Just like pharmaceutical drugs, AI-powered healthtech needs to be validated through testing – and one company is on a mission to provide the data needed for validation studies
Spotted: When a pharmaceutical company needs to test a new drug, it employs a Contract Research Organisation (CRO). These are companies that specialise in designing clinical trials. But in today’s medical landscape we have a new tool that also needs testing: artificial intelligence (AI). AI in the healthcare market is skyrocketing, growing at a rate of 167.1% from 2019 to 2021. And vast amounts of medical data are needed to train and test all of those algorithms. Obtaining high-quality data can be tricky, creating demand for a ‘CRO-for-AI’.
This is the niche that startup Gesund is hoping to fill. Gesund provides AI companies with curated medical data and tools to help them test and refine their algorithms to optimise patient outcomes, and secure regulatory clearance.
Gesund has built a ‘pipeline’ of data that comes from existing agreements with hospitals and other clinical sites. But the real novel value of Gesund’s technology is not so much the data itself, but how it is delivered.
Machine learning in other sectors mostly uses cloud technology supplied by third parties. But this model breaks down in the healthcare industry, as medical data comes with stringent regulatory requirements, and most hospitals don’t use the cloud, relying instead on secure networks protected by firewalls. Gesund’s platform is based on a technology stack that resides within hospital premises or a private cloud. Its ‘validation dashboard’ therefore acts as an easy-to-use intermediary between secure hospital networks where data is stored, and the AI companies who need that data for trials.
AI companies upload their models onto the secure Gesund platform. Gesund then runs appropriate data from its network of clinical partner sites through those models, returning validation insights to the AI company.
Although only founded in 2021, Gesund has built momentum with a recent $2 million (€1.8 million) seed funding round.
Written By: Matthew Hempstead