The Healthcare Crisis That AI is Uniquely Suited to Fix

Healthcare is currently in the throes of a capacity crisis.

I’m not talking about a developing-world or even a developed-world capacity crisis. I’m talking about a global capacity crisis fueled by escalating demand on healthcare systems that simply don’t have the infrastructure or the number of providers necessary to meet the needs of patients who now, more than ever, are living longer and sicker.

The aging population – a so-called “silver tsunami” – has become a serious concern worldwide. In America alone, some 11,000 seniors each day age into Medicare, the federally funded program that pays the healthcare bills of people 65 and over. Complicating matters further is that an overwhelming majority of these individuals enter Medicare already living with chronic health conditions, many of them expensive to treat. The National Council on Aging reports that about 95 percent of adults aged 60 and older have at least one condition, while almost 80 percent have two or more. At the same time, the United States is facing a national shortage of healthcare workers that’s expected to climb to 100,000 by 2028. And many people don’t know that the U.S. is short on hospital beds. Total admissions at U.S. hospitals now exceed 34 million nationwide but the number of staffed beds has been declining since 1975. In the United Kingdom, doctor shortages mean that the average general practitioner has to care for 17 percent more patients today than nine years ago. And as the European Union’s older population grows, the total number of hospital beds dropped by 170,000, or 7 percent, between 2012 and 2022. Put plainly, even the wealthiest countries in the world currently do not have the resources to confront the silver tsunami and the immense wave of demand that comes with it.

Even as advances in biotech have been exceeding expectations for cures and relief of suffering, these innovations are also adding to our healthcare delivery challenges. Each new diagnostic test and therapeutic add to the tasks and cognitive load required of our short-staff clinicians. And as top clinicians focus on the hottest new treatments, fewer and fewer are focused on the basics of prevention, chronic illness, and care coordination.

The good news, I believe, is that a different type of innovation can fix today’s care-delivery challenges – innovation that simplifies the patient experience, reduces overhead costs, and streamlines care pathways. But that’s going to require the healthcare system to accelerate the successful adoption of cutting-edge technology, like artificial intelligence, to bring about unambiguous benefits for clinicians and their patients alike. Just as important, for systemic adoption to truly take root, it will also take a change in mindset among many practitioners. 

As a primary care physician, professor of medicine at Harvard Medical School, and former CEO of the Massachusetts General Physicians Organization, I have devoted much of my 30-year career to studying and tackling care-delivery issues and exploring how technology can be deployed to address them. Most recently, I served as the inaugural National Director of Transformation at National Health Service England, functioning as its chief digital officer where I oversaw care-improvement infrastructure, technology and data strategy and services, biotech innovation, and life science research. While at the NHS – where my team and I successfully expanded services to more than two-thirds of the 60 million people living in England via the NHS self-service App­ – I saw firsthand how instrumental technology can be in not only solving the existing capacity crisis but in also lowering the cost of care, which in America grew 7.5 percent in 2023 to $4.9 trillion (or $14,470 per person), accounting for 17.6 percent of GDP. 

I recently had the opportunity to serve as a panelist at Health@Davos in Switzerland where I joined other professionals in healthcare, public health, and scientific research to discuss how AI can address global health challenges, including the world’s capacity crisis. We discussed how AI can be leveraged to improve quality of care, enable personalized medicine, as well as detect and manage diseases. For example, in the area of diagnostics, a company called Skin Analytics recently received the European Union’s Class III CE mark – which is the EU’s highest level of regulatory scrutiny – for its medical device DERM, making it the world’s first legally authorized autonomous AI for detecting skin cancer. In its announcement, the company shared that DERM achieved a 99.8 percent accuracy rate in ruling out cancer, surpassing the performance of human dermatologists. Adoption of this innovation could significantly lower the unit cost of detecting skin cancer. Another innovation by a company called Cognoa, uses short videos of children to screen for autism. Diagnostic tools like these are being developed at a rapid rate today, but adoption by professionals has been slow. So while we can expect to see more and more tools come on the market that will lessen the burden on providers and open up access for patients, clinicians will need to embrace these tools if they have a chance at fixing our capacity crisis.   

In 1970, the late author Michael Crichton published a book called “Five Patients,” which was inspired by his experiences as a student at Harvard Medical School in the late 1960s. In the book, Crichton describes some early uses of computers in medicine. He predicted that computers would soon play a significant role in clinical decision making. Well, the future that Crichton prophesied has been slow in coming, but it is finally here. Today, we’re seeing a proliferation of AI-driven clinical decision support tools. In healthcare, these innovations can be especially useful in reducing what’s called “excess cognitive load,” or the overwhelming amount of data physicians must sort through. Simply put, there is way too much information for clinicians to assess in order to make good decisions. This is where AI has the greatest potential. For AI, there’s no such thing as too much information. These tools can quickly sift through vast amounts of data to zero in on the critical details and help inform clinicians’ decisions – oftentimes at the point of care. I’m excited to be working with Red Cell startups like Trase Systems, an AI agent platform that’s helping to pioneer this field with tools that can recommend care pathways to help clinical teams close care gaps, utilize data mining to flag and triage patients for more targeted hands-on care, and assist providers to make better decisions concerning patient treatment. 

In addition, AI tools can help to empower patients with more “self-service” options for routine care needs. For example, Harvard University, Massachusetts Institute of Technology, and others have collaborated to create OpenEvidence to aggregate and synthesize clinical information. It’s like an encyclopedia of all medical knowledge that’s now available to the public, which means anyone can type in a medical question and get a response that is both astoundingly accurate and clearly documented. Such innovations are a huge democratization of the expertise of physicians to help patients determine what they can manage on their own versus what requires medical intervention. 

These are just some of the exciting technological developments that can help bring about the transformation that our healthcare system so desperately needs. But that will only happen if those individuals who deliver care test and use these technologies to full effect. In “Five Patients,” Crichton noted that the only people who can change healthcare are the ones providing the actual care because it is their work that must change. Given the risks involved in healthcare, physicians will be understandably conservative in their adoption of innovation. And, yes, the roots of opposition may also lie in fear for one’s job. Such “guild protectionism” must take a backseat to innovation because it all boils down to this: The way healthcare is currently being delivered is not working and is inadequate to meet the demands of our society. If we are to overcome today’s capacity crisis, we have to understand that maintaining the status quo in a system that is failing is a risk we can no longer afford.