Skip to content Skip to navigation

The Enemy Is Us

July 1, 2006
by James Feldbaum
| Reprints
Institutions should look inward if frustrated by a lack of CPOE uptake in their facilities.

The coup de grâce for resistance to computerized provider order entry, CPOE, came when the elected board of our 800-bed community hospital announced the date mandating its implementation. It was the natural next step in our migration away from paper and into the electronic medical record and was an integral part of a larger patient safety initiative.

By the end of the two-year staged rollout, more than 400 providers were entering their own orders electronically.

This accomplishment makes us conspicuous on a national electronic battlefield littered with casualties. Our battles have been hard fought and our successes, hard won. We surmounted obstacles of software, hardware, and hospital budget cutting. We won the confidence and cooperation of our physician workforce and overcame considerable resistance from some quarters. Yet, despite our victories, today we still face the same enemy to patient safety that has plagued us from the start — us.

Admittedly, there were threats from external forces. There were software and hardware glitches that tested the endurance and talent of our system administrators. There were software programs that failed to deliver and hardware systems that were too rapidly outgrown. There were upgrades and there was downtime. Now, however, armed with powerful and capable software, we have yet to achieve the quality and safety improvements that are at our fingertips. The enemy was, and still is, us.

What's the problem?

Let's forget the past and the various symptoms of rebellion against CPOE: physician threats to sue for restraint of trade or to move their patients to a competing hospital; orders recklessly placed on scraps of paper and inserted into charts unseen (and often missed); and the sudden glut of verbal, telephone, and fax orders. We survived those skirmishes and today CPOE reigns in our institution. So why are our patients not as safe as they deserve to be?

First, we must ask if a computerized order is inherently better than a handwritten order. There is little dispute about the illegibility of physician handwriting. In one 24-hour period in our institution, over half of all handwritten orders had legibility problems that required clarification. Illegibility is a recognized source of medication error and delay. A well-crafted electronic order is complete, unambiguous, and legible and can be processed in milliseconds.

However, an incorrect order will, regrettably, move with the same rapidity as a correct order. A rushed or distracted order session, a confusing physician/computer interface, poor physician training, carelessness, or a slip of the mouse can result in an electronic order incorrectly identifying drug, dosage, or patient. The U.S. Pharmacopoeia's 5th annual study of medical errors reports that errors in electronic order submission have actually exceeded errors made by handwriting.

Provider distraction accounts for more than 50 percent of such errors. Human error can, will, and does migrate into electronic order sessions. The computer, like a scalpel, needs to be utilized with skill, clarity of attention, and precision. In America's healthcare system, serious medication errors occur in up to 10 percent of patients admitted to a hospital, many with tragic consequences. Until physicians recognize that cutting-edge computer technology can cut for ill as well as for good, our patients will be no safer.

Secondly, we must ask if we are fully utilizing the computer's capacity to reduce errors and improve the quality of medicine. A critical component of CPOE is computerized decision-making support (CDS) available in real-time. Most fundamentally, CDS can perform drug allergy checks, drug-laboratory checks and drug-drug interactions. More sophisticated CDS can be employed to suggest treatments on the basis of laboratory findings or specific diagnosis.

Avoid the over-alerting

The computer can effectively page or send email to a provider when requested clinical parameters are met or critical alerts are generated. Although our ability to craft alerts is limitless, our tolerance for them is not. Physicians will resent intrusions into their order sessions by clinically irrelevant, overly abundant alerts, and alert fatigue will confound compliance with critical computer-generated warnings and suggestions.

A culture must evolve where alerts perceived as non-critical are tolerated and meaningful alerts are heeded if the power of the computer is to be translated into improved patient safety. The quest for perfection should not devalue the present exasperatingly less-than-perfect state of safety warnings.

Furthermore, we must ask if we conform to a culture of best practices. Do we apply the very best medical evidence to our decision-making processes at the time of ordering? Medical information doubles almost every five years, and despite a provider's best efforts, our evidence-based knowledge decays at a similar rate. The days of "that is the way I have always done it" are over.

A well-designed package of CDS modules can warn of errors of omission as well as commission. Real-time reminders about corollary orders (reminders to order blood sugar tests if diabetes medications have been ordered, for example) can be employed. Entire sets of diagnosis and treatment-specific order sets can be created or purchased that contain state-of-the-art protocols that can be periodically updated to keep pace with advances in knowledge.

It can be done