanent methods (LAPM), closer to rural communities Educationalists (FAWE) • Christian Health Association Objective 2: Generate demand for FP services of Kenya (CHAK) • CHVs conduct household visits and community sensitization meetings and distribute information, Consortium members also education, and communication (IEC) materials worked closely with the Minis- try of Health in both counties. • CHVs provide referrals to facilities for reproductive health and FP services requiring medical intervention

Objective 3: Create an enabling environment for FP • Project activities are monitored and project achievements and lessons learned are disseminated • PWK partners engage with county government officials and other partners in dialogue on FP issues • Advocacy acitivities promote county budget allocations to support FP and sustainability of project

June 2016 In light of these trends, the Kenyan government EVALUATION OVERVIEW tried to reposition FP as a development priority, making the National Coordinating Agency for The evaluation assessed the implementation, out- Population and Development (NCAPD) a semiau- comes, and sustainability of the PWK project. It also sought to identify key lessons learned and recom- Evaluation Goals tonomous agency within the Ministry of Planning and National Development in 2004. It also added mendations for future FP programming in Kenya. The 1. Document project imple- a new line item for FP commodities to the 2005 study relied primarily on the collection and analysis of mentation Ministry of Health (MoH) budget, launched a in-depth qualitative data, triangulated with program 2. Assess the project’s influ- National Reproductive Health Strategy in 2007, monitoring and other secondary data. Qualitative data ence on FP knowledge, and developed a Strategy for Improving Uptake of from female youth (ages 15-24), women (ages 25-49), attitudes, and practices LAPM in 2008. During this period, the government men (ages 18-59), and CHVs were collected through 3. Provide insights into key also launched the Community Health Strategy focus group discussions (FGDs). In-depth interviews factors affecting the sus- (2006), which sought to improve access to and use (IDIs), sometimes in groups, were conducted with tainability of the project of health services, including FP, at the community CHEWs and health facility staff. Among these respon- 4. Identify key learnings and level. The strategy organized Community Health dent types, 64 FGDs and 30 IDIs were conducted recommendations for Volunteers (CHVs) into Community Units, linked across 16 facility catchment areas (8 in Busia and 8 in future FP programming in Kenya to a health facility and supervised by Community Siaya). IDIs were also conducted with 8 county and Health Extension Workers (CHEWs). In doing so, 6 national level stakeholders, and 7 project staff, for a it began phasing out the parallel CBD system that combined total of 114 FGDs and IDIs. relied on CBD agents. KEY FINDINGS: IMPLEMENTATION Against this backdrop, the PWK project was devel- oped in 2009 as part of a larger national effort by In collaboration with the MoH, the project the Kenyan government, donors, and NGOs to recruited and trained a new workforce of 600 counter the stall in fertility decline in Kenya. The CHVs integrated into MoH structures to deliver PWK project, initially launched as a three-year FP services at the community level. To support demonstration project (2009-2012), is a communi- “Some of us although we rapid scale-up and sustainability, the project lever- used to be health workers ty-based FP initiative implemented in two large rural before we were trained by aged existing CHVs and recruited new CHVs through counties in western Kenya: Busia and Siaya. These Packard, there were some local governance structures. CHVs were trained on counties were selected because of their high fertil- areas where we were still shy. the MoH’s CHV curriculum by MoH trainers, using So Packard really enlightened ity rates and low CPR in 2009 and the absence of us and helped us to be coura- the basic and FP-specific modules for new CHVs donor or other FP initiatives. From 2012-2015, proj- geous and to talk about reality and the FP module for existing CHVs. CHVs demon- ect activities were intensified and scaled up in Busia concerning sex. . . .” strated a strong understanding of the project’s goals, - CHV and Siaya during the project’s expansion phase. and felt the training equipped them with the knowl- PWK project goals were to (1) increase routine use edge and confidence needed to deliver FP infor- of contraceptives among women of reproduc- mation, counseling, and commodities. Many CHVs tive age and (2) reduce preferred family size and became passionate champions for FP, speaking influence fertility intentions among women and openly about their own experiences with LAPM and men. The project adopted an integrated, sus- other methods. CHVs felt supported by CHEWs and tainable model of FP service provision that aligns facility staff, reporting that they generally received with the existing policy priorities and systems necessary commodities and guidance. mentioned above, and addresses both demand- Over 160 providers were trained on delivery of LARC and supply-side barriers to increase uptake of FP methods, with some facility staff receiving training services. The project has three key components: on commodity management. A seven-day training “We had been offering shal- (1) improving the supply of FP services by training low counseling on FP at the on provision of long-acting reversible contraception CHVs to provide CBD services, conducting mobile facility, so after that training (LARC) was offered to facility in-charges and other outreaches, and training facility staff, (2) increas- I learned how to counsel a providers identified by the county Reproductive Health mother properly.” ing demand for FP services through household Coordinator. Providers praised the comprehensiveness - Facility staff visits, community meetings, and distribution of and quality of the training, believing it improved their information, education, and communication (IEC) ability to provide counseling on FP methods, while materials, and (3) improving the policy environ- also allowing them to provide new contraceptive ment through advocacy to promote budget methods, specifically implants and intrauterine contra- allocations for FP and sustainability of the project’s ceptive devices (IUCDs). Facility staff also appreciated interventions. the commodity management training, though imple- mentation of this training was more limited.

2 Mobile outreaches were a key vehicle for in outreaches in Siaya, resulting in high levels of Snapshot of increasing contraceptive options and access, but condom distribution. Key Program Outputs were logistically challenging. Originally envi- Referrals by CHVs facilitated use of FP services at CREATION OF FP WORK- sioned as outreach events in communities, mobile facilities, including during outreaches. CHVs acted FORCE outreaches were conducted at facilities to allow as a bridge between communities and facilities • 600 CHVs recruited and for provision of LAPM. In Busia, a dedicated team through their provision of referrals for facility-based trained to provide FP ser- of providers conducted all outreaches; in Siaya, vices reproductive health and FP services. CHVs provided existing facility providers were supplemented with more than 78,000 referrals, 67,000 of which were • 164 facility staff trained to additional MoH staff. Outreaches involved col- provide implants and IUCDs for LAPM. Referrals expedited services at facilities laboration between CHVs and facility and project and ensured that they would be provided for free. SUPPLY AND REACH OF staff, and were logistically challenging. However, CHVs followed up with referral clients to inquire SERVICES they were effective in meeting unmet demand for about side effects and method satisfaction. • > 400,000 community contraceptives. The project conducted an average members participated in FP of 10 outreaches per month, though frequency at a sensitization meetings Method mix and perceived quality of care given facility varied. improved at facilities, but privacy and cost con- • > 210,000 people reached through household visits cerns remain. Many women reported improvements in the quality of FP services at facilities over the last • > 4 million condoms and KEY FINDINGS: SUPPLY OF FP >190,000 pill cycles pro- SERVICES few years, noting in some cases improved availabil- vided through CBD ity of different FP methods. However, many women • Approximately 67,000 CHVs reached a diverse clientele—includ- continue to have concerns about the privacy and referrals by CHVs for LAPM ing women, men, and youth—using different cost of FP services at facilities, and other facility staff • > 31,000 clients served strategies. On average, CHVs conducted 60 and community members noted that occasional through outreaches household visits each month—totaling more than stock-outs still occur, particularly of injectables. • > 153,000 couple years of 210,000 visits from 2013 to 2015. One-on-one protection generated by interactions helped CHVs build relationships with The project increased the supply of short- and project activities community members, and were appreciated for long-term methods. Across the various meth- the privacy and confidentiality they offered. CHVs ods provided, the project generated over 153,000 also used a number of community platforms— couple years of protection (CYP) from 2013 to including barazas, women’s and youth group 2015 (Figure 1). Implants provided by the project meetings, churches, and theater performances—to accounted for almost half of the project’s overall provide FP information. CHVs held over 8,000 contribution to CYP. Despite the low conversion community meetings attended by 400,000 factor for condoms (120 condoms equals one CYP), women and men during the project’s expansion the project distributed so many condoms that they “In the past we would only get account for the second-largest share of the project’s injectables, pills and condoms phase. Women’s and youth group meetings were but the rest were not there found to be particularly effective. contribution to CYP. but now you can get all the methods [at facilities].” CHVs were a convenient, and sometimes pre- KEY FINDINGS: KNOWLEDGE, - Female beneficiary ferred, source of short-term methods. CHVs pro- ATTITUDES AND INTENTIONS vided over 4.2 million condoms and nearly 200,000 cycles of pills across 186,000 client interactions CHVs shifted attitudes in support of FP. The from 2013 to 2015. CBD served women, men, project’s slogan, “Jamii ndogo: Jimudu,” which con- and youth, with CHVs employing tailored delivery veyed the message that smaller families are easier to approaches for different clientele, often meeting manage, seemed to be effective with beneficiaries. with women in secret and using euphemisms to In addition, most FGD participants understood that communicate with youth. Many women noted that FP can be used for spacing as well as stopping child- CHVs were their preferred or main source for FP bearing and that use of FP can have both health and “Nowadays when you go to services, due to the convenience and confidenti- economic benefits. Beneficiaries and CHVs noted a baraza, the chief will tell ality CBD offered. Male CHVs were as effective as that community members and leaders encourage you to talk about FP. All these leaders have welcomed FP, female CHVs in reaching new CBD clients and do women to seek out the CHV and use FP services, so it has changed people’s not appear to have targeted males only. especially if they observe that a woman has been perceptions in the village.” giving birth relatively frequently. - Facility staff Mobile outreaches complemented CBD by expanding access to injectables and LAPM. CHV efforts to engage youth have made youth Outreaches served more than 24,000 female FP more open to discussing FP topics. CHVs tailored clients. Women appreciated the convenience of their strategy to reach youth, leveraging youth group outreach locations and that services were pro- meetings and meeting youth privately outside their vided free of charge. More LAPM were provided homes. CHVs felt that, over time, it became easier in Busia, whereas over 7,000 males participated and more acceptable to talk to youth about repro-

3 Figure 1. Couple years of protection attributable to the PWK project years of protection attributable to the PWK project 80,000 ion t c e 60,000 o t p r

f o

40,000 s a r e

y 20,000

pl e u

o 0

C Pills Condoms Injectables Implants IUCDs BTL Vasectomy

Source: Authors' calculations based on APHRC Endline Report, 2016. Data from October 2012 to May 2015. Note: Pills and condoms from CBD and outreaches, injectables from outreaches only (facility-based provision is not solely attributable to the project), all other methods from outreaches and facilities. Couple years of protection (CYP) calculated using USAID's conversion factors and 10 CYP per bilateral tubal ligation (BTL) and vasectomy.

ductive health and FP, with parents becoming more The impressive increase in uptake of implants is open to CHVs talking to their children about these one of the project’s notable successes, setting topics. CHVs commented that male youth have a project counties apart from other areas and high demand for condoms to prevent pregnancy likely driving increases in CPR. Project activities and sexually transmitted infections. contributed to a large increase in implant use over the course of the project. Implants accounted for CHVs had less success in increasing male support Figure 2. almost all of the increase in the modern CPR in the for FP, but the project’s messaging on the eco- Use of FP methods project counties during the project period, while use FP methods to tnomicarget a benefitsnd engag eof m FPen ,seemed women toan resonated CHVs with project ofco untotheries methodsalso had hremainedigh essentially flat (Figure obssomeerved men.that m Despiteale supp theor tproject’s for FP i sintention still low ,to targetus e and lower rates of injectable use 60 2). In 2014, project counties also had higher rates of witandh so engageme men men, fearin womeng that f eandma lCHVse met hobservedods can that surrounimplantding co untusei eands, f ulowerrther sratesuppo ofrt ininjectableg the use than 50 caumalese pe supportrmanent for in fFPert iisli tstilly, b low,irth dwithefec somets, and men project’dids co surroundingntribution to counties, increases further in supporting the ag e 40 undfearingesirab lthate side female effect smethods. Outside can of ccauseondom permanent CPR inproject’s the proj econtributionct areas (Fig tour eincreases 3). in the modern

c en t 30

r disinfertility,tribution ,birth CHV defects,s’ effor tands to undesirabletarget men wsideere effects.

e CPR in the project areas (Figure 3). P 20 notOutside effectiv eof, wcondomith man ydistribution, CHVs, in cCHVs’luding efforts male to e project helped women 10 CHtargetVs, p menrefer rwereing t onot cir ceffective,umvent u nwithple asmanyant CHVs, FP metThehod projectwith the helpedir fertil iwomenty inten talignions, theirbut choice of 0 intincludingeractions wmaleith mCHVs,en du preferringring hous etoho circumventld visits. there isFP sti lmethodl room f owithr im ptheirrove mfertilityent intentions. Women 2009 2014 Hounpleasantwever, CHV interactionss and bene withficia rmenies n oduringted th athousehold em phasemphasizedized that CH thatVs aCHVsnd fa candility facilitystaff h estafflped helped chanvisits.ging However, economi cCHVs condi andtion beneficiariess and econom inotedc that them choose the best method to meet their needs. Any modern method them choose the best method to meet their argchanginguments fo economicr small fam conditionsilies seem tando b eeconomic Most women were able to differentiate the appro- Pills or condoms needs. Most women were able to differentiate inflargumentsuencing so mfore smallmen’s families attitud eseems tow atord be FP influencing, and the apprpriatenessopriatene ssof o variousf variou methodss method sby b yfertility intention. Injectables 49 somemen e men’sven vo attitudeslunteered toward for vas eFP,ct oandmie 49s. men evenfe rtilityThis inte nappearstion.  tois haveappea translatedrs to have into an improved Implants volunteered for vasectomies. translatematchd into between an impro women’sved match fertility between intentions and IUCD KEY FINDINGS: USE OF FP SERVICES methods adopted. Implant use was more common women’s fertility intentions and methods BTL KEY FINDINGS: USE OF FP SERVICES adoptedamong. Impla womennt use w awhos m owantedre comm too ndelay amo childbirthng for at least two years, and sterilization was more Source: 2009 data from Project monitoring data suggest that the project women who wanted to delay childbirth for at common among women who did not want any APHRC Baseline Report, conProjecttribute dmonitoring to large se cdataular isuggestncrease sthat in C thePR .project lea st two years, and sterilization was more 2013. 2014 data from Kenya contributed to large secular increases in CPR. commomoren amo children.ng wome However,n who did the no tmajority want an yof women Demographic and Health During the PWK project period, Kenya for whom a LAPM could be appropriate still choose Survey, 2014. expDuringerience thed a PWKdram projectatic inc rperiod,ease in Kenyamoder experiencedn CPR m ore children. However, the majority of women a dramatic increase in modern CPR and decreasef or whoam short-term a LAPM cmethod.ould be appropriate still Note: Currently married and decrease in unmet need for FP and fertility women ages 15-49. rateins .unmet Nation needally, tforhe FPCP andR in fertilitycreased rates. by 1 4Nationally, c hoose a short-term method. perthecent CPRage pincreasedoints (fro bym 3914 percentageto 53 percen pointst), wit h(from KEY FINDINGS: IMPROVING THE 39 to 53 percent), with LAPM use rates more than POLICY ENVIRONMENT doubling (from 8 to 17 percent). These dramatic changes in FP indicators were mirrored in PWK Devolution presented risks and opportunities for project areas and other regions of Kenya. Project FP in the PWK project counties. Under devolution, monitoring data on new and repeat FP clients and responsibility for health budgets was transferred their uptake of methods provide plausible evidence to newly created counties, which were expected that the project contributed to some of this secular to finance government provision of FP services. increase in the modern CPR.

4 Figure 3. Use of injectables and implants in Busia, Siaya, and other rural counties in project regions

50

e 40 g

t a 30 n e c

r 20 e

P 10 0 Injectables Implants Injectables Implants Homa B ay Wester n Kakam ega Nyanza Migo ri region: Busia region: Siaya Vihi ga Kisii Source: Kenya Demographic and Health Survey, 2014. Note: Current use among currently married women ages 15-49.

Busia Whereas Kenya had previously self-financed 50 per- at the local and international levels, advocacy and cent of its FP commodity costs, in the first year that policy engagement efforts were mainly focused at “So that’s why we’re say- ing if all this [devolution] is counties set their own budgets for health (2013), the county level. As a result, the project is not as happening, FP could lose if there were no budget allocations for commodity well known among national-level stakeholders. we don’t keep the tempo, the procurement in any counties. On the other hand, advocacy.” because devolution allows counties to determine - National NGO stakeholder KEY LEARNINGS their own priorities, the PWK project partners saw potential for the governments of Busia and Siaya CHV provision of FP services, including CBD, can Counties to become leaders in FP, setting an exam- be cost-effectively integrated into existing health ple for other counties to follow. system structures and processes. The project effectively leveraged existing MoH training modules Progress on prioritization of FP in Busia and Siaya and trainers, and structures under the Community reflects the efforts of many actors, including the Health Strategy to recruit, train, and support CHV PWK partners and the county MoHs. Recognizing provision of FP services. CHVs became a widely the benefits of coordinating advocacy efforts, PWK respected and used source of FP information, partners worked together with CSA, FAWE and “Other partners come with counseling, and short-term methods—and were well CHAK to engage county-level government and pre-determined strategies, but supported by CHEWs and facility staff. Although the Packard came with an open MoH officials in dialogue about their projects and project demonstrated that CBD can be integrated mind and wanted to hear FP policy and budget priorities more broadly. MoH from our experiences. They into a broader set of CHV services, multi-tasking staff credited PWK partners for their role in securing would place the information across too many service tasks could potentially out there and let us critique. It the FP line item in ’s budget, but felt strain CHVs and jeopardize their ability to provide was a bottom up approach.” that other actors, especially champions within the high quality FP services. Sustainability of the CHV FP - County MoH staff county administration, were more influential in the services may require continuation of CHV stipends. decision to include CHV stipends in the county budget. In Busia County, FP and health in general Tailored approaches are needed to reach different do not have such strong champions; therefore demographic groups with FP services. As commu- progress has been slower, despite concerted nity members, CHVs were able to create safe spaces efforts by partners and the MoH. Recently, the PWK to meet with youth and women whose husbands partners began laying the foundation for potential or families were not supportive of FP. CHVs were replication of PWK project activities in neighbor- also able to use informal, unplanned interactions to ing counties by hosting advocacy meetings with deliver FP information and commodities. “[Referrals] have an impact government officials in Bungoma, , and because once you give a Counties. Increasing awareness and accessibility of LAPM, client a referral it also gives particularly implants, can have a dramatic impact her courage to go to a certain facility for a certain service.” The project partners used a variety of mech- on uptake. CHVs and facility staff counseled anisms to disseminate evidence and learning. - CHV women on the full range of FP options, and implants These include conference presentations, project emerged as the most popular LAPM. The project reports posted on partners’ websites, and journal used several vehicles to make LAPM services more articles. Although dissemination activities occurred accessible, including through referrals issued by

5 CHVs, outreaches (particularly in Busia), and train- tracking and improvement, but CHVs and CHEWs ing of facility staff (especially in Siaya). noted that they spent a considerable amount of time on reporting—time that could have been used Highlighting the economic benefits of small to provide services to communities. Reporting families can be an effective approach to shifting burden may also have discouraged facility staff from FP attitudes, but shifting practices may be more providing complete data. challenging. By emphasizing that smaller families can be easier to manage and support, and that FP Targeted engagement and advocacy efforts are can be used for birth spacing rather than just stop- critical for securing budget allocations for FP, but ping childbearing, the project was able to generate they may not be sufficient without a local cham- support for FP among women and community pion. Despite progress in advancing the FP agenda leaders. The project was not able to eliminate and securing resources for FP—in part due to the widespread anxiety about FP leaving women advocacy efforts of PWK partners and other actors— infertile, and many people continue to believe that county MoH stakeholders stressed the importance young women should marry and have one or two of a local government champion for securing ade- children before starting FP. Refined and intensified quate funding for FP service provision. IEC efforts may be needed to build greater trust in reversible contraceptive methods. KEY CONSIDERATIONS Although men used CBD for condoms, CHVs had Project partners should seek county support and limited success reaching men through IEC activ- “[FP] is just the way you can funding for ongoing trainings, in addition to the ities, or increasing male support for FP. CHVs’ write a sentence and then put current advocacy on funding for FP commodities a comma, this doesn’t mean pessimism about their ability to change strong and and CHV stipends. Initial and refresher trainings full stop.” pervasive disapproval of FP among men, combined - Male beneficiary will be needed for replacement and existing CHVs with the lack of an organized platform to engage and facility staff, respectively, to ensure high quality groups of men in discussions about FP, impeded service provision, particularly as FP guidelines evolve the project’s success in changing the culture of and new FP methods become available. male resistance to women’s use of FP. Incorporating injectables into CBD could be a Elimination of stock-outs at facilities may require cost-effective approach to increasing the CPR. more county-level coordination and support. The The MoH already endorses CBD of injectables in project’s commodity management training helped underserved areas with high maternal mortality rates facility staff track existing inventory, but did not and low facility coverage. Although neither Busia nor improve their ability to forecast commodity needs. Siaya County qualifies as a high-need area based on PWK partners played an important role in reducing these criteria, allowing and enabling CHVs to pro- stock-outs by providing buffer stocks and facilitat- vide one of the country’s most popular FP methods ing redistribution of commodities across facilities. could spur further increases in routine contracep- Confidentiality and cost concerns continue to be tive use. PWK partners could advocate for broader barriers to use of FP services at facilities. Despite inclusion of injectables in CBD using evidence from perceived improvements in the quality of facili- other parts of Kenya and the PWK project. ty-based FP services, women still have concerns County MoHs should consider different about the privacy and confidentiality of facility approaches to outreaches. The project’s services. In addition, FP services are not always demand-responsive approach to organizing out- provided for free at facilities, as is required by the reaches was logistically challenging and may not MoH, and transportation costs and waiting times be sustainable. However, an alternative approach to are often non-trivial. facility-based outreaches, such regularly scheduled Monitoring systems need to be right-sized to the visits by MoH medical officers, may improve the capacity of health workers collecting the data. potential for sustainability and help ensure that rural The project’s ambitious monitoring system gen- communities have adequate and reliable access to erated monthly data that were useful for program the full range of contraceptive options.

The authors gratefully acknowledge the support and guidance of the PRH team at the Packard Foundation and the PWK team members at APHRC, MSK, FHOK, GLUK, and CSA who gave input on the design of the evaluation and answered so many of our questions about the PWK project. We would also like to thank our local consultant, Rose Olayo; the data collectors and note-takers at the GLUK for their hard work and commitment to this effort; and colleagues at Mathematica for their valuable contributions. We are greatly indebted to everyone who participated in the focus groups and interviews we conducted for the evaluation, and particularly the Ministries of Health in Busia and Siaya Counties for their contributions to the evaluation.

Masthead photo on p. 1 by Jonathan Torgovnik courtesy of the Hewlett Foundation. 6