I did two things last week which at first seemed unrelated. First, I finished reading the book The Checklist Manifesto by Atul Gawande. Second, I attended a roundtable discussion on the topic of utilization management, with a focus on the history and use of pre-authorization.

In The Checklist Manifesto, Gawande describes his quest to increase safety in medicine, specifically surgery. He uses examples from aviation and construction to come to the conclusion that checklists are the key. Even experienced, skilled practitioners can benefit from the use of written guides.

Pre-authorization is the insurance practice of requiring approval for expensive medical treatments and testing before they are done, with cost savings as the main goal. Presumably, authorizers use appropriateness algorithms to determine if a test will be approved or not.

Like I said, these might seem unrelated. One is focused on safety and the other appears to be about money. But these are actually examples of two solutions aiming at the same problem: getting doctors to do the right thing.

Doing the right thing is really difficult in healthcare, for two basic reasons*:

• It’s hard to know what the right thing is. Despite all the guidelines that exist, much of healthcare still falls under “miscellaneous” without a clear algorithm to follow.
• Even when we know what the right thing is, medical knowledge has grown beyond a single person’s ability to ingest and use in practice. Recalling and correctly applying the information is a challenge.

There is a lot to dissect with just these two issues, including things like the state of medical research, habit, tradition, and culture. For now, let’s focus on the second challenge: actually doing the right thing once it is already accepted medical knowledge. How can doctors better access and use medical knowledge in their practice? More education and memorization is not the answer, so let’s focus specifically on the technology gap – how can technology be used to help doctors better access and use medical knowledge? Once we have the knowledge or the best practice, how do we put it into the workflow?

If the goal is to deliver knowledge in a usable way, we have to consider two features: the timing and channel of the delivery of information. The knowledge must be delivered at the appropriate time and in a manner in which the recipient can actually use it.

For example, think about your car GPS. It wouldn’t be useful if it gave directions the way people sometimes do: “Whoops – you should have made a left back there.” It has to deliver the instructions at just the right time – long enough ahead for you to execute the instruction but not so long that it confuses you. Similarly, it has to be done in the right way, using the correct channel. If it sends you a phone text with each direction, you’d have to take your eyes off the road, so it gives you audio instructions or perhaps projects the directions on a heads-up display. If knowledge is not delivered at the time of need or in a manner that is effective, then it has little value.

Pre-authorization suffers from – among other things – a problem with timing. The word means that it comes before something, but that’s from the perspective of the insurance company. From the perspective of the doctor, it comes after a key step: after the decision has been made to order a test or prescription. That’s too late. That late timing and the potential need for the doctor to circle back and change the plan can strain the doctor-patient relationship, cause distrust on the part of the patient, and create inefficiency in the system.

Checklists can be very effective, as Gawande points out, but paper checklists have a limited channel of delivery. You have to pick it up and refer to it in the middle of the workflow. This can work very well as a pre-operative checklist, but it might not work as well in a primary care doctor’s office.

Much of healthcare is performed as a meeting involving a patient, a doctor, and the space between them. The essential doctor-patient interaction is a face-to-face conversation, a basic human social activity. It’s hard to insert technology into that without making it awkward.

It’s tough to get both the timing and the channel right. Most of the solutions we have prioritize one over the other, with either sub-optimal results or ignored solutions. We have great clinical decision tools inserted at the point of order entry after the doctor has already made the decision, so doctors find ways to game them in order to support their existing plan of action. We have algorithms in books, on cards, and embedded in apps, and yet doctors still trust their memory over these tools, partly because they are difficult to use in the moment. When doctors do focus on their computers during a patient visit, even just for documentation, we have found that it worsens the human interaction.

The challenge may be lessened for doctors such as radiologists or telehealth physicians who work in front of computers, because information can potentially be delivered to them through the screen at just the right time. While there has been progress in this area, these solutions don’t translate easily to the non-screen interactions of their colleagues.

Perhaps new technologies such as virtual or augmented reality might hold the answer. A lot has to change in both the technology and the cultural norms of a patient visit, however, before these are realistic options.

Solving this problem will take time and thought. But the first step in improvement is determining what problem you are trying to solve. In The Checklist Manifesto Atul Gawande tackles the problem of to how make medical care safer and his answer is the checklist. Even if you accept his premise (and I do), we still have a challenge before us. Given the way that doctors and patients interact, how do we manifest the checklist?

*Gawande categorizes these as errors of ignorance in which we don’t know enough and errors of ineptitude in which we don’t use what we know. It’s interesting to note that “we” can refer to either the medical community at large or a single practitioner. We have both types of problem, involving both types of “we.”