CPOE fails Leapfrog's test

Leapfrog has released a study based on a simulation tool (fake patients, simulated inappropriate order entry) showing that, across many hospitals and many EMR brands, CPOE does a poor job of intercepting errors. For anyone who uses CPOE in the real world that's a no brainer---you don't need a study.

Margalit Gur-Arie, guest blogging at Kevin MD, explains the report and notes that the more complex the situation the more poorly CPOE performs:

For basic adverse events, such as drug-to-drug or drug-to-allergy, an average of 61% of events across all systems generated appropriate warnings. For more complex events, such as drug-to-diagnosis or dosing, appropriate alerts were generated less that 25% of the time.

And for the far more common and complex judgment errors, (eg dangerous drugs such as heparin given for weak and sloppy indications, antibiotic selections in sepsis that are too narrow based on patient attributes, local resistance patterns and clinical circumstances) CPOE doesn't do squat.

On the other side of the spectrum are the simple alert triggers which are often inappropriate, such as flagging the combination of antiplatelet agents and low molecular weight heparin for acute coronary syndrome as “duplicate therapy”, leading to alert fatigue.

We have a long way to go before CPOE is a patient safety tool.