How Touchkin keeps tabs on your health by tracking your phone usage

Touchkin is a platform that uses AI and machine learning to provide personalised care solutions.

For Jo Aggarwal and Ramakant Vempati, one of the biggest concerns they had while working abroad was the health of their family living back in India. The husband and wife duo were working in high-flying corporate jobs with organsations like Pearson Learning Solutions and Goldman Sachs International in the UK.

Living away from family, the couple realised that remote care giving was a huge global challenge, not just in India, but abroad as well, especially when it came to understanding the emotional and mental well-being of the families living in their home countries. So, in 2012, the two of them moved back to India to take care of their family.

They soon discovered that there were many others who had family care responsibilities in India, but could not move back home, for any number of reasons. Looking closely at this problem, one that they had faced firsthand, the duo felt that Big Data and mobile technology could be harnessed to create a better way to care for parents and loved ones while being away.

“We wanted a world where you could Google how your loved ones are, and be there for them virtually and practically,” says Jo. That was when they came up with the idea of Touchkin.

The initial product offering

Touchkin is an emotionally intelligent platform that uses AI and machine learning to provide personalised care solutions via a happiness buddy, which is currently being offered as a B2C product.

Touchkin was initially intended to help families take care of each other, by creating a platform that helped identify issues early on by establishing the context of a user through the study of their phone usage patterns.

They, however, realised that there was a much stronger need to do this in a formal healthcare setting, as it would be helpful in identifying risky behaviour. They also found studies from prominent institutions like Harvard and MIT that had pointed to a link between mobile phone usage and depression, as well as other behavioural health conditions.

“We set about testing this in the real world, in semi-urban India. Last year, we ran a pilot in Aurangabad, Maharashtra, with Dr. Archana Sarda, who is the founder of Udaan, an NGO that supports over 500 underprivileged people with Type 1 diabetes,” says Jo.

Running tests

Over a period of three months, 50 Type 1 and Type 2 patients were given the Touchkin solution, and the doctor was provided alerts to follow up in case of activity or mood alerts.

They found that close to 60 percent of the patients were suffering from moderate to severe depression, which doctors typically never monitor, or even ask about.

“Research shows that people with diabetes who also have depression find it harder to manage their condition, take medicines on time, and stick to their diet or exercise plan. These then cause serious complications over time,” explains Jo.

Unfortunately, such behavioural conditions are very often missed–and even if they are diagnosed, few doctors would know what to do. The business case for catching these issues early was thus compelling.

After the pilot, they found that they were able to identify depression in diabetes patients at a higher level, and their mobile and chat-based interventions actually reduced the severity of symptoms by 50 percent over the period of the study.

Current offering

Patients felt supported and engaged, and reported reduced stress levels and recorded increased activity and adherence.

So, Jo and Ramakant shifted their focus to helping people improve their mental resilience in an accessible and non-threatening way, embedded in familiar everyday contexts that actually improve physical health.

“We are now implementing this through our products: Wysa, your ‘happiness buddy’ on the B2C side, and custom behavioural health solutions for enterprise clients and partners on the B2B side,” adds Jo.

Today, Touchkin’s platform personalises engagement to a user’s lifestyle and emotional state, creates buy-in, and helps them change behaviours. This may be aimed at feeling less anxious, or improving one’s activity, or even quitting smoking.

Solutions provided

The idea is to create a conversational, personalised experience focused on their goals. As a by-product, each person also has developed key skills for managing their mental health.

There are three components to this solution–a predictive engine, an AI chatbot, and a coach network. Coaches form a key part of the solution. The other two components, meanwhile, are designed to make the coach-led intervention proactive, cost-effective, and feasible at scale.

The predictive engine provides the ability to passively and non-intrusively collect data and create user context and, with proper training, proactively identify potential issues. It collects sensor data from the phone and interprets more than 50 variables in a resource-efficient manner.

With this user context, the engine is able to raise ‘alerts’ for a health issue. Currently, the engine has been trained to detect depression in people with diabetes, and it can also track anomalies in a user’s behaviour, like disturbed sleep or low activity.

The AI chatbot is a ‘virtual coach’ or a conversational interface for customer or patient engagement, and for the delivery of behavioural and lifestyle therapy. The chatbot can be used anonymously by users and in use cases where there are very high concerns about privacy. Jo explains,

“Here, our IP consists of a rules/simulation engine that is able to flexibly create conversations responding to alerts in any desired use case, domain-specific AI capability that recognises user sentiment and emotion especially relevant for behavioural health, and a content library for evidence-based techniques that is optimised for delivery through a chat format.”

Coaches are non-specialists, like diabetes educators, who are trained and certified to deliver behavioural support in a cost-effective manner. Coaches use proactive alerts to decide whom to reach out to, and what to check for based on user context and engagement.

The chatbot, in turn, gives them performance support and recommended responses to ensure consistency of approach. They also get a repository of techniques to refer users to using shortcuts and predefined programmes. The specialist consultations help them tackle really hard issues.

The stigma of mental health

However, one of the biggest problems is that people don’t report mental health problems due to ignorance, denial, or social stigma, thus making it even more difficult to get the right access to support.

The World Bank and WEF estimate that mental health issues are expected to cost the Indian economy one trillion dollars over the next eight years, six times the national expenditure on all healthcare for the last 19 years.

Today, with the need to create awareness for seeking help in such cases, several organisations like The White Swan Foundation, The Live Love Laugh Foundation, and HealthEminds, among many others, are striving to break the stigma associated with mental illnesses. There is also the online platform YourDost, which helps people speak out on the problems they face.

Growing space

“At Touchkin, we believe that mental health is like any other fitness–it is for everyone, not just patients. We want to take it out of a ‘specialist’ mode where people see themselves as mental health cases and see psychologists. Instead, we go to high-risk groups and use these techniques to help them achieve their goals,” says Jo.

In short, the focus is on linking the support to non-mental health goals. Unlike consumer technology, healthcare is less about the number of users and more about careful design and engagement and efficacy results.

There are two things that have worked for the team from the business point of view–firstly, they are using passive sensing rather than requiring user input. Jo adds that they can help improve patient motivation using best-practice evidence-based behavioural management techniques.

“All other technology interventions tend to serve motivated patients, while the problem lies with the unmotivated group. This is whom we want to serve,” says Jo.

The team claims to have over 20,000 users for the platform, with 100,000 conversations and counting. On the enterprise side, the team is also running pilots using their behavioural health platform with global use cases–healthcare companies, insurance firms, and pharma majors in India and in international markets.

Funding and future

“In these projects, we are helping use AI-enabled behavioural management techniques to drive physical health outcomes; each of these has a billion-dollar use case, so proving evidence of impact is most important over the next 12 months,” Jo says.

Touchkin was recently selected as one of six startups globally to be a part of the InsurTech accelerator set up by Zurich-headquartered Swiss Re, one of the world’s leading reinsurance companies.

Touchkin had raised an angel round of funding of $400,000 in January last year, and will be raising a pre-Series A round sometime this year. Speaking of future plans, Jo adds,

“Monetisation in the Indian market needs patience. This is common for the healthcare industry, and will take time. For now, we are running research, pilots, and impact in the Indian market, while focusing on the international market for revenue. At this stage, being based out of India is good for us, as we are in the early stages, but over time, serving international healthcare may need a stronger build-out of our international team.”

Website