Today, about 10% of all prescriptions filled in the U.S. trigger a concurrent safety alert to be sent to the dispensing pharmacist. Of these alerts, 88% are simply overridden by the pharmacist. In roughly one third of the cases, pharmacists were already aware of the problem; for another third, they did not believe that the problem existed; and in another third, they thought that the problem was insignificant.
This high volume of alerts in the absence of action diffuses the value of the more significant messages, contributing to “signal overload” by dispensing pharmacists. This problem is beginning to be addressed by several of the new and recent initiatives described later (see “Emerging Solutions”). cialis professional 20 mg
Retrospective drug use review (rDUR) systems typically involve written communications to the prescriber after the prescription has been dispensed. Given the lag time between prescribing, dispensing, and receipt of this information, rDUR systems are typically used to address patterns of care or sub-acute concerns that require long term follow-up.
In an unpublished survey by Express Scripts of 471 pre-scribers who received rDUR letters, 60% found them useful, 25% were ambivalent, and only 15% did not find them useful. Thirty-eight percent of surveyed physicians reported changing treatment plans after receiving the rDUR intervention. A six-month pre-rDUR and post-rDUR intervention claims analysis demonstrated up to a 22% decrease in prescribing selected drug categories compared with randomized control groups without rDUR.
Critical Database Selection Is Required for Checking Electronic Drug Interactions
There is little agreement as to which drug-drug interactions (DDIs) are the most clinically important. Abarca reviewed five leading drug compendia and discovered that most DDIs were identified in only one or two of these sources and were absent from the others. Of 2,372 interactions listed in at least one of the five sources, only 65 were listed in all five sources. More recently, the same researcher and his colleagues surveyed the “high-severity” DDIs listed in four leading DDI compendia. There was very poor concordance in the listings between compendia; nearly three-fourths of the DDIs were found in only one or another of the compendia. Only 2% of the DDIs were identified in all four compendia.
Cavuto et al. reported the results of presenting paired prescriptions for terfenadine and ketoconazole, a combination that has been established as presenting an increased risk of tor-sades de pointes. These paired prescriptions were presented simultaneously to 50 pharmacies in the Washington, DC, area and to seven more pharmacies across the U.S. Both prescriptions were filled 35% of the time despite electronic drug-interaction software. Either the software programs did not identify the drug interaction, or the pharmacists chose to dispense the medications even though they had received an alert.
Hazlet et al. presented 16 well-documented DDIs in six fictitious patients to 516 community pharmacies in Washington State, representing nine different DDI software programs. The software was unable to detect 35% of these DDIs. More surprising than the variation among software vendors was the variation within given programs—this can be attributed to inconsistent software installation and operations at different locations.