About Risk Modeling

A risk prediction model is an analytical tool that uses health risk factors to estimate the likelihood of a patient experiencing a specific health outcome during their treatment or in the future, such as an adverse event—which is defined as a complication that can occur during medical care. The ability to reliably predict potential risk factors that can occur after surgery helps in:

  • Surgical decision-making
  • Counseling of patients and families
  • Planning how to use available resources
  • Measuring the quality of care

Using Artificial Intelligence to Model Risk

The Center for Outcomes & Patient Safety in Surgery (COMPASS) aims to leverage artificial intelligence (AI) to predict and model surgical risk and determine the best interventions, or treatments, for high-risk patients. AI-powered medical technologies can analyze large amounts of data. Additionally, they enhance how medical professionals diagnose and treat patients and identify areas of patient care that require improvement, with the goal of improving patient outcomes.

Below are examples of ways that COMPASS is accomplishing this mission.

Predictive OpTimal Trees in Emergency Surgery Risk (POTTER)

POTTER is a highly accurate and user-friendly clinical decision support tool used to aid medical providers in predicting the likelihood of emergency surgery mortality (death) and morbidity (a disease or a symptom of disease). It was developed by a team of Massachusetts General Hospital and Harvard Medical School (HMS) surgeons, along with mathematicians from the Massachusetts Institute of Technology (MIT).

View all publications related to POTTER

Trauma Outcomes Predictor (TOP)

TOP is an AI interactive smartphone tool used to predict outcomes for patients undergoing trauma and emergency surgery. It was developed by Mass General and HMS surgeons, along with mathematicians from MIT.

View all publications related to the TOP tool