Latest News

Building A Credit Score For Health: Sapiens Data Science’s Approach To Precision Medicine

By Benjamin Ross

March 20, 2019 | SAN FRANCISCO—We are quickly moving toward a world where people can be in control of their mortality, Bradley Perkins, Co-Founder and CEO of Sapiens Data Science, said in his keynote presentation during the recent Molecular Medicine Tri-Conference in San Francisco.

Perkins’s company offers a digital platform that provides a personalized 10-year mortality risk assessment based on twelve health risk factors. The idea was to give a kind of credit score to an individual’s overall health, an apt approach when considering one of Sapiens’ co-founders was once the CEO of credit score company, FICO.

“We transform this 10-year mortality risk into an age and sex-adjusted score,” Perkins said. The concept of a score is an easily consumable way to present an individual’s performance against these twelve risk factors, with a broken-down report on each risk available for the user as well.

Perkins says the platform is an in silico approach to precision medicine that was impossible a decade ago. “Publicly available data has changed dramatically,” Perkins said. “We have control over a lot more data... This provides new opportunities to build new platforms that are controlled and operated by individuals.”

The Sapiens health platform leverages data taken from the Bill and Melinda Gates Foundation’s global surveillance infrastructure, which improves the understanding of the causes of mortality and disability, for pretest probabilities of 10-year mortality risk.

The vision is not to simply give out scores, but to encourage individuals to take control of their health. Each of the twelve risk factors the platform focuses on are within the realm of an individual’s control, said Perkins. “And if these things alone can do a good job predicting your mortality risk without consideration of previous diagnoses or family history, that’s an empowering observation for people.”

The risk factors, determined by the Institute for Health Metrics and Evaluation (IHME), are categorized into metabolic, behavioral, and dietary risks.

Sapiens used data from a cohort of over 5,000 people that were matched with death certificate data and referenced the cause of death back to the risk factor categories. The results, Perkins said, revealed the prevalence of these risks when predicting the death and rate of mortality within the cohort. “[The prevalence] ranges from a high level of exposure to risk associated with low physical activity (87% of the American population do not meet the recommended clinical guidelines for physical activity), to a low rate of smoking tobacco at only 22.5%,” he said. “And it’s actually worse than that because these things don’t run independently, they run in big groups.” The majority of people are associated with at least six or seven of these predictive risks, Perkins said. In fact, all of the people in the cohort Sapiens analyzed were associated with at least one of the twelve risk factors.

Perkins said the specificity and sensitivity of using the twelve risk factors combined has a high sensitivity to 10-year mortality prediction on par with similar tests used in medicine today.

Sapiens was able to apply this data into their data science platform, which aggregates and analyzes data on behalf of individuals in real time. The data, Perkins said, will come primarily from an individual’s electronic medical record (EMR), which is downloaded onto their Apple HealthKit.

“Everything about this we’ve made medical grade,” said Perkins. “We want people to feel comfortable that when they take this information to their physician, that it will be highly recognizable and will be crossed with the latest clinical guidelines.”