Johns Hopkins University, Washington, DC
Craig Evan Pollack , Pamela R. Soulos , Jeph Herrin , Xiao Xu , Nicholas Christakis , Howard Forman , James B. Yu , Brigid K. Killelea , Shi-Yi Wang , Cary Philip Gross
Background: Advanced imaging tests (AIT) such as magnetic resonance imaging (MRI) and positron emission tomography (PET) scans were adopted rapidly into breast cancer practice despite unproven benefits. We sought to examine the extent to which social contagion among physicians explains the adoption of AIT. Methods: We used SEER-Medicare data to construct physician patient-sharing networks during T1 (2004-2006) and determined the potential impact of these networks on the adoption of AIT during T2 (2007-2009). Our T1 patient-sharing networks included all surgeons, radiation oncologists, radiologists, and primary care providers who cared for women with breast cancer during the 3 months prior to diagnosis through 9 months after. We then generated peer groups consisting of physicians who most frequently shared patients with one another using the Girvan-Newman method. Finally, for surgeons whose patients did not receive AIT during T1, we used hierarchical logistic regression to examine the subsequent use of AIT among their patients diagnosed in T2 (N = 3,219 patients). Our key explanatory variable was use of AIT in the corresponding surgeon’s peer group in T1; we controlled for patient sociodemographic and clinical characteristics and included random effects for the surgeon, peer group, and hospital referral region. Results: Overall, 16.4% of patients received AIT during T1, increasing to 34.6% in T2. Of the 735 surgeons who did not have patients who received AIT during T1, over half (52%) had patients who received AIT during T2. Patients whose surgeons’ T1 peer groups had the highest use of AIT ( ≥ 25% of patients received AIT in T1) were more likely to receive AIT compared to patients whose surgeons were in a peer group where no patients received AIT during T1 (36.1% vs 21.8%, adjusted Odds Ratio 2.92, 95% Confidence Interval 1.53,5.60). T1 peer group explained 4.7% of the variation in T2 AIT use, compared to 15.9% for the surgeon and 3.3% for hospital referral region. The patterns were stronger for MRI than PET when modeled separately. Conclusions: Physician patient-sharing networks and the phenomenon of social contagion may offer important opportunities to better understand—and potentially intervene upon—the adoption of cancer technologies.
Disclaimer
This material on this page is ©2024 American Society of Clinical Oncology, all rights reserved. Licensing available upon request. For more information, please contact licensing@asco.org
Abstract Disclosures
2024 ASCO Genitourinary Cancers Symposium
First Author: Michael Leapman
2023 ASCO Annual Meeting
First Author: Eva Lengyelova
2020 ASCO Virtual Scientific Program
First Author: Alan Baltz
2021 Genitourinary Cancers Symposium
First Author: Ashanda Rosetta Patrice Esdaille