Skip to main content Back to Top


Study Finds Medicaid DUR Ineffective

Kate Traynor

A study described in the Sept. 17 Journal of the American Medical Association (JAMA) concluded that retrospective drug-utilization review (DUR) has no effect on clinical outcomes among Medicaid beneficiaries and questioned the widespread use of the reviews.

Sean Hennessy, Pharm.D., Ph.D., lead author of the JAMA report, said that although DUR programs "seem like a good idea, and they are being done by well-intentioned professionals for the right reasons, the best evidence is that [DURs] don't do what they're supposed to do."

Retrospective DUR programs are designed to identify and reduce medication errors by defining circumstances in which specific drugs or drug combinations should not be prescribed. Through DUR, prescription claims data are screened for appropriateness, and reports of exceptions are generated for prescriptions that were dispensed in violation of the criteria. DUR personnel may send so-called alert letters to physicians whose prescriptions are flagged by the system, with the intent of improving prescribing practices.

After analyzing data from a DUR software program used by six states with a combined Medicaid population of nearly 6 million beneficiaries, Hennessy and his colleagues at the University of Pennsylvania School of Medicine found that retrospective DUR did not change the rate of hospitalization among Medicaid beneficiaries. This finding held true even when the analysis was restricted to exceptions that led to the greatest number of alert letters.

The researchers also found that the frequency of drug–drug, drug–disease, and therapeutic duplication exceptions did not change after the reviews were implemented, an indication that DUR did not change prescribing practices for the study population.

The Omnibus Budget Reconciliation Act of 1990 required all state Medicaid programs to implement retrospective review by 1993, but the legislation was enacted without subjecting DUR to scientific study.

Thus, Hennessy said, retrospective review has been "implemented without any evidence that DUR programs work." In the JAMA report, he and his colleagues urge policymakers to "consider withdrawing the legislative mandate" for retrospective DUR.

Earlene E. Lipowski, associate professor at the University of Florida College of Pharmacy, credited the idea behind Hennessy's work.

"I think it's long overdue," she said. "Certainly, we ought to be monitoring the impact of these programs." But, she added, "I'm not quite ready to dismantle all of these programs just based on the results of one study."

"We've put a lot of up-front costs into developing the infrastructure to conduct these programs," Lipowski noted. "I think it would be nice to challenge people to show how they can be effective, and how can they be more effective than they currently are. And in the event that that evidence isn't forthcoming, then we might go back and look at the mandate and whether or not such a mandate ought to be part of any future government program."

Hennessy's study focused on DUR exceptions that accounted for 80 percent of the alerts issued during the study period. Lipowski, who is a member of the Florida Medicaid Drug Use Review Committee, said that this approach may not have given appropriate weight to the exceptions with the greatest clinical effects.

In the case of drug–drug interactions, "it may mean these [drugs] are used together frequently, and the adverse events are infrequent," she said. "Otherwise, we wouldn't find so many cases in which these drugs were used together."

Lipowski said that DUR programs could benefit from research that examines drug interactions that "we actually see leading to patient harm and negative outcomes" instead of interactions that are based on pharmacologic studies.

She praised a Canadian study reported earlier this year in JAMA that examined drug-related toxicity leading to the hospitalization of seniors who had been treated with glyburide, digoxin, or an angiotensin-converting-enzyme (ACE) inhibitor. The study identified clinical harm resulting from the coadministration of one of these drugs with a medication known to cause adverse events, such as cases of hypoglycemia in seniors who took glyburide and cotrimoxazole.

This study also caught Hennessy's attention.

"This is exactly the kind of research that should be shaping the interventions that we make," he said. He speculated that other ways to improve DUR could include decreasing the time lag between the writing of a prescription in violation of a criterion and the receipt of an alert letter by the physician.

"Getting a letter several months after you've written a prescription is probably not a very good way to convey the message," Hennessy said.

He said that computerized prescriber order-entry systems "have a lot of promise" for issuing timely information about appropriate prescribing, but should be evaluated for their potential to introduce errors into the prescribing process.

Possible improvements aside, Hennessy strongly recommended that "any new [DUR] program would need to be shown to be effective before going out and implementing it in large numbers of people."

"You've got to do [DUR] right," Lipowski said. "You have to carefully monitor what you're doing to see if you're getting the results you had anticipated. If you don't get the results you've anticipated, you've got to go back and ask yourself why."