Imputation based methods for incidence rate estimation of rare diseases from federal and state mortality data supplemented with disease registry data
MetadataShow full item record
Incidence rates require measurements of the number of new cases and the size of the population at risk over a specific time period to be computed. When this information is missing, incidence during a given time interval cannot be estimated. The United States releases federal death certificate data on a yearly basis that consists of data that is missing precise dates of birth and death for each individual. Further, certain rare diseases share ICD codes with other diseases and the data is lacking information about time to onset of disease, adding additional complications for death and disease incidence rate estimation. This dissertation develops imputation-based methodology for use with the aforementioned data supplemented with state death certificates and disease registry data to estimate disease specific incidence rates. When deaths due to two or more diseases are reported under a single ICD code, methodology to obtain type-specific mortality incidence rates are developed. Mortality incidence rates are obtained using the federal death certificate data supplemented by state death certificate data, obtained by the Population Health Observatory (PHO) of the School of Public Health and Health Professions. Finally, disease registry data containing information regarding time to onset of a specific disease is used in conjunction with the federal and state death certificate data to obtain estimates of disease incidence rates. Bootstrap methods of inference are proposed and applied to the problem of estimating the incidence of death due to Krabbe Disease (KD) and the incidence of KD. The behavior of the proposed methodology is examined in a Monte Carlo simulation study.