One of the more popular columns in The New York Times Magazine is “Diagnosis” – adapted as a docuseries in 2019 on Netflix – where Lisa Sanders, M.D., describes patients who present with troubling symptoms and no clear diagnosis. These articles describe cases where it takes a while for doctors to figure out what’s going on, and sometimes the answers come from the unlikeliest of sources.
Today you can read these articles knowing that artificial intelligence (AI) is going to enable doctors and researchers to solve these cases more quickly and easily. That should be no surprise. AI seems like a natural fit for complex cases. But what might surprise you is that it won’t be long before you start hearing about AI in just about every medical specialty. That includes on-demand care settings like urgent care, retail clinics and virtual medicine.
What Is Artificial Intelligence?
AI is often confused with more primitive rules-based or algorithmic-based systems that asynchronously follow specific clinical protocols to move through a visit using if-then or branch-tree logic. AI differs from these systems in that it gives machines access to much larger amounts of health data and allows them to “learn” for themselves with each encounter.
A skeptic might view AI as an amazing solution for which there is no known problem, especially when it comes to low-acuity healthcare situations like urgent care. But it is a faulty assumption that the problem has to be a complex scientific calculation for AI to be applied, a notion promulgated by the association of big machines, like IBM Watson, with artificial intelligence.
To the contrary, there are now several companies going after the low-acuity healthcare market. Many are chatbot systems with AI back ends, as listed here on the Traxcn site. But the best way to explore AI in the on-demand market is to look at specific applications. At UCP Merchant Medicine, we use Intellivisit Concierge, one of the systems mentioned on the Traxcn site. The applications discussed below are real-world examples of Concierge at work in the on-demand world.
Nearly all of our clients are health systems, many of them using the Epic EMR and many with a footprint of 10 or more urgent care centers along with a telemedicine offering. One of the biggest reasons UCP Merchant Medicine chose to work with Intellivisit is that Concierge is integrated with Epic out of the box. And it is a full integration, meaning Concierge appears as one of the tabs in Epic’s top menu bar. It is also part of Epic’s AppOrchard.
The Front Door
If you were a fly on the wall of any health system C-suite, you know one conversation (and problem) that comes up over and over is finding a one-stop approach to directing patients for the most appropriate and efficient care.
What is the most common approach for health systems? More often than not it is a central phone number, first telling patients, “If this is a medical emergency, please hang up and dial 911.” Then the patient hears a list of options that has boiled down everything a hospital does into a list of single-digit numbers. After hearing that list, callers are lucky to remember which number is the best option for their problem.
And that’s just the central phone number solution to figuring out what to do. There’s always the website, which, by the way, is one of the options you are given when you use the central phone number. On that website you see the following options: Find a Doctor, Find a Location, Pay My Bill, Walk-in Care, Urgent Care, COVID-19 Resources, Request a Medical Record, and Contact Us (which sends you into these same options along with a dizzying number of additional choices).
As healthcare leaders attempt to make the patient experience better, sometimes the result is a process that has too many features. Think about it. In the last 10 years we have added patient queue management systems; text-based satisfaction surveys; telemedicine integrated with urgent care; COVID testing and vaccinations; price transparency scripts for our front desk staffs; follow-up appointment scheduling, including in-home nurse visits. As we add these features to bring about more convenience, we sometimes fall short on integrating them seamlessly for the patient.
You get the point. The front door to most health systems is not a friendly place and the systems built to better serve patients can be a twisted web. But COVID-19 has changed that landscape. Patients have been forced to learn the benefits of a digital front door. And in the not-too-distant future, perhaps as soon as later this year, the digital front door will be opened by a very friendly, voice activated, AI-based system. In fact, this is already at play at a number of health systems, albeit somewhat under the radar.
Based on some simple questions and a recognition of your Internet device, an AI-based front door will find your medical record, billing and insurance history, and current location. All of this will save you time and aggravation by having you simply confirm your identity and reason for coming to the digital front door. In the case of Intellivisit’s Concierge, each question presented is based on the answer to the previous question, until very quickly you are directed precisely to where your problem can be solved, whether it’s virtual, urgent, or emergency care.
In most cases, you will have gone through a Concierge digital medical interview before leaving home. If that interview results in a low Emergency Severity Index (ESI) score and no lab tests are needed, you may never leave home and instead initiate a virtual visit that solves the problem on the spot. That includes sending a prescription to your preferred pharmacy or even having the medications delivered to your front door. It’s also possible Concierge will direct you to an urgent care center, where you are greeted by a staff member who knew you were on the way, knew your chief complaint, had authorization to start labs, and had advised the provider on duty about your situation. Or with a high enough ESI, you might be directed to call 911 or go directly to a local emergency department.
Let’s take another scenario where you just walk into an urgent care center. In this case a staff member greets you, escorts you to an exam room and conducts the medical interview in person using Concierge as the guide. At the end of that interview, Concierge creates an ESI score and point-of-care tests are ordered as part of the standing order protocols established by the site’s medical director. An initial set of differential diagnoses and treatment suggestions are also prepared for the provider. All of this is done based on your initial medical interview prior to meeting with a provider.
Among the urgent care sites currently using Intellivisit Concierge, in 95 percent of cases the provider chooses from the top-five diagnoses suggested. In 66 percent of cases, the provider chooses the top diagnosis suggested by Concierge.
So how is the AI system coming up with these suggestions for point of care tests and differential diagnoses? Two things are going on in the background. First, with each question the system is searching through more than 10 million patient years of data covering ages 1 to 99. This data includes the patient’s own medical record, so the system can flag drug allergies, preventive services that are due or suspicious trends from prior visits. Additionally, the system accesses other health-related data covering 40,000 symptoms, 4,000 diagnoses, 7,000 medications, lab and test results, demographics, drug allergies and patient outcome surveys.
Second, with each encounter over time, the system is designed to “learn,” thus improving the suggestions presented to the provider. It bears repeating, the system is not replacing the provider. It simply makes suggestions to the provider based on orders of magnitude more data.
Benefits of AI
One of the primary benefits of AI in urgent care is what we refer to as elevated clinical IQ. The moment providers click on the Concierge tab in the EMR, they have a head start on what’s going on with their patient. There is no substitute for an in-person medical exam, but an AI-driven medical interview that connects to a vast amount of healthcare data provides the kind of advanced triage not previously available. This translates into two additional benefits: increased provider capacity and reduced care variability.
Provider using the Concierge tool experience time savings of 10-50 percent. So at a minimum, a provider who sees 40 patients per day feels like they have seen 36. Or said another way, a provider who typically sees 36 patients a day, can see 40-54 using AI-enhanced urgent care. Increased provider productivity also means less stress over the course of a shift. It can mean closing the doors on time, which impacts the entire clinic team. This makes leadership and culture advancement realistic and effective rather than empty words about vision and mission.
For providers, one of the most attractive features of Concierge is the automatic creation of a medical note within Epic. Just like the suggested diagnoses, this medical note is considered a draft until the provider approves it as is or edits it to his or her satisfaction. In either case, the reduction in charting time is significant. SmartSets within Epic also aid in the creation of a medical note, but this feature isn’t linked to an AI back end. These are usually originated by the provider based on their most common patient scenarios and their habits around creating a SOAP (subjective, objective, assessment, plan) note.
But perhaps more importantly, AI-enhanced urgent care will result in reduced care variability. We have found this applies to experienced providers as a group, but especially between experienced providers and new grads. Reduced care variability not only translates into better clinical outcomes, but it also leads to higher patient satisfaction scores, especially provider specific patient satisfaction scores.
Over time, we expect AI-enhanced urgent care will also significantly reduce door-to-door times for patients. This is the result of the advanced triage process, pre-ordering tests based on the medical interview, giving providers a head start on their medical exam, and reducing charting times. All of this will impact patient satisfaction scores, which are already aided by reduced care variability.
It is becoming increasingly clear that we are moving to a “digital-first” front door. The technology was already in place for this to happen, but COVID-19 created a perfect storm for a huge number of healthcare consumers to experience a digital front door for the first time.
What does “digital first” mean in practical terms and, further, what does “relationship” mean in the digital age? Digital-first means patients will seek access to a health system by starting the process through digital means, most likely their smart phone, tablet or desktop computer. This device effectively becomes a patient’s decision support tool, in which they use artificial intelligence as the first step in making health care buying decisions.
In terms of relationship development, the sooner a health system can become integrated with a patient’s digital lifestyle and meet them where they are in terms of health needs, the more effective that health system will be at managing them over their lifetime. Digital-first makes it possible to create stickiness with the patient, such that the health system is positioning itself to aid the patient in reducing their costs by selecting appropriate care over time. And by moving toward more common digital consumerism and retail principles, which are at play in most every other customer situation, a patient’s relationship with the health system feels less foreign.
AI has the potential to smooth out this experience from the very start. And ironically, a smooth digital front door will smooth out those who come in through the physical front door. More than half of those patient visits could be new patients or those who are looking to reconnect after breaking away for a period of time. Building these relationships through positive, predictable experiences will be the key to competitive differentiation in a post-COVID world.
Applications of tools like Intellivisit make this possible, and will become a core strategy of health systems over the coming years. COVID-19 has altered the health care industry and shown that patients are ready, willing, and able to adapt to a digital-first front door.
Brandon Robertson is managing director at UCP Merchant Medicine; Weston Johnson is chief administrative officer at UCP Merchant Medicine; and Tom Charland is founder and president of Merchant Medicine.