Telemedicine is becoming widespread and coincides with the rise of conversational artificial intelligence (AI) tailored to health care. More than 71% of healthcare providers now have robots for good reason: studies show that hospitals could save more than $ 100 million a year if they offered such applications and that people are about 53% more likely to frequent places where they can send messages.
But not all robots are created equal, especially with regard to health services. In an effort to make a difference, the UserTesting research firm based in Mountain View, California, interviewed a group of 500 users for their opinion on the top five health bots: Ada , Mediktor, Your.MD, Symptomate and HealthTap. Respondents were given a list of symptoms of colds and food poisoning and were asked to use the robots to obtain a diagnosis and then to evaluate their experience in several categories, including "ease of use", "Speed," "credibility," etc. "Aesthetics" and "enchantment". The questions they answered were used to generate quantitative percentages ranging from 0 (worst) to 100 (best).
"Consumers expect a lot from their digital experiences. Whether ordering a ride, managing finances or seeking medical advice, customers expect a fast, easy and quality customer experience, "said Janelle Estes, head of analytics at UserTesting. "Our study shows how important it is to leverage human knowledge to understand consumer sentiment and preferences to create a superior customer experience. This can ultimately determine the success or failure of new generation applications such as chatbots and conversational AI.
So how are the bots? Not particularly good, taken as a whole. None have succeeded brilliantly, and many still have some progress to make in terms of usability and the ability to solve complex conditions, according to UserTesting.
Credibility is another obstacle that robots have not yet overcome. Interviewees at the UserTesting Survey expressed concern that some of the tested bots do not appear to comply with the 1996 Portability and Health Insurance Liability Act, which provides for confidentiality and data security to protect information medical patients. Grammar errors in the application and links to shady online resources did not contribute to the perceived credibility and led some survey participants to say that they would need additional information before being persuaded to follow the suggestions of the robots.
A report by Pegasystems published in October 2018 echoes these sentiments. He revealed that 65% of people preferred to talk to a person rather than a chatbot because of the perceived lack of "intelligence" by the latter and that 18% thought that the robots were "ineffective" or "annoying". (That said, 49% rated strongly when it came to providing basic information.)
This is not just bad news for health robots, though. Nearly 46% of those surveyed said that prior to the study, they had consulted a robot to look for medical symptoms. Overall, 73% said they found the "chatbots" in the UserTesting study "useful", while only 12% said they had "no help".
"The combined benefits for both consumers and healthcare providers indicate that consumers are likely to begin to see more and more of the conversation interfaces as providers build on them for better serve patients, "UserTesting writes in his report. "The intrinsic convenience and empathy that these apps can provide is driving consumers to embrace the world."
Not surprisingly, some analysts predict that interactions with robots will exceed 2.8 billion by 2023 and that by 2025, the global robot market will reach $ 1.23 billion.
That said, bad user experiences threaten such continued expansion. According to a study by the New York School of Medicine, 50% of people who download a health-related mobile app eventually give up because the experience is "too cumbersome".
"Chatbots have the potential to transform the way patients engage in their health systems and relieve some of the overwhelmed staff," said Michael Larner, an analyst at Juniper Research, cited by UserTesting.