Human error responsible for drug injuries in hospitals, study says.

A significant number of patients at two Boston hospitals suffered drug-related injuries as a result of errors by doctors, nurses, and pharmacists, a team of reseachers Harvard University has found.

In a six-month study of procedures for administering drugs to patients in the two hospitals, the researchers documented 334 errors that led to 264 preventable "adverse drug events" (ADEs) and potential ADEs. Potential ADEs involved errors that were caught and corrected or errors that could have resulted in adverse reactions but did not.

To decrease the number of errors, the researchers recommended that hospitals shift their focus from educating and disciplining individuals who make mistakes to improving the procedures used to order and administer drugs.

"This systems approach is based on the concept that although individuals make errors, characteristics of the systems within which they work can make errors more likely and also more difficult to correct," the researchers wrote. They added that "while individuals must be responsible for the quality of their work, more errors will be eliminated by focusing on systems than on individuals."

The study, conducted in 1993 by a team known as the ADE Prevention Study Group, included 4,031 adult patients in 11 units at Massachusetts General Hospital and Brigham and Women's Hospital. The group recently published its findings in two articles in the Journal of the American Medical Association. (David W Bates et al., Incidence of Adverse Drug Events and Potential Adverse Drug Events, 274 JAMA 29 (1995); Lucian L. Leape ct al., Systems Analysis of Adverse Drug Events, 274 JAMA 35 (1995).)

Of all ADEs reported in the study, 1 percent were fatal (none of them preventable), 12 percent were life-threatening, 30 percent were serious, and 57 percent were significant. Of the life-threatening and serious adverse drug events, 42 percent were preventable.

Errors were most likely to occur at the ordering stage. For example, physicians prescribed the wrong drug or the wrong dose or ordered that it be administered at the wrong interval. In other cases, doctors prescribed drugs to which patients had known allergies or failed to take into account potential interactions with other drugs. Almost half the errors leading preventable ADEs and potential ADEs occurred at this stage.

About one-quarter of the errors occurred when nurses administered the drug to the patient. Mistakes included giving the wrong drug or the wrong dose and using the wrong technique. In other cases, nurses administered the drug to the patient at the wrong time or missed a dose completely.

Two major causes of the mistakes, whether committed by physicians or nurses, were lack of knowledge about the drug and lack of information about the patient. Another common cause of error was that a drug was misidentified, which often occurred when hospital personnel confused two drugs having similar names or packaging.

The researchers recommended greater use of computers in hospitals to make information about drugs and patients more accessible to physicians, nurses, and pharmacists. Computers could serve as a central repository of information and would reduce human error, they said.

Databases could include information about drug side effects, interactions, and dosages, for example. Computers containing patient information - such as laboratory test results, blood chemistry, and drug treatments - could be programmed to block inappropriate prescriptions. For example, a computer could alert the physician or pharmacist if a dose was too large for an elderly patient or if the patient was allergic to the drug.

The study team compared errors to symptoms of disease. "[T]reatment of the error or symptom does not correct the underlying malfunction," the researchers wrote. "In both errors and symptoms, `cure' requires attacking the underlying causes. It is modification of these underlying causes, or systems failures, that is most likely to be successful in reducing errors."