The “Big Picture”: Maturity of Artificial Intelligence in Healthcare
The current pandemic provides an opportunity to re-evaluate the maturity of Artificial Intelligence in Healthcare.
For some time now, AI solutions are being developed and deployed across the healthcare ecosystem. In the biopharmaceutical sector, they are finding applications in drug discovery, clinical trials, predictive modeling to identify at-risk patients, and even supply chain management. Within health systems, AI solutions are being embedded to aid with diagnosis, predict outcomes, and enable continuous monitoring inside and outside of the hospitals. In health insurance, AI is being used to perform actuarial risk-modeling at individual and population levels, developing hyper-personalized insurance offerings, all the way to fighting fraudulent claims. Across all segments, AI is being deployed to drive efficiency in back-end operations, be it financial management, inventory management, even talent management.
The current pandemic has seen AI-enabled solutions joining the fight. We have seen the emergence of mobile technology such as new police helmets that detect fever up to 16 feet away, robots sensing people without masks, even phone apps that alert you if you are in proximity of someone with the infection. Chatbots are offering medical consultation. Intelligent drones and robots are helping with care and food delivery. AI is being deployed for the rapid discovery and development of drugs and vaccines. AI-based models are helping to forecast potential spread and impact as social distancing measures are eased, and enable reopening of the economy.
While these developments are good news and will continue to enable a shift of healthcare from where the hospitals and clinics are to where the people are, how do we evaluate the overall maturity of AI in healthcare?
In our book, Redefining Innovation: Embracing the 80–80 Rule to Ignite Growth in the Biopharmaceutical Industry, we used the SAE International’s J3016 framework that defines the level of driving automation to introduce the concept of Smart Integrated Medicine. The same frame is highly valuable in understanding the maturity of Artificial Intelligence in Healthcare.
Most vehicles on the road in the US and Europe today qualify to have Level 1 capabilities — lane assist, distance control, or emergency braking. However, the driver remains in full control at all times. Quite a few are embedding features that would qualify as Level 2. For example, cars that steer themselves in tight parallel parking spaces. However, they still require the driver to brake and accelerate in most cases. Waymo has been working on cars that would have qualified as Levels 3 to Level 5; however, that development is on hold under the current crisis. Given that we are at a very early stage of development in autonomous driving, the auto insurance policies today are still written, assuming Level 0 automation — the lowest common denominator when it comes to cars globally.
Using the above framework, we can make the case that AI technologies in healthcare, broadly speaking, today at best resemble Level 1 maturity of autonomous cars. In most cases, these technologies can provide decision-making assistance, while the human — the physician, the researcher, the operator — must still do the primary decision-making. The current pandemic is proof. Despite all the development in AI, we were not able to predict the onset of the pandemic. While it is finding use in helping to prevent the spread, detect the infected, or develop a cure, it is primarily assistive, except for the chatbots that are acting as virtual doctors. Even in a rush to find the cure, given lack of underlying data, AI is finding limited use in the development of the vaccine or a biopharmaceutical solution, or proving the efficacy of existing options such as HCQ. While Google and Apple have announced a partnership to assist with contact tracing, the concern for infringement on individual privacy seems to be winning over the need for the collective health of the society.
Healthcare is more personal than autonomous vehicles. Despite the use of autonomous vehicles to transport COVID-19 tests, we are a few years away until we see Level 5 vehicles all around us. I would bet and say that it will take even longer before we see Level 4 or Level 5 healthcare solutions. Beyond an evolution in policy, insurance, and legal models, it will take a lot of “trust” and grappling with “privacy” and “safety” issues for us to adopt such solutions fully.
That said, we should see accelerated investment in AI-enabled technologies as we continue to shift away from the physical architecture and continue to build the technological architecture of healthcare — a future where care is delivered where the person is, not where the hospital or clinic is. Incumbent healthcare companies should map out how they can leverage AI to enhance their current business model and operations. At the same time, they should also develop a robust understanding of potential healthcare business models AI could enable — and make strategic choices regarding their role in such a future. While we may be a few years away from Level 5 AI in healthcare, it will fundamentally impact the provision of care in the near term.
— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
Ruchin Kansal
Founder and Managing Director, Kansal & Company
𝘞𝘦 𝘤𝘰𝘢𝘤𝘩 𝘵𝘳𝘢𝘯𝘴𝘧𝘰𝘳𝘮𝘢𝘵𝘪𝘷𝘦 𝘭𝘦𝘢𝘥𝘦𝘳𝘴, 𝘤𝘰𝘯𝘴𝘶𝘭𝘵 𝘸𝘪𝘵𝘩 𝘤𝘰𝘮𝘱𝘢𝘯𝘪𝘦𝘴 𝘢𝘯𝘥 𝘤𝘰-𝘤𝘳𝘦𝘢𝘵𝘦 𝘸𝘪𝘵𝘩 𝘪𝘯𝘷𝘦𝘴𝘵𝘰𝘳𝘴 𝘵𝘰 𝘣𝘶𝘪𝘭𝘥 𝘵𝘩𝘦 𝘧𝘶𝘵𝘶𝘳𝘦 𝘰𝘧 𝘩𝘦𝘢𝘭𝘵𝘩𝘤𝘢𝘳𝘦.
𝘍𝘰𝘳 𝘮𝘰𝘳𝘦 𝘪𝘯𝘧𝘰𝘳𝘮𝘢𝘵𝘪𝘰𝘯, 𝘱𝘭𝘦𝘢𝘴𝘦 𝘷𝘪𝘴𝘪𝘵 https://kansal3c.com/
Follow us on Linkedin
Copyright © 2020 Kansal & Company, All rights reserved.
