.png)
44 – Topics on Dimension Reduction
Modelling Survival After Diagnosis of A Specific Disease Based on Case Surveillance Data
Ruiguang Song
Centers for Disease Control and Prevention
Gengsheng Qin
Georgia State University
Kathleen McDavid Harrison
Centers for Disease Control and Prevention
Xinjian Zhang
Centers for Disease Control and Prevention
H. Irene Hall
Centers for Disease Control and Prevention
Motivated by a study assessing the impact of treatments on survival of AIDS (Acquired Immune Deficiency Syndrome) patients, we developed a semi-parametric method to estimate the life expectancy after diagnosis using data from case surveillance. With the proposed method, the life expectancy is estimated based on the traditional non-parametric life table method, but the age-specific death rates are estimated using a parametric model to derive more robust estimates from limited numbers of deaths by single year of age. The uncertainties associated with the semi-parametric estimates are provided. In addition, the life expectancy among people with the disease is compared with the life expectancy among those with similar demographic characteristics in the general population. The average years of life lost is used to measure the impact of the disease or the treatment on the survival after diagnosis. The trend of impact over time can be evaluated by the annual estimates of life expectancy and average years of life lost in the past.