Background In resource-constrained settings, Community Health Workers (CHWs) are the first point of contact between communities and the health system, as providers of maternal and newborn health services. However, little is known of the quality of community-based postnatal care (PNC). We assessed the content of PNC provided by CHWs and women’s experiences of care in two Kenyan counties. Methods We used a cross-sectional, mixed methods design to examine the quality of PNC services provided by CHWs. Trained observers attended PNC home visits to assess technical quality using a 25-item checklist covering four PNC domains: infant health warning signs, maternal health warning signs, essential newborn care, and breastfeeding. The observers completed an 8-item communication quality checklist. We conducted follow-up surveys with observed PNC clients to assess their experiences of care. Finally, we used in-depth interviews with CHWs and focus group discussions with observed PNC clients to understand the experiential quality of care. Results Observations suggest shortcomings in the technical quality of PNC home visits. CHWs completed an average of 6.4 (standard deviation SD=4.1) of the 25 PNC technical quality items. CHWs often lacked essential supplies, and only six percent carried all four of the CHW job aids and tools specified in the national guidelines for maternal health at community level. However, CHWs completed an average of 7.3 (SD=1.1) of the 8 communication quality items, and most PNC clients (88%) reported being satisfied during follow-up interviews. Higher technical quality scores were associated with older mothers, better communication, longer visit duration, and CHWs who carried at least three job tools. CHWs expressed a strong sense of responsibility for care of their clients, while clients underscored how CHWs were trusted to maintain their clients’ confidentiality and were a valuable community resource. Conclusion This study identified gaps in the technical quality of CHW PNC practices, while also recognizing positive elements of experiential quality of care, including communication quality, and trusting relationships. This study also demonstrated the strength of the CHWs’ role in establishing linkages between the community and facilities, as long as the CHW are perceived as, and enabled to be, an integral part of the PHC network in Kenya.
The research presented here is part of the Frontline Health project, which focuses on ways to measure CHW performance and identify the operational considerations that affect the institutionalization of community health in an array of settings. Using a cross-sectional, mixed methods design, this study examines the quality of PNC services provided by CHVs to women and young infants (under 60 days) in two counties in Kenya. The research team captured data from three sources: direct observation of PNC home visits to assess technical quality, follow-up surveys with observed PNC clients to assess experiential quality, and a series of in-depth interviews (IDIs) with CHVs and focus group discussions (FGDs) with observed PNC clients to develop a comprehensive understanding of the technical and experiential quality of the care, including contextual factors that influence quality of care. This study uses three data sources: direct observation of PNC home visits to assess technical quality, follow-up surveys with observed PNC clients to assess experiential quality, and a series of in-depth interviews (IDIs) with CHVs and focus group discussions (FGDs) with observed PNC clients to develop a comprehensive understanding of the technical and experiential quality of the care, including contextual factors that influence quality of care. The study was conducted in rural settings in Kilifi (coastal) and Bungoma (western) Counties. Both counties reflect a range of commonly observed barriers to care, including geographical access constraints and cultural vulnerabilities. Kilifi has a neonatal mortality rate of 26 deaths per 1000 live births, a skilled delivery rate of 52.3%, an under-five mortality rate of 141 deaths per 1000 live births, and a maternal mortality ratio of 540 deaths per 100 000 live births. Bungoma has a neonatal mortality rate of 33 deaths per 1000 live births, a skilled delivery rate of 44%, an under-five mortality rate of 145 deaths per 1000 live births, and a maternal mortality ratio of 360 deaths per 100 000 live births [29]. Kilifi and Bungoma Counties were purposefully selected in collaboration with County Health Management Teams to include a functional CHS where CHVs conduct routine visits, collect data, and receive some form of incentives. Additionally, the selected sites had several collaborative projects between non-governmental organizations (NGOs) and the government being implemented to strengthen local CHS through capacity building and improve access to health information. From 29 July to 23 August 2019, we shadowed 68 CHVs – 32 in Kilifi County and 38 in Bungoma County – as they conducted PNC home visits with postnatal women whose infants were less than 60 days old. CHVs were linked to 12 facilities representing different tiers of the health system (eg, referral hospital/one sub-county hospitals, health centers, dispensaries); three quarters were female. We observed between 1-3 PNC home visits with each CHV. During observations, trained research assistants with social science backgrounds completed an observation guide without intervening during the home visit. Observers measured technical quality of care using a checklist of 25 key health items that were adapted from the national CHV Manual [28] and are routinely completed by CHVs during a PNC visit. These items comprise four domains: infant health warning signs, maternal health warning signs, essential newborn care, and breastfeeding. Observers directly assessed whether CHVs completed these 25 items during PNC home visits. The observations also assessed CHV’s communication with their clients, whether follow-up actions were taken when an infant was identified as unwell, and whether CHVs were properly equipped with four essential job tools: referral slips for follow-up care, a visit schedule register, the CHV household handbook, and a set of community health job aids. Within two weeks of completing each observation, we conducted follow-up surveys with the observed PNC clients. During the follow-up interviews, we collected basic demographic information and assessed the client’s experience of the observed CHV visit, including the four PNC domains and their communication with the CHV. Interviewers asked women about their perceptions of the 25 health checks assessed during the observations and asked a series of communication quality items about the social interaction and attitudinal aspects of care. Interviewers also asked women about their satisfaction with the care they received from the CHV using a Likert-scaled item with four ordinal responses ranging from very dissatisfied to very satisfied. To better understand the contextual factors that influence CHVs ability to plan and conduct PNC visits, we held IDIs with 12 of the CHVs who were observed during a PNC visit and eight of the CHEWs who supervise them. We also conducted two FGDs in each county (ntotal = 4), with a total of 29 postnatal women whose CHV home visits were observed for this study. Trained qualitative interviewers facilitated IDIs and FGDs using pre-developed interview guides. We present cross-sectional data from the technical and communication quality checklists as observed during PNC home visits. Specifically, we examined the proportion of visits at which each item in the checklists were completed and produced summary scores for each domain. As some checklist domain scores might improve over time, while others – particularly those in the neonatal care and breastfeeding domains – would logically be expected to decrease as the child ages, we disaggregated the observation data by the number of contacts the mother had received from a CHW at the time of the observation: first contact, second contact, and third or greater contact. We used simple linear regression to assess whether there is any relationship between aggregated quality scores and contact number. Finally, we conducted multivariate linear regression analyses to examine factors associated with observed quality of care, controlling for county, CHV sex, maternal age, whether the infant was unwell during the visit, visit duration, and whether the CHV had carried at least three job tools to the visit. As CHVs conducted multiple PNC home visit observations, we used cluster analysis methods to adjust for intra-class correlation. All data were analyzed using Stata 14 (Stata Corp, College Station, TX, USA). For qualitative data, we thematically analyzed the transcripts, field notes, and observations using NVivo 12 (QSR International, Victoria, Australia) software. We developed and refined an initial coding framework based on the process of open coding and progressive categorization of issues. We used charting processes to summarize key themes and concepts including references and quotations that we compared across sites for similarities and differences to support identification of key issues around quality of care for PNC clients.
N/A