How chatbots help healthcare focus on patients and save costs

With Facebook and Skype allowing to build integrated chatbots, rising of connected devices and patients’ demands for better care, it’s only logical that the chatbot market grows. Last year, segments of these little helpers generated $36.5M in revenue. According to BIS Research report, this number will grow up to $498.1 in 2029. Let’s explore what value chatbots can provide for both sides of the main in-healthcare interactions.

Chatbots help patients learn what their symptoms mean

Everyone’s been in this situation. You cough all night, or woke up with a sore throat or noticed some new things on your skin — and you go to Doctor Google. Circumstances of time and space collide, you’re remembering something else from the other night, and lead yourself towards dire thoughts of some unspeakable diagnosis.

If you do not do it, you’re in 20% of the Internet population which prefers real health advice instead of search engine. Other people tend to diagnose themselves with cancer, half of the time, or some other illness, and, on the Internet, their first diagnosis is correct only in 34% of cases. And the scarier the diagnosis, the more we believe in it. Not exactly the brightest idea for our health.

So, healthcare providers start to act to beat this Internet diagnosis trend and use the power of connectivity to provide people with chatbots symptom-checkers.

For instance, Alexa, Google Home, and Cortana rely on Symtomate. A chatbot accepts descriptions of symptoms, then ask additional questions to figure out the most fitting diagnosis, and then offers to users next steps.

Basically, that’s how all symptom-checker chatbots work: users have to tell what they feel, elaborate a bit, and see what can do about their diagnosis. You might think that what Google does, but the two are different. Most of the chatbots that are used by healthcare organizations are based on data from real medical cases and medical knowledge from trusted sources and journals. Partnering with software companies, clinics and other healthcare facilities aim to stop patients from diagnosing themselves wrongly and trusting in pseudoscientific snake-oily things.

Surely, chatbots can’t replace patient-doctor communication and AI-powered symptom checkers are not clinical tools. That’s why the next steps, often, are intended to lead patients toward making tests or booking appointments. Symtomate and Ada do just that.

Chatbots help drive awareness and knowledge

We’ve already established that people want a search engine to answer questions related to their health. Chatbots can help not only with tailoring symptoms to the most likely diagnosis but in offering evidence-based, scientific and sympathetic answers to questions people worry about.

Roo, Planned parenthood’s bot, is created to provide sex education. Sex-related questions might be awkward and the lack of sex ed is dangerous, so it’s only logical to use the internet to deliver that knowledge.

Chatbots help clinicians and patients with communication and appointments

So symptoms-checkers may suggest user book an appointment or to talk to doctors, no matter the field of healthcare the chatbot is operating in. Mental health chatbots we’ll talk about later also offer to talk to therapists often, if patients wish for that — or if they have suicidal thoughts.

For doctors, chatbots that listen to patients’ symptoms first, and scheduling a visit second mean they have an opportunity to know what to say and what to do with patients beforehand. Drawing out preliminary suggestions on patients’ state, doctors can shorten the appointment “figuring-out-what-happened” time and, if their chatbot is connected to EHRs, they’ll only need to complete the profile with the final health assessment.

Such technology is good for both of them: doctors save their time to focus on analytics and decision-making and patients feel engaged and know the moment they enter the room. Smooth.

According to the BIS report we’ve mentioned before, medical triage chatbot segment will grow the most, and they will likely replace non-emergency hotline services. That will also benefit hospitals, EMS, and patients: no unnecessary movements and financial spendings.

Chatbots help with health monitoring and management, for both patients and providers

One of the reasons healthcare chatbots will get so popular in the further decade is the rising availability of wearables and connected devices. Processing life signs from wearable, they notify users about abnormalities and events out of usual patterns in their data, offering suggestions and insights on what may be going on. It’s a very good, explanatory and simple self-assessment tool that will help people understand how their daily activities and thoughts affect their physical state and, driving awareness, maintain healthy habits and track their progress.

Even without wearables, chatbots that accept regular user input on their physical state and providing them with visualization of their state in the long-term, in return, are still valuable for self-management. They are useful for providers, too. Sensely’s Molly, for instance, is a “virtual nurse” patients have to talk to every day — and doctors on the “other side” monitor reports about their answers. Using bot’s triage algorithms and Sensely’s empathetic and engaging interface, Mayo Clinic — company’s partners, — can deliver highly personalized treatment and adjust it, in case something happened, without a need for patients to leave their home.

Chatbots are of use in intervention science, making healthcare truly accessible

Alison Darcy, CEO and founder of Woebot, a company with a chatbot that uses cognitive-behavioural therapy tools to help users deal with depression, anxiety, and other mental health struggles, nailed their value in her SuperBot 2018 talk. The company decided to conduct a clinical trial, launched Woebot for the first time, and during the first release day, CBT became accessible for more people than a doctor could help during his or her whole practice. “Two-third of people who need to be in front of clinicians will never get there,” she said.

Mental health is a huge issue for the world, and, as it happens with everything we need to improve in our life, people need to think about it regularly. For CBT to work, people with depression have to use its “therapeutic lenses” regularly, recognizing destructive thoughts every time they pop up. So Woebot teaches them it, offering exercises, knowledge, and tracking tools.

Companies develop scaled mental health tools for people with chronic conditions, too. Chatbots like Wefight’s Vik help patients overcome the loneliness that is walking hand-to-hand with a chronic illness.

Another thing chatbots sometimes help with is grief. Replica’s co-founder, Eugenia Kuyda, created a chatbot on the basis of digital footprint her friend Roman left behind after he was killed in an accident. Digital version of her friend helped her deal with the loss. Griefbots are a bit controversial subject: on the one hand, they can really help people mourn. On the other hand, our cognitive override mechanism can break when we’re stressed. As humans, we believe first and think and analyze later, and in the griefbot case, we may want to postpone admitting our loved ones died because we continue the conversation with a version of them.

Chatbots have great power, and guess what comes with it

The success recipe for healthcare chatbot (and for any chatbot, honestly) is combining deep domain expertise, high-quality, rigorous data science and great user experience. Without these, a path to the dark side lies.

Good user experience means using emotional human language, training bots so they talk kind and sympathetic. Otherwise, it can only make things worse.