Cancer cases lost to follow-up due to death, changing residence or incomplete information can cause bias in case-control studies. This study quantifies this loss in the California Cancer Registry by the length of time from diagnosis, reporting time to the registry, geographical area and site of diagnosis. Study subjects were cancers diagnosed in 1996 and1997 from three regional registries in agricultural areas of California (Region 2, Central; Region 3, Sacramento; and Region 6, North) and reported to the California Cancer Registry by May, 2000. We calculated the time between diagnosis and report to the registry for all 60,868 cases, and we randomly selected 8000 cases from the 11 most common types of cancer to identify cases, which were interviewable, which we defined as alive and residing at the same residence as at diagnosis. The median time between diagnosis and reporting to the registry was 373 days overall. It differed with respect to type of cancer, geographical region and race/ethnicity. Based on the registry data only, we estimated that 66% of the cases were interviewable three months after diagnosis, 36% were interviewable after 1 year and 15% after two years. We found differences in the natural reporting time to the CCR by type of cancer, geographical region and race/ethnicity. We also showed that interviewablity decreases with time from diagnosis, thus loss to interview could affect risk estimates in registry based case-control studies.
Learning Objectives: Recognize that the lag time between cancer diagnosis and reporting to the registry may cause substantial loss of subjects due to death, moving and incomplete follow-up. Identify characteristics of subjects that may potentially bias a case-control study.
Keywords: Cancer, Methodology
Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: California Cancer Registry
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.