Examining small area estimation techniques for public health intervention: lessons from application to under-5 mortality data in Uganda.

In Uganda, estimates of under-5 mortality are available only at national and regional levels. None exist at decentralized levels of governance or district level. Using small area statistical techniques in a Hierarchical Bayesian Framework, we applied a modeling approach to determine whether we could learn how to target health interventions to reduce under-5 mortality at the district level. Our modeling approach has an advantage over the commonly used Standardized Mortality Ratios, as it estimates the relative risk of under-5 mortality for a particular district. Using data from Uganda's Demographic and Health Survey in 2006, we were able to estimate relative risk of under-5 mortality for each district. Our findings reveal the evidence of district-to-district variations in under-5 mortality with potential spatial clustering. We believe that this information will be useful in Uganda, as interventions can be targeted at districts with higher relative risk of under-5 mortality. Discussion of these results at district level could increase funding for primary health-care activities. Our analysis also suggests the utility of small area techniques for other countries and other health problems.