Analytics & Reporting

Analytics & Reporting

Analytics & Reporting

To better understand performance in our risk contracts, evaluate the effectiveness of population health management programs, create innovative tools, and identify trends in both specific cohorts and broadly across the Mass General Brigham network, our analytics groups rely on robust data managed in the Mass General Brigham Enterprise Data Warehouse (EDW).

This data infrastructure, coupled with a strong analytics team, allow us to support both patients and providers in an effective, meaningful way.

Learn more about our activities:

Program Performance Analytics

In order for programs to run as effectively and efficiently as possible, we rely on performance reports and dashboards to understand program performance. These actionable insights allow Population Health to target the “right” patients and adjust programs to maximize their impact. Some interventions that help maximize program potential include:

  • Developing more precise risk stratification
  • Identifying patients at highest risk for future high cost episodes
  • Identifying and reducing areas of overuse
  • Improving capture of quality measures
Utilization Reports

Understanding how and where patients receive care is vital for understanding population and provider level trends that impact health delivery and outcomes. Utilization reports provide information to help identify key insights in the way care is being used and delivered, with the goal of optimizing site of care, improving outcomes and the quality of care, while reducing costs. Utilization of services is an important driver of health care spending. They also serve to identify those who are doing exceptionally well within our system, which allows our teams to examine and scale best-practices across our network. For Example:  Population Health identified a marked decrease in Skilled Nursing Facility Admits/1000 at one of our hospitals after the implementation of a Transitional Care Management pilot. The longitudinal utilization report allowed us to identify declining trend at a specific site as compared to other sites—as well as an attributable reason for the shift. Over time, the improvement has attenuated, but it has not returned to prior high levels. We can extrapolate from this, and recommend other sites with High Skilled Nursing Facility admits/1000 use the Transitional Care Management program as one management option for appropriate patients. Carefully charted utilization data allow teams to make interpretations and recommendations for targeted interventions. It also helps us identify areas for attention and, if necessary, carefully create and test new programs to address those hot spots. This allows Mass General Brigham to run efficiently and at the highest quality possible.

Predictive Analytics, Algorithms, and Machine Learning

Population Health, in partnership with Health Catalyst, has leveraged the Enterprise Data Warehouse (EDW) to create a machine learning algorithm for Adult High-Risk Care Management patient identification. The goal was to create a more focused list that better represents the patients who are eventually enrolled in the program. The new algorithm models past selection decisions, and calculates a probability that a patient would be selected for enrollment in iCMP.  Variables were selected for the machine learning model by testing each variable and determining its impact on predicting High-Risk Care Management enrollment. Models were developed separately for Commercial, Medicare and Medicaid payers; and the cut-off points for High-Risk Care Management eligible patient probabilities vary across these populations.

While High-Risk Care Management is the most robust predictive tool in our program, one that has grown and developed over the past five years, Population Health is working to identify more opportunities for predictive modeling to enhance patient care and identification.

© 2022 My CMS All rights reserved.