"As a health care provider, I've always been interested in how my connections with my colleagues influenced the care my patients received." That's how Craig Pollack, MD, MS, MHS, describes the roots of his interest in his latest research topic: using doctors' social networks as a lens on the treatment they provide. That, and his focus on cancer treatment disparities as a Robert Wood Johnson Foundation (RWJF) Clinical Scholar (2006-2009) at the University of Pennsylvania, led him to become the lead author of an article in Health Services Research examining the ways treatment of prostate cancer among Medicare patients might reflect the social networks of their doctors.
Pollack and former colleagues from the University of Pennsylvania, including Katrina Armstrong, MD, MSCE, co-director of the University of Pennsylvania Clinical Scholars program and an RWJF Physician Faculty Scholars Program alumna (2001), examined Medicare claims data for more than 5,300 male patients, 65 or older. The patients all lived in one of three cities, and all had been diagnosed with prostate cancer in 2004 or 2005.
The claims data allowed the researchers to identify social networks—clusters of doctors who treat the same pool of patients—and then analyze the treatment their patients receive. That analysis revealed significantly different treatments among different clusters of doctors, with prostatectomies much more common among some groups than others. Pollack points out that because prostate cancer has no single "right" treatment, the varying treatment approaches doctors and patients opt for are best thought of as "differences," not "disparities." For example, because prostate cancer is generally slow-growing, a watch and wait approach may be in order, and not evidence of poor treatment.
However, by examining the differences in treatments, and relating those differences to the social networks of the doctors, Pollack and his colleagues concluded that it might be possible to identify a number of "highly connected" doctors in specific communities."It is possible," they write, "that the highly connected physicians may serve as opinion leaders and be used in interventions to facilitate the spread of norms and quality standards."
The researchers note that understanding the relationships among physicians is increasingly important in the context of health care reform, because care delivery models emphasized by reform—medical homes and accountable care organizations (ACOs), for example—rely greatly on providers' communication with each other about their patients and their practices. "Network analysis may provide a way to model these relationships," they write. "For example, the approach may be used to demonstrate the extent to which physicians in the same ACO are currently clustered with one another in caring for patients, describe how differences in network structures may impact the ability of different ACOs to improve quality and control costs, and elucidate how reforms lead to changes in clustering over time."
Further, the researchers observe that claims data is an important tool in such analyses, because it allows researchers to avoid costly and time-consuming surveys of doctors. More than that, "claims data allow a much wider view of how all the pieces fit together and how they're linked to each other," Pollack says. "You can see the traces that patients leave as they travel through the health care system, most notably, all the doctors they see. … This type of analysis is important in trying to understand the variations in the care that patients get. Not only does it allow you to examine those variations across regions, it helps you dig down to see what variation looks like within areas and within cities. And it helps in understanding the extent to which structures of health care delivery may be associated with those differences."
The study was published online on October 18, 2011, in the journal Health Services Research. It is part of a special issue on "Bridging the Gap Between Research and Health Policy" that features research articles from current and former RWJF Clinical Scholars. The print edition will be published in February 2012.