By Noel J. Guillama-Alvarez
There’s much to share about AI in healthcare as we pursue our mission of empowering consumers through AI. Our goal is to make Electronic Health Records (EHRs) more accessible, helping consumers take control of their health data and advance scientific research. In this pursuit, we closely follow developments in healthcare, AI, and even cryptocurrency, as they all play a part in shaping our future. We have several blogs in the pipeline, but new, fascinating updates keep pushing them back. One thing we are confident about is that real change in healthcare will come from the consumer side. It’s the only way forward.
In a blog last December, we highlighted a statement from a respected researcher at the 2024 Radiological Society of North America Convention. He stated that “800,000 Americans die or are seriously injured by misdiagnosis,” and believed that AI could help address this problem. (See the link below.) That number was staggering to me, and I’m still trying to fully grasp its implications.
We regularly see reports on AI in clinical applications, most of which offer cautionary advice. Beyond the three areas we’ve emphasized—Radiology, Pathology, and Genetics—AI is still a work in progress in clinical settings. I recently read a report that discussed AI in clinical applications, noting that doctors (and society at large) need to “define acceptable error margins, address hallucinations, and mitigate variability in performance.”
Exploring Healthcare Research and AI’s Role
I recently had the chance to explore research papers in depth, and since we plan to add substantial research content to our platform, this is highly relevant. The sheer volume of research in healthcare, wellness, diagnostics, pharmaceuticals, and genomics is mind-boggling. I had been estimating about 20,000 papers per year (or about 60 per day), but many people I shared this with thought that number seemed too high. So, I dug deeper and came across a paper titled “The Landscape of Biomedical Research” that left my brain reeling. Here’s the abstract:
The number of publications in biomedicine and life sciences has rapidly grown over the last few decades, with over 1.5 million papers published annually. This massive growth makes it difficult to keep track of new scientific works and to have an overview of the field’s evolution. We present a 2D atlas of the entire corpus of biomedical literature, offering a unique perspective of life sciences research. We base our atlas on the abstracts of 21 million English articles from the PubMed database, using PubMedBERT and t-SNE to embed the data into 2D. We use our atlas to study topics like the emergence of COVID-19 literature, the rise of machine learning, and gender imbalances in academic authorship. We also provide an interactive web version of the atlas for exploration and further research insights.
In the graphic from this study, we can see the key areas of research like cancer, physiology, immunology, biochemistry, and more. Interestingly, one of the graphics shows the rapid rise of COVID-19 research (highlighted in yellow) while another visualizes research involving machine learning (shown in orange). Topics like cancer biomarkers, tumor imaging, and healthcare data are key subjects in the analyzed papers. A specific mention of Convolutional Neural Networks (CNN) stood out as a method used in many of these studies.
The point I’m making is that there’s an overwhelming amount of information out there. How much of it is actually relevant to a doctor at any given moment is a significant challenge. With so much data, AI can easily take a doctor and patient on a wild goose chase. We continue to ask: In most cases, is AI truly helpful for doctors? Is it worth their time? Is it more accurate? And, perhaps most importantly, is the time and money investment worth it in the most complex cases?
The Doctor’s Dilemma: AI’s Information Overload
Just think about how much new information is available to every doctor in their specialty each month. I’ve seen some general medical research published in a handful of journals, and it’s already overwhelming—let alone when factoring in AI applications. The desire for AI in clinical work is real, but as we’ve discussed, the question remains: What failure rate are we, as a society, willing to accept?
FDA’s Concerns and AI Regulation
The U.S. Food and Drug Administration (FDA) has expressed concerns over the rapid growth of AI in healthcare. As of December 2024, over 1,000 FDA-approved products feature AI components, with many more under review, including those related to AI-driven drug and biological product development. It’s still unclear how the FDA will respond under new leadership and changing mandates, but it’s safe to assume that submissions will only increase as large language models (LLMs) continue to evolve. Additionally, Medicare has already approved a few AI-driven products for reimbursement.
Excluding LLMs, over $20 billion has been invested in AI-focused startups in the past 18 months, many developing AI agents. From my perspective, there are too many players, too little traction, and a lack of a clear value proposition. I know doctors personally, and they will take a risk on free tools. But sometimes, “free” can end up being “too expensive.” Meanwhile, government bodies, universities, large corporations (both payors and healthcare providers), and hospitals continue to fund AI research. However, many of my colleagues are frustrated by the “mandates” from above, as they’ve been burned by promises of relief and solutions before—specifically with Electronic Health Records (EHRs).
AI and the Consumer: Empowering the Patient
AI is arguably the most groundbreaking technology of this century, capable of processing vast amounts of data that humans cannot. I was recently reflecting on the leap from my old mechanical calculator (used in high school) to the HP 41 CV I used in college and work. Now, we have AI tools like Alexa and ChatGPT that can answer complex questions almost instantly. AI is set to impact nearly every knowledge-based field, healthcare included.
I checked the latest data on ChatGPT-3, which operates with 175 billion parameters. Healthcare, with its hundreds of millions of medical records and billions of transactions annually, is the most data-rich industry in the world. In the U.S. alone, the total size of healthcare data is about 3,000 exabytes, segmented across 500 EHRs. If you enter data into an AI model, there’s a chance it could be retained indefinitely.
What we need is a system that allows consumers to aggregate their own health data, have it reviewed by healthcare-specific AI, receive feedback, and—when needed—use the information to better communicate with healthcare providers or challenge insurance decisions. Empowering the consumer with knowledge and data is the only way healthcare will ever truly change.
About HealthScoreAI ™
Healthcare is at a tipping point, and HealthScoreAI is positioning to revolutionize the industry by giving consumers control over their health data and unlocking its immense value. U.S. healthcare annual spending has exceeded $5 trillion with little improvement in outcomes. Despite advances, technology has failed to reduce costs or improve care. Meanwhile, 3,000 exabytes of consumer health data remain trapped in 500 fragmented USA EHR systems, leaving consumers and doctors without a complete picture of care.
HealthScoreAI seeks to provide a unique solution, acting as a data surrogate for consumers and offering an unbiased holistic view of their health. By monetizing de-identified data, HealthScoreAI seeks to share revenue with consumers, potentially creating a new $100 billion market value opportunity. With near-universal EHR adoption in the USA, and advances in technology, now is the perfect time to capitalize on the data available, practical use of AI and the empowering of consumers, in particular the 13,000 tech savvy baby boomers turning 65 every single day and entering the Medicare system for the first time. Our team, with deep healthcare and tech expertise, holds U.S. patents and a proven track record of scaling companies and leading them to IPO.
Noel J. Guillama-Alvarez
https://www.linkedin.com/in/nguillama/
+1-561-904-9477, Ext 355
https://www.biorxiv.org/content/10.1101/2023.04.10.536208v1.full?utm_source=chatgpt.com
https://oxiohealth.io/healthcare-ai-radiology-and-ia-800000-deaths-and-disabled/
https://systematicreviewsjournal.biomedcentral.com/articles/10.1186/s13643-018-0819-1
https://en.wikipedia.org/wiki/Convolutional_neural_network
https://ourworldindata.org/grapher/scientific-publications-per-million