This study proposes to use central cancer registry data (SEER) linked with secondary claims data (Medicare) to provide epidemiologic information on the occurrence, characteristics, and risk factors for second primary cancers (SPC) in the elderly. The elderly currently represent a large population at high risk for cancer. Studying this population is important for several reasons. The demographics of the elderly are changing. In 30 years approximately 20% of the population will be 65 and older. They have the highest age-adjusted cancer incidence, the largest morbidity (60% of incident cases), and mortality burden (60% of cancer-related deaths). Considering the cancer risk in this group, focus on the elderly is an important step in studying SPC. The overall purpose of this project is to assess the risk and impact of SPC in the elderly population to identify selected cancers on which to focus more in depth future studies. This will be done through the accomplishment of three goals. First, the risk of selected SPC will be assessed. The second goal will be to identify risk factors that predict the occurrence of these second primaries. The third goal is to evaluate outcomes and factors that predict outcomes related to second primary cancers in the population. The results of this study will provide information that will permit the generation of hypotheses regarding the increasing phenomenon among a previously unstudied group. The study will generate clinically relevant information on SPC and outcomes that may relate to potential prevention of second primaries. The long term objective of this project is to use the information gained from this preliminary study to identify population subsets among elderly cancer patients on whom to perform additional studies. The populations identified will be those with initial primary cancers having the highest risk of second primary incidence, poor outcomes, or which are potentially preventable. Future studies on population subsets would include measures of genetic contribution (such as family history) and lifestyle variables (such as smoking history) that are not feasible to collect for the larger sample used in this analysis of the entire SEER population.
