HHS - Quantify Prediction Error, 2-time Scale, Risk Mode - MRAS

The Department of Health and Human Services (HHS), through the General Services Administration (GSA), is conducting market research to explore capabilities related to advancing risk prediction methods in cancer survival studies. This initiative is spearheaded by the Surveillance Research Program (SRP) of the National Cancer Institute (NCI), which supports various cancer surveillance programs, including the Surveillance, Epidemiology, and End Results (SEER) Program. SEER covers nearly half of the U.S. population and provides crucial data on cancer incidence, survival, treatment, and mortality trends. The overarching goal of this contract is to enhance statistical models that quantify prediction errors in competing risk scenarios, thereby delivering more clinically relevant information for physicians and patients alike.

The contract seeks expertise in extending methods for continuous-time competing risks survival models on two time scales: age for non-cancer deaths and time since diagnosis for cancer mortality. Current models often utilize net survival measures which do not account for non-cancer causes of death an important consideration for patients with comorbidities who may face higher risks from other causes. By improving these models to incorporate competing risks such as other-cause death alongside cancer death probabilities, SRP aims to provide tools that better inform clinical decisions regarding treatment plans.

A significant aspect of this work involves developing measures to assess predictive accuracy and quantify prediction error across multiple competing risks over time. While existing research has primarily focused on single cause predictions at specific time points, this project will explore methodologies that offer comprehensive estimates across all potential outcomes. These enhancements are expected to be incorporated into reports or web-based tools used by healthcare professionals and patients to evaluate the likelihoods of dying from cancer versus other causes or surviving altogether.