the health care community is focusing attention on reducing preventable medication risks through a "systems approach" to risk management. Recent reports from the Institute of Medicine (ION) and the Food and Drug Administration (FDA) have highlighted problems of medication errors and risk in the delivery of health care. 1, 2 In addition to raising awareness of this problem, these reports have heightened the public's awareness of medication risks and spurred initiatives to improve risk management in health care delivery.
||hile pharmacotherapy significantly improves and prolongs the lives of many patients, medications are not without risk. Although some risks associated with medications are unavoidable or unknown, others are known and preventable (see Figure 1). Increasingly,
Growing Awareness of Medication Errors
The IOM report To Err is Human, published in 2000, focused public attention on medical and medication errors. 1 This report cited data showing that approximately 44,000 to 98,000 American die each year as a result of medical errors, with approximately 7,000 deaths attributable to medication errors. 1, 2 According to another study cited in this report, preventable adverse drug events were experienced by nearly 20% of hospitalized patients. 1, 3
These statistics demonstrate that medical errors have a huge public health impact: they represent the eighth leading cause of death in the United States and are associated with estimated direct and indirect annual costs of $17 billion to $29 billion. 1, 2 The IOM report stressed that medical errors should not be viewed as the fault of specific individuals, but rather the result of system-wide flaws.
In 1999, FDA issued a report, Managing the Risks from Medical Product Use, which described a detailed risk management model for preventable adverse events. 2 This report began with a review of current problems in the management of the risks of medical products. While some risks of medications are known and unavoidable, this report cited data showing that about half of adverse drug events in an urban tertiary care hospital were preventable. 2, 4 This report focused on preventable risks and presented various strategies for their prevention through formal, system-wide risk management programs.
A spate of recent FDA-initiated drug withdrawals (12 in 4 years) has underscored the atmosphere of reduced tolerance for medication risk (see Table 1). Within one recent 12-month period, five drugs were withdrawn for safety reasons: fenfluramine, dexfenflramine, terfenadine, mibefradil (Posicor-Roche), and bromfenac sodium.5 In the case of mibefradil (Posicor-Roche), the withdrawal was spurred not by risks inherent to the drug, but because the drug had the potential to cause dangerous interactions with a large number of other medications. 2
Traditional Approaches to Risk Management
One of the key roles of FDA review has been to identify risks of medications during the premarketing period and ensure that these risks are adequately communicated in product labeling. Data collected during the premarketing period is limited by the number of patients enrolled in clinical trials (usually a few thousand, compared with millions of patients who may take the drug following FDA approval). However, when FDA approves a medication for use in the United States, the drug's benefit-to-risk ratio is considered to be well quantified and, upon approval, FDA establishes expectations for the product's safety and efficacy. FDA attempts to communicate known risks associated with the product to health care professions through a variety of mechanisms, including the product labeling.
In some cases, however, a safety concern associated with a medication is not identified until after the medication has been marketed. When these risks are identified, FDA works with the manufacturer to communicate the newly identifies risks to the health care community and general public, using tools such as press releases, label changes, Dear Health Professional letters, and patient package inserts. However, the identification of a new hazard can trigger a variety of emotional responses, and some consumers may respond to the new information with anger. These emotional responses may contribute to a crisis atmosphere, creating pressure to shorten the timeframe for regulatory decision making.
Limitations of Traditional Risk Management Tools
FDA and other authorities have pointed out that traditional risk communication tools, such as Dear Health Professional letters and product labeling changes, do not lead to desired changes in physicians' behavior. FDA has stated that many drug withdrawals have resulted, in part, from the health care system's inability to manage the known and preventable risks associated with these products. FDA concluded that the risks were not being properly managed and that the medications were unsafe under routine conditions of use. A brief overview of two recent withdrawals- cisapride and troglitazone- illustrates some of the important limitations in traditional risk management strategies.
Cisapride (Propulsid-Janssen) was approved in 1993 for nocturnal heartburn, and was widely used, with approximately 5 million outpatient prescriptions filled in 1995. 6 During the postmarketing period, cisapride was found to be associated with a risk of QT prolongation and arrhythmia when administered to patients with specific conditions or when coadministered with certain other medications. In June 1998, label warnings were strengthened through the use of a "black box" warning to emphasize the contraindication of coadministering cisapride with other drugs known to prolong the QT interval. FDA also sent 800,000 Dear Health Professional letters, issued a talk paper on the subject, and posted information on its Web site.7 These warnings were updated in January 2000 to recommend that physicians perform an electrocardiogram and certain blood tests prior to prescribing cisapride. In this update, FDA noted that an analysis of 270 adverse event reports, including 70 fatalities, found that approximately 85% of the cases occurred in patients with identifiable risks.8
Despite these measures, traditional risk communication efforts had little or no impact on physicians' prescribing practices. An analysis of the research databases of two managed care organizations and a state Medicaid program revealed virtually no change in cisapride prescribing habits (based on computerized patient records) before and after FDA's 1998 regulatory action (see Figure 2). These findings generally indicated the failure of traditional risk communication tools to change physician prescribing behavior.
As of December 31, 1999, the use of cisapride was associated with 341 reports of heart rhythm abnormalities, including 80 reports of death. Cisapride was withdrawn from the market on July 14, 2000.9 However, cisapride remains available to a small number of patients through a restricted distribution program.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -