3234.0: Monday, October 22, 2001: 4:30 PM-6:00 PM | ||||
Oral Session | ||||
| ||||
| The 1999 Institute of Medicine (IOM) report "To Err Is Human: Building a Safer Health Care System" viewed medication errors and other health care misadventures in a systems context, suggesting that in aggregate these are the eighth most common cause of U.S. deaths and thus a critical public health problem. Recommendations were made to improve identification and monitoring of errors, and reduce their occurrence. Specifically, the Agency for Healthcare Research and Quality (AHRQ) was asked to "fund and evaluate pilot projects for reporting systems, both within individual health care organizations and collaborative efforts among health care organizations," and the Food and Drug Administration (FDA) to "work with physicians, pharmacists, consumers and others to establish appropriate responses to problems identified through post-marketing surveillance…." Post-marketing surveillance, of course, has long been used to identify relatively rare adverse drug reactions, most not due to medication error. This session deals with statistical issues in monitoring health care errors and other relatively infrequent medical mishaps. Statistical challenges and possibilities for misadventure in analyses of medical errors will be reviewed, as will statistical issues raised by studies at an AHRQ-funded Center for Education and Research in Therapeutics oriented to cardiovascular therapeutics, including their underuse or overuse. An FDA researcher will report on using a data mining technique, currently employed for detecting adverse drug effects from post-marketing surveillance data, to systematically detect synergistic multiple-level drug interactions, a preventable medication error problem difficult to assess systematically with current methods. | ||||
| See individual abstracts for presenting author's disclosure statement. | ||||
| Learning Objectives: At the conclusion of the session, the participant should be able to: 1. Identify the types of statistical problems inherent in epidemiologic analyses of the risks of adverse drug reactions and adverse medical events, and the implications for analysis, reporting, and critical appraisal of the literature. 2. Describe a Bayesian data mining approach to detecting adverse medical events through identification of outlying cells in large, sparse contingency tables, and its potential application to detection of medication errors. 3. Discuss statistical issues arising in design and analysis of post-marketing studies of arrhythmic medications, of medical devices used in patients with ischemic or valvular heart disease, as well as in studies of why important medications are underutilized. | ||||
| Peter B. Imrey, PhD | ||||
| Peter B. Imrey, PhD | ||||
| Introductory Remarks | ||||
| Adverse statistical events in the study of adverse medical events A. Russell Localio, JD, MPH, MS, Jesse A. Berlin, PhD, Cynthia D. Mulrow, MD, MSc | ||||
| Detecting adverse drug reactions: the statistical reality Elizabeth DeLong, PhD | ||||
| Application of screening algorithms and computer systems to efficiently signal combinations of drugs and events in FDA's spontaneous reports Ana Szarfman, MD, PhD | ||||
| Discussion | ||||
| Sponsor: | Statistics | |||
| Cosponsors: | Socialist Caucus | |||
| CE Credits: | CME, Health Education (CHES), Nursing, Pharmacy, Social Work | |||