Association between Degrees of Separation in Physician Networks and Surgeons' Use of Perioperative Breast Magnetic Resonance Imaging

Yongren Shi, Craig Pollack, Pamela R. Soulos, Jeph Herrin, Nicholas A. Christakis, Xiao Xu, Cary P. Gross

Research output: Contribution to journalArticle

Abstract

Background: Perioperative magnetic resonance imaging (MRI) is frequently used in breast cancer despite unproven benefits. It is unclear whether surgeons' use of breast MRI is associated with the practices of other surgeons to whom they are connected through shared patients. Methods: We conducted a retrospective study using Medicare data to identify physicians providing breast cancer care during 2007-2009 and grouped them into patient-sharing networks. Physician pairs were classified according to their "degree of separation" based on patientsharing (eg, physician pairs that care for the same patients were separated by 1 degree; pairs that both share patients with another physician but not with each other were separated by 2 degrees). We assessed the association between the MRI use of a surgeon and the practice patterns of surgical colleagues by comparing MRI use in the observed networks with networks with randomly shuffled rates of MRI utilization. Results: Of the 15,273 patients who underwent surgery during the study period, 28.8% received perioperative MRI. These patients received care from 1806 surgeons in 60 patient-sharing networks; 55.1% of surgeons used MRI. A surgeon was 24.5% more likely to use MRI if they were directly connected to a surgeon who used MRI. This effect decreased to 16.3% for pairs of surgeons separated by 2 degrees, and 0.8% at the third degree of separation. Conclusions: Surgeons' use of perioperative breast MRI is associated with the practice of surgeons connected to them through patient-sharing; the strength of this association attenuates as the degree of separation increases.

Original languageEnglish (US)
JournalMedical care
DOIs
StatePublished - Jan 1 2019

Fingerprint

Breast
Magnetic Resonance Imaging
Physicians
Patient Care
Surgeons
Breast Neoplasms
Medicare
Retrospective Studies

Keywords

  • breast neoplasms
  • magnetic resonance imaging
  • Medicare
  • surgeons

ASJC Scopus subject areas

  • Public Health, Environmental and Occupational Health

Cite this

Association between Degrees of Separation in Physician Networks and Surgeons' Use of Perioperative Breast Magnetic Resonance Imaging. / Shi, Yongren; Pollack, Craig; Soulos, Pamela R.; Herrin, Jeph; Christakis, Nicholas A.; Xu, Xiao; Gross, Cary P.

In: Medical care, 01.01.2019.

Research output: Contribution to journalArticle

Shi, Yongren ; Pollack, Craig ; Soulos, Pamela R. ; Herrin, Jeph ; Christakis, Nicholas A. ; Xu, Xiao ; Gross, Cary P. / Association between Degrees of Separation in Physician Networks and Surgeons' Use of Perioperative Breast Magnetic Resonance Imaging. In: Medical care. 2019.
@article{d40d078d7a18443eb0e971c2554cd86c,
title = "Association between Degrees of Separation in Physician Networks and Surgeons' Use of Perioperative Breast Magnetic Resonance Imaging",
abstract = "Background: Perioperative magnetic resonance imaging (MRI) is frequently used in breast cancer despite unproven benefits. It is unclear whether surgeons' use of breast MRI is associated with the practices of other surgeons to whom they are connected through shared patients. Methods: We conducted a retrospective study using Medicare data to identify physicians providing breast cancer care during 2007-2009 and grouped them into patient-sharing networks. Physician pairs were classified according to their {"}degree of separation{"} based on patientsharing (eg, physician pairs that care for the same patients were separated by 1 degree; pairs that both share patients with another physician but not with each other were separated by 2 degrees). We assessed the association between the MRI use of a surgeon and the practice patterns of surgical colleagues by comparing MRI use in the observed networks with networks with randomly shuffled rates of MRI utilization. Results: Of the 15,273 patients who underwent surgery during the study period, 28.8{\%} received perioperative MRI. These patients received care from 1806 surgeons in 60 patient-sharing networks; 55.1{\%} of surgeons used MRI. A surgeon was 24.5{\%} more likely to use MRI if they were directly connected to a surgeon who used MRI. This effect decreased to 16.3{\%} for pairs of surgeons separated by 2 degrees, and 0.8{\%} at the third degree of separation. Conclusions: Surgeons' use of perioperative breast MRI is associated with the practice of surgeons connected to them through patient-sharing; the strength of this association attenuates as the degree of separation increases.",
keywords = "breast neoplasms, magnetic resonance imaging, Medicare, surgeons",
author = "Yongren Shi and Craig Pollack and Soulos, {Pamela R.} and Jeph Herrin and Christakis, {Nicholas A.} and Xiao Xu and Gross, {Cary P.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1097/MLR.0000000000001123",
language = "English (US)",
journal = "Medical Care",
issn = "0025-7079",
publisher = "Lippincott Williams and Wilkins",

}

TY - JOUR

T1 - Association between Degrees of Separation in Physician Networks and Surgeons' Use of Perioperative Breast Magnetic Resonance Imaging

AU - Shi, Yongren

AU - Pollack, Craig

AU - Soulos, Pamela R.

AU - Herrin, Jeph

AU - Christakis, Nicholas A.

AU - Xu, Xiao

AU - Gross, Cary P.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Background: Perioperative magnetic resonance imaging (MRI) is frequently used in breast cancer despite unproven benefits. It is unclear whether surgeons' use of breast MRI is associated with the practices of other surgeons to whom they are connected through shared patients. Methods: We conducted a retrospective study using Medicare data to identify physicians providing breast cancer care during 2007-2009 and grouped them into patient-sharing networks. Physician pairs were classified according to their "degree of separation" based on patientsharing (eg, physician pairs that care for the same patients were separated by 1 degree; pairs that both share patients with another physician but not with each other were separated by 2 degrees). We assessed the association between the MRI use of a surgeon and the practice patterns of surgical colleagues by comparing MRI use in the observed networks with networks with randomly shuffled rates of MRI utilization. Results: Of the 15,273 patients who underwent surgery during the study period, 28.8% received perioperative MRI. These patients received care from 1806 surgeons in 60 patient-sharing networks; 55.1% of surgeons used MRI. A surgeon was 24.5% more likely to use MRI if they were directly connected to a surgeon who used MRI. This effect decreased to 16.3% for pairs of surgeons separated by 2 degrees, and 0.8% at the third degree of separation. Conclusions: Surgeons' use of perioperative breast MRI is associated with the practice of surgeons connected to them through patient-sharing; the strength of this association attenuates as the degree of separation increases.

AB - Background: Perioperative magnetic resonance imaging (MRI) is frequently used in breast cancer despite unproven benefits. It is unclear whether surgeons' use of breast MRI is associated with the practices of other surgeons to whom they are connected through shared patients. Methods: We conducted a retrospective study using Medicare data to identify physicians providing breast cancer care during 2007-2009 and grouped them into patient-sharing networks. Physician pairs were classified according to their "degree of separation" based on patientsharing (eg, physician pairs that care for the same patients were separated by 1 degree; pairs that both share patients with another physician but not with each other were separated by 2 degrees). We assessed the association between the MRI use of a surgeon and the practice patterns of surgical colleagues by comparing MRI use in the observed networks with networks with randomly shuffled rates of MRI utilization. Results: Of the 15,273 patients who underwent surgery during the study period, 28.8% received perioperative MRI. These patients received care from 1806 surgeons in 60 patient-sharing networks; 55.1% of surgeons used MRI. A surgeon was 24.5% more likely to use MRI if they were directly connected to a surgeon who used MRI. This effect decreased to 16.3% for pairs of surgeons separated by 2 degrees, and 0.8% at the third degree of separation. Conclusions: Surgeons' use of perioperative breast MRI is associated with the practice of surgeons connected to them through patient-sharing; the strength of this association attenuates as the degree of separation increases.

KW - breast neoplasms

KW - magnetic resonance imaging

KW - Medicare

KW - surgeons

UR - http://www.scopus.com/inward/record.url?scp=85065329973&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85065329973&partnerID=8YFLogxK

U2 - 10.1097/MLR.0000000000001123

DO - 10.1097/MLR.0000000000001123

M3 - Article

C2 - 31008899

AN - SCOPUS:85065329973

JO - Medical Care

JF - Medical Care

SN - 0025-7079

ER -