You Wanna Help? Stop Helping!

I recently attended the second annual KLAS Arch Collaborative meeting in Salt Lake City. If you’re not aware of the Arch Collaborative, it’s a group of hospitals, healthcare systems, and clinics which survey their clinical users about the technology they use to do their jobs. The Collaborative members then benchmark against peer institutions to see how they’re doing. Some crazy Arch Collaborative findings:

• Even if you’re running the same electronic health record (EHR), your physicians may be very satisfied or very dissatisfied. Translation: It’s not just the software, it’s how you configure, deploy, support, and govern the EHR.
• Physicians builders (doctors with a smidge of extra under-the-hood EHR training who can make minor changes in the system) drastically improve the experience for their colleagues – especially those in the same specialty.
• EHR personalization for physicians makes all the difference in the world. Minor changes in how the doctor sets defaults, reads reports, and orders tests and medications can cause major changes in happiness and efficiency.

(Want to read more about the KLAS Arch Collaborative? Check out this article just published in Applied Clinical Informatics.)

The Salt Lake City meeting was organized to encourage KLAS Arch Collaborative members to learn from one another via formal presentations, panels, and small group discussions. There were physicians, nurses, administrators, and C-suite members in attendance, which made for interesting and joyously unpredictable conversations. One session I attended was on using information technology to help fight the opioid epidemic.

The United States is in the midst of dealing with opioid addiction like we’ve never seen before. We’re beginning to understand how we got here, and we are well on our way to making it easier for physicians to do the right things when dealing with pain (see this and this). That’s what we started discussing: how to configure the EHR so docs can easily do the right thing. Such tactics as:

• Changing ordering defaults to automatically give a minimum supply of pain meds
• Configuring clinical decision support to alert physicians to a patient with a history of opioid addiction or a dose of meds that may be too high
• Offering a prescription for naloxone, a potentially life-saving antidote to opioid overdose

The group of informaticians and clinical operational leaders started to discuss how states across the nation are responding to the opioid crisis. We quickly came to the conclusion that, as the saying goes: the road to hell is paved with good intentions. Case in point: most states have set up prescription drug monitoring programs (PDMPs) to act as central clearinghouses for potentially-dangerous medicines. Whenever certain prescriptions (e.g. opioids) are filled by pharmacies, those pharmacies send pertinent data to the PDMP, chiefly patient demographic information and prescription details. PDMPs help decrease the practice of doctor shopping (moving from doctor to doctor to get far more opioids than is safe.) The states then regulate the requirement that physicians query the PDMP before prescribing certain doses or courses of opioid medications. If the doctor is about to prescribe a pain med, but sees the patient got a similar pain med just a few days ago from a different doctor on the other side of town or the other side of the state, the physician can reconsider the pros and cons of the order.

Sounds great, right? End of story. Stop reading. Nothing to see here, folks. Move along.

While the regulators and politicians really and truly want to do the right thing, they sometimes don’t understand what the right thing is. It’s great to legislate that physicians query the PDMP to ensure they don’t over-prescribe dangerous opioids. One state did just that. A hospital in that state interfaced to the state’s PDMP database so the patient’s opioid history seemed to magically appear in the EHR just at the right time: when the doctor was prescribing. Boom. A huge win. But . . . state regulators later decided that they needed to know if/when doctors were checking the database, so they effectively nixed the interface and instructed doctors to manually login to the PDMP website itself. Naturally, this is a workflow killer, as the doc needs to leave the comforts of the EHR → open a Web browser → navigate to the PDMP site → enter a username and password different than the one used for the EHR → enter identifying information to find the patient → and then reap the rewards of learning if the patient is doctor shopping (newsflash: most aren’t!)

Dear reader, you’ll never guess what happened. Ok, maybe you already have guessed. Compliance with querying the database plummeted after the interface was shut down. While physicians wanted to do the right thing, it became onerous to do so. Hence, maybe a doctor just checked if the patient seemed like they were potential abusers. That’s a bad idea, of course, because one can never know who is a potential abuser!

This PDMP interface debacle is just one of the regulations-gone-bad topics we discussed at the Arch Collaborative session. While we all pledged to become more involved with the political process so that we can effectively lobby our regulators and representatives, this is not something most of us ever wanted to do. It would be great if lawmakers better understood the implications of their actions. Is it appropriate for politicians to mandate PDMP reporting and querying? I think it is. But exactly how that happens may not be best left to folks who aren’t doing the work day in and day out.

As a senior surgery resident said to me on my first day as a third-year medical student: “You wanna help me? Don’t help me.”

Do you have stories of well-intentioned regulation gone bad? Tweet at me at @CraigJoseph.