More Hospitals Report Medication Errors, But USP Finds Few Changes
Little change. In both years, the most frequently identified causes of error were staff performing inadequately and not following protocol.
Staying steady at 3% was the reported percentage of errors that harmed or killed patients. For the second consecutive year, insulin, heparin, and morphine were the products most commonly associated with potential and actual errors. And distractions, workload increases, and inexperienced staff members remained the most common factors contributing to medication errors.
In 2000, 184 hospitals reported 41,296 medication-error records to USP, up from 6,224 errors reported by 56 facilities in 1999.
Pharmacist Diane Cousins, USP's vice president of practitioner and product experience, said she was surprised that many of the trends in medication errors reappeared the second year, given the much larger number of hospitals reporting.
"I would have expected some difference somewhere along the line, whether it was the severity of the errors or whether it was the types or causes," she said. "And yet, I think that's probably one of the significant findings of this report. Because what we believe this is indicating to us is it's a call to action for the areas that need attention in health care."
The report, "Summary of Information Submitted to MedMARx in the Year 2000: Charting a Course for Change," summarizes medication-error information submitted to USP's MedMARx system, a Web-based database for hospitals to anonymously report and track medication errors in a standardized format. About 3% of the nation's hospitals participated in the program in 2000, according to USP.
The report said that the distribution of where errors occurred in the medication-use process was similar for both years, and that the administration phase continued to be the most frequently reported point at which errors originated.
But, according to USP, the administration phase is the closest to patient care and the least difficult stage at which to identify errors. As such, its high visibility may obscure the actual phase at which an error arose.
Although 97% of actual errors did not result in patient harm, 69% of those errors reached the patient.
Staffing and computer entries are on the rise. Staffing issues became one of the five most frequently cited reasons for medication errors in 2000.
"We weren't sure [hospitals] would document this as diligently as they have," Cousins said.
She said six staffing issues accounted for 33% of the factors that contributed to medication errors, an increase of 6 percentage points from 1999.
The New England Journal of Medicine reported in May that shortages of hospital nurses decrease the quality of patient care.
Cousins noted that computer entry errors moved to fifth place, up two places, among the most frequently reported reasons for medication errors.
"An important thing about computerization is that it isn't computerization alone that helps reduce errors. It's good computerization," Cousins said. "Computerization is becoming so prominent, we want to be sure that we don't just put these systems up and not have adequate training for staff that are expected to use them effectively and accurately. So I think that is a concern for us more and more-that we are certain that not only are [systems] designed well, but people are trained well on them."
One encouraging finding in the 2000 report, Cousins said, is that the actions taken by the hospitals were documented for 60% of the harmful errors, up from 38% in 1999. The most frequently documented action was informing the staff member who made the initial error. In 1% of the cases overall, hospitals reported changing a policy or procedure.
"I don't think that any low percentages necessarily indicate that these things are not happening, but more so they indicate that hospitals are not documenting [actions taken] as part of their medication-error reporting system," Cousins said. "I think they are taking action."
Cousin said that oftentimes hospitals wait to see how a trend develops over time before they respond to a single report, but the records for the early errors will not include the later action.
USP plans to add a "batch update" option to MedMARx so users can "go to all those records that were identified in a trend and update them all at once," Cousins said.
No release date has been set for the new option.
Next time, Cousins said, USP plans to examine the data in new ways.
"We will try to drill down further to find out more," she said.