UK-based roadtohealth, which makes a health improvement app called Quealth, has announced it has completed a clinical validation study of the dementia risk assessment tool within the app.

Quealth, which is free and available on both iOS and Android, guides users through lifestyle modification behaviors to reduce the risk of five major non-communicable but sometimes preventable diseases: type 2 diabetes, cardiovascular disease, COPD, dementia and six forms of cancer.

While not a diagnostic tool, the app uses Quealth’s proprietary algorithm to calculate risk based on the user’s answers through a series of questions covering family medical history, biometrics and lifestyle. Each user gets a “Quealth Score,” ranging from 1-100, and then provides educational content and coaching based on research into the science of behavioral change.

“Quealth’s risk algorithm has been optimized and validated to an excellent level of predictive accuracy for dementia,” Paul Nash, Quealth’s head of clinical governance, said in a statement. “It has a 72 percent chance of distinguishing individuals at risk of developing dementia from those who are not at risk. This is comparable with leading international risk scores.”

To validate the dementia risk assessment, researchers at the University of Nottingham’s School of Medicine compared the performance of Quealth with other clinically validated risk assessments, including one that investigates the association between cardiovascular disease, one looking at the risk factor from diabetes, and another from Australian National University.

The company has been steadily working on validating all of their disease risk predictors in an ongoing program at the University of Nottingham, and said they have successfully validated all but the one for cancer, but they expect to finish that study soon. 

“We correlate against the world’s databases, which have validated risk assessments based on longitudinal studies, and that puts us in a strong position,” roadtohealth CEO Alistair Wickens told MobiHealthNews in an interview. “This is a big one for us, because dementia is very difficult to validate.”

Wickens’s goal is to so firmly establish the validity of Quealth’s risk predictor that it becomes a known metric in its own right.

“I do generally want it to be the case where it is common to ask ‘what’s your Quealth score?” he said.

The company is actively pursuing partnerships with large insurers for incentivized employee wellness programs, and most recently partnered with global insurance company Reinsurance Group of America. They also work UK insurer Aviva and several large companies including Samsung, Lenovo and Validic.

Wickens explained how Quealth uses AI to motivate behavior change through lifestyle coaching and education. With machine learning, the app can personalize how it delivers coaching to each person over time in a way that an established program cannot, he said. Additionally, it can address thousands of people at a time.

“If people are serious about trying to tackle the science behind what motivates people to change, there is really no system of humans big enough to cope with it, which is why I think you need to go down the route of algorithmic health coaching,” Wickens said. “We understand that, what makes it different and able to personalize health is to get very granular data based on AI. How do I understand the nuances of how to speak to you versus someone else in your office?"

While lifestyle modifications to lower the risk of developing dementia are still not as clear as say, losing weight to prevent type 2 diabetes, Wickens said a primary aim of getting people to use the tool is to carry out more longitudinal studies with larger numbers of people.

“We have some ongoing and are starting to draw conclusions, but it’s still hard to really quantify and draw exact correlations at this time,” he said. “This biggest thing is our ability to manipulate data and intelligence in a way we weren’t able to before, that a traditional doctor didn’t have access to or didn’t know. And we can draw conclusions from clinical or medical history because AI allows us to reach beyond that existing, current patient in front of us to thousands and thousands of others and look at the best treatments. The fundamental difference of data is the big thing – it’s more than data; it’s the ability and speed with which you can interpret it and transfer it to others.”