AI's role in identifying, treating metabolic conditions

I am a fifth-generation western Kentucky native, and I am told one of my great-grandfathers was the only person enlisted in the Civil War for the Union from Marshall County. I did my pediatrics training at Washington University at the St. Louis Children’s Hospital. I was fascinated during my interview process with far-flung children’s hospitals like Seattle Children’s. In the end, I connected with my program officer Jim P. Keating and decided to stick close to home around Paducah, Kentucky. It made sense to stick close by in a large medical specialty group after residency, but after seven years I decided to do a solo house call practice in St. Louis, Missouri. Concierge medicine was heating up, but I think the industry thought it a bit odd. Back in 2009, I got a phone call from Mayo Clinic asking me to present at the Transform Symposium about “customer experience.” Shortly after that, Brian Dolan from MobiHealthNews called me “the first iPhone doctor” for seemingly having my medical practice rigged up to my iPhone while pulling e-commerce payments.

Status quo medicine was never my thing.

I decided to try my hand at rural solo practice in the very county my great-grandfather was from. The move back to Kentucky was stark. I went from making house calls in several of the most affluent zip codes in the country, to treating families in the least. The thing that impressed me the most about the region after having been gone for such a long time was the sudden deaths of high school friends in their mid- to late 40s. Obesity was rampant. Everyone had a couple of chronic diseases. I didn’t recognize my friends. Something horrible happened while I was gone.

Patients were coming in, eager and just so thrilled that I was even there. I would tell teens and their families what to do during their checkups. Families who never drink water. Families who dislike vegetables. A year later, they would return only to have gained 10 pounds rather than lose any. Then, a patient came in and had lost some weight using one of the consumer apps. I could see the value, but noticed that I didn’t truly provide any of that value. Luckily, Doximity and Healthtap came calling with adviser roles and a full-time telemedicine offer I couldn't refuse. Before I knew it, I was recruiting our team to solve this blind spot of health care: the time in between visits in which patients progress from obesity, prediabetes, and pre-hypertension into full-blown chronic disease.

Today, there is an app for everything. Wearables are ubiquitous. Consumer health mobile programs, employer-led programs, and payer-led programs are available for patients to improve their health. But where is the primary care provider? We are drowning in problems and struggling with burnout. We are expected to hold responsibility for outcomes, but without any prior training (or tools) in affecting health behavior change.

Robert Wilson, MD, put his problem succinctly in this recent episode of the Startup Health Now Podcast. Medical practices are inherently reactive, and the challenge of embedding proactivity into that system is a top priority:

“A lot of primary care providers are not happy with the status quo. We don’t like the fact that we have to wait until someone is sick in order to really impart our clinical judgment. So, we would love to find ways to integrate prevention and risk reduction and all of these things into our daily practice. Especially family medicine. We take care of the whole family. When we impart our advice, it trickles down through the entire family. So, if we have a system of apps or integrations that will allow us to give that information in a more timely manner, that's going to be something much more meaningful.”

I attended a telehealth conference in 2022 led by Eric Thrailkill and his team at the Nashville Entrepreneur Center. I sat in the audience trying to drill down further on the provider problems evidenced by the leaders of medical schools in Nashville. Several themes related to patient-generated data emerged from the noise.

There is a firehose of patient-generated data coming in, but:

1. It’s not filtered.
2. It’s not actionable.
3. It doesn’t fit within my workflow. (I am not getting this information at the right time.)

Then there are patient problems, described succinctly in this infographic from the U.S. Centers for Disease Control and Prevention. Heart disease, cancer, chronic lung, stroke, Alzheimer’s, diabetes, and chronic kidney disease: These are the leading drivers of our nation’s $4.1 trillion in health care costs.

The challenge in primary care is to systematically learn about the entire population, educate the population on their risk, quantify their motivation level, then stratify patient populations into groups. Everything has to fall together at the right time. You must get actionable data to primary care physicians before their patients leave the office. Then, they can implement the provider-led interventions you have created to engage patients throughout this "blind spot" of health care when chronic diseases are progressing.

I am a pediatrician, so the solution to our chronic disease problem has been obvious to me and my fellow primary care colleagues for a long time. The Affordable Care Act is working in reverse to impact hospital costs:

• Waiting until a 74-year-old heart failure patient is heading home from the hospital to enlist in chronic disease management or remote monitoring is waiting too long.
• Waiting until a 64-year-old patient has been diagnosed with diabetes to then focus on their med adherence data and analytics is waiting too long.
• Waiting until a 59-year-old patient with obesity develops diabetes and hypertension and then “managing their condition” with medications and specialist visits is waiting too long.
• Waiting until an overweight 42-year-old develops the full picture of metabolic syndrome to then discuss pharmacotherapy is waiting too long.

We need to embed prevention across our entire patient population and automate as many of those processes as possible with algorithms and artificial intelligence to make our days easier, engage our patients, and show them we care.

Today’s patients are ready to live happy and healthy lives in their homes, free of metabolic disease. Now is the time for providers to lead the way by offering engagement programs and provider-led interventions that will create drastic cost reductions early in our rising risk populations.

Natalie Davis, MD, is a pediatrician and chief medical officer of PreventScripts. She is a graduate of Murray State University, The University of Kentucky College of Medicine, and Washington University Pediatrics at St Louis Children's Hospital. She became obsessed with mobile technology and its potential to scale health across populations while serving on advisory boards at Healthtap and Doximity.