Tuesday, May 18, 2010

Paper Kills

BIDMC implemented Computerized Provider Order Entry in 2001. As we approach the ten year anniversary of our implementation, it's great to see the recent press on the Stanford study demonstrating a 20% decrease in mortality after implementing CPOE at Lucile Packard Children’s Hospital.

I've written about our experience and my top 10 approaches to make CPOE successful.

Our CEO has blogged about it.

Medscape has a great summary of the effort.

The bottom line - of course bad software implemented poorly can cause new errors. Of course change in workflow can cause unintended consequences.

However, now that the industry has broad experience with electronic ordering (it's a meaningful use requirement for 2011), I think we can conclude that questioning the wisdom of implementing CPOE is like asking if a parachute works against gravity - I do not want to be in the control group of that clinical trial!

We need to be careful to design clinician friendly user interfaces, embrace web-based systems that require little training, and incorporate enough decision support to keep patients safe but careful to avoid crying wolf too often, creating alert fatigue.

As I wrote in a recent blog, paper-based medication ordering killed my grandmother. Unreadable orders, drug-drug interactions, and prescribing errors in the elderly cause harm.

The Stanford study now gives us the objective evidence we've been waiting for. We can use CPOE with confidence and finish the implementation in those community hospitals which do not yet have it.

CPOE is the medication version of a parachute. I would not want to write medications without CPOE any more than I would want to jump from a plane without a parachute.


Michelle W said...

Great non-alarmist post on CPOE. In this debate on EHRs, we shouldn't loose sight of the fact the paper this tool replaces isn't the most reliable either; it is merely the form some people feel most comfortable (and thus, safer) with.

Upgrading to a newer tool should all be about how well such a transition can affect patient care, and thus the clinician's practice.

Nuclear Fire said...

CPOE is a generic term. It is about as useful to say "drugs are good. We should use drugs to treat diseases." True, but we should use drugs proven to work. Some drugs have been; others have been shown to cause harm. "right drug, right dose, right route."

The same is true for medical software. We need to speak in specifics about specific software. In the referenced study, the specific software was not reported, only that it was commercially available from Cerner. On Cerner's website, I find 8 different products listed under "Cerner CPOE solutions." Are they all great? Are most of them great, but one is dangerous? Are Cerner's product just as good as anothers or does Cerner's decrease mortality while Epic Systems increases it?

If we continue to talk in generics, companies will continue to market "me too" products based on the "fact" that CPOE is good for XYZ reason. When people use these systems and have negative outcomes, the whole system will get blamed. Worse, patients will be hurt on untested systems.

Selfishly, I hope we move to specifics soon, so I will have more data with which to decide which systems to buy.

Disclaimer: I have no financial or professional interest in any company involved in making healthcare IT products.