Mental health disorders are an issue for more than nine hundred million people worldwide. The impact on this group is tremendous but the impact on the Health Industry that is trying to serve them is overwhelming. That’s because the Health Industry relies on mental assessments to identify and assist people who are in need. Unfortunately, today’s diagnosis accuracy rates are very poor. In fact, according to a study conducted at the Johns Hopkins Bloomberg School of Public Health, 66 % of such diagnoses are incorrect. Misdiagnoses have huge financial costs. They also take a toll on patients’ health, and sometimes, their lives.
To change this we have to change the way mental health is assessed.
First, providers need to be able to do more assessments, earlier in the medical journey process. Right now, people are simply waiting too long before seeking help. Sometimes they wait because they don’t want to admit to a problem but often, they wait because mental health screening is not a common process even when visiting a clinician. However, it is known that the earlier the screening occurs, the easier and less costly it is to help someone. There is no need to wait for a crisis to happen, the process can become proactive but unfortunately the current standard of care is about being reactive to a crisis.
But doing assessments alone is not enough.
it will lead to increasing demand without addressing the supply side. Which is why we should always think about putting the clinicians (and anyone working in the mental health industry) in a position to succeed.This means simplifying mental data analysis, freeing up staff from routine check-ins and facilitating the triage process.
A further difficulty today is that any app, device, hardware manufacturer, or health provider will often build their own, siloed, diagnoses systems. 99% of the time, this system is not even a core component of their primary value proposition. For these companies, it adds needless complexity and expenses, not to mention the likelihood of data biais.
Focusing on assessments.
Okaya is focused on solving the mental wellbeing assessment and workflow problem, and building the best-of-breed wellbeing diagnosis API. For over 2 years, the team at Okaya has been exploring how to use the technologies available in today’s environment, coupled with an innovative use of machine learning algorithms. The core idea driving our development activity was to see if it was possible to recognize and quantify an individual’s mental wellbeing from a short video stream.
For example, in the field of fatigue, PERCLOS is one of the Gold Standards and is based around calculating blinks. But the challenge in a real-world environment is that the algorithm has to be able to measure variations due to pitch and yaw of the face, as well as light exposure.
Using voice bio data and contextual analysis Okaya’s ML algorithm can deliver an assessment that is over 80% accurate using the PERCLOS comparison. We are now starting IRB studies to further map our analysis to known standards such as PHQ-9 and Epworth. The best part is that we do not only consider visual inputs but also pay attention to voice, and external data, in doing our assessments. In future blogs we’ll dive into these areas more.
API First. Assessment-as-a-Service
Implementing Okaya in your current architecture is very simple: Your user checks in (either via our app or directly via API integration). A check-in is taking a 15 second video selfie. At that time, the user can opt-in to share other information with us. Then we use this longitudinal information along with different data sources to see how the patient is trending.
By incorporating Okaya into your processes, your company can go beyond getting the usual structural benefits of an API and open new strategic avenues.
You now have the power to not only better help your users by taking immediate actions based on the assessments; you can also be more strategic in how you develop new subscriptions, services, and offers to keep your customers longer on your platform.