The healthcare technology startup Medopad, which developed a tool for clinicians to track the vital signs, has used large populations from China to improve the predictive analytics of their artificial intelligence (AI).
Training AI involves machine learning of factors taken from samples of patients, but large amounts of participants are difficult to find in most countries. In response, Dan Vahdat, CEO of Medopad, traveled to China where large populations are easier to access.
“We started going there nearly four years ago,” said Vahdat. “We wanted to get there simply because of the scale of numbers. If you have scale you have access to an asset that nobody has. It gives you an edge against other companies in terms of the predictions you can do.”
Medopad can collect data through wearable fitness trackers or glucose monitors. The information it collects is then analyzed to predict and suggest diagnoses to physicians. However, the AI must be trained with machine learning on actual samples of patients to gain its knowledge of predictive analytics.
“Depending what kind of studies we want to run we can target 10,000 or 100,000 [people],” Vahdat adds. “I don’t think anyone can access 100,000 [people] outside of China.”