Novel Immunotherapy Agents in Oncology: Generalizability of Trial Results and Drivers of Clinical Utilization

Thumbnail Image


Publication or External Link





Cancer is the second most common cause of death in the United States after heart disease. Novel immunotherapy agents such as nivolumab and pembrolizumab have become an essential, albeit extremely expensive, component of oncology care since their first approvals in melanoma in 2014 and lung cancer in 2015. However, little is known about differences between immunotherapy clinical trial participants and the real-world patient population, or about the drivers of provider utilization of these agents.

The first objective of this dissertation used the SEER-Medicare linked database with claims data from 2014-2016 to conduct two aims analyzing potential disparities between Medicare beneficiaries on active treatment for melanoma and lung cancer and Medicare clinical trial participants. Aim one compared the characteristics of Medicare patients on active cancer treatment to Medicare patients on active cancer treatment clinical trials. Aim two compared Medicare patients receiving the novel immunotherapy agents nivolumab or pembrolizumab to Medicare patients participating in trials of these two immunotherapy agents. Because of the demographic differences in the melanoma and lung cancer patient populations, these aims were analyzed separately in melanoma and lung cancer. As hypothesized, patients in clinical trials were significantly younger and had fewer comorbid conditions than patients undergoing active cancer treatment not in clinical trials. Underrepresentation of non-White and female patients in clinical trials was hypothesized, but these results were less consistent.

The second objective used Medicare Open Payments data from 2016 and Medicare provider utilization data from 2017 to analyze 1) if industry payments promoting nivolumab or pembrolizumab were positively associated with whether a provider was a high utilizer of the agent, and 2) among these high utilizers, if industry payments were positively associated with greater utilization amounts. The hypothesized results, that industry payments were associated with greater likelihood of high utilization and more utilization among high utilizers, were seen in some of the analyses but not consistently throughout the study. Through unique analyses of recent datasets, this dissertation advances our understanding of potential disparities in clinical trial representativeness and the generally positive relationship between promotional payments and provider utilization of immunotherapy agents in the Medicare cancer patient population.