Inside this Issue:

Download the Newsletter

Editors:

Joseph Cummings
Charles Dizard

Informatics Today is a publication of the University of Pittsburgh Department of Biomedical Informatics

The University of Pittsburgh is an affirmative action, equal opportunity institution. Published in cooperation with the Department of University Marketing Communications.

UMC64565-0608

Developing a System to Detect Adverse Drug Events in the Nursing Home Setting

Steven Handler, MD, MS, CMD, a doctoral fellow in the Department of Biomedical Informatics, is leading a research effort to explore the use of a clinical event monitoring system to detect and reduce the number of adverse drug reactions (ADRs) among nursing home residents. ADRs can be defined as unintended or noxious responses to a prescribed drug. A clinical event monitor is an automated clinical decision support system that provides feedback to health care professionals that a potential ADR is present, based on information available in an electronic format.

The Centers for Disease Control and Prevention's National Health Care Surveys report that, in 2004, there were 16,100 nursing homes in the United States, with 1.73 million beds and 1.49 million current residents. The chance of nursing home placement is currently 45 percent for those older than 65, and the number of people requiring nursing home placement is expected to double by 2020.

In nursing homes, it is estimated that ADRs occur between 1.19 and 7.28 times per 100 resident months. This translates into approximately 2 million ADRs occurring annually in U.S. nursing homes and has an estimated cost of more than $7.6 billion annually. Data from large studies suggest that nearly half of these ADRs are preventable and as many as 80 percent are associated with problems with medication monitoring.

Comprehensive chart review has been the standard for detecting and researching ADRs. This process can be time and cost intensive, but a functional clinical event monitoring system is a tool that has been shown to detect ADRs accurately and in a cost-effective manner.

The design of a clinical event monitoring system requires an agreed-upon set of standards or signals that the program will use to determine the presence of a clinical event. Handler and his colleagues approached this design challenge by intensively searching the published literature for potential signals that represent a potential ADR. This search of 29 publications yielded a list of 80 potential signals. The 80 signals were placed in four categories: laboratory and medication combination signals, medication concentration signals, antidote signals, and Resident Assessment Protocol (RAP) signals.

With the list of potential ADR signals in hand, Handler and his colleagues enlisted the help of a panel of experts in geriatrics. The panel's task was to reach a consensus on which of the 80 signals would most likely be associated with a potential ADR in the nursing home setting. The panel consisted of physicians, pharmacists, and advanced practitioners such as physician assistants and nurse-practitioners.

The panel of experts completed a two-round Delphi survey via the Internet and reached consensus on 40 signals. Of these, 15 were laboratory and medication combination signals, 12 were medication concentration signals, 10 were antidote signals, and three were RAP signals. The complete results of this work were published in the May 2008 issue of the Journal of the American Geriatrics Society.

This consensus list is the first step toward developing a system to detect and reduce ADRs in the nursing home setting and represents the basic science of clinical event monitor development. Handler; Shyam Visweswaran, MD, PhD, an assistant professor in the Department of Biomedical Informatics; and Melissa Saul, director of the Clinical Research Informatics Service, have recently completed the initial evaluation of the clinical event monitor using the 37 signals applied to data from a single University of Pittsburgh Medical Center-owned nursing home. The complete results of this study have recently been accepted for publication in the Proceedings of the American Medical Informatics Association Annual Symposium. This research was supported in part by the National Institutes of Health, Roadmap/NCRR /University of Pittsburgh Multidisciplinary Clinical Research Career Development Award and the American Federation for Aging Research.

Handler is an assistant professor in the Department of Medicine's Division of Geriatric Medicine and has a secondary faculty appointment in the Department of Biomedical Informatics. Handler plans to continue developing and assessing clinical decision support systems, primarily in the nursing home setting, that focus on medication monitoring. He has recently received funding from the Commonwealth Fund to determine the prevalence and use of health information technology for clinical care processes in the nursing home. This grant also will help to determine why nursing homes lag behind hospital and ambulatory care settings in their adoption and use of health information technology.