Good morning. Mr. Chairman and Members of the Committee, thank you for the opportunity to speak with you this morning about adverse drug events. I am Michael R. Cohen, a pharmacist and president of the Institute for Safe Medication Practices (ISMP). ISMP is an independent nonprofit organization that works closely with practitioners, regulatory agencies, health care institutions, professional organizations and the pharmaceutical industry to provide education about adverse drug events and their prevention. Our focus has been on medication errors, including errors related to medical devices. March of 2000 will mark my 25th anniversary of working with a system for health care practitioners to voluntarily report medication errors and hazardous conditions to a national program. Practitioners who report medication errors do so altruistically, in an effort to make colleagues aware of situations that have caused patient harm. The program originated in a column in a professional pharmacy journal in which practitioners were provided with name-blinded descriptions of errors, suggested prevention strategies, and encouragement to voluntarily report medication errors. Reporter confidentiality has always been assured and maintained, and prompt feedback has been routinely provided to reporters through professional publications and personal communications. Therefore, the program has gained the trust and respect of practitioners throughout the nation.

In 1994, an agreement was reached with the United States Pharmacopeia (USP) to operate the USP-ISMP Medication Errors Reporting Program and today ISMP provides independent review of all reports. Also, as FDA MEDWATCH partners, ISMP and USP share information with the FDA and have ongoing dialog with the agency about errors that relate to drug nomenclature (proprietary and nonproprietary names), labeling, packaging and medical device design. Information about medication errors and other adverse drug events, along with prevention recommendations, is shared with the medical community through a biweekly publication, ISMP Medication Safety Alert! and through 16 journal publications that reach nurses, nurse practitioners, pharmacists, physicians and physician assistants. In addition, our communications reach world regulatory authorities and pharmaceutical manufacturers internationally through regular publication in the WHO Pharmaceuticals Newsletter and through ongoing columns in Australia, Canada and a monthly feature in the Journal of the International Pharmaceutical Federation.

Medication Error v. Adverse Drug Reaction

I will first offer some thoughts to clarify today's discussion. As noted in the GAO Report Adverse Drug Events: Incidence and Treatment Costs Are Uncertain, adverse drug events stem from two different sources: adverse drug reactions and medication errors. Adverse drug reactions are previously known or newly detected side effects of drugs that may occur in the course of error-free medication use. Medication errors, in contrast, are preventable events in the medication use process - errors in the prescribing, dispensing, administration and use of medications. This distinction is important in our discussion, as minimizing adverse drug reactions and medication errors require different mechanisms for identification, evaluation, and improvement.

Let me explain. Heart valve damage believed to be associated with the concomitant use of fenfluramine and phentermine for weight loss is a frequently cited adverse drug reaction. Before recognition of the potential link between these products and the adverse cardiac effects, the heart valve damage would appropriately be categorized as an adverse drug reaction -- the medication combination was prescribed intentionally, dispensed correctly, and used by patients according to instruction. Nonetheless, this error-free use resulted in adverse drug events and medical harm - harm related to an adverse drug reaction, not a medication error.

A medication error, by contrast, involves a preventable event somewhere in the medication use process that sometimes causes patient harm. Such an event could occur in a hospital where a nurse or pharmacist misreads a physician order and, subsequently, dispenses or administers an incorrect dose - either too high and yielding unsafe toxicity or too low and yielding no benefit. Similarly, patients may erroneously use medications in the home. For example, an asthmatic patient may use an inhaler incorrectly or patients may fail to comply with their physician's directions for medication use. While each of these examples is a medication error, patients are not always harmed by such errors. If any degree of patient harm result from a medication error, the event is classified as an adverse drug event.

Distinguishing between the two causes of adverse drug events is important in developing effective solutions. While some recommendations can address both sources of the problem, other solutions must be unique to the source ---- medication error or adverse drug reaction. Because my work has focused primarily on the prevention of medication errors, my remarks will be limited to this problem.

Medications are an unbelievable blessing, but humans must safely prescribe, prepare, dispense, and sometimes administer these drugs. Yet, humans are fallible, so errors do happen. Medication use has been complicated by the large number of new drugs we see every year -- by the fact that today's patients are typically older and sicker, requiring ever more complex treatment strategies -- and even by the proliferation of over-the-counter products. And the GAO report is correct in concluding that we do not have a good sense of the frequency with which medication errors occur or when those errors cause harm. However, there is no denying that much can and should be done to improve medication safety, even in light of this fact.

Role of Reporting

One primary mechanism to improve our understanding of errors and their causes is to stimulate health care professionals, health care systems, and patients themselves to report medication errors. However, ISMP suggests a broader scope than focusing on just those errors that cause harm, adverse drug events. Reports of errors and "near misses" that, luckily, have not resulted in harm, and hazardous conditions that could lead to harm, are as necessary for learning about errors and their causes as reports of errors that have actually resulted in patient harm. Such a broad scope of data provides content experts with better information to gain new knowledge about causes of errors and share that knowledge with the entire health care community to prevent those errors from happening again.

Stimulating and supporting reports of adverse drug reactions is also important. In the immediate aftermath of an adverse drug event, it may be difficult to identify whether a medication error, an adverse drug reaction, or both caused the event. Therefore, reporting systems should coordinate their efforts to ensure that the appropriate experts are involved in the evaluation. Today, ISMP serves a primary role in ensuring that the USP Medication Errors Reporting Program is well coordinated with the FDA MEDWATCH program to provide a better chance at identifying the sources of adverse drug events.

With any reporting system, the following issues must be addressed:

• the primary goal of the system;
• active participation by health professionals and consumers;
• system design and structure;
• the scope of reports requested;
• confidentiality and legal discovery of reported information;
• expert analysis and use of results; and
• communication of the knowledge learned from the reporting system.

Reporting systems must go beyond counting errors to focus instead on learning about the events and developing and disseminating solutions. A representative sampling of error reports is more valuable than a large volume of reports because most errors are not typically "individual" in nature, but rather are repetitive events resulting from system failures that can be prevented. Therefore, you don't need to know about all errors to develop effective safety strategies. For example, our experience with potassium chloride is instructive. We knew over 15 years ago from scattered reports of patient deaths with potassium chloride concentrate injection, that the drug should not be available in undiluted form in patient care areas. I'm happy to report that over the past two years, over 90% of US hospitals replaced this drug with pre-diluted forms or had their pharmacies dilute it for clinical use. In addition, it is important to understand that massive collection of data could easily overwhelm the system, slowing analysis or making it impossible to read through reports and perform proper analysis.

To stimulate participation in reporting programs, voluntary, non-punitive reporting has proven to be an effective method for obtaining needed information about errors. Existing mandatory reporting systems, which are inherently punitive in nature, have suppressed reporting and discouraged the open discussion of errors, which is necessary to develop and disseminate appropriate safety strategies. Conversely, voluntary reporting has been far more successful at garnering a sufficient representative sampling of error reports, providing expert analysis, and disseminating high-leverage safety strategies effectively. Therefore, voluntary reporting programs will likely be more effective than state mandatory reporting programs.

The Institute of Medicine report recommends mandatory reporting to state agencies. We believe that it will be extremely difficult or impossible for state agencies to acquire sufficient expertise to analyze reports appropriately. Many root causes of error are buried deep within healthcare systems and require expert evaluation to extract the most important information about where the failures have occurred. For example, a recent case occurred in Denver where three nurses were indicted for their role in a baby's death. The grand jury knew only that together the nurses gave a tenfold overdose and that they changed the route of administration from intramuscular to intravenous. However, our examination of the case revealed at least 54 different failures in the system, all aligning at the same moment in time, that set up this error. At trial, once the jury understood the extent of the problem a not guilty verdict was returned.

Because mandatory reporting systems typically receive information on a wide array of medical issues, it is virtually impossible to provide content experts to analyze such disparate data effectively to develop safety strategies for error prevention that can be extrapolated nationwide. As a result, targeted voluntary programs that are national in scope and focus on specific categories of medical error (in this case medication error) will likely be more effective.

To be successful, reporting must come from front-line practitioners, be unrestricted, and include serious errors, "near misses," and hazardous situations. For that reason, we believe that the Institute of Medicine report contains a serious oversight. The report recommends that state-administered, mandatory programs receive reports of serious and fatal events caused by error, and voluntary systems, such as the MERP, receive reports of "near misses" and less harmful errors only. Although the IOM clearly acknowledges the immense value of voluntary practitioner reporting in learning about errors, they fail to provide a channel for practitioners to report serious errors to voluntary reporting systems. Only front-line practitioners who are involved in the event can provide the details necessary for effective analysis and selection of strategies to reduce or prevent the likelihood of recurrence. Further, reporting from health care professionals in all practice settings is necessary in order to address the fragmentary nature of adverse event information noted in the GAO report. Medication errors occur in many settings outside of hospitals-in long-term-care facilities, in physician's offices and other ambulatory care settings, in retail pharmacies, in rehabilitation centers, and in the patient's home. To maintain patient confidentiality, patient identity should not be included in reports. Following expert analysis, provider and reporter identities could be removed from error reports. Moreover, federal legislation should extend state peer review protections to all information submitted to a reporting program. This level of confidentiality/legal protection is necessary to enhance the current program and to keep reporting prolific, honest, and effective. Additionally, providers must continually be encouraged to disclose medication errors to patients/families honestly and promptly.

The ISMP MERP has been successful in collecting reports about medication errors, analyzing the reports, and publishing our findings about the causes of errors and safe practice recommendations. But we need cooperation from government and the health care community. The reporting program itself does not and should not have the power to mandate implementation of recommendations that stem from error analysis. Healthcare providers and others (including pharmaceutical and device manufacturers and services companies) must be held accountable for the implementation of appropriate safety practices - before more tragic and preventable medication errors occur! This means safer manufacturer labeling and packaging, failsafe device design, and computer software that captures prescriptions for unsafe drug orders. In addition, findings from the reporting program must be shared with regulatory agencies, accrediting bodies, federal agencies, and others - to assist those bodies in establishing and enforcing feasible safety recommendations as standards that must be upheld by all participants in the healthcare process. These bodies must also be held accountable for enforcing implementation of proven safety practices and monitoring compliance through on-site surveys. We must also gain the cooperation of professional organizations and regulatory agencies to improve awareness of current voluntary reporting systems and enhance participation. The airline industry has done an excellent job in making the voluntary Aviation Safety Reporting System successful. Because of this program, there is continual learning on how to minimize the impact of human error. On a typical coast to coast flight many human errors are made. Yet, because of system safeguards, these errors do not interfere with a safe landing. On several occasions while waiting for a flight, I have asked airline pilots about the program. They not only know of it, many stated that they have reported to it and usually they produce a copy of the ASRS reporting form.

Any reporting program must include a formal and effective communication program that reaches the entire healthcare community and consumers, so that all participants in the healthcare process can use the information gained through the reporting system to help ensure patient safety. Consumers are often forgotten in our attempts to reduce error and adverse drug events, but they play an essential role both in reporting and minimizing errors. An accurately prescribed and dispensed medication may still result in error and an adverse event if the patient is not armed with sufficient information and ongoing oversight to use the medication correctly.

Beyond Reporting: Additional Recommendations to Reduce Error and Improve Medication Use

The GAO report accurately identifies the need for a more coordinated effort to improve medication use-reducing medication errors and identifying and reducing the incidence of adverse drug reactions. ISMP has identified several highly critical issues upon which to focus attention and make better use of the powerful technology known as medication. In addition to reporting these issues are:

• the need for systematic practitioner review of all drug product proprietary names and packaging before applications are submitted to FDA
• the need for patients themselves to understand their important role in error prevention
• the need for continuous quality improvement activities to address safety issues in our nation's health care practice sites
• the need for improved technology

Practitioner Review

Practitioner review for all drug products prior to FDA approval is clearly warranted if you subscribe to the positive error-prevention techniques found in Failure Mode and Effects Analysis (FMEA). FMEA is a technique that is used successfully in many other industries, such as the transportation and nuclear power industries, to guard against human errors by providing a systematic analysis of the potential for user error, to prevent them before they happen.

Some have predicted that full implementation of practitioner review in the United States within the context of Failure Mode and Effects Analysis could prevent virtually all of the medication errors linked to trademarks and packaging that practitioners report to USP, ISMP and FDA.

Patient Education and Involvement

Another important component is improving patient understanding of their important role in safe medication use and error prevention. About 25% of reported medication errors stem from having the wrong drug administered due to confusion of proprietary and nonproprietary names. An educated patient or caregiver can be a crucial last check on the safety of any medication. Patients must understand how crucial their role in ensuring safety.

Patients and caregivers must insist upon full explanations of new medications while in the doctor's office, and consistently demand that medication purpose and full instructions are written on each new prescription. Such clarity will help patients be better informed and help pharmacists ensure that their interpretation of the prescription is consistent with the prescriber's intent. This can help immeasurably in avoiding common misunderstandings on drug orders.

Further, patients must demand that their health care practitioners have access to crucial patient information that can help prevent errors. Too many serious preventable adverse drug events stem from insufficient information about the patient's clinical condition. Currently the availability of this information is often fragmented. For example, a physician treating a patient may be unaware of a medication prescribed by another physician. This is especially true in ambulatory care. Studies show that many prescriptions are written for drugs that are inappropriate for the patients' individual health conditions. If better clinical information about patients, including laboratory data, current therapy, information about chronic diseases, organ function, allergies, and more, is available at the pharmacy level, the pharmacist can screen drug orders appropriately, be a vital safety check within the medication delivery system, and prevent untold numbers of errors. If we are to provide safer care, our nation's health care practitioners simply must have improved access to critical patient information. Harvard researchers (Leape LL et al. Systems analysis of adverse drug events. JAMA 1995;274:35-43) showed that over 40% of adverse drug events can be tied to a lack of critical patient information and drug information at the time of prescribing, dispensing and administration of medications. The availability of web sites or "smart cards" where patients could voluntarily have confidential clinical information available for use by their health care practitioners must be a part of the overall effort. Access must be in full control of the patient.

Continuous Quality Improvement

Data from the ISMP Medication Error Reporting Program reveals that medication errors are repetitive in nature. Yet, as mentioned earlier, it seems that too many organizations and individuals within the health care community do not believe it will happen to them. It is not until a serious error hits close to home before aggressive prevention efforts are undertaken. This "head-in-the-sand" attitude is unacceptable, as such attitudes could be responsible for thousands of easily preventable deaths. The development and implementation of continuous, organization-wide, quality improvement (CQI) efforts should be the highest priority in all health care settings. Such quality improvement efforts must be aimed specifically at preventing well-known and repetitive categories of errors, eroding patient confidence in our health care system. In addition, errors known to have occurred within an organization, including ambulatory care pharmacies, long term care facilities, home care agencies, etc., must be recorded, and regularly reviewed and analyzed. Organizations must be held accountable to assure that CQI programs are operating to address problems that have been revealed through internal and external reporting programs.

Informational tools like ISMP's Quarterly Action Agenda, which is a readily available list of medication error problems compiled from our nation's reporting program, can be a backbone of any CQI effort. ISMP is prepared to assist state boards, health departments, accreditation agencies, regulatory authorities, and individual organizations in using such informational tools to develop effective CQI strategies that can successfully stop repetitive errors.

Improved technology

As noted in the GAO report, research (Bates DW et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA 1998;280:1311-16) shows that over half of prescribing errors can be prevented through computerization physician order entry (CPOE). However, an ISMP survey of our nation's computer systems shows that fewer than 13% of U.S. hospitals even have the capability for CPOE. Further, computerization of prescribing in ambulatory care is thought to be under 5%. As noted, about 25% of medication errors reported to our program result from drug name confusion. Continued reliance on written and verbal prescribing is a large part of the problem. Drastic improvements are also needed from software and information vendors. Our research (ISMP Medication Safety Alert! February 10, 1999 - www.ismp.org) shows that most in-use software used in hospitals today does not alert users to prescribing or processing errors in an accurate and efficient manner. System vendors and organizations must jointly accept responsibility for implementing systems that merge patient and drug information in a way that consistently prevents adverse events. Such systems will also improve detection of adverse drug events.

Bar coding systems are often recommended to assure accuracy of drug dispensing and administration. Although we have begun to see such systems develop, mainly in ambulatory care pharmacies which use bulk containers, many manufacturers do not place bar codes on unit dose packaging, which is more commonly used in hospital and long term care practice. Thus, hospitals have been slow to implement bar coding systems without such support because each unit dose package must be repackaged and relabeled with bar codes within the facility. This in itself increases the possibility of error. Further, there are several different types of bar codes. This makes it difficult for technology vendors to market their systems. Currently, the pharmaceutical industry has not agreed upon a single standard method of bar coding. While many within the industry are working to standardize bar code systems for unit dose packaging of pharmaceuticals, pharmaceutical manufacturers must join in this effort by assuring that all drug packaging has a standardized, readable bar code or other machine-readable code.

Improving medication use and reducing adverse drug events by reducing medication error is essential to providing quality health care. Efforts to improve reporting and implement high-leverage solutions to address the causes of errors identified through reporting efforts are important steps in improving medication use. Physicians, nurses, pharmacists, other healthcare professionals, and patients must all be involved collaboratively to yield the greatest benefit.

Clearly there is a need for expansion of independent voluntary practitioner reporting programs which are effective in gathering information about errors but could reach more people and have even greater impact if resources were available.