Electronic prescriptions do not reduce errors: study

It seems like a completely logical use of computerized medical records: having doctors fill out prescriptions electronically instead of scribbling their instructions on slips of paper.

A new Toronto-based study, however, found that giving doctors access to electronic prescribing did nothing to cut the rate of potentially catastrophic prescription mistakes, the most desired result of e-prescriptions.

The findings suggest the computerized system added its own, unique errors, and that Canada needs to rethink the hugely expensive push to digitize health information, the lead researcher says.

"I think this is a thing we do because it seems to be the modern thing to do, because medical care is relatively fragmented and it seems to be a way to co-ordinate it, and we live in the computer age, so why shouldn't we try?" said Dr. Merrick Zwarenstein, who headed the study at Toronto's Sunnybrook Health Sciences Centre.

"But is the literature showing the difficulties and the challenges and the relatively modest benefits that each of these things obtains?"

Such systems seem to work well only when implemented painstakingly and gradually, modifying off-the-shelf software to fit the needs and culture of the physicians who use them, said the senior scientist with Sunnybrook and the Institute for Cllinical Evaluative Sciences.

Even with billions of dollars being spent on the field, that typically does not happen now, he said.

A top official with Canada Health Infoway, the federal-provincial body that is trying to spearhead the adoption of electronic health records across the country, said the Toronto research may offer some helpful lessons, but argued the bulk of evidence points to e-prescribing being a success where it has been put in place.

"Off-the-shelf systems have also been able to show benefits quite quickly," said Jennifer Zelmer, a senior vice-president with the agency. "The research is consistent — e-prescribing definitely can improve patient safety."

There is no question that the traditional, handwritten process of ordering medication for patients can cause serious, even fatal problems. Previous research has suggested that adverse drug events — where medicines inadvertently make patients sick — occur in up to 3% of outpatient visits, with more than half blamed on prescription errors. A 2004 Canadian study estimated adverse events caused as many as 22,000 preventable deaths in hospitals annually, with 24% of the events triggered by drug mix-ups.

At Sunnybrook, the researchers convinced 33 doctors to participate in the study of a commercial program where prescribers use a handheld device to pick the patient and drug names from drop-down menus. The machine either zaps the prescription electronically to pharmacies that are part of the system, or prints out paper versions for patients to take with them.

The trial lasted 66 weeks, with the electronic system available for 44 weeks, and traditional paper and pen for the remainder.

Even when the e-prescribing was available uptake was relatively modest — about 8% of prescriptions were filled out electronically. Regardless, when the two periods were compared, the rate of prescription errors was the same for both — about 6% In fact, there were actually significantly more cases where pharmacists had to call the doctor's office to clarify prescriptions during the electronic period.

It is difficult to know exactly why e-prescribing did not help, but it is likely doctors made missteps like tapping on the wrong patient or drug name in the device menus, said Dr. Zwarenstein.

"It's designed to fix certain old sources of errors, but it creates new ones," he said. "Those are not errors that would have occurred in the same way on hand prescriptions."

As for the reluctance of doctors to use the system — even when provided with ample training and support — he said many likely doubted there was much of a problem with paper prescriptions, and saw the digital process as taking up more time in already busy schedules.

Dr. Zwarenstein said most other studies of e-prescribing to outpatients were not rigorous, controlled trials, and likely biased toward a positive result. The only solid evidence that computerized prescriptions can reduce errors has come from two U.S. hospitals where the systems were implemented over periods of as long as 40 years, with changes to the software implemented along the way, he said.

He said the 6% error rate his study identified means electronic prescribing must be made somehow to work, and the impetus for ushering in other types of electronic health records is, if anything, even greater. But the physician argued the concept has to be introduced only after careful consultation with the people who will use it.

"We have to do this, but how to do it right is no small task."

http://www.nationalpost.com/news/Electronic+prescriptions+reduce+errors+study/4524330/story.html