By Jasmine Williams
Advances in technology from voice technology to telehealth services improve care delivery and make tasks more convenient and simpler for healthcare providers and patients alike.
The mental health chatbot is an emerging healthcare tech tool that is filling a significant gap in care. While it may sound like something out of a sci-fi movie, mental health chatbots are more science than fiction. They use artificial intelligence (AI) technology to interact with individuals looking for mental health help.
Many people need mental health assistance, but there’s a shortage of psychologists and other mental health professionals to help. While chatbots don’t replace the care a real person can give, they can help fill the care gap. COVID-19 has especially increased the demand for mental health services and the need for chatbots. Mental health professionals are also starting to refer patients to chatbots to provide additional support at a lower cost between live therapy sessions.
How AI can support mental health
How can a therapy chatbot assist someone looking for mental health support using artificial intelligence?
AI technology’s main function is to complete and perform tasks that typically need human intelligence. AI tech keeps advancing and is used in many different industries around the world — including healthcare.
With a mental health chatbot, various preset responses can reply based on the information the user gives. In development, the chatbot “learns” how to respond based on conversations with real people. It can respond accurately to thousands of different questions or statements with speech that sounds like a human — not a robot. It may not be the traditional patient-on-the-sofa interaction, but the chatbot can be useful.
Tips for creating mental health chatbot content
You can still provide patients with quality mental health care while using a chatbot. Although chatbots may seem impersonal, the questions and responses a patient receives are written by real, caring people. The bots just help facilitate the interaction when a human isn’t available. With that in mind, developing responses takes careful deliberation. In particular, you’ll want to make the experience as caring and personable as possible. Keep these ideas in mind when adding mental health chatbots to your services.
Ask questions
As in human conversation, asking questions can narrow down a generic response to a specific issue or problem. Suppose the therapy bot isn’t sure of the best response to someone entering a simple statement. In that case, it can continue to ask questions until it has something it can build on.
The app Woebot asks questions as soon as you enter the app, such as, “If you don’t mind me asking, what’s your gender identity?” When you answer “female,” it then lets you know it has a program for new moms and asks if you’ve given birth in the past three months.
Chatbots never stop learning — as they receive new responses or questions, they use it as training for future interactions.
Some people facing mental health struggles need a place to vent. Consider building in responses like, “What triggered your anxiety today?” or “It’s understandable that you feel that way — everyone has tough days.” These chat responses can help engage the user and prompt them to share their emotions and feelings while also making them feel understood.
Provide useful resources and tips
Individuals struggling with issues like anxiety and depression may need resources that can help. Share helpful articles on what they’re experiencing, tips to overcome daily anxieties, hotline numbers and more.
Keep the resources the chatbot provides simple. By providing easy-to-access resources, the user may feel less overwhelmed.
Find out how serious the user’s issues are
Ask users how serious their mental struggles are as soon as they log in. Woebot asks the question, “How often have you felt depressed over the past two weeks?” as soon as the chat begins. If someone’s symptoms are severe and have lasted longer than several days, the chatbot could immediately refer them to a mental health crisis hotline or 911.
An appropriate response to include in the database for this situation could be, “I want to help you and make sure you’re ok. Please call this hotline to speak directly with someone who can help you deal with the feelings you’re having right now and remind you you’re not alone.” Most chatbots aren’t yet able to handle mental health emergencies — that’s best left to humans.
Want to learn more about how to write mental health content? Check out this blog post.
How to write content that ‘speaks’ to the chatbot user
It may seem daunting to write content for a chatbot that’s supposed to help a wide range of people. Keeping these tips in mind can help make sure you’re providing content that’s helpful and compassionate.
Understand the user’s biggest struggle
Start by asking the user exactly what they’re feeling at that moment. It could be fear of the unknown, anxiety caused by workplace tension, or grief due to a family member’s loss. Create responses that correspond to specific issues or scenarios. This will make it easier to develop answers that feel more human and tailored to each individual’s issues.
You’ll also want to provide options the user can select. That will help the chatbot guide the conversation and respond appropriately.
A conversation could go like this:
Chatbot: Hi Amber, welcome back! How are you feeling today?
Amber: Not that great.
Chatbot: I’m sorry to hear you’re not doing well today. It’s not always easy to discuss, but can you tell me what has you feeling down?
(Provide click options of: I’m stressed. I’m anxious. I’m depressed. I’m sad.)
Chatbot: Anxiety can feel very overwhelming. Is your anxiety caused by an upcoming event, one of your relationships, or overthinking?
Amber: Overthinking.
Chatbot: That’s very tough. But, you’re not alone. Here are a couple of blogs that discuss anxiety caused by overthinking and how to deal with those anxious feelings. If those aren’t helpful, here’s a list of ten things you can try that might put a stop to those never-ending thoughts.
The chatbot can provide responses and content that match the user’s struggle more closely by asking questions that require specific answers. The therapy bot responses were empathetic to the user’s feelings and felt personalized by responding with the user’s name.
Ask if current events are impacting the user’s mental health
Current events, such as the riot at the U.S. Capitol on January 6, are known to increase mental distress. Create questions to find out if current events are adding to the user’s mental struggles. Certainly, news and social media coverage of the COVID-19 pandemic and racial tension are examples of issues that have triggered more people to seek mental health care for anxiety, depression and other conditions. Include relevant information and articles to help ease anxieties and fear about coping with current events.
Remember to use plain language
Mental health issues can affect any person regardless of race, religion, age, or income. Use words and phrases that are easy to understand, read and pronounce. For example, rather than “Would you like to communicate with someone on the hotline?” say, “Would you like to talk with someone on the hotline?” Choosing user-friendly words allows the chatbot to help more people.
The number of people in need of mental health assistance will continue to grow during the COVID-19 pandemic and beyond, outnumbering the human professionals available to help. Therapy bots may be an answer to your healthcare system’s need for a listening ear in the mental health space.
Mental health content support. If you’re looking to create mental health content for your patients, WriterGirl can help. We’re a reliable content partner with a team of writers that are 100% dedicated to healthcare. Drop us a line to learn more.
—
Editor’s note: Karla Webb contributed to this blog post.