In our passion for all things technology driven, we have computerized, digitized and onlined everything we can think of. Sometimes the results have been brilliant, absolutely brilliant. Sometimes the results have been less than impressive, but often, a la Goldilocks, the results have been somewhere in between.
Consider the e-prescribing software now available for doctors. Computer-generated prescriptions are trending, trending, trending — an apparently green-friendly approach that’s supposed to eliminate errors, save time and money, promote world peace, you name it, this does it. Really? Let’s have a look.
Older adults take lots of prescription drugs and the elderly are particularly at risk for adverse drug events (ADEs) because of this. And not only are there ADEs to worry about, the elderly are often prescribed potentially inappropriate medications (PIMs) that can result in dangerous falls and other problems.
To get around these kinds of hazards, primary care physicians have been working with e-prescribing software that contain embedded treatment algorithms and triggers that pop-up and warn the doctors (without having to push anymore buttons) about potential PIMs. This makes it much easier for the physician to change their medication recommendations and decisions at the point of prescribing, in theory making the whole process more efficient and safe.
Doctors who tested this approach liked it, to a point. Many would consider using such software in their daily practices, but with some modifications: they disliked the repetitive alerts or receiving triggers on content they already knew about; they wanted the data to be scrupulously up-to-date and accurate; they wanted any alerts to be brief, very focused, and able to be absorbed in 30 seconds or less. The work continues.
On the patient side of computer-generated prescriptions, electronic prescribing systems are again promoted as the way to reduce errors and adverse events, problems pretty common in the ambulatory care setting. A new study finds, however, that 1 in 10 computer-generated prescriptions contains at least one mistake.
Researchers looked at 3,850 computer-generated prescriptions received over a 4-week period by a commercial pharmacy chain with stores in Arizona, Florida and Massachusetts. Independent reviewers audited the prescriptions for errors. When an error was found, it was classified by type, and any mistake that had the potential for harm was tagged as a potential ADE.
Here’s what they discovered. A total of 452 prescriptions – nearly 12 percent of the total – had one or more of 466 errors. Of these errors, 163 – that’s 35% – were classified as potential ADEs. They calculated the rate of prescriptions containing ADEs as 4.2%. Finally, more than half of these potential ADEs were significant, though none in this study was deemed life-threatening. To break it down further,
- 40.3% of the errors were found for anti-infective prescriptions, followed by nervous-system drugs and respiratory-system drugs.
- Nervous-system drugs had the highest rate for potential ADEs at 27%, followed by cardiovascular drugs and anti-infectives.
- Omitted information was the most common cause for errors (60.7%) while ADEs came in at 50.9%.
- Error rates varied considerably among the various computerized systems used, from 5.1 percent to 37.5 percent.
The researchers concluded that implementing any computerized prescribing system without rigorous testing, comprehensive functionality and appropriate processes in place to ensure its meaningful (correct) use does not in any way reduce medication errors. In short, technology and nifty software alone will not solve the problem, despite the marketing hope and hype.
Both computer-based and provider-based interventions are necessary to reduce the mistakes and errors associated with computer-generated prescriptions. This means these systems need to be chosen with educated care and come with extensive, on-going training (beyond what a vendor might offer), timely updates, calculators and specific drug-decision support.
Clearly, then, the work continues.