Dana-Farber Institutes New Drug Ordering Rules

August 1, 1996

SAN DIEGO--If a cancer patient dies because of a medical error, such as an incorrect dosage, it is only human nature to blame the tragedy on one incompetent person. But finger pointing is not a useful approach, says George Demetri, MD, since generally the problem has multiple causes, and a system overhaul may be in order.

SAN DIEGO--If a cancer patient dies because of a medical error,such as an incorrect dosage, it is only human nature to blamethe tragedy on one incompetent person. But finger pointing isnot a useful approach, says George Demetri, MD, since generallythe problem has multiple causes, and a system overhaul may bein order.

Dr. Demetri is associate director in the Office of Medical Affairsfor the newly formed Dana-Farber/Partners Can-cerCare System,composed of the Dana-Farber Cancer Institute and adult oncologyoperations from Mass General and the Brigham & Women's Hospital.

Dr. Demetri spoke at the cancer care conference sponsored by theSociety for Ambulatory Care Professionals and Health TechnologyAssessment of the American Hospital Association.

As everyone in the room was aware, Dr. Demetri and his colleagueslearned their lessons the hard way--in the face of a firestormof negative publicity after a patient on an experimental protocoldied from chemotherapy overdoses at Dana-Farber. Another patientsuffered severe cardiac failure after receiving the wrong chemotherapydoses.

"There is a clear complexity of treatment errors in mostmedical mistakes," he said, whether a wrong leg is amputatedor an incorrect drug dose given. "One single individual actingalone cannot, in general, cause these errors," he said.

Citing liberally from the pioneering work on medical errors fromDr. Lucian Leape, Harvard School of Public Health, Dr. Demetrisaid that most treatment errors represent simultaneous failuresof multiple and interacting clinical systems.

"I think there are great lessons to be learned in studyinghow the systems failed, where the safety overlaps were insufficientto prevent the error, and how to improve our clinical fail-safesystems," he said.

In their work on medical errors, Dr. Leape's group at Harvardhas broken down the possible contributing factors into severalcategories, he said, such as drug ordering by physicians, thedrug order fulfillment system, clinical monitoring, and the clinicalcare environment.

This systematic analysis was useful to the multidisciplinary groupof physicians, nurses, and pharmacists who reviewed the situationat Dana-Farber.

In response to the errors, Dana-Farber immediately implementednew clinical checks and balances: All high-dose chemotherapy ordersare now reviewed by a separate committee. The institution beganusing new computer software to block possible high-dose errors,and new educational initiatives were launched for all caregiversregarding research studies. What's more, all chemotherapy ordersare now reviewed and co-signed by staff attending physicians.

One area of concern is how to put new protocols online in a waythat minimizes risk. Generally, he said, "there is one principalinvestigator who really knows the regimen, but probably at least10 people who wind up implementing it."

Another "set up for error" can occur when a repetitivedrug order is transcribed, he said. Institutions must have safeguardsto catch mistakes that may be repeated over several days due todrug

ordering and order fulfillment systems. This is a critical issuefor oncology, since chemotherapy is typically given in repetitivecycles. "Often there is one decision point for the chemotherapyorder and then no other check," he said. "That one decisionpoint can multiply many times."

Another fact that surfaced during the institutional review atDana-Farber was that nurses typically have no back up to catchtheir mistakes.

He pointed to a study from the Leape group, which reported ondrug errors that occurred during a 6-month period at two Bostonhospitals in 1993. Of all the admissions, 7% of patients sufferedfrom an adverse drug event. The majority of drug errors occurredin the process of physician ordering and nursing administration--eachabout 40%.

"The errors were balanced between physicians and nurses,"Dr. Demetri said, "but about half of the physician errorswere intercepted--86% by nurses and 12% by pharmacists."Only 2% of the administration errors (generally occurring withinnursing) were intercepted.

"The doctors are on the high wire with a net below them,but the nurses operate without a net," he said. "Weshould keep this study in mind when we review systems in thesecost-conscious times and ask, for example, if it is possible forone nurse to care for 60 cancer patients. These data suggest theanswer is no."