Human-Centered AI Implementation

Happy New Year! I attended the HIMSS AI in Healthcare Forum Conference in beautiful #SanDiego, #California, local techies now call it the new Gene Town, in #USA.? I’m highlighting a few snippets of my interactions with a few movers and shakers and subject matter experts about how they feel and are acting about AI in healthcare. This article about AI is not written by an AI, but simply to provide a snapshot of the current conditions of the industry and market forces for AI implementations.? It was a well organized conference and I really enjoyed learning from other participants.?

Dr. Harvey Castro , MD, MBA, CEO of Medical Intelligence Ops delivered the opening keynote on GenAI with specific use cases where GenAI like ChatGPT is making an inroad in clinical decision support.? A few good use cases he pointed out are: 1) Overdose diagnosis in Emergency Medicine; 2) Personalized patient communication with empathy and accuracy; 3) Intelligent hospital discharge.? Dr. Castro mentioned that in 2023 most health systems in the US are undertaking GenAI related projects and expects that in 2024, there will be more deployment activities.? Dr. Castro emphasized the transformative potential of GenAI, stating, 'By integrating advanced AI into healthcare, we're not just changing medical practices; we're enhancing the very nature of patient care, making it more personalized, efficient, and above all, human-centered.'

I interacted briefly with Dr. Shreya Shah , MD, Medical Informatics Director and Dr. Christopher Sharp , MD, CMIO at Stanford Health Care .? Both spoke on a panel titled- 'How Stanford Health Transformed Patient Care By Combining Compassion With Innovation.’? They are very optimistic about the role of AI in care delivery and currently use AI to communicate with patients incorporating empathy and sensitivity.

I ran into Dr. Sujit Dey , PhD, the Founder and CEO of CIPRA.AI .? Sujit and I worked at two different companies in the past.? CIPRA collects structured and unstructured data using wearables and IoTs (eg iPhone, Fitbit, Glucose Monitor) from Hypertension and Diabetes patients.? It also collects lifestyle data such as diet including sleep data.? Using AI and data, CIPRA helps both caregivers and patients with precise, personalized, and proactive decision support plans, as well as aiding the prevention strategies and scenarios.? The solution significantly improves the patient engagement, clinical workflow, data collection, and reduces the cost of care in the long run including reversal.? CIPRA is deployed at multiple hospitals and clinics globally and will have an official launch in #India in January, 2024.?

During breakfast, we discussed the current status/role of AI in clinical care delivery.??Dr. Aaron Jacobs , MD, the Executive Director of Population Health at UNM Health in #Albuquerque, #NewMexico, thinks that AI already surpassed the cognitive knowledge of a human doctor, and there is no limit to what AI can do in the medical field.? He added ‘how we contain AI so it doesn’t harm humans.’? My answer was, we need to deploy strict policing, meaning federal regulatory interventions.? As we bring new technologies to the market, the regulatory piece of the technology is mostly behind the curve, and AI is not an exception, the explosive intake of AI in the marketplaces is ahead of the regulatory framework, no doubt.? As we have seen on television, even Mr. Sam Altman , CEO of OpenAI is asking the regulatory bodies to intervene.? While we continued our talk with the human brain and AI, Dr. Reid Conant, MD, CMIO, Microsoft Health wanted to know the current limitations of AI in terms of massive data crunching and simulation. I mentioned that one of the major AI bottlenecks in Health and Life Sciences now is the GPU/TPU limitations.? Until quantum processors are available, the full utilization of AI will not be realized.? In the meantime, companies like NVIDIA are developing subsystems such as the NVIDIA DGX Platform, a super computer targeted for boosting technologies like GenAI.? In a recent interview with Andrew Sorkin at the NYT DealBook conference, Mr. Jensen Huang , MS, CEO and Founder, NVIDIA said the DGX systems is the first AI supercomputer in the industry that will accelerate drug discovery.

There was a consistent underlying theme of Ethics, Equity, and Human-Centered AI in a few sessions.? What does it mean?? For example, the Stanford University HAI (Human Centered AI) Institute describes its mission as- ‘The mission of HAI is to advance AI research, education, policy and practice to improve the human condition.’? Fair enough, how do we bring that conviction to human-centric care models, without biases, corrupted historical data, structural inequality but encompassing SDOH, fairness, privacy/security/safety, compliance, equitable access, and more. Dr. Forrest Pascal , PhD, Principal, AI/ML Model Governance at Kaiser Permanente presented a ‘Best-in-class Health Equity Governance Model’.? While he discussed Kaiser’s undertaking on creating an equitable care delivery model powered by AI, he captivated the audience highlighting the risk factors for not having a solid AI/ML Governance model and scenarios where AI can exploit the model.? He added ‘The question we should be asking is- how do we put guardrails on the humans that are in the loop?? Just adding an human-in-the-loop recursively adds myopic or narrow perspective without concern for broader implications. We will never get rid of bias 100 percent, but we can put 100 percent of our perspiration into assuring that harmful bias has been mitigated.’

I had some impromptu discussions with the delegation from the Ministry of Health, #Singapore.? Being one of the best Health Systems in the world, reaching nearly 50/50 public/private coverage for a multi-ethnic population of 6 Million people, the Ministry is focused on regulating AI deployment with human guardrails.? Mr. Raymond Chua , MBA, the Deputy Director General of Health was on a panel- ‘Healthcare Innovation, Regulation, And Policy In The Age Of AI: A Global View.’? The Ministry is promoting the enhancement of care with AI utilizing human controls.

During a lunch roundtable discussion, we spoke about food and lifestyle in medicine.? I cited Steve Jobs, Apple Co-founder, who once said, ‘eat your food as medicine otherwise you have to eat medicine as your food.’? A well known Cardiologist in #Florida, Dr. Pradip Jamnadas , MD talks about food, fasting and microbiome chemistry for preventing/reversing autoimmune diseases like diabetes and hypertension.? I mentioned Ayurveda, an Indian Yogic Science which is a several thousand year old medical practice that promotes food as medicine and suggests a diet and lifestyle for maintaining and reversing adverse health conditions.? Ayurveda also emphasizes fasting for better health management.? Dr. Jennie Luna , MD, an Endocrinologist and an Obesity Medicine Specialist at Scripps Health , was a participant at the conference and said she practices fasting and discusses fasting with her patients. Being grown up in the Indian Subcontinent, I saw yogis embracing Ayurveda and fasting who lived 70+ years without modern medicine.? Other participants were curious about the science behind it, and how AI is deciphering the phenomena.? I mentioned my friend, a serial entrepreneur, Mr. Naveen Jain , MBA, CEO and Founder of Viome raised close to $200M to hack microbiomes with AI for disease prevention and reversal. 'Leveraging our proprietary mRNA sequencing tool in combination with the best-in-class AI, Viome is able to zero in on the root causes of health symptoms that act as a precursor for chronic disease,' said Naveen. 'By offering individuals comprehensive health scores as well as personalized food and supplement recommendations, we can help users live long and healthy lives.'

About the Author:

Mr. Pronoy Saha , BSEE/MBA founded Health Technology Forum , (HTF) to address the unmet healthcare needs of underserved and underrepresented populations (5B+ people are medically underserved) utilizing technologies like AI/ML and Big data.? Big data lakes and oceans are being filled with health data; a whopping 30% of data created today in the USA is health data.? We promoted innovation, innovation, innovation in our first decade of operations, currently we are promoting implementation, implementation, implementation.? As digital health technologies are becoming more reliable, scalable, and deployable we continue with our journey of promoting technologies into the care delivery framework for closing the care gap for the billions.? HTF is currently engaged with the Neighborhood Health Enablement (HTF NHE) initiatives promoting Health Kiosks for augmenting and scaling primary, preventative, routine, and wellness care delivery in the neighborhood.