Drug interactions and the problem with default settings

At some hospitals, pantoprazole is not on the formulary. So, when a patient takes clopidogrel and pantoprazole, the drugs are automatically switched to clopidogrel and omeprazole. For a proton pump inhibitor (PPI), the default is omeprazole.

The Medical Letter, the FDA and UK’s National Guidelines for Stroke recommend if patients take clopidogrel and need a PPI, avoid omeprazole and use pantoprazole instead. The concern is that omeprazole prevents activation of clopidogrel (clopidogrel is a pro-drug), and so, the drug might not work.

At some hospitals, the drug-drug interaction reporting can be adjusted by various people and committees. Reporting of the clopidogrel-omeprazole drug interaction can be disabled.

If it is true that clopidogrel in these patients is in an inactive form, then the patient should be at increased risk of having vascular events. These patients should also have more hemorrhagic complications. This might occur because the anti-platelet and anti-coagulation regimen is adjusted while the patient takes omeprazole. Later, the omeprazole might be stopped or switched to pantoprazole, and the hemorrhagic risk increases. This could all happen without the physician being particularly aware.

This is not just about clopidogrel and the PPIs, the point is that decisions have been taken out of the physician’s hands. Defaults are being set according to a general recipe, and reporting of drug interaction is altered at the whim of who knows who.

Even though hospitals talk about the importance of process and process improvement, it is hard to change process. Physicians are left having to correct errors of the process, without being exactly aware these errors even exist. It is difficult to recognize things when they are unexpected.

Decisions ought to be in the hands of the doctor (decentralized), but we all know that drug companies, hospitals, fellow doctors, insurance companies and various governmental agencies want to control our decisions. This can happen in overt or covert ways. There is even a book, Nudge, that explains how to control other people’s decisions covertly, using defaults and other behavioral tools.

This is OK provided people are nudged to make good decisions. It would be fine if it worked to make sure patients taking clopidogrel got pantoprazole and not omeprazole, but it doesn’t. These are tools and can be used by anyone, for any purpose, even for no obvious purpose at all.

One point is that it is important to recognize that defaults in the hospital system can work to make your patient sicker. This is not just clopidogrel-omeprazole, it’s patients with Parkinson’s who automatically get Phenergan and Haldol as prns, and it’s patients on topiramate admitted with kidney stones who are continued on the drug. It’s a problem with defaults.

Another point is that other people can monkey with drug-drug interactions reporting, so you need to have a simple system you control and trust.

Finally, the last point is a warning that other entities will read Nudge and thereby learn to use behavioral techniques in order to influence your decisions, probably to the detriment of your patients.

Bradley Evans is a neurologist. 

email

Comments are moderated before they are published. Please read the comment policy.

  • http://www.facebook.com/shirie.leng Shirie Leng

    Bradley, I totally get you on this. As an anesthesiologist, I write post-op pain orders for my patients in the PACU. If someone tells me they get nauseous with percocet, which pretty much everyone does, that is entered as an “allergy” in the computer. The result of this is that ALL opioids as well as tylenol are red flagged and you have to manually override this warning in order to order any pain medications. Put aside the fact that nausea is not an allergy. I am put in the position of ordering something the computer says the patient is allergic to, a legal hazard if ever there was one.

Most Popular