P3.106: Predictors of Under-Five Care-Seeking in Migori County, Kenya: Results of a Population-Based Cross-Sectional Survey

Background: Despite improvement, the under-five mortality rate in rural Migori county, Kenya (73 per 1000 live births) remains higher than the national (41) and global rate (37.7). Most causes of under-five mortality, including respiratory infection, diarrhoeal disease, and malaria, have well-established treatments that prevent mortality. Additionally, many children do not die in a healthcare facility but rather at home. Increasing care-seeking for these illnesses is effective in targeting under-five mortality. The Lwala Community Alliance operates an innovative Community Health Worker (CHW) program, inclusive of traditional birth attendants who are often excluded from the healthcare system, that includes professionalization through enumeration, training, and consistent supervision. This training involves primary identification of acute respiratory infections and malaria in the home with subsequent facility referral. In the present study, we aim to describe care-seeking for acute respiratory infection (ARI) and its associated factors in a rural Kenyan community served by the Lwala Community Alliance.

Methods: A cross-sectional survey containing care-seeking history for children under the age of five with ARI symptoms was administered both in areas served by the Lwala Community Alliance and in control areas. Sampling was based on a modified version of the WHO Expanded Programme of Immunization (EPI). Multivariable logistic regression was conducted to investigate factors associated with care-seeking. This regression included presence of Lwala programming, CHW visitation, level of education, size of household, marital status, and wealth quartile.

Results: 7,775 children under the age of five were included in the study. Households in wards that received Lwala programming were more likely to be visited by a CHW (58.1 vs. 28.2%, p< 0.001). Care-seeking for ARI symptoms was increased in areas receiving Lwala programming compared to control areas (68.0% vs. 65.2%). Receiving Lwala programming was also associated with receiving a malaria RDT (p = 0.024) and treatment with an antimalarial (p = 0.002). A marital status of being separated was associated with decreased odds of care-seeking (aOR = 0.107; 95% CI = 0.0132-0.8604; p = 0.036). Odds of care-seeking for ARI symptoms were increased in the areas receiving Lwala programming, although not significantly (aOR = 1.088; 95% CI = 0.6564-1.8042; p = 0.743).

Conclusion: As compared to control areas, the areas serviced by Lwala programming had greater rates of care-seeking for ARI symptoms and a significant increase in diagnosis and treatment of malaria, revealing success in the implementation of malaria training among CHWs and its effect on care of these patients. Having parents who were separated was associated with decreased odds of care-seeking and identifies an area in which targeted programming is needed. This localized data reveals areas for intervention as well as generalizable data that other organizations working in similar environments in sub-Saharan Africa may utilize.