Results-Based Financing Impact Evaluation Household Survey 2017, Endline Survey - Burkina Faso

Abstract
---------------------------

The endline survey of the impact evaluation for Health Performance-Based Financing (PBF) in Burkina Faso was conducted in six regions of Burkina Faso from April to June 2017.

The main objective of the impact evaluation is to assess the impact of the PBF intervention on quality of care and health care utilization for Maternal, Newborn and Child Health services, while the baseline survey provides the empirical foundation for the assessment of impacts on a large variety of indicators. The baseline data collection for the impact evaluation took place between October 2013 and March 2014.

The impact evaluation innovates by assessing the combination of PBF with two sets of additional measures to enhance equity, community-based targeting of the ultra-poor and subsidization of health services provided to them; and community-based health insurance. PBF was implemented at the district level in both primary- and secondary-level health facilities - with control districts in the same regions. The study used a blended experimental and quasi-experimental design, including both randomization of the standard PBF and PBF plus the additional equity measures at the facility-level within PBF districts to test the effects of these equity measures on outcomes of interest (experimental part), and a comparison of PBF districts to matched control districts to test the effects of PBF (pooled across all intervention arms with and without equity measures) compared to status quo (quasi-experimental part).

Geographic coverage
---------------------------

Centre-Nord, Nord, Sud-Ouest, Centre-Est, Boucle du Mouhoun, Centre-Ouest

Analysis unit
---------------------------

Households and individuals (members of sample households)

Universe
---------------------------

Selected districts in the 6 regions.

Random sample of 15 households in a randomly selected village from the catchment area of each health facility in intervention districts, as well as from the catchment area of a random sample of health facilities in the control districts for an intervention-control facility ratio of approximately 3:1. Household eligibility criterion: At least one currently pregnant women and/or at least one woman who has ended a pregnancy in the two years preceding the survey.

Kind of data
---------------------------

Sample survey data [ssd]

Sampling procedure
---------------------------

The survey covered a total of 24 districts across 6 regions, 12 intervention districts and 12 control districts. All health facilities in intervention districts were included in the study. In control districts, a random sample of health facilities was drawn for an intervention-control facility ratio of approximately 3:1.

In 8 intervention districts across 4 regions (Batié, Diébougou, Ouargaye, Tenkodogo, Kaya, Kongoussi, Gourcy, and Ouahigouya), health facilities were randomized to receive either the standard PBF (T1), PBF with systematic targeting and subsidization for the ultra-poor (T2), or PBF with systematic targeting, subsidization for the ultra-poor, and additional provider incentives to provide services to the ultra-poor (T3). In intervention districts in the Boucle du Mouhoun region (Nouna and Solenzo), health facilities randomized into 2 groups: standard PBF (T1) or PBF plus community-based health insurance and systematic targeting and subsidization for the poor and (T4). In the intervention districts in region 6 (Kuodougou, Sapouy), all health facilities received the standard PBF model (T1). As referral hospitals to all facilities in their respective districts, all hospitals (CM, CMA, CHR) in districts where randomization happened received T2. The randomization happened during public ceremonies in each region. All health personal from the regional level (Direction Régionale de la Santé (DRS), District level (Equipe Cadre de District, (ECD)) and local level (health facilities and district hospital) were invited to participate. Detailed information about randomization is available in the attached report.

Surveyed households were selected using a cluster sampling technique. Within the catchment area of each primary-level health facility, one village was randomly sampled. Within each sampled village, 15 households were randomly sampled from all households fulfilling the following criterion: The household contains at least one currently pregnant women and/or at least one woman who has ended a pregnancy in the two years preceding the survey.

Mode of data collection
---------------------------

Computer Assisted Personal Interview [capi]

Research instrument
---------------------------

Questionnaires were based on HRITF Impact Evaluation Toolkit templates. The following questionnaires were used:
1) Household Questionnaire - Questionnaire - C1
Depending on a section, the questionnaire was administered to the head of household or the best informed household member, and the mother or the main guardian of children younger than 15 years old.

2) Household Female Questionnaire - Questionnaire - C2
Depending on a section, the questionnaire was administered to all female members of the household 15-49 years old, women who had delivered a baby within the two years preceding the survey, and a mother or main guardian of children younger than 5 years old.

In addition to collecting data through household questionnaires, the survey teams performed the following tests/measurements:
1) Rapid Diagnostic Tests (RDT) for malaria for all children less than 5 years old and pregnant women present in the household during the visit,
2) RDTs for anemia for all children younger than 5 years old and pregnant and nonpregnant women who had delivered a baby in the preceding 2 years present in the household during the visit,
3) Weight and height of all children less than 5 years old present in the household during the survey team's visit.

Response rate
---------------------------

100%

摘要
---------------------------
2017年4月至6月，在布基纳法索的六个地区对基于健康绩效的融资（PBF）影响评估的终期调查得以实施。
影响评估的主要目标是评估PBF干预对孕产妇、新生儿及儿童健康服务质量及医疗保健利用的影响，而基线调查则为评估对大量指标的影响提供了经验基础。影响评估的基线数据收集于2013年10月至2014年3月进行。
此次影响评估的创新之处在于，通过评估PBF与两套额外措施的组合，以增强公平性，针对极度贫困人口开展社区基础上的目标定位，并对其提供的医疗服务进行补贴；以及社区基础上的医疗保险。PBF在初级和二级卫生设施区均以区级为单位实施——并在同一地区的控制区设有控制组。研究采用了混合实验和准实验设计，包括在PBF区内对标准PBF及PBF加额外公平措施的设施级随机化，以测试这些公平措施对感兴趣结果的影响（实验部分），以及将PBF区与匹配的控制区进行比较，以测试PBF（汇总所有带有和不带有公平措施干预臂）与现状相比的影响（准实验部分）。
地理覆盖范围
---------------------------
中心北部、北部、西南部、东部中心、莫胡恩环、西部中心
分析单元
---------------------------
家庭和个人（样本家庭的成员）
总体
---------------------------
六个地区中选定的区。
随机抽样
---------------------------
从干预区每个卫生设施的集水区随机抽取的15个家庭的随机样本，以及从控制区随机抽取的卫生设施的集水区中抽取的随机样本，干预区与控制区的干预-控制设施比约为3:1。家庭资格标准：至少有一位目前怀孕的妇女和/或至少有一位在调查前两年内结束怀孕的妇女。
数据类型
---------------------------
样本调查数据 [ssd]
抽样程序
---------------------------
调查覆盖了6个地区中的24个区，其中12个干预区和12个控制区。干预区内的所有卫生设施均包括在研究中。在控制区，随机抽取了卫生设施的样本，以实现约3:1的干预-控制设施比。
在4个地区的8个干预区（巴蒂埃、迪布古、乌阿加耶、滕科多戈、卡亚、孔古西、古尔西和乌阿希古亚），卫生设施被随机分配以接受标准PBF（T1）、带有对极度贫困人口的系统化目标和补贴的PBF（T2）或带有对极度贫困人口的系统化目标、补贴以及为提供服务给极度贫困人口提供的额外提供者激励的PBF（T3）。在布克勒穆胡恩地区的干预区（诺乌纳和索莱佐），卫生设施被随机分配到两组：标准PBF（T1）或PBF加社区基础医疗保险和系统化对贫困人口的目标定位和补贴（T4）。在第6个地区的干预区（库杜古、萨波尤），所有卫生设施均接受标准PBF模式（T1）。作为各自区内的所有设施的转诊医院，所有医院（CM、CMA、CHR）在进行了随机化的区均接受了T2。随机化在每个地区的公共仪式上完成。所有来自区域级（卫生区域局（DRS）、区级（区框架团队，ECD）和地方级（卫生设施和区医院）的卫生人员都被邀请参加。有关随机化的详细信息可在附带的报告中找到。
调查的家庭是通过聚类抽样技术选定的。在每个初级卫生设施的集水区内，随机抽取一个村庄。在每个抽样村庄内，从所有满足以下标准的家庭中随机抽取15个家庭：家庭中至少有一位目前怀孕的妇女和/或至少有一位在调查前两年内结束怀孕的妇女。
数据收集方式
---------------------------
计算机辅助个人访谈 [capi]
研究工具
---------------------------
问卷基于HRITF影响评估工具包模板。使用了以下问卷：
1) 家庭问卷 - 问卷 - C1
根据部分内容，问卷由户主或最了解的家庭成员，以及15岁以下儿童的母亲或主要监护人进行填写。
2) 家庭女性问卷 - 问卷 - C2
根据部分内容，问卷由15-49岁的所有女性家庭成员、在调查前两年内分娩的妇女以及5岁以下儿童的母亲或主要监护人进行填写。
除了通过家庭问卷收集数据外，调查团队还进行了以下测试/测量：
1) 对所有5岁以下儿童和在访问期间家中出现的怀孕妇女进行疟疾快速诊断测试（RDT），
2) 对所有5岁以下儿童和在访问期间家中出现的怀孕和非怀孕且在两年前分娩的妇女进行贫血快速诊断测试（RDT），
3) 对所有在调查团队访问期间在家中的5岁以下儿童进行体重和身高测量。
回应率
---------------------------
100%