Background: Failure to meet health outcome targets among developing countries, including those
related to child mortality, has been attributed to weak health systems. Yet there is little research on
on health system determinants of mortality. Moreover, many indicators of health system
performance are poor or underutilized. In this study, we use a novel approach to quantify the
readiness of the health system to deliver services and investigate the relationship between
readiness, health service utilization rates and child survival in a rural region of northern Ghana.
Methods: Using data from the Ghana Emergency Obstetric and Newborn Care (EmONC) facility
survey, we constructed an index of health system readiness for facilities in the Upper East Region
from an analysis of principal components (PCA). We used this measure to examine the influence of
health system readiness on infant (<1 year), child (1-4 years), and under-5 mortality among 8,917
children born between 2000 to 2011 using Cox proportional hazard models adjusted for known
socio-demographic confounders. Additionally, we assessed the association between health system
readiness and health service utilization rates including current modern conceptive use, and ANC
and skilled birth attendance at last pregnancy using logistic regression.
Findings: In general, children under 5 year of age exposed to hospital and community-level
facilities with higher readiness scores experienced significant survival advantages over children
exposed to lower scoring facilities. Lower infant mortality was strongly associated with improved
readiness at the hospital level (HR: 0.58; 95% CI: 0.37, 0.90; High vs. Low). Functional communitylevel
facilities were associated with a reduction in child mortality (HR: 0.56; 95% CI: 0.33, 0.95).
Facility readiness was also associated with increased probability of ANC attendance (OR: 1.29;
95% CI: 1.02, 1.64) and current modern contraceptive use (OR: 2.12; 95% CI: 1.55, 2.92).
Interpretation: Our findings suggest that increased facility readiness and the scale up of primary
health care services have had a positive impact on the survival of children in northern Ghana. We
believe our results may be applicable to other low resource settings. These data provide evidence
for the child survival benefits of increased investment in stronger health systems. Additionally, we
have shown that a novel index of health system readiness can be constructed using principal
component analysis of service provision data. This measure could be used to monitor health
system progress and should be studied further.