• Q&A With Steven Waldren, M.D., M.S.

    2019 Preview Includes Artificial Intelligence, Machine Learning

    January 02, 2019 10:18 am Sheri Porter – Last October, the AAFP's top techie earned a new title and a whole bundle of new responsibilities. Family physician Steven Waldren, M.D., M.S., long recognized as an expert in his field, was named an AAFP vice president and the Academy's chief medical informatics officer.

    waldren qa

    As always, many projects and initiatives are percolating around the Academy, and Waldren's area is no exception.

    He sat down recently with AAFP News to pin some specifics to the work he'll be doing on behalf of the AAFP and its members in 2019. What follows is the result of that Q&A.

    Q. What are a couple of key phrases that members will hear a lot in 2019?

    A. Artificial intelligence (AI) and machine learning. When I speak to physicians at meetings and conferences, I always ask if they have ever used either of those, and very few people raise their hands. But when I ask if they've ever searched Google or used Gmail, every hand goes up. Most are surprised when I tell them they've all used artificial intelligence and machine learning.

    Q. How can these technologies be used to enhance family medicine?

    A. We need technology to decrease the amount of administrative burden physicians deal with on a daily basis. For instance, right now, physicians create more than 90 percent of the patient record. When they create a note, they start from scratch; even if there's a template and boxes to check, the physician is doing that work.

    What if instead we had a self-documenting record where the computer, the staff and the patient generated 90 percent of the record, and the physician was able to take his or her skill to do the other10 percent? And what if there were no boxes to be checked? Instead, the work was actually recorded each time the physician said something or took an action?

    Q. Can you give a specific example?

    A. There's been a lot of growth in the past year in conversational AI. Today, I can say, "Hey, Google, is it going to rain tomorrow?" and Google will respond to me and say, "Yes, rain is predicted tomorrow in Leawood." But how can we take that technology and place it in the workflow and exam room of the physician?

    Well, wouldn't it be great if the physician could simply ask the computer -- "Hey, computer" -- to pull up the last mammogram and the last abnormal mammogram? There already are a couple of start-up companies trying to do that in the health care space. We're currently talking with some of them in terms of focusing their work on decreasing the administrative burden.

    Q. What's the AAFP doing to speed up this work and get products into the hands of family physicians?

    A. The AAFP is looking to create an innovation laboratory where we take these emerging tools and offer them to practices that are interested in testing and deploying them.

    The outside innovator would implement a tool inside a real family medicine practice, and the AAFP would monitor progress and then conduct a post assessment. Testers will answer followup questions about a tool's effectiveness, its adoptability, and what changes need to be made. The AAFP would then give that feedback to the developer to make product adjustments. The end goal would be accelerating the time to market by providing those best practices to the industry.

    Q. What can the Academy do to get innovators in the market excited about developing new products and services that really support family medicine?

    A. That's another big area the AAFP is addressing in 2019. There are some gaps that the market is not really interested in filling, so the AAFP is creating a challenge program -- akin to an XPrize program(www.xprize.org) -- to explain a gap we want to be filled. We'll then push companies to compete with one another.

    The AAFP Board of Directors has authorized funding of these programs during the next couple of years and will give awards to the companies, products and solutions that are the best in that space. We'd love to have five or six companies competing to deliver the best self-documenting record to tens of thousands of family physicians.

    Q. Can you give an example of a specific product that's already developed and close to being tested in a medical clinic?

    A. We're exploring a partnership with a company that is affiliated with Apple and working on a product related to the Apple Watch. The scenario goes something like this: When the physician walks into the exam room, the watch starts to record the audio. As the physician talks to the patient about what's going on -- the diagnosis, the care -- the watch takes that recording, transcribes it, and uses artificial intelligence to create the initial note. Another company we are talking with provides a digital assistant that today works like Alexa, where you can tell it commands and it will create portions of the note. For example, "Pull my last exam for this patient into the note; change lung exam to normal."

    So, the intent is that the physician never touches the keyboard or the mouse, never has to worry about the computer, and when he's done, the note is on the screen. The physician simply verifies that it is correct, tweaks it as necessary and it's done.

    Those are actual products, and in a couple of weeks, I'll be going to San Francisco to talk with the companies' CEOs about this whole idea of an innovation lab and what the AAFP is trying to do.

    Q. What will the AAFP be doing in 2019 to push the needle on interoperability?

    A. We're joining a new effort, which is a combination of the Healthcare Services Platform Consortium(www.hspconsortium.org) and the Clinical Information Interoperability Council.(www.hl7.org) The goal is to establish standard models of clinical data so that services and apps could be developed that wouldn't have to be recreated for each individual EHR (electronic health record) or health care organization. We have joined this effort because they want this effort to be physician-led.

    We combined work with those two organizations so that we have a physician-led effort to define the clinical data we need to take care of patients -- data that would be nationally recognized and could be deployed across all different EHRs and other IT products so they can start to interoperate.

    It's not about pushing around data and 200-page documents; this is about pushing out clinical information that can then be processed by the receiving system and integrated into that EHR without special effort.

    Q. The AAFP has a lot going on with outside entities. Why are all these innovative companies willing to work with the Academy?

    A. Family physicians have been aggressive in putting new products and services in their practices to deliver better care. They're more open to that than physicians in other specialties, and that has been an asset. And the AAFP has committed a fair amount of resources toward driving technology.

    It's also important to understand that the AAFP can't do this work in a vacuum. The federal government spent $36 billion and was unsuccessful in getting the type of technology that we really need in health care. So as a single organization of a single specialty, it's hard to imagine that we could make a big impact alone, but I think we can with strong, well-aligned partnerships.

    Q. How will you define success at the end of 2019?

    A. A couple of things. First, I want to see the AAFP deploy some innovative solutions in typical family physician practices and see the positive benefits that result in those practices -- for instance, outcomes that really decrease administrative burden.

    The other thing that I'm looking forward to is the result of our engagement with leaders in the AI/machine learning community to articulate the needs of family physicians. I want to make sure that those companies focus on augmenting and expanding what family physicians do in a way that improves the family medicine experience.

    I think the biggest challenge facing our specialty 10 to 15 years down the road will be in keeping AI and machine learning on the right track -- assisting physicians in our work and not replacing or diminishing the patient-physician relationship.