J. Marc Overhage, MD, PhD

Presentation

“Big” Methods for Big Data: Lessons Learned from the Observational Medical Outcomes Partnership

Abstract

Using drug safety questions as a test case, the Observational Medical Outcomes Partnership (OMOP) systematically evaluated the application of well established epidemiologic methods to observational healthcare data. The OMOP investigators learned a variety of important lessons including that scale of big data dramatically increases the chance of identifying false positive associations based on the application of traditional statistical criteria for significance; that false positives are increased further by the biases found in observational healthcare datasets and that the associations discovered are heavily dependent on the  specific statistical method employed and detailed design parameters chosen for those methods. These issues demand that we reinvent our approach to interpreting the results obtained from analysis of big data in healthcare.

Biography

Marc OverhageDr. Overhage is the Chief Medical Informatics Officer for Siemens Healthcare where he leads innovation and content efforts.  Prior to joining Siemens he was the founding Chief Executive Officer of the Indiana Health Information Exchange and was Director of Medical Informatics at the Regenstrief Institute, Inc., and a Sam Regenstrief Professor of Medical Informatics at the Indiana University School of Medicine.

He has spent over 25 years developing and implementing scientific and clinical systems and evaluating their value.  With his colleagues from the Regenstrief Institute, he created a community wide electronic medical record (called the Indiana Network for Patient Care) containing data from many sources including laboratories, pharmacies and hospitals throughout Indiana. The system connects nearly all acute care hospitals in the state and includes inpatient and outpatient encounter data, laboratory results, immunization data and other selected data.  In order to create a sustainable financial model, he helped create the Indiana Health Information Exchange, a not-for-profit corporation. In addition to advancing interoperability, Dr. Overhage has developed and evaluated clinical decision support including inpatient and outpatient computerized physician order entry and the underlying knowledge bases to support them.

Dr. Overhage was a principle investigator for the Observational Medical Outcomes Partnership (OMOP) with systematically evaluated the methods commonly used in observational research and continues that work in the Observational Health Data Sciences Initiative (OHDSI) which is a multi-stakeholder, interdisciplinary collaborative to create open-source solutions that bring out the value of observational health data through large-scale analytics.

Dr. Overhage is a member of the Institute of Medicine, a fellow of the American College of Medical Informatics and the American College of Physicians.  He received the Davies Recognition Award for Excellence in Computer-Based Patient Recognition for the Regenstrief Medical Record System. Dr. Overhage received his BA, with High Honors, in Physics from Wabash College and his PhD in Biophysics and MD from Indiana University School of Medicine.  Dr. Overhage was a resident in internal medicine, a medical informatics and health services research fellow and then chief medical resident at the Indiana University School of Medicine. He practiced general internal medicine for over 20 years including the ambulatory, inpatient and emergency care settings.