The pathway to child survival in the Birhan Cohort, Ethiopia, 2018-22.

Background: Child mortality remains a concern in Ethiopia despite the significant achievements in the past three decades. Proper implementation of the existing low-cost interventions can prevent two-thirds of the deaths. Understanding illness recognition, care-seeking behaviours, and barriers that caregivers encounter during a child's illness along care pathways is imperative. We aimed to describe illness recognition and reactions of caregivers of children <2 years, and factors associated with severe illness or death related to the care pathways, including the child, caretaker, household, and health system.

Methods: We conducted a prospective cohort study using an open birth cohort of Birhan field site, from December 2018 to November 2022. The analysis included newborns followed up to two years old who had an illness episode and for whom data were available for the mother-child dyad. We extracted and linked data on community follow-up and morbidity visits, clinical signs and symptoms of illness at health facility visits, verbal autopsy of deceased children, and maternal health and healthcare. We used descriptive and logistic regression analyses.

Results: Of 3969 eligible children enrolled in the Birhan Cohort, 1397 (37.8%) had at least one episode of illness during the first two years of life. Of those, 108 (8%) experienced a severe illness or died, of which the majority (n = 76; 70.4%) were newborns. Most sick children (n/N = 714/1187) did not get treatment from a formal source; 53.1% (n/N = 684/1289) of those with mild or moderate illness and 27.8% (n/N = 30/108) of the severely ill or deceased. The mean delay in care-seeking was 5.9 (standard deviation (SD) = 10.6) days for those with mild or moderate illness, and 1.7 (SD = 0.58) for the severely ill or deceased. Only 4.8% (n/N = 27/559) of children sought care from a health post (HP), and 68.1% (n/N = 94/138) of children were referred for further care. Only 68.4% (n/N = 13/19) of the severely ill or deceased children were referred, of which 3 (4.9%) accepted the referral. Compared to a newborn, being a young infant (adjusted odds ratio (aOR) = 0.05; 95% confidence interval (CI) = 0.008-0.27) and a child (aOR = 0.03; 95% CI = 0.005-0.17) were associated with a reduction in the odds of severe illness or death. Children who sought care from an HP had a higher risk of severe illness or death than those who consulted a government hospital (aOR = 19.6; 95% CI = 2.71-142.40). Belonging to a rich family resulted in a reduction in the odds of the outcome compared to a poor household (aOR = 0.15; 95% CI = 0.02-0.94).

Conclusions: Illness recognition and care-seeking were low in the Birhan field site, and when care was sought, it was delayed. Care was sought from an HP in rare cases. Health workers did not refer about a third of severely ill or deceased children for further care. Being a newborn, consulting a HP rather than a hospital, and belonging to a poor family had a significantly higher risk of severe illness or death. Strategies should be devised targeting the modifiable factors identified at individual, family, or community and health facility levels to improve child survival.