This proposal has two primary aims: (1) to improve the understanding of the association between patient- reported functional health, comorbidity, and sociodemographic factors and other cause mortality in older men newly diagnosed with prostate cancer;(2) to develop a prototype tool for calculating individualized risk of other cause mortality in this population. Prostate cancer is the most common non-cutaneous cancer in American men and primarily afflicts those age 65 and older. However, most men are diagnosed at early stages with tumors that most often have an indolent course. Guidelines recommend that patients only pursue aggressive treatment if they have >10 year overall life expectancy. Within 5 years of diagnosis, only 11% of American men die due to their prostate cancer, while the majority of patients diagnosed with prostate cancer die of other causes. While validated calculation tools have been developed for clinical use in predicting prostate cancer related mortality, no validated tool has been developed from existing models that identify variables associated with the risk of dying of other causes (OCM). As men in the U.S. live longer and the population above age 65 is rapidly growing, individualized predictions of life expectancy are necessary given the substantial heterogeneity in individual health status. Nearly three quarters of this aging population may have multiple comorbid conditions. Previous work has shown that decrements in patient-reported functional health may be more strongly associated with OCM than the presence of most individual comorbidities. These findings have promise for improving the approach to accounting for the impact of multiple conditions. This predoctoral research training proposal seeks funding to explore the generalizability of prior work demonstrating the association of patient-reported functional health and OCM. The proposed work will help to identify the variables most strongly associated with OCM in older men with prostate cancer, utilizing data from 4,510 subjects in the linked Surveillance Epidemiology and End Results- Medicare Health Outcomes Study (SEER-MHOS) database. This database contains detailed information on cancer characteristics, treatment, cause of death, baseline comorbidities, sociodemographic information, and functional health measures. This proposal will evaluate and improve upon the performance of other cause mortality prediction models through modern statistical techniques to assess predictive model performance. After identifying key predictors of OCM, we propose to develop a prototype clinical risk-calculation tool that estimates personalized risk of 10-year OCM, adapting validated techniques for developing risk calculators. This study will help to establish the utility of patient- reported functional health measures in improving the accuracy of OCM risk estimation in older men newly diagnosed with prostate cancer and will make progress towards a clinically useful OCM risk estimation tool. Completion of this work will help to better identify older patients who would most likely benefit from aggressive treatment vs. those who may not, as they may be more likely to die of other causes.
Prostate cancer is very common in older men but, since most patients will die of causes not related to their prostate cancer, it is recommended that patients only undergo aggressive treatment for their cancer if they are expected to live for 10 years or more. This study will use patient-reported health information to estimate patients'risk of dying o causes other than prostate cancer. The results can help clinicians and patients make more informed decisions about treating prostate cancer.
