Medicare Shared Savings Program: Key Factors Triggering Success

Document Type



Media is loading

Publication Date



Presentation: 17:58


The CMS MSSP Innovation Program incentivizes healthcare providers to form accountable care organizations (ACOs), which coordinate patient care and share in the financial savings achieved through improved care coordination and reduced healthcare costs. The program generates vast amounts of data on patient outcomes, healthcare utilization, and cost savings, which can be used to evaluate the effectiveness of different care coordination strategies and identify areas for improvement. Providers face challenges in implementing clinical initiatives due to their tight schedules and the need to prioritize time and focus on factors that trigger shared savings. Big data analysis is crucial in the CMS MSSP Innovation Program, as it can identify patterns and trends in healthcare utilization and outcomes that can inform policy decisions and improve the overall quality of care. Exploratory data analysis (EDA) was conducted to identify important features, and various classification trees using different algorithms were created to select the optimal model. Logistic regression models were then used to finalize and select the optimized predicting model. The analysis shows the importance of reducing costs by eliminating avoidable utilization and monitoring key indicators such as specialist TCOC strategy, PCP to SPC ratio, chronic disease management, SNF admissions per 1000 patients, and benchmark expenses. The impact of the COVID-19 pandemic on ACOs and a higher percentage of Medicare and Medicaid dual-eligible patients among the ACO is not statistically significant from the analysis. Additionally, certain key indicators affect physician-led and hospital-based ACOs differently, and having a higher number of specialists in the ACO network has a more negative impact on physician-led ACOs than hospital-based ACOs. Overall, the analysis demonstrated a rigorous approach to data analysis and model selection, which can be used to develop effective predictive models in a wide range of fields. The findings can inform policy decisions and improve the overall quality of care in healthcare systems.



This document is currently not available here.