Health Services Research – News

During her research years, Helen M. Johnson, MD has been conducting clinical research in the Division of Surgical Oncology with surgical oncologists Drs. Jan Wong and Nasreen Vohra and statistician Dr. William Irish. Our projects focus on identifying prognostic variables that predict outcomes in patients with breast cancer. A recent project (https://www.ncbi.nlm.nih.gov/pubmed/31297648) focused on clarifying the relationship between age at diagnosis of breast cancer and mortality. Age at diagnosis is often included in analyses about breast cancer outcomes, but different investigators use different approaches. Some include age as a continuous variable i.e. a spectrum of values, while others include it as a categorical variable i.e. grouped into categories such as decades or intervals such as <40, 40-60, and 60+. Statistically speaking, these groups are biologically arbitrary and classification of research subjects into such groups risks reduced accuracy of statistical models. Our group used advanced modeling methods to explore the statistical relationship between age at diagnosis and breast cancer mortality. We used data from adult women with breast cancer registered in the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) database, a population-based registry of cancer patients. Our models adjusted for factors such as breast cancer stage and subtype. We observed a quadratic relationship between age at diagnosis and adjusted risk of breast cancer mortality: the risk decreased from age 18 to 45 and increased thereafter. We argue that statistical models seeking to estimate breast cancer prognosis should include age as a continuous variable and that investigators should ensure appropriate modeling of this variable in order to provide the most accurate predictions possible. This project sparked a second study, which will be presented at an upcoming conference, the Society of Surgical Oncology’s International Conference on Surgical Cancer Care in Boston, MA in March 2020.