JAMA study only shows that no technology cures problems in isolation
CPOE will survive
While everyone in healthcare information technology is flapping over the latest of two studies to describe some pitfalls of electronic prescribing, a collective deep breath and some reflection may be in order.
Yes, a study published in the Journal of the American Medical Association of an early-stage computerized physician order-entry system at a single hospital may be a sort of cautionary tale for those rushing into big-ticket IT (April 2005,
p. 1). But like an earlier study that linked CPOE to medication errors, this one is far from a repudiation of the whole notion of paperless systems. The authors clearly did not intend for it to be taken that way, a point of view that wasn't fully reflected in some national newspaper coverage. Nowhere in this study does it say we should stick with hand-scribbled prescriptions, the source of most medication errors.
Reading the study, I was left a little surprised at the defensive reactions from IT proponents. Anyone who has been reading this magazine or IT periodicals knows of some of the expensive stumbles, including the infamous one at Cedars-Sinai Medical Center in Los Angeles. Poor interfaces with legacy systems and troubles with physician buy-in are common, and Cedars-Sinai is hardly alone in throwing its system away. That the JAMA study found that the software involved doesn't match the way that clinicians practice is equally unsurprising. Researchers found that as many as 20 screen pages were needed to review some seriously ill patients' medications. Medication changes that took place automatically were problematic from the get-go.
As noted in our coverage, however, the 1997-era CPOE model at the hospital in question -- the Hospital of the University of Pennsylvania in Philadelphia -- has already been replaced. Another study -- of how the second and third generations of software fare -- would be more valuable.
That any form of CPOE can cause so many medication errors is at once understandable and regrettable, and therein lies the cautionary part. We have to view any such system as just one component of a comprehensive commitment to quality of care and patient safety. All systems and processes have to be continually adapted, analyzed and updated. A new data system can't simply be uploaded and used without regard to how clinicians should be working, nor should doctors be allowed to try to adapt these systems to their individual ways of doing business. Everyone has to come together to make IT work toward a common goal of better patient care.
We have a national process under way to ensure that all healthcare IT systems work off of common technical standards that enable them to share data easily -- the whole notion of interoperability. The biggest IT vendors have signed on to this strategy. To say that much work remains to be done toward a national electronic medical record is a laughable understatement, but what this work is about is providing a blueprint for providers to follow in adapting next-generation systems, thus reducing the kinds of mistakes found by the JAMA researchers.
Finally, let's not completely overlook the JAMA findings. The deep concern many feel for the state of healthcare quality has many ready to embrace anything that sounds like a panacea. CPOE has been sold as the end of medication errors, and clearly it is not. The task is to proceed toward widespread adoption of interoperable, universal healthcare IT as quickly as possible but with a renewed awareness of the fallibility of machines and their operators and a relentless attention to outcomes.
Send us a letter
Have an opinion about this story? Click here to submit a Letter to the Editor, and we may publish it in print.