Development of an effective illness severity measure and assessment of the impact of perceived illness severity on formal careseeking for fatal illnesses of neonates and infants in six sub-Saharan Africa countries and Pakistan.

Early careseeking for sick children can make the difference between life and death. Verbal autopsy (VA) studies of the cause of death typically ask about severe symptoms such as seizures and possibly mild or moderate symptoms such as rash, but without examining the relationship between caregivers' perception of illness severity and appropriate careseeking. Verbal and social autopsy (VASA) is a newer method that builds on VA by also examining social factors related to death. From seven VASA studies conducted in Africa and Asia we developed a 2-sign method based on activity level and feeding behavior and a multiple sign method of identifying mild, moderate and severe illness of neonates and 1-11-month-olds. We then examined the relationship of caregivers' perception of their child's condition at illness onset and several covariates to seeking formal health care during the fatal illness. The 2-sign and multiple sign methods effectively distinguished mild, moderate, and severe illnesses, respectively, of neonates and 1-11-month-olds. Careseeking was almost uniformly decreased for severely ill neonates (8.4%-41.8% vs mild: 15.0%-66.7% and moderate: 30.5%-68.5%, p = 0.12- < 0.001), but multivariate analysis revealed that older age in all six African countries (AOR 1.11 [95% CI 1.02, 1.21], p = 0.02 to 1.10 [1.04, 1.16], p < 0.001) and moderate illness in three (4.83 [1.06, 21.96], p = 0.04 to 4.35 [1.59, 11.93], p = 0.005) were associated with careseeking, while severe illness was no longer significant. Similar to neonates, older age in three of five countries (1.26 [1.01, 1.58], p = 0.046 to 1.10 [1.03, 1.16], p = 0.003) and moderate illness in one (2.24 [1.17, 4.30], p = 0.016) were drivers of careseeking for 1-11-month-olds. Careseeking was increased in some countries for infectious illnesses but not for intrapartum- or prematurity-related conditions. Child mortality studies should assess severity level and caregivers' response at various illness stages. Because older infants have more specific illness signs, the 2-sign method should be used only for neonates. Behavior change messages encouraging careseeking for moderate illness signs should be developed. The 2-sign method can serve as a practical tool for this purpose for illnesses of neonates. Effective interventions may require overcoming local barriers to careseeking and bringing delivery and newborn care closer to communities to prevent and treat early onset neonatal illnesses.