Today we introduce our official paper detailing the methodology, performance, and evaluation of the Large Medical Model. We released it to pre-print on Arxiv our paper titled “Introducing the Large Medical Model: State of the art healthcare cost and risk prediction with transformers trained on patient event sequences”

Here’s the to the published abstract on arXiv and a PDF of GenHealth’s Large Medical Model paper.

Abstract

With U.S. healthcare spending approaching $5T (“NHE Fact Sheet” 2024), and 25% of it estimated to be wasteful (“Waste in the US the health care system: estimated costs and potential for savings”, n.d.), the need to better predict risk and optimal patient care is evermore important. This paper introduces the Large Medical Model (LMM), a generative pre-trained transformer (GPT) designed to guide and predict the broad facets of patient care and healthcare administration. The model is trained on medical event sequences from over 140M longitudinal patient claims records with a specialized vocabulary built from medical terminology systems and demonstrates a superior capability to forecast healthcare costs and identify potential risk factors. Through experimentation and validation, we showcase the LMM's proficiency in not only cost and risk predictions, but also in discerning intricate patterns within complex medical conditions and an ability to identify novel relationships in patient care. The LMM is able to improve both cost prediction by 14.1% over the best commercial models (Milliman, Cotiviti, etc.) and chronic conditions prediction by 1.9% over the best transformer models in research predicting a broad set of conditions. The LMM is a substantial advancement in healthcare analytics, offering the potential to significantly enhance risk assessment, cost management, and personalized medicine.

Our thoughts

This isn't just an incremental improvement; it’s a transformative shift in healthcare technology, setting a new benchmark for precision, efficiency, and personalized care.

For too long, traditional systems built on statistical analytics like those from Milliman, Cotiviti, and Hopkins have dominated healthcare cost and risk prediction. But with GenHealth.ai's LMM, you can now incorporate thousands of attributes from a patients history. Drawing on trillions of healthcare events and data from over 140 million patients, the Large Medical Model merges cutting-edge neural network transformers with a healthcare-native approach. Healthcare organizations now have an AI model that doesn't just recognize general patterns but delivers precision insights for every patient, every interaction, in real time.

Transforming Healthcare: From Administration to Patient Care

What sets the Large Medical Model apart is its versatility. Whether it's automating prior authorizations, streamlining population health analytics, or identifying fraud, waste, and abuse, GenHealth.ai’s LMM is designed to make healthcare work smarter.

The true potential lies in its ability to power personalized patient care at scale. In seconds, it can offer more accurate, actionable insights than ever before, helping plans, providers, and patients make decisions with better-than-human precision. Ultimately, this will lead to faster treatment, fewer errors, and a dramatically improved patient experience.

State of the Art Performance (SOTA)

The numbers speak for themselves:

• Unmatched accuracy in predicting healthcare costs and clinical risks, blowing past industry standards.
• Chronic condition prediction like never before – giving healthcare teams an edge in managing long-term patient outcomes. One model that can predict any condition or event, and usually do it better than purpose built models
• Immediate, actionable insights at an event level and over time, making it easier for insights to move from research to real-world application.

Building the Future of Healthcare – One AI at a Time

We are not stopping here. We have already creating new applications on top of the LMM, making its superior power accessible to health plans and providers alike. From our co-pilot for population health analytics called G-Mode to automation tools for prior authorizations and utilization management, GenHealth.ai is ensuring that everyone – from data analysts to frontline healthcare providers – can harness the full power of generative AI.

With the Large Medical Model, GenHealth.ai is setting a new standard for AI in healthcare. It’s not just about making processes more efficient; it’s about unlocking the future of personalized medicine. This is the beginning of a whole new era in health and longevity.