From the director ... doctors ignore computerized drug safety alerts.

When it comes to improving patient safety it sometimes seems as if doctors can be part of the problem than part of the solution. Just how much part of the problem was revealed in a study recently published in Archives of Internal Medicine. Researchers found that doctors who used a computerized prescribing system in their office practices paid scant attention to the built-in safety alerts warning them of potential harm to a patient.

Computer alerts of a possible serious interaction with another drug were ignored on average 90% of the time. Warnings of possible allergic reactions fared a bit better--they were disregarded only 77% of the time. The researchers reviewed over 3 million prescriptions "written" electronically by nearly 2,900 physicians across three states using a common electronic prescribing program.

So what did the researchers conclude? The authors lay most of the blame for the wholesale inattention to the warnings on inadequate design. They suggest that physicians may suffer "alert fatigue" from systems that produce too many "high severity" warnings and that customization by clinical specialty might help increase compliance. Unfortunately the study was not designed to answer the most important question; whether or not patient harm actually occurred as a result of physicians ignoring warnings.

Next November will mark the tenth anniversary of the Institute of Medicine's landmark report on medical errors, To Err is Human. In that report, the Committee on the Quality of Health Care in America, of which I was a member, recommended that the health care system quickly move from a paper-based drug prescribing system to one that was computerized. We believed that electronic prescribing, especially when it included decision support, such as drug interaction alerts, would significantly reduce medication errors.

Since our initial IOM recommendation, the move to computerized prescribing (known as computer physician order entry or CPOE in hospitals) has had its ups and downs. The problems have generally centered on physician resistance, usually based on the claim that such systems are too disruptive to work flow, can cause new kinds of errors and are costly to implement.

I would argue that the weight of available evidence supports adoption. According to a January 2009 Agency for Healthcare Research and Quality report, computerized physician order entry systems incorporating clinical decision supports (like drug interaction alerts) do improve patient safety and help reduce costs. Across the country some hospitals have seen medication errors reduced by as much as two thirds.

I have no doubt that today's systems can be made better and more user friendly, However, to my mind, the continued resistance of physicians to their adoption, as exemplified by the Archives study, is indefensible. Such physician intransigence costs thousands of patient's lives each year due to medication mistakes that we know how to prevent.

Others seem to agree; Medicare, for one, will require electronic prescribing by all physicians within the next several years. Medicare will not only pay more to physicians who shift to electronic prescribing but also, at some point down the road, pay less to those who do not. The continued resistance to a significant potential improvement in medication safety is, I would suggest, one among many examples of physician insistence on "the perfect" as being the enemy of the "good."

Arthur A. Levin, MPH, Center for Medical Consumers [C] 2009

Print Get the Mag Weekly Updates