Chatbots are playing an increasingly important role in the healthcare industry, their adoption spurred by a spike in the demand for medical care brought on by the COVID-19 pandemic and the maturity of the technology’s underlying artificial intelligence (AI).
As it’s been widely reported, the pandemic has created the paradoxical conditions of creating a great deal of patients and paperwork for healthcare professionals while at the same time limiting people’s exposure to healthcare due to quarantines and social distancing. Parts of the healthcare industry had over the past few years begun to deploy chatbots — or “conversational assistants” — to remind patients to take medication, collect insurance or health information, or provide fairly generic information about disease symptoms, nutrition or treatments.
The COVID-19 pandemic has created even greater opportunities for chatbots use in the healthcare industry. These include reducing time patients spend filling out paperwork in doctors’ offices, scheduling appointments, providing emotional support for those living in quarantine and even helping patients determine whether their COVID-19-like symptoms warrant a trip to the doctor.
Whereas early chatbots used pre-scripted dialogue and didn’t perform well when users deviated from that script, today’s chatbots are taking advantage of AI to improve their performance the more they are used. A key component of AI chatbots is natural language processing (NLP), which enables a chatbot to more accurately interpret a patient’s written or spoken questions as well as extract important information from the dialogue.
The worldwide COVID-19 pandemic — and the resulting societal push to put as many services online as possible — has created a tremendous opportunity for healthcare chatbots. In March, Microsoft, the U.S. Centers for Disease Control and Prevention (CDC) and a number of major healthcare providers launched the “Coronavirus Self-Checker” chatbot to guide users through self-assessment and, if necessary, connect them to additional resources and medical professionals.
Healthcare provider Providence was the first to make the Coronavirus Self-Checker chatbot available, via its website. The app asks a number of questions based on CDC guidelines and, depending on the answers, gives an option to contact a doctor or participate in a virtual video visit. Within just a few weeks, the chatbot had created more than 40,000 sessions, with more than a million messages sent between people and the chatbot, the Wall Street Journal reported.
With so many patients unable to see their doctors in person, chatbots have become a safer, more convenient way to interact with a variety of medical professionals. A June 2020 New York Times article, for example, detailed one Houston native’s reliance on the Replika chatbot as an antidote to loneliness and mental stress placed on her while she remained quarantined at home. According to the Times, half a million people downloaded Replika during the month of April alone, at the height of pandemic. The pandemic has marked a distinct turning point for the app, originally launched in 2015 by San Francisco start-up Luka to make restaurant recommendations.
One of the main motivations behind healthcare chatbots is to ease the burden on primary care doctors and help patients learn to take better care of their health. Many people who make an appointment for a colonoscopy, for example, cancel it or fail to show up. The problem is particularly extreme in vulnerable or disadvantaged populations — studies show that as many as 40 percent of these patients don’t follow through with the procedure.
In 2019, Northwell Health deployed a chatbot to help overcome the fear and stigma of colonoscopy tests, an important procedure for early diagnosis of colorectal cancer, which accounted for 13.5 percent of all U.S. cancer deaths in 2017, according to the CDC. The New Hyde Park, N.Y., healthcare provider launched a chatbot to help reduce no-shows for colonoscopies at the company’s Long Island Jewish (LIJ) Medical Center and Southside Hospital.
Northwell’s Colonoscopy Health Chat, based on Conversa Health’s automated conversation platform, uses AI to address misunderstandings and concerns about the exam. The platform delivers information in a responsive, conversational way over email or text. Different chat options, available in English or Spanish, educate patients on the benefits of the test and what to expect before, during and after the procedure. The program also provides date and location reminders as a patient’s appointment draws near.
Northwell also uses Health Chat to help reduce hospital readmissions for high-risk patients after a heart attack, stroke or other serious condition, and to prevent and manage side effects and other problems in patients receiving radiation therapy for head and neck, breast or prostate cancer.
In general, the healthcare industry offers a number of use cases for chatbots, whether providing patients with information, offering conversational support or delivering ideas for therapeutic solutions.
There are many other opportunities for the healthcare industry to tap as well. Healthcare insurance companies also have several good options for putting chatbots to good use, starting with those that make the insurance process easier to navigate. Geolocated chatbots can guide people through hospitals and allow them to ask questions based on the section of the hospital where they are located. Chatbots could also be more widely deployed for tracking prescriptions and medication use, as well as enabling doctors and patients to share health diaries.
Chatbots offer healthcare organizations the opportunity to extend their resources, reaching a greater number of people with the help of interactive conversations that can help identify symptoms, manage medications and monitor chronic health problems. The goal at this time is not to fully diagnose patients via virtual assistants but rather to guide patients to the right resources and help healthcare professionals better understand a patient’s needs.
Healthcare chatbots offer healthcare professionals and their patients the following advantages:
Projections as to the size of the healthcare chatbot market in the coming years vary greatly, but many agree it will soon be worth at least hundreds of millions of dollars. A 2019 market intelligence report by BIS Research projects the global healthcare chatbots to generate more than $498.1 million by the end of 2029, up from $36.5 million in 2018. A January 2020 market research report by Meticulous Research projects the healthcare chatbots market to reach $703.2 million by 2025, mainly attributed to improving internet connectivity and speeds, the continued adoption of smart devices, limited availability to doctors, AI advancements and increasing demand of virtual health assistance. Factors that could hold back the market include data privacy concerns, some companies’ lack of expertise in chatbot development and mistrust in medical guidance delivered via an app.
Botpress is an inclusive and open-source conversational AI platform for developers who wish to create chatbots for healthcare or any number of other industries. Our platform’s natural language understanding supports more than 20 languages, and the conversation studio allows you to seamlessly translate from one language to the other without creating multiple chatbots.
Botpress supports developers through a framework that allows developers to access and build on common features and methodologies, speeding development time and resulting in better coding standards. Frameworks also act as middleware allowing developers to connect to many important related services through a single API call.
If you want to develop your own chatbot or wish to learn more about our platform and services, don’t hesitate to contact us or Request a demo.
Build impressive chatbots, without having to be an NLP expert. Botpress' NLU strategy supports you in creating a conversational interface.
Learn how to create a chatbot that uses an action to call the Giphy API and provides a gif to the user.
The chatbot development ecosystem is still relatively immature, which means it’s not as easy as it will eventually be...