I stumbled across a story on Amazon’s “secret” team that is supposedly looking at healthcare, including electronic health records and virtual visits. It’s supposedly called 1492 (if they chose that as an homage to Christopher Columbus, they had better rethink some of the cultural baggage around his “discovery” of North America). It sounds like they’re exploring interoperability as well, along with figuring out whether they can use the Amazon home-based devices like Echo in a healthcare capacity.
I’ve been a big fan of Atul Gawande ever since “The Checklist Manifesto” and enjoyed reading a transcript of a recent interview with Tyler Cowen. His opening comments on artificial intelligence were realistic and balanced, which was refreshing given the hype we’re used to seeing with headlines like “Dr. Watson Will See You Now.” He concisely explains how challenging it can be to fully understand what the patient is telling you.
Those of us in the trenches know this, but folks on the technology side underestimate the power of the story vs. data points. Patients often point to problem areas or sources of pain and have trouble explaining whether the problem is more external or internal. Some can’t offer descriptive words at all. Then there is the issue of individual perception of pain or problems. Of course, algorithms could probe into that, but there could be hundreds of questions needed to include or exclude various decision points.
He disagrees with the IBM Watson decision to address this problem and notes that the issue is complicated by the fact that the patient data changes over time. Not only discrete data, but the patient’s perceptions change, as does the patient’s willingness to bring new symptoms to the clinician’s attention and also the understanding of the interviewer. He sees technology as more of an adjunct.
I think most of us caring for patients agree. I’m tremendously fond of clinical decision support and systems that help me ensure I’m not missing anything I should be thinking about with complex patients. I think automated checklists are fantastic, and rather than making me practice “cookbook” medicine, they are helping me deliver the same quality care to every patient every time, regardless of how rushed or distracted I might feel at any given moment. They help level the care we deliver when we are trying to see patients in six-minute increments rather than the 30 minutes many of us wish we had.
He specifically mentions Isabel, which I’ve had available in a couple of EHRs that I’ve used in the hospital setting. Isabel prompts you to think about diagnoses you may be missing in rank order based on the data.
Cowen asks his thoughts on the potential of gene editing with CRISPR, which he finds concerning due to the “unpredictable things that people will discover that you can try to do with gene editing.” When those edited genes are propagated in living organisms, they can spread rapidly, and he doesn’t “think we’ve thought through that in the least.” There’s also the risk that people will want to genetically select against characteristics that they feel are undesirable without fully understanding the implications. On the other hand, he notes that many conditions are the result of the interaction of multiple genes and aren’t something that CRISPR will be able to significantly modify.
Gawande also goes on to talk about safety in the operating room and how the rise of procedures where the patient is awake is changing culture. That patient can now be part of the team and not just a passive participant. These procedures have been common in neurosurgery, where brain mapping is needed to try to protect the speech and movement centers while working on other areas. He notes that he’s seeing them in non-brain surgeries, where the team can interact with the patient about their medical issues and goals for the surgery.
Other patients don’t handle awake surgeries very well, so he does note that sometimes you have to adjust on the fly. I know this firsthand since I once had a procedure under “light sedation” and the surgeon asked the anesthesiologist to put me out a bit more because apparently I would not shut up and was getting sassy with the scrub nurse, who I recognized as having hazed me during medical school.
He notes that while checklists have been effective in reducing errors, there are still barriers to success because people either check the boxes by rote or end up not using the checklists at all. The first problem is something that I’ve seen in many organizations I’ve worked in. It can be as simple as running out of a supply and discovering that someone initialed an inventory form just hours before that the exam room was fully stocked.
As a busy urgent care, that’s a major concern in our practice, but fortunately we don’t have a lot of problems with people falsifying their inventory checks. One of our execs is a former Naval officer and “gundecking,” where someone says they did something that they really didn’t do, is a cause for termination. Leadership makes it clear that when you falsify logs, you undermine our mission of care delivery and it is not tolerated.
The idea of people blindly marking a surgical checklist is frightening. He mentions that organizations can take checklists to extreme, taking one 19-item checklist to an 81-item level that was unusable. Administrators rather than clinicians had bloated the content, which essentially led to people ignoring it.
They go on to explore the disconnect between healthcare and health outcomes. He notes that data from coverage expansions like the Massachusetts healthcare reforms has shown that some interventions are more powerful than others – namely primary care, chronic illness care, and mental health care. He also notes the difference between death reduction and changes in quality of life. Still, we’re not getting the biggest killers under control, like high blood pressure. Organizations like Kaiser have been able to improve outcomes through more assertive management of barriers to care.
I see issues with coordination of care and comprehensiveness of care daily, as patients come to the urgent care for situations that would be better handled by a primary care physician. Some days I struggle with the fact that I’m part of the problem – perhaps if we weren’t as accessible, or convenient, or fast, patients would put more pressure on their primary care physicians to re-engineer how they’re delivering care. I still see plenty of physicians who don’t leverage the technology they have in front of them or who refuse to change their office policies and procedures to better support their patients. I have experienced botched prescription refills, botched appointments, and general chaos when trying to get care myself.
The interview also covered the state of medical education, the FDA, and his thoughts on indie music, but I’ll have to leave you hanging for my summary of those topics. Tune in to next week’s Curbside Consult for the rest of my recap.
Email Dr. Jayne.