logo

Using AI-Powered Chat, Mental Health Support is Accessible

Posted by Marbenz Antonio on August 3, 2022

Creating a mental health resource chatbot with a deep neural network. | by Alisha Arora | Medium

Theresa Nguyen, chief program officer at Mental Health America, responded, “What isn’t a barrier? “when asked what challenges stand in the way of people discovering proper mental health options. “For someone who is having mental health issues, these barriers could include things like shame, access, and even the cost of solutions.

Nguyen was once someone who need those resources for herself before she became an advocate for easily available mental health services. She battled depression from an early age, but she was unable to locate the resources she needed to understand what was “normal”. She claimed that in the absence of solutions, she felt guilty and believed it was simpler to conceal her difficulties than to confide in someone.

She didn’t want to bother her family with her mental health issues as her parents were working, immigrant Vietnamese parents. She confided in a close friend initially, then the National Suicide Prevention Line, but she had trouble getting help or finding a solution.

Enter Mental Health America 

Theresa began working with Mental Health America in Orange County, California, in 2005 as a rehab specialist. Nguyen currently works with Mental Health America (MHA), a community-based nonprofit organization that supports people with mental illness and promotes general mental health, and has more than 15 years of expertise in the field. The mission of MHA is to advance mental health as an essential component of general wellness.

I loved the fighting spirit of the organization,” Nguyen said. ““The people that I spoke to in the MHA network talked about the importance of self-determination, that my voice is the most important voice in this journey.”

Theresa established the MHA’s online screening program, which allowed website visitors to take a mental health assessment to learn more about a potential mental health condition and locate resources for the best course of action. The number of screenings provided by the program increased from a few thousand per week to 1.5 million in 2020 and 5 million in 2021.

The texting service

Her most recent project is an adaptable texting platform for mental health intervention powered by AI that she is working on with Northwestern University and the University of Toronto. In the previous year, 79% of those who screened positive for serious depression using MHA’s online assessment did not wish to use medication or psychotherapy. The development team for the texting application driven by AI was led by these worrying figures. The program tries to provide assistance and support with few access barriers, allowing anyone to receive assistance at their comfort level.

Even if someone today wanted to get help, care access and cost issues can usually keep them from doing so. There may not be many places where you can find mental health professionals, and not everyone has access to the technology needed to look for assistance. With the messaging platform, users can experiment with different approaches to treating their mental health concerns without worrying about expense or where to get care. They also don’t have to plunge into engaging with another person right away.

“It’s immediate, you just type some things. If you hate it, you can leave it. If you like it, you stay engaged and it’s low cost. You don’t have to navigate the system of care or insurance to get immediate support,” explains Nguyen.

Machine learning assistance

Since most users stop using mental health apps after two weeks, according to a recent study, the texting platform presents a more promising option. The platform works by sending out pushed text messages all day long, allowing users to participate whenever they’re ready. The AI bot can be communicated without data or an internet connection because it uses SMS. A text discussion is started after registering on the MHA website, and the user responds to questions. The platform then makes use of machine learning to target particular engagement interests and discover more about the user’s activities, such as the subjects they are interested in and the best informational format. To assist the bot to be more personalized, users can also provide optional information about their demographics, experiences, and interests. Mental illness can have radically unpredictable ebbs and flows. Nguyen expects the platform would make it easier to plan care routines around these severe changes.

“I might not open an app or I’m not going to call my friend, but if I’m lying in bed in the middle of the night, and I’ve told this platform, ‘Hey, I really like getting texts at 10 PM.’ And it sends me texts at a time when I know that I’ll probably need it. That’s the kind of experience that I can plan for when I’m well, but when I’m not well, it’ll just come because I need it.” 

Theresa thinks the bot can lessen the shame and condemnation that usually follow the decision to seek assistance—emotions that she feared as a child. She continues by saying that the development of digital technology is particularly significant for offering tools to children and teenagers who might have to deal with more concerns about what is supposed to seem normal.

Making your community a success

Untreated disease and psychosis can continue for years at a time. Nguyen hopes for the service to bridge the gap between perceiving the need for care and obtaining it.

“The impact that this texting program could have on the population at large is incalculable,” says an anonymous user of the program. “I know I would benefit from it and I am sure that thousands of others would, too. A lot of young adults and teenagers suffer from mental health issues because, well, frankly, the world sucks, and anything that is well-researched and proven to be beneficial, things such as this texting program will definitely impact my peers for the better.”

It can be difficult to decide to seek treatment. It can be challenging to find a provider with who you can identify.

“It’s hard to find a provider period, right? Let alone someone you trust,” Theresa said.

Users will be able to accept a privacy policy when signing up for the texting platform, and all identifiable information is encrypted to protect against tampering. Users can be confident that their privacy is secured because no personally identifiable information will be saved. It can be the first time someone seeking assistance has disclosed important personal information. Others might be talking about dealing with a mental condition for the first time. Before their initial visit, users are supposed to have access to the platform’s knowledge and abilities, allowing them to move on to further assistance when they are at ease. Theresa refers to it as positioning users for success.

The constant availability of the messaging platform is comforting, particularly at a time when healthcare providers are suffering from burnout as a result of the rising demand for their services. The asynchronous bot is useful since it won’t become exhausted, according to Nguyen. To add a sense of community to the platform, the team is experimenting with crowdsourcing user-generated advice and encouraging words.

The website, according to interviewees who used it, helped individuals feel validated, try out new talents and activities for their mental health, and connect with others. “There are so many people with so many different ideals and perspectives on what solutions we could come up with, an anonymous user said on community development. And considering the number of attendees, I liked that. Although we are all unique, we also have a lot of the same problems. It is quite helpful.”

Next steps 

Through a grant from the AI for Accessibility initiative, Microsoft provided the project’s early financial and technical support. The platform will be made available to 10,000 people as the next step in the study process. The team’s ultimate objective is to make the texting platform available to millions of visitors to the MHA screening website each year.

Since she was a young girl who struggled in silence with her mental health, Nguyen has gone a long way. She is aware of her family’s diagnosed and undiagnosed mental illness as an adult because it was just not mentioned. Her struggle with mental health is ongoing, but she feels more prepared to receive treatment and is appreciative of the readily available technology that enables her to do so. She believes the platform will provide users the knowledge, skills, and confidence to advocate for themselves and their voice as the patient, and she wants people to be able to readily get care now at their speed.

“I think what’s cool is that we’re giving power back to people and allow them to seek help at a level they are comfortable with,” explains Nguyen.

 


Verified by MonsterInsights