Fewer e-prescribing alerts could save lives

NEW YORK —More than a million Americans become sick, injured or die every year because of medication errors, many of which result from physicians’ stereotypically illegible handwriting on prescription pads.

Electronic prescribing has long been touted as a solution to the problem, especially e-prescribing systems that automatically alert prescribers to potential interactions between drugs. According to a study published last month in the Archives of Internal Medicine, however, doctors override more than 90% of drug interaction alerts. Nevertheless, the systems still prevent a significant number of harmful drug interactions, and they could go further if needless alerts were eliminated.

“Electronic prescribing alerts in ambulatory care may prevent a substantial number of injuries and reduce healthcare costs in Massachusetts,” the researchers wrote in their conclusion. “Because a few alerts account for most of the benefit, e-prescribing systems should suppress low-value alerts.”

Harvard Medical School associate professor and Dana-Farber Cancer Institute researcher Saul Weingart and colleagues used data from 279,476 alerted prescriptions written by 2,321 ambulatory care clinicians in Massachusetts who used ZixCorp’s PocketScript e-prescribing application between Jan. 1 and June 30, 2006. PocketScript allows physicians to submit prescriptions to pharmacies by computer, creating a patient medication profile based on previous prescriptions, checking newly prescribed medications against a database and alerting physicians to potential low-, medium- or high-severity interactions.

An expert panel of four physicians and three pharmacists estimated the likelihood and severity of potential drug interactions that might occur with each alert, along with likely harm to the patient and healthcare utilization needed to address them, using a sample of 100 common drug interaction alerts. The researchers then created cost estimates using published sources and payer data, such as the average cost of a medical hospitalization or doctor visit due to an adverse drug event.

Based on the estimates, the researchers found that, even though physicians overrode 91% of drug interaction alerts, the alerts they accepted likely prevented 402 adverse drug events. These included three potential deaths, 14 cases of permanent disability and 31 cases of temporary disability. The alerts also resulted in 34 fewer emergency room visits, 39 fewer hospitalizations and 267 fewer doctor office visits, producing a cost savings of $402,619. In all, the researchers estimated, it took 331 alerts to prevent one adverse drug event, and 10% of alerts may have accounted for 78% of cost savings.

One problem, however, was “alert fatigue,” and the researchers wrote that eliminating unneeded alerts would help prevent it. “[We] believe that the technology’s ability to prevent [adverse drug events] makes it worthwhile, and our findings suggest that significant efficiencies could be gained by reducing overalerting,” the researchers wrote. “Doing so would mitigate alert fatigue, thereby increasing the percentage of clinically significant alerts accepted and the number of [adverse drug events] averted.”

The researchers also acknowledged several limitations in the study, such as the exclusion of redundant alerts and formulary adherence and allergy alerts, as well as not taking into consideration cost savings resulting from formulary adherence, prevention of allergic reactions, use of generic drugs and legible prescriptions.

Nevertheless, streamlining the alert system would produce potential benefits, they found. “Our study suggests that drug alerts have the potential to prevent harm and reduce healthcare costs,” the researchers wrote. “To do so, however, clinicians need relief from alerts with little clinical value.”