Back to Blog

Introducing AI: The newest member of your healthcare team

By Dr. Nirav R. Shah, MD, MPH, Chief Medical Officer, Sharecare

We have a perfectly designed $3.9 trillion healthcare system. It’s perfectly designed to deliver more healthcare, not enable better health. But what is it that people really want? Better health. 

The reality is that we simply do not have enough clinicians to meet demand using traditional, office-based treatment. And in today’s care environment, physicians are burning out in droves, with much time and labor spent on work that isn’t actual care. The system is broken, but there’s reason for hope, and I believe we’re at an inflection point. 

During the COVID-19 pandemic we’ve seen the acceleration of telehealth, home care, and many other long-overdue changes to how we access and experience healthcare. And now, we have the opportunity to build a healthcare system that can leverage all that is remote, digital, and mobile, yet still be personal, accessible, caring, and affordable. Digital solutions powered by artificial intelligence (AI) can help augment care and, in turn, help us live healthier lives.

I recently met with Tom Taulli, a contributor with Forbes, to discuss some of the challenges and opportunities with AI in healthcare – an important and evolving topic that’s worth exploring.

Applying AI within healthcare

There are a number of healthcare applications where AI is working well today, and that’s because the right tools are matched to the right jobs. AI can work well when it’s informed by large data sets. Put simply, by collecting a significant number of answers to a single question – such as “When is a patient likely to have a health incident?” – AI can ‘learn.’ 

As an example, AI vision is helping radiologists proficiently read CT scans. With AI vision, a machine can quickly scan for abnormalities, which are then confirmed by a human. Yet for certain types of tumors, humans perform better than machines. As a pair, AI and radiologists perform better together than either alone. But not all healthcare problems are amenable to AI solutions. Humans are still better than machines on many tasks, and it’s best to consider AI applications in the areas where humans are prone to error – often, those are the routine, the boring, and the redundant tasks that machines don’t get tired of doing. By task-shifting to automation, clinicians can save time and bandwidth for their most unique capabilities: empathy, compassion, and care.

Reducing burden on the care system 

In healthcare, the early use cases of AI are already taking on some of the more mundane and repetitive tasks and helping clinicians complete them more efficiently. This increases patients’ access to their providers and allows those clinicians to practice at the top of their scope. For example, in the near future, physicians may have digital scribes capable of listening to healthcare encounters and writing three separate notes — for clinical documentation, for the patient’s reference, and for billing purposes. AI also will help clinicians understand what is in the data, summarizing and prioritizing a patient’s pertinent issues across the vast stores of information from the patient, their medical record, and other sources that give useful clues (e.g., a child’s report card). Think of AI as an extender of the care team, helping the doctor, nurse, therapist, and social worker with a different set of tasks – tasks for which AI can be optimized.

Unlocking the value of health data

The amount of health data an individual can generate is growing, and collectively, it could hold answers to better treatments and health outcomes. Machine learning needs data to build models, and the future of data collection will occur more often at the source – simply defined as the places where data is generated, such as on mobile phones and smart devices. By leveraging video, chat, voice, and photos from smartphones (with user permission, of course), we can get closer to a real-time, ultra-personalized, and 360-degree view of health. It’s also important to note that progress in healthcare does not have to come at the price of privacy. In fact, AI-based health solutions, when built and used thoughtfully, can give people more transparency and more choice. For example, through technologies like edge AI, federated learning, and zero-trust infrastructure, we can preserve privacy by sharing data in limited ways while democratizing the clinical learnings more broadly.

Digital approaches driven by AI have boundless potential to extend the clinical team, unburden clinicians, and help deliver on the promise of improved health for all. The mid-and long-term prospects for AI in healthcare are exciting to watch, and at Sharecare we’re excited about bringing that limitless potential to how you experience care.