Our findings demonstrate that there are many models of care for orphaned and separated children and that each plays a valuable and important role in the response to the orphan crisis in sub-Saharan Africa. This is supported by what others have found in Malawi and South Africa [20, 24]. However, we identified several important differences in institutional, family- and community-based care environments with respect to their ability to uphold children’s rights and provide basic material needs, while identifying opportunities to strengthen family-based orphan care.
First, family-based care environments in UG County are significantly less likely to be able to provide for children’s basic material needs than CCI’s and may not be able to provide a standard of living adequate for the child’s physical, mental, spiritual, moral and social development. This is consistent with other research from sub-Saharan Africa indicating that extended families are stretched and unable to meet their care-taking expectations including providing, food, clothing and other basic needs for children [7, 9, 15, 42]. Second, Kenya’s constitution prohibits corporal punishment ; yet the majority of families (77%) and a quarter of institutions report using this method of enforcing discipline. The use of corporal punishment is an ineffective means of discipline, is linked to child abuse, may have long-term negative effects on the child [44–47], and violates the Convention on the Rights of the Child Article 19 and 37  and therefore is a fundamental breach of children’s human rights . Third, despite reports in other sub-Saharan African settings stating that children in institutional care lack links to family, and community [26, 27, 49], CCI’s and CBO’s in UG County worked to facilitate family connections through their programs and ensuring that children in their care have knowledge of and where possible, contact with, their families. Fourth, very few children in both institutional and family-based care have birth certificates and therefore are lacking their right to their identity. Children without birth certificates have difficulties accessing education, obtaining identity documents and other items necessary to transition to successful adulthood [27, 42]. Fifth, while religion plays a central role in Kenyan culture , the rights of children to freedom of thought, conscience and religion may be infringed upon in CCI’s as the majority are faith-based environments and over half have compulsory religious education. Lastly, the majority of institutional care environments limit admission to children below the age of 12 and therefore orphaned/separated adolescents are lacking care options when no extended family is able to take on this responsibility and are being subject to discrimination due to their age. Yet, these institutional care environments are an important place of last resort for orphans and non-orphans alike as it has been documented that over 90% of non-orphaned children living in CCI’s were admitted due to maltreatment and the majority of orphans due to extreme destitution .
We recognized several opportunities to strengthen family-based models of care. First, there lacked households providing foster care in UG County. Foster care by non-relatives may be a viable solution to enable children to remain living in family-based care environments that has been overlooked, yet plays a role in many other sub-Saharan African countries . Second, CBO’s and CCI Plus models of care should be expanded and augmented, as these programs provide support to families caring for orphans and enable children to remain in family-based settings. Augmenting programs providing school fees, uniforms, psychosocial support, and economic development opportunities can build capacity with families to ensure they can meet the needs and protect the rights of children in their care. Lastly, religious institutions provide extensive support to institutional care providers; yet they have not reached out into the communities to support households caring for orphaned and separated children which represent a missed opportunity to ensure children remain in family-based care. Additionally, findings from other studies indicate that expanding the government Cash Transfer to Orphans and Vulnerable Children program may have the ability to build capacity for family-based care and protect orphaned children living in poor households [36, 52]; as children living in households receiving this transfer in UG County have improved school attendance, nutritional status, and future outlook on life .
There are several strengths to our findings. The first is that these data come from a well-defined geographic area, and are representative of orphaned and separated children in Uasin Gishu County, Kenya. Second, we were able to randomly sample households to obtain a representative sample and reduce selection bias. Third, we comprehensively recruited and assessed all CCI’s and a large proportion of CBO’s providing care in UG County. Lastly, we used empiric methods to measure the ability of different care environments to uphold children’s rights.
There may be limitations to our findings. Many of the outcome measures are self-reported and therefore may be subject to various kinds of bias, including reporting bias such as interviewer or social desirability bias. We tried to minimize these potential biases through the deployment of Community Health Workers to conduct the household level data collection and by using unannounced household audits. However, highly sensitive data concerning issues such as corporal punishment may still be biased by self-report and validated measures of discipline were not utilized in the data collection instrument. Although the site assessment was administered to the head of household or facility Director, children were not asked the same questions and may have responded differently to questions regarding their basic material possessions, corporal punishment or other characteristics of their care environment. However, for many of the questions posed in the site assessment the head of household or facility Director would be able to provide the most accurate answer regarding the care environment’s characteristics. Additionally, because the facility and household level data were collected on the unit of the household/facility and not per individual child, we were not able to disaggregate the data separately for the 275 (9.2%) of non-orphaned children living in the same care environments as the orphans. These data may also be subject to random misclassification bias if there were errors in recording or data entry. Lastly, because there are so few registered foster families, we were unable to systematically invite foster families to participate in the study.