Tapping into advanced analytics is enabling healthcare providers to reach more patients with improved outcomes
The Role of AI in Healthcare
The art of science in certain branches of medicine or complex disorders like neurology and epilepsy, respectively, are very complicated. In the case of neurology, many symptoms can overlap or there can be ambiguity in a patient’s symptoms—and there are clear consequences of ambiguity. Beyond the symptoms, treatment plans for epilepsy are also complex, and to put it simply, issues of uncertainty are less common in a practice like cardiology, where the heart is a much simpler organ than the brain.
However, the proliferation of artificial intelligence applications in healthcare have the power to produce enhancements of efficiency and efficacy that are currently unimaginable. Opportunities exist in healthcare to apply artificial intelligence practices for the enhancement of outcomes and efficiency while reducing complexity of specialized disorders.
AI Can Pave a More Clear Path
Healthcare operates at the intersection of the patient, caregiver, and the treatment plan. But today healthcare has been empowered by technology and data. When doctors ask questions to their patient that results in data. The doctor takes a patient’s responses, analyzes and interprets this “data” and then recommends a course of action. The more data that the physician has access to, and the higher the frequency, resolution, and cadence of that data, the better the results for the patient. As such, there are two benefits where AI affects healthcare—better matching of a patient with a therapy pathway and better matching of a patient with a physician or care team.
If the patient is presented with a controversial or severe course of action, many patients will seek a second opinion under the belief that a different doctor will analyze the same data differently. Then the patient (who is often untrained in the medical field) will have to choose the doctor who they think can most accurately predict the most positive outcome for the patient. However, AI can become a digital second opinion so the patient doesn’t have to seek a second doctor and a second office visit. AI can interpret and analyze the information based on historical data for a different course of action.
Secondly, whether it be due to circumstance (relocation, changes in insurance providers, a bad experience, etc.) or seeking a second opinion, people migrate from physician to physician. Physicians have come to expect it; it’s a human behavior pattern. AI has the potential to pattern-match a patient with a clinical path forward by connecting them with the clinician that is best suited to their needs. Believe it or not, depending on a patient’s symptomatology, there is the “right” doctor to see—AI can predict that with the highest degree of certainty.
Examples of AI Applications in Healthcare
With AI’s potential to unlock a more seamless patient experience, imagine how it can transform care in specialized disorders. Consider the patient journey in epilepsy, for instance. If a patient has a seizure once, does the patient see a neurologist or their GP? At what point does the GP refer the patient to a neurologist? And how many cycles of “trial and error’ are attempted before the neurologist enlists the support of an epileptologist? On average, a patient lingers with a neurologist for 8 years before being referred to an epileptologist. It should be 18 months. Epilepsy has many causes, symptoms, and treatments. With the power to match a clinician to a patient by levels of specificity, AI can and should look at data collected from the entire continuum of care to predict what epilepsy subtype the patient may have and then route them to the right caregiver quickly and efficiently. The chronic lack of urgency in epilepsy management could be the difference between frequent seizures and the other risk factors associated with epilepsy and better outcomes for the patient and reduced healthcare costs.
Beyond affecting the individual patient experience, AI also has the potential to provide a broader population with disease-modifying drugs. In the case of epilepsy, 65% of epilepsy patients can achieve freedom of seizures from pharmacologics and an additional one-third of patients don’t respond to traditional medications. Of the remaining patient set, 8% can achieve seizure freedom with surgery and non-pharmacologics. That leaves behind 20% of the population with epilepsy with no known therapy that can provide relief. No one has studied this population or organized the medical data into clusters of similarities to target for future therapy. By unlocking AI and data science, data on this 27% can be collected, monitored, and genetically screened to find small bundles of member groups that previously were uncategorized. Physicians can then identify similarities for future drug research.
Standardizing Data to Inform Better Results
On a more macroscale, metrics and outcome measurement is an issue across healthcare as a whole. For every disease group there are different metrics that matter with 0% agreement across fields about what metrics matter the most. To continue moving the needle in streamlining patient care, data needs to be interoperable and transferable. There needs to be coalitions that are working to define endpoints and outcome measures. Consider epilepsy again, of course being seizure-free is the goal, but what about other debilitating side effects like the impact on mental health or the increased risk of heart issues, diabetes, and obesity?
One of the leading healthcare measurement organizations, International Consortium of Health Reported Measures, is doing just that. ICHOM is working to unlock the potential of value-based healthcare by defining global Standard Sets of outcome measures that matter most to patients and are driving worldwide adoption of these measures to create better value for all stakeholders.
ICHOM recently announced the kickoff of their Working Group for the Epilepsy Patient-Centered Outcomes Set, marking an important step toward promoting data quality and availability, and strengthening the delivery of high-value healthcare for epilepsy patients.
Conclusion
The collaboration between physicians and technology has the power to yield an output that is trusted by physicians and regulators. This has a domino effect on care access (which is not a constant around the country), improved affordability of quality care, and most importantly, a more clear path to effective patient treatment.
Leo Petrossian is the Chief Executive Officer of Nile, an epilepsy management platform that replaces trying with knowing in epilepsy care through a patient app and health care provider portal. Leo is an experienced entrepreneur and digital health innovator who often speaks on medtech innovation and investments.
