“High-quality evidence from observational data: the case of preoperative stress testing”
Matthew Pappas, MD, MPH, is an Assistant Professor, academic hospitalist, and health services researcher at the Cleveland Clinic. His current research, supported by NIH, uses a variety of machine learning techniques and causal effect models in an effort to understand the effects of preoperative tests and perioperative interventions on outcomes after noncardiac surgery. Ultimately, he hopes to make medical decision-making, including preoperative decision-making, simultaneously more personalized and more rigorously evidence-based.
In this talk, Dr. Pappas will discuss preoperative stress testing, including its rationale, conceptual shortcomings, and limitations. Using patient-level electronic medical record data from nearly 160,000 visits to a dedicated preoperative risk assessment clinic, he will discuss recent and forthcoming publications regarding the costs and consequences of preoperative stress testing, along with the diagnostic and prognostic value they offer. He will demonstrate the use of sophisticated machine learning techniques (such as vectorization and convolutional neural networks) and advanced causal effect models to answer clinical questions, including how to minimize biases and address missingness in clinical data, and how to use clinical data to generate rigorous clinical evidence.
Click here to join the virtual event: Join virtual meeting
Password: 808670