Q&A: AI ready to assist, not replace, human beings in asthma, allergy care

Key takeaways:

• Artificial intelligence (AI) may improve genetic screenings to prompt more selective allergy testing.
• AI may track activity and warn patients when asthma attacks are imminent.
• Coding and billing also may benefit.

As breakthroughs in artificial intelligence power more effective and efficient treatment elsewhere in medicine, many asthma and allergy specialists may wonder how these innovations could impact their day-to-day practice.

Maria Gil, MHA, specializes in how data, technology and artificial intelligence (AI) can transform health care solutions as a partner with Genpact, which is a global professional services firm that serves eight of the top 10 global health care companies.

While working with hospitals, payer-provider networks and other health care companies, Maria has integrated cutting-edge technology including AI with people processes to improve patient and customer experiences and solve industry pain points.

Healio spoke with Maria to find out more about how AI will impact allergy and asthma care.

Healio: Clinicians use a variety of tools to diagnose asthma and allergy. Is it possible for AI to assist clinicians in these diagnoses?

Gil: As I’m not a clinician, I can’t go into diagnosis specifics. But what I can say is that AI has the potential to accelerate time to diagnosis and rule out issues.

Maria Gil

Years ago, when I was leading digital health at AXA Asia, I was looking at what companies were doing in the space of diagnosis and triage. I remember talking to a company that had trained its AI on a series of medical textbooks. There are a lot of issues with that going directly to a consumer, but having a comprehensive, up-to-date decision assistant for clinicians is definitely possible.

With regard to allergy assessments, we won’t be able to replace the need for a skin test. But we might be able to improve our genetics screenings to prompt those tests in certain populations, rather than waiting for individuals to ask for them due to an allergic reaction.

Ultimately, the role of AI will be to assist humans, rather than to fully replace them, to make sure we are using the most up-to-date information and can analyze data more deeply and assess results with greater accuracy.

Healio: Treatment adherence is a challenge once patients leave the clinic and are responsible for taking their medication appropriately. Can AI play a role in monitoring patient adherence and helping patients stick with their treatment regimens?

Gil: Adherence will always be a two-way street. We can only interfere as much as individuals will let us.

Several years ago, I reviewed an AI-based technology for diabetes management that was installed on the phone. It would look at activity patterns and geofencing to prompt trigger-based reminders to check blood sugar or take insulin. Basically, a diabetes patient could get an alert at a restaurant to take their insulin because they were about to eat, or near a playground when combined with activity data suggesting that they had just exercised. However, if patients don’t give permission for the AI to integrate with the geofencing data, or if they choose to ignore reminders, there is no amount of intelligence that will help.

We can make smart inhalers that will notify us when we need to re-order. We can better analyze activity data to understand what would trigger an allergic reaction. We can set up alerts based on local weather patterns that can inform us that allergy season is starting.

But whether we are talking about remote monitoring for asthma, or a million other conditions, we cannot forget the two-way street. Yes, there is a lot of data that we can better leverage to better engage populations and drive healthier behavior. But those individuals still have to choose to take those recommendations, and that is not something AI can help with.

Healio: Asthma and allergy care may involve collaboration with primary care providers and other specialists. Are there any areas where AI can facilitate this collaboration?

Gil: AI can help in the digitization, collation, and summarization of content. From the introduction of OCR, or computer vision, technology has been able to read and input paper documents into the digital world. Leveraging generative AI, now we can parse that document and pull out unstructured information with even greater accuracy. Furthermore, we can assign information to patient files and summarize notes with greater accuracy than ever before.

That said, AI is one of many tools available to facilitate the collection and sharing of information to improve collaboration. It can be thought of as a short cut or grease on the wheels to help us better translate our thoughts into words, to do research faster and to ideate more broadly, especially with the power of large language models (LLMs). But we also need to be cautious not to overly rely on it and lose the human touch.

While collaboration can be addressed with AI, as well as by a variety of other technologies, and there are still challenges that are independent of technology. There will be a great deal of change management that will need to take place in health care to facilitate its adoption. No different than the introduction of the electronic medical record, the broader introduction of AI will be met with resistance and reluctance until it is well understood, tested and fully appreciated for the value it brings.

Healio: Are there any areas where AI can improve practice management processes for asthma and allergy practices, such as coding, billing, patient communication or even marketing?

Gil: It is much easier to see the more immediate impacts of AI on practice management. My hypothesis is that coding and billing is going to go through rapid transformation in the next 3 to 4 years, with advances across revenue cycle management, from more accurate coding to real-time payment tracking.

Beyond that, AI can be unleashed to streamline pre-authorization and appeal denials, as well as to ensure patients receive the treatment they need, without spending hours waiting on hold for the payer to resolve an issue. If this wasn’t health care, I would say we could move faster, but the industry is fragmented and generally slow to move. We are already seeing these improvements in other industries with AI dramatically transforming task-based processes.

Healio: Research into treatment for asthma and allergy is ongoing. How can AI facilitate this research?

Gil: AI is already impacting research. For example, AI algorithms can help determine which molecules are most likely to be safe and effective human therapies, optimizing drug discovery. In early 2020, Exscientia announced the first-ever AI-designed drug molecule to enter human clinical trials. The AI-fueled pipeline has been expanding at an annual rate of almost 40%. AI-enabled drug discovery holds the promise of dramatically changing pharmaceutical R&D by improving productivity, broadening molecular diversity, and improving chances of clinical success. Asthma and allergy are not immune — pun intended.

Healio: Are there any other areas where AI may have an impact on asthma and allergy care?

Gil: Ultimately, we are moving toward a generation of tech-enabled humans. This shift will apply to every aspect of health care, from diagnosis and treatment to adherence and remote monitoring to back-office operations. The question will reside more with how we choose to adopt and embrace technology, rather than its inevitability, which is no different from the smartphone.

Imagine a world where AI makes a call to the doctor to make an appointment on your behalf, saving you time. It is then used at the doctor’s office to call your insurance to verify benefits, allowing the front office staff to focus on you when you arrive, after receiving an AI-generated alert from your phone that you’re almost there. Imagine the interaction with the physician, where they are assisted by AI to make the most fact-based diagnosis, making your care more personalized than ever, allowing the physician to spend more time with you, rather than dictating notes.

Tech-enabled does not mean less human. In fact, tech-enabled could create more time for us to have more meaningful interactions, leaving machines to be on hold for the next available operator in our stead.