A Network Analytic Approach to Modeling Variation in Prostate Cancer Care

Associate Name: Craig Evan Pollack

Funding Source/Period of the Grant: NCI K07 07/05/11 - 06/30/16


Unexplained variation in the clinical management of prostate cancer is recognized as an important problem in cancer treatment. To investigate potential explanations for this variation, the proposed project uses a social network analytic approach to model the connections that exist between physicians caring for patients with prostate cancer. This approach explicitly recognizes the clinical reality that many patients face: they see multiple physicians for their screening, diagnosis, treatment, and follow-up care across a range of clinical venues and at various times. Not only does each physician contribute to the type and quality of care patients receive but also the relationships between physicians within a geographical region may be an important and modifiable determinant of care.
The specific aims of this study are: (1) to map the connections between PCPs, urologists, and radiation oncologists as a network and examine whether specialists'positions in the network structure are related to their patient's prostate cancer treatment;(2) to determine whether the network structure helps explain observed differences in prostate cancer treatment between black and white men;and (3) to investigate the determinants of the network structure through a survey of PCPs about their referral decisions. To achieve these aims, this project uses SEER-Medicare claims data and a survey of PCPs. The SEER- Medicare sample consists of men with localized prostate cancer diagnosed from 1995-2005 and their associated PCP and specialist providers. The position of specialists in the cancer care networks will be quantitatively determined and the positions will be related to patient-level variation in the type and quality of cancer care and to racial differences in care. The survey of PCPs will elucidate the factors important in their referral decisions. Understanding observed variations may suggest ways that the network structure may be leveraged to improve quality in cancer care and reduce disparities in outcomes. The candidate, Craig Pollack, MD, MHS is an Assistant Professor in the Division of General Internal Medicine at Johns Hopkins School of Medicine with a joint appointment in the Department of Epidemiology and membership in the Sidney Kimmel Comprehensive Cancer Center. His long-term goal is to become an independent, grant-funded health services researcher in the field of cancer control. His short-term goals are to develop expertise in cancer prevention and control, network analysis, advanced health services research methods, and organizational theory and provider behavior. Dr. Pollack has assembled a mentoring team that includes national leaders in cancer research, urological cancer, health disparities, and network analysis. The proposed project provides Dr. Pollack with the training, mentoring, and resources to become a successful independent investigator dedicated to improving cancer care.

(from NIH RePORT)

Photo: "prostate ribbon" by Bill Heffer is licensed by CC BY 2.0