Using Correcting Sibling Survival to Calculate 5-Year Mortality Rates and Adjust for Invisible Mortality

Benjamin Shapiro, Institute for Health Metrics and Evaluation, University of Washington
Austin E. Schumacher, University Of Washington

Corrected Sibling Survival (CSS) is the gold-standard among demographers for measuring adult mortality from questionnaires on sibling survival where nonresponse bias among sibships with no eligible interviewees is thought to render observed mortality an underestimate of true mortality. The Global Burden of Disease (GBD) study processes large scale surveys, such as those from the Demographic and Health Surveys system, through an adaptation of CSS that involves inferring female mortality among families that are not represented within the survey. The plan, for future GBD developments, is to transition to generating age-specific mortality rates from survey data, but the zero-survivor adjustment limits the granularity for which hypothetical deaths can reasonably be inferred. The purpose of this presentation is to provide an exploratory analysis detailing how we will adjust the sibling survival process to match the granular criteria: we will shift CSS to produce mortality rates among 5-year age groups limited to individuals reported in a survey; we then adapt the method outlined by Feehan and Borges (2021) to implement a correction factor that reintegrates zero-survivor mortality into the full mortality rates. We demonstrate this process with data from the 2021 Burkina Faso DHS.

Keywords: Data and Methods, Social network methods, Families, Unions and Households

See extended abstract.