Understanding Physician Network Dynamics
https://www2.deloitte.com/us/en/insights/industry/health-care/physician-networking-for-performance-improvement-in-health-care.html
Receiving goods and services at a low cost is something everyone can get behind. As such, one key component of health care is to optimize cost while maintaining, or even improving, quality. Interestingly, studying the structure of physician networks may provide insight on how health care costs can be brought down. A physician network can be described as ties between physicians based on sharing patients, information, and clinical decisions. This article from Deloitte Insights highlights the fact that “physician centrality” (clustering coefficient) is actually associated with lower costs. The differential impact between physicians with varying centrality values are not negligible. In fact, 15% of costs for a hip fracture replacement equates to nearly $2,000. In addition, adjusted behavior based on this model analysis over the course of two years saved Medicare more than $384 million. Using this information, health systems should be able to improve performance by influencing care patterns, whether its from the health care themselves or physician behavior.
The application of networks in the health care system occurred to me because of another class I am currently taking, PAM 2350: The U.S Healthcare System. At a glance, this article clearly relates to material from this class because it analyzes a network structure and the impact it can have on costs. The mathematical model can be defined as nodes being people and edges being connections. Based on the data trying to be aggregated and analyzed, people can either be producers (physicians) or consumers (patients).
Notable properties that manifest themselves in this study are that the “rich get richer” and the “long tail” phenomenon. The following diagram clearly exemplifies these ideas. Each node is a physician; the node size represents the number of patients seen by the physician while the node color represents the physician’s specialty. From the interactive display on the website, we see that there were only a few physicians who accumulated many patients, while the majority of the nodes only had a few patients. It was interesting to see the use of network analysis in the scope of the development of health care management.
