The primary suspect, Phillips says, is a beginning-of-the-month increase in pharmacy workloads and a consequent increase in their error rates.
“Government assistance payments to the old, the sick and the poor are typically received at the beginning of each month. Because of this, there is a beginning-of-the-month spike in purchases of prescription medicines,” Phillips says. “Pharmacy workloads go up and – in line with both evidence and experience — error rates go up as well. Our data suggest that the mortality spike occurs at least partly because of this phenomenon.”
Phillips and his coauthors examined all United States death certificates from 1979 through 2000 to analyze the 131,952 deaths classified as fatal poisoning accidents from drugs. A small number, 3.2 percent, of the deaths were from adverse effects of the right drug in the right dose. The vast majority, 96.8 percent, resulted from medication errors – the “wrong drug given or taken,” or “accidental overdose of drug,” or “drug taken inadvertently.”
The study excluded deaths from overdose of street drugs or from intentional poisoning (suicide or homicide).
The beginning-of-the-month mortality spike was particularly pronounced in people for whom the mistakes proved rapidly fatal – those who were dead on arrival at a hospital, died in the emergency department or as outpatients. In this category, deaths jumped by 25 percent above normal.
But could it be that the mortality spike is due not to pharmacy error but simply to the increased number of people buying, then consuming drugs?
To test this, Phillips and coauthors ran analyses on populations of the elderly and the poor. If increased consumption alone was to blame, the researchers reasoned, mortality would be highest in the groups relying on government assistance and therefore purchasing their medicines at the start of the month.
The beginning-of-the-month spike was similar across groups, however. The spike was as evident in the young and well-off as in the elderly and poor, suggesting the problem was at least partly due to an increase in pharmacy error at the beginning of the month.
Phillips notes that the National Center for Health Statistics database used in the study did not contain highly specific clinical information – no information on prescription type, dosage, days supply, etc. – and he urges further research with data richer in this kind of detail.
To reduce the medication-error death rate, Phillips suggests that pharmacies (that don’t already do so) consider increasing staffing levels at the beginning of the month and that government officials consider spreading assistance payments out over the entire month.
“Even in the absence of policy changes or further research,” Phillips says, “it is appropriate for both patients and clinical staff to be especially careful to check the accuracy of their prescriptions at the beginning of each month. If this is done, it seems plausible that some lives will be saved.”
Phillips’ coauthors are Jason R. Jarvinen, sociology student, UCSD, and Rosalie Phillips, executive director of the Tufts Health Care Center, Tufts University.
From UC San Diego