We are surrounded by technological innovation throughout our daily lives, and yet, as a society, we are complacent when it comes to the “old world” software model that underpins our US healthcare system.
The new artificial intelligence introductions rocking the consumer world are rippling through other industries, and medicine is not immune to the possibilities of AI. Artificial Intelligence unlocks the ability to begin to predict diseases before they become chronic, it can empower doctors to provide highly personalized care to an individual patient while also scaling programs on a population level.
That said, the notion of scaling healthcare operations alongside automation of both administrative and clinical workflows has held great promise for decades, with little systemwide change. Whether the industry discussion focused on automation via system integration technology, classic machine learning, or more advanced deep learning neural networks (aka ANN’s, which include Large Language Models, or LLMs, like ChatGPT), the motivation for and barriers to implementing these automations have remained much the same.
In other words, for decades there has been technology available to advance the efficiency and scalability of healthcare, but the environment has not always existed for it to be successful. While deep learning ANN’s have made it possible to automate all kinds of pattern recognition, and LLM’s bring incredible new capabilities to generate and consume human language, the barriers in our industry – namely data privacy and fee for service business models (in addition to other mis-aligned incentives across the ecosystem) – exist much in the same way they have in the past.
So how do we overcome these barriers that, despite technological progression beyond many of our expectations, prevent us from effectively scaling up our most respected (and expensive) healthcare resources? We must create a system-wide environment that not only welcomes innovation but is agile enough to integrate new technology as it enters the market. I see three areas of focus for building an infrastructure designed for ethical innovation in the practice of medicine:
The respectful use of data
If artificial intelligence is the engine that’s powering technology, data is its fuel, and the higher quality the fuel, the better that engine runs. Pretty much every generative AI company is clamoring for high-value data, which has led to many debates around ownership, including in the courts, within the Federal Government and even as a point of contention with the current writers and screen actors strike (WGA and SAG-AFTRA) in Hollywood.
In healthcare, data ownership becomes all the more complex – is the medical IP owned by the doctor or the health system? Does the EMR vendor who houses the data have a right to it? And what about the patient’s ownership of their own health data? Though we can debate who owns the data until the cows come home, what we must agree on is the respectful use of that data to fuel innovation. There have been a number of missteps recently, including the sharing of medical data for marketing purposes with Big Tech companies. A patient’s privacy is and must always be viewed as sacred, but in order to build a healthcare ecosystem that can adapt to technological advancements, we must agree on a process to utilize this valuable data en masse.
With the right data, artificial intelligence can unlock completely new models of care, including impact-driven preventive medicine at scale, stopping and even reversing the chronic diseases that plague our society. It can be a powerful tool for growth in our industry, but we must stop focusing on the “land war” of who owns the IP, and start focusing on how best to respectfully use it to drive innovation.
AI in favor of the Doctors
Though many view AI as a tool to drive efficiency – which often means replacing people with automation – I believe AI’s most powerful use in medicine is in partnership with doctors. A computer can’t visit a patient’s bedside, it can’t create an emotional bond that drives the “doctor/patient” relationship. We call the practice of medicine “health care” for a reason – you can’t remove that human element. Health care is highly complex and the human factor is deeply entwined in a doctor’s work. The real challenge to cultivating that human factor of medicine is the massive amount of paperwork expected of providers, which takes valuable time away from the actual care of patients. AI has the ability to remove much of the administrative noise of practicing medicine, freeing up doctors and clinicians to focus on cultivating that human element of care.
When you look at the traditional medical practice model, there has always been a line between what is controlled by doctors and what is controlled by the technology vendors. This power struggle is prevalent throughout our industry, affecting every level of the system. It also leads to the misalignment of incentives, which ultimately slows the innovation process. Bringing me to my third point...
Securing a seat at the table
To continue to work at the forefront of healthcare innovation today (and decades from now), we must align incentives across every part of our ecosystem, facilitating adoption of new care models and the adaptations to new technologies. I call this alignment “full stack,” a term drawn from the software domain that means the ability to manipulate every component across a suite of technologies. As an organization in healthcare, that means the ability to build and evolve technology alongside the care and business models. This greater alignment means greater access to high quality data; opportunities for cross-collaboration between doctors, engineers, data scientists and AI experts; even the adoption of value-based care models becomes easier.
By adopting a full stack strategy, we no longer become spectators in the development of new technologies, but rather an active partner in its implementation within our industry. Artificial Intelligence is here to stay, and we in the healthcare sector should be leading the charge on how we integrate this technology into our industry.
AI is a new category of innovation from what we’ve seen in healthcare in the past, and we as an industry don’t have a great track record when it comes to tech innovation and adoption on a system-wide level. The practice of medicine is complex – far more so than many consumer sectors – and you can’t just cut and replace what already exists. We must incentivize the adoption process of wrapping new tools around workflows and empowering our doctors as active partners in this transformation. Together as an industry we should be embracing innovation but actively engaging in its adoption to ensure alignment across our businesses. We must be so much more than just participants in this change, we must be economic partners in the transformation.
Sanjit Mahanti has over 15 years of technology and innovation experience in the healthcare space. His dynamic and evolving expertise includes executive leadership positions in academic healthcare, population management and technology. During his tenure as Chief Business Development Officer for Keck Medicine of USC, he was instrumental in the creation of the organization’s innovation management function. Today, Sanjit serves as co-founder and head of business development for Akido a health tech company and medical network leveraging data and artificial intelligence to transform the healthcare experience for all, focusing on addressing the systemic inequities that lead to chronic illness and vulnerability.