Health AI 35: What to know, who to watch

• AI-enabled hardware and medical devices - More and more consumers are getting comfortable using medical devices at home. Large language models (LLMs) can combine the biometric data collected by those devices with the latest medical knowledge to help predict disease and other health events in real-time. For example, IVP portfolio company Hello Heart combines a blood pressure monitor with AI to give patients a personalized plan for improving heart health.
• Academic search engines powered by AI - The internet is full of useful medical information, but most of it is far too dense for the average person to understand. Companies like Consensus and Elicit are democratizing medical knowledge by allowing consumers to ask medical questions in natural language and then responding with simple, clear answers linked to peer-reviewed publications.
• AI-native agents and coaches - AI chatbots are a new way for people to get personalized, science-based insights. In September 2023, IVP portfolio company WHOOP partnered with OpenAI to launch WHOOP Coach, which offers consumers tailored answers to health questions based on their unique biometric data and the latest medical research. Companies like Thrive AI, Rosebud and Slingshot AI are taking an AI first approach to mental health.

Providers - Reducing the administrative burden on healthcare workers

Today, physicians spend more than 40% of their time on manual administrative tasks — time they can’t spend providing patient care. Generative AI can help reduce that burden, but only if it focuses on the right workflow — one that’s ready for automation, where providers are willing to try something new and where a point solution could eventually become a platform.

• Workflow automation - To save physicians time, IVP portfolio company Abridge is automating patient notes and streamlining workflows that have traditionally been done by hand. Other workflows ripe for AI disruption include patient engagement and clinical decision making. Harnessing providers with comprehensive medical information at the point of care reduces research time; Hims & Hers built MedMatch to aggregate data and the knowledge of the medical community, helping providers deliver personalized care in an intelligent and informed way.
• Medical imaging - Companies like Rad AI are revolutionizing how radiologists operate by automating the reporting process via dictation. Others can analyze medical images with greater accuracy and speed than humans, and then communicate insights and flag outliers to providers.
• Coding - Most hospital systems are looking for two things: growth and efficiency. Medical coders analyze physician notes and lab results to assign correct diagnosis and procedure codes for insurance billing, yet coding errors cause U.S. hospitals to lose billions every year. SmarterDx and CodaMetrix are using AI to ensure that relevant information and diagnoses are accounted for, helping close a hospital system’s revenue integrity gap.
• Automating pharmacy operations - The 340B Drug Pricing Program helps hospitals serving vulnerable communities manage rising prescription drug costs. However, the program’s requirements are complex and mistakes can lead to financial penalties for physicians and pharmacies, which companies like Plenful guard against. Additionally, companies like Latent make the approval process easier by simplifying prior authorization and reducing prescription turnaround times.

Pharmaceutical companies - Making drug discovery faster and cheaper

Before pharmaceutical companies can test a new drug, they have to analyze enormous amounts of data to determine what might work. Generative AI can shorten development timelines by up to 75% and cut R&D spending in half — but only if companies focus on tasks and data that are ready to be structured and synthesized.

• New drug development - Generative AI can make it easier for researchers to not only optimize drug candidates to find the right protein structure to maximize efficacy but also to design the correct drugs from scratch. This allows pharma companies to have shorter development cycles and get more effective candidates to clinical trials faster. Companies like Cradle Bio are speeding up development for protein engineering, which has the potential to transform the treatment of everything from the common flu to cancer.
• Enabling AI in the R&D process - To accelerate AI development, companies are bridging the gap between AI researchers and leading medical institutions. By overseeing FDA clearance processes and managing commercial sales, Bunkerhill Health shortens the timeline from research to practical application in healthcare.

Payers - Reducing the burden of regulation and costs

Every U.S. health plan has to comply with different rules and codes for every provider it works with—regulations that can change whenever Congress passes a new law. Generative AI can track these changes and help payers adjust in real-time. AI can also help payers in healthcare by automating and optimizing tasks like claims processing, fraud detection, and risk assessment.

The best solutions will not only offer seamless integration into existing workflows, but also prioritize data accuracy, scalability and the ability to adapt to future regulatory changes.

• Prior authorization workflows - Prior authorization requires physicians to get insurance approval before they can prescribe a specific medication or deliver a certain type of care. It’s a labor-intensive and error-prone process with long lead times due to doctor and nurse shortages. Companies like Tandem Health and Cohere Health are speeding up the process, saving payers time and money.
• Enabling value-based care - More health plans are looking for providers who can prove they’re actually making patients healthier — but this value-based care can lead to complex risk contracts. Companies like Garner Health are helping health plans save money by incentivizing members to visit high-quality doctors and offering care navigation for self-funded employer plans.