AI Talk: AI and future of health care, part 1

Jan. 21, 2022 / By V. “Juggy” Jagannathan, PhD

Last year, 2021, was a tumultuous year to say the least. And we are starting 2022 in the throes of the fifth wave of the COVID-19 pandemic. We are witnessing unprecedented changes to our health care ecosystems.

With that theme in mind, to kick-start 2022, I am taking a stab at a 10,000-foot level characterization of key issues that the health care industry faces and how artificial intelligence (AI) can impact the solution trajectory. In this three-part series, we will explore the following themes: interoperability of health systems, drug costs, hospitalization costs, affordable access to care and addressing health equity.

Part I. AI and future of health care: Interoperability and drug costs

Interoperability. An apt analogy that characterizes the state of our ability to share data in health care is the story of the “Tower of Babel.” Data about patients is distributed across a variety of health care facilities, from hospitals to clinics to payers to pharmacies to registries, using a variety of incompatible standards. The ability to exchange, aggregate, process and understand data about patients is a critical infrastructure piece for providing efficient, effective care and for deploying AI solutions as well.

In a recent blog, the Centers for Medicare & Medicaid Services (CMS) underscored the need to “journey toward achieving a health care system where patients, providers, and payers can easily exchange information to bolster better care outcomes.” CMS has mandated several rules which require seamless transfer of information across payers, providers and patients starting this year (though the pandemic has delayed the enforcement of the rules, for now). The rules require the ability to share a standardized dataset labeled “United States Core Data for Interoperability (USCDI).” CMS has also indicated how that data can be shared – using protocols such as FHIR. This year, 2022, will be a transformative year for interoperability as these rules become enforced.

Drug costs. One of the big drivers of costs in the health care system is the ever-increasing costs of drugs. This excellent report from Congressional Budget Office (CBO) last year provides a detailed analysis of the research and development costs that underlie drug development. The average cost to develop a new drug ranges from $1-2 billion!  Read the full report for all the factors that go into that cost, but here I will highlight a few factors where AI can make a huge difference.

  • Drug discovery is one major area of expense. AI has incredible potential when it comes to identifying molecules which can impact specific targets in human body. Google parent company Alphabet has just launched a new company, Isomorphic Laboratories, based on its DeepMind research on protein folding with the express purpose of reimagining the drug discovery process with AI. A recent survey explores the many ways deep learning techniques can help with drug discovery.
  • Clinical trials are another major expense. All drugs go through multiple, progressively more complex phases before they are approved by Food and Drug Administration (FDA). We all have become quite familiar with some of the lexicon used during this pandemic – phase I, phase II, phase III trials, etc. This CBInsights article details various ways in which AI can turbocharge clinical trial enrollment and monitoring. Clinical trials recruiting participants or patients seeking to enter a clinical trial can leverage electronic health record (EHR) data (story above) to automatically provide matching.
  • Monitoring of patients in clinical trials, as well as adherence to medication regimens can be accomplished through tele-visits and a variety of remote monitoring technologies. One novel technique which can potentially streamline and considerably reduce the cost of a clinical trial is the use of digital twin technology. Take a look at this CBInsights article on this front. The core idea here is that a digital twin model can create a baseline of what will happen to the patient without treatment, and it can be compared to how the treatment impacted the patient. Whether the FDA will approve of such approaches, only time will tell. If models can be used in even earlier phases, they could help cut the costs of conducting such trials significantly.

AI use is pervasive in drug discovery and development. This recent survey article published last fall goes through a gamut of applications that are now enabled by AI.

I am always looking for feedback and if you would like me to cover a story, please let me know! Leave me a comment below or ask a question on my blogger profile page.

V. “Juggy” Jagannathan, PhD, is Director of Research for 3M M*Modal and is an AI Evangelist with four decades of experience in AI and Computer Science research.