Say hello to Molly, Florence, and Ada—they’re just a few of the helpful, smart algorithm-powered chatbots taking their place in health care. Chatbots are computer programs designed to carry on a dialogue with people, assisting them via text messages, applications, or instant messaging. Essentially, instead of having a conversation with a person, the user talks with a bot that’s powered by basic rules or artificial intelligence. Chatbots are already widely used to support, expedite, and improve processes in other industries, such as retail, and now, the technology is gaining traction in health care, where it is helping patients and providers perform myriad tasks.
MEET THE CHATBOTS
For clinicians, chatbots can streamline the interaction with electronic health records and decrease documentation burden. “Modern electronic health records rely on clinicians to either type or dictate consultation notes, which is associated with clinician burnout, increased cognitive load, information loss, and distraction from other tasks in the clinical encounter,” says Liliana Laranjo, a Research Fellow at the Centre for Health Informatics at Macquarie University in Australia (right). “Conversational interfaces can employ recent advances in speech recognition, natural language processing, and artificial intelligence to provide clinicians with tools to automatically document elements of the spoken clinical encounter.”
Clinicians can also use chatbots to quickly and easily retrieve information about drug interactions and side effects. Safedrugbot, for example, is a chat messaging service that helps doctors and other health professionals obtain information about the safety of drugs for women who are breastfeeding.
For patients, there are chatbots that can book appointments and perform other customer service tasks. Bertalan Meskó, Director of The Medical Futurist Institute (right), points to the company Kore.ai, which offers smart bots for health care facilities. “The digital assistant can connect patients to the right contacts, provide or change appointment details, assist with refilling prescriptions or paying bills, and deliver test outcomes,” says Meskó.
Chatbots are also becoming the first contact point for primary care. People can now turn to chatbots to identify symptoms and recommend further actions. These chatbots are not designed to offer diagnoses, but to guide people and help reveal if they are seriously ill and should see a doctor.
Babylon Health, Ada Health, and Your.MD all offer chatbots that can check one’s symptoms and recommend courses of action. These chatbots ask questions of patients and compare the answers to cases from their medical database. Then they provide possible diagnoses, offer actionable health information, and help book the necessary doctor’s appointments. Sensely goes a step further: their virtual medical assistant, Molly, appears as a virtual persona on the screen that patients can communicate with using either text or speech.
Other chatbots are designed to monitor a person’s health status and help manage chronic conditions. For example, Florence functions as a personal nurse; she can remind you to take your medicine, provide instructions if you forget to take a pill, monitor your health, and find and book doctor’s appointments. She also helps motivate you to achieve your health goals and provides medical information.
A SMART HEALTH CARE TOOL
“One of the nice things about a chatbot is that it’s very accessible; you can use it on your smartphone,” says Christopher Lovejoy, a Doctor at St. George’s Hospital and the Clinical Data Science and Technology Lead at Cera Care, both in London (right). “People like chatbots for their accessibility and the convenience of using it.”
Chatbots also have the potential to improve access to health care by delivering advice or counseling remotely at lower costs. This could ease the burden on medical professionals. “There are so many simple medical questions out there that do not require the full attention of a physician but letting them remain unanswered leaves the concerned people nervous, confused, and clueless,” says Meskó. “Patients should not have to leave work and travel to their doctor’s office just to get a ten-second response to a simple question.”
With chatbots improving patient experiences, saving money, and increasing efficiency, is it possible that they will replace human workers? Sandeep Reddy, a Professor at the Deakin University School of Medicine in Victoria, Australia, says that chatbots will complement human health care professionals, not replace them (right). “Even with advances in chatbot technology, there will be a need for patients to seek medical care supervised by human clinicians,” he says. “While healthcare chatbots work well for screening medical conditions and referring serious patients for medical care, they cannot replace the medical environment and equipment that serious medical conditions require.”
Laranjo agrees with Reddy. “Most likely, chatbots will be able to help with routine tasks, streamlining the work of clinicians and facilitating self-care for patients,” she says. “As a result, the work of clinicians will certainly change and adapt, ideally leaving more time available for them to focus on patient care and the art of medicine.”
Although chatbots may prove useful for preliminary information, it is important that patients not use them to replace human doctors. Chatbots still have limitations, both in their algorithms’ ability to recognize and analyze human conversational inputs and in their inability to recognize symptoms of certain medical conditions that require visual and tactile examination. Lovejoy says that while chatbots are helpful, it’s important to discourage public over-reliance on them. “Someone might say to a chatbot, ‘I’m short of breath,’ but as a clinician, I can examine ten people with shortness of breath and obtain a great deal more information,” he says. “One of them I might not be concerned about at all; others I might be severely worried about. A chatbot runs the risk of not picking up that nuance.”
Chatbots are still at a relatively early stage of development but because of their versatility, it’s likely that we will continue to see an expanding role for them in the health care sector. Reddy believes that advances in machine learning, computer vision, virtual faces, and natural language processing abilities will enhance chatbots’ capability to recognize and analyze human conversations, screen medical conditions, and appeal to humans even more. “Chatbots will increasingly be used by stretched health care services and insurance organizations to screen their patient populations so only serious patients are referred to hospitals and primary care clinics,” he says. “Further, doctors will recommend all their chronic condition patients, especially those who require nutritional, exercise, and medical management advice, to download and use chatbot applications to help them manage their conditions.”
Meskó also predicts wider use of health care chatbots. “In the future, we will be able to share our health data and medical records with these chatbots to help them make the wisest decision,” he says. “And I’m sure software developers and tech companies will recognize the huge potential for the use of ‘gamification,’ as well,” he continues. “For instance, if you could collect points or rewards for eating healthy foods or exercising, patients might be motivated to live healthier lifestyles.”
According to Meskó, chatbots could even one day be able to detect certain medical conditions based only on our voices. “It’s an exciting time because there is a lot of research that has gone on and a lot of potential uses that remain unactualized,” says Lovejoy.
Originally published in IEEE Pulse by Mary Bates on May 28, 2019