Healthcare Analytics:
Demystifying Predictive Analytics

Data mining, machine learning, regression modeling, statistical learning - it’s enough to make anyone’s head spin! Don’t be impressed by the jargon. A few key principles will help you stay cool amidst the predictive analytics hype.

In this two part series, we will take a case-based approach to the common concepts behind those methods as they apply to everything from recommending movies to predicting emergency room visits or heart attacks, ranking the safest hospitals, profiling surgeons by their complication rates, and recommending personalized treatments or tests.

We will focus on why most prediction models have major limitations and their common pitfalls, why they rarely provide ‘actionable’ insights (despite what their proponents tell you!), how they can be harmful, why simple models are often better than complex ones, how much and what kind of data we really need to build them, and how they should be tested before being used to make decisions.

These sessions are NOT about how to build prediction models, but we will mention some additional resources and tools so as to further demystify predictive analytics.

A basic understanding of statistics is helpful but not required.

Click Here for On Demand

Who Should Attend
Educational Objectives
Instructor