Like ‘em or loathe ‘em, many physicians seem to have some strong feelings about the use of chatbots in healthcare, according to a new research study.
The study, which appears in the Journal of Medical Internet Research, looked at the benefits, challenges and risks to patients associated with the use of healthcare chatbots. The study draws on responses from 100 practicing physicians based in the U.S.
As a preface to this discussion, it should be emphasized that just 30% of physicians who responded to the survey had direct personal experience with the use of chatbots for health-related issues. Also of note is that just 18% of physicians reported that their patients used chatbots, while 24% said that their patients didn’t use them and 58% didn’t know whether their patients did or did not.
Still, responding physicians thought healthcare chatbots were likely to become more important over time, with 42% of physicians agreeing that chatbots were either very important or somewhat important in healthcare, and 44% stating that they would be very likely or somewhat likely to prescribe the use of healthcare chatbots within the next five years.
Many respondents had positive feelings about using chatbots for administrative tasks, including a belief that chatbots would be beneficial for scheduling doctor appointments (78%), locating health clinics (76%) or providing medication information (71%).
In addition, 54% of respondents agreed that healthcare chatbots could help patients better manage their own health, along with improving access to and timeliness of care (53%), reducing travel time to see their healthcare provider (52%), preventing needless visits to healthcare providers (49%) or leading patients to disclose more information than they would to healthcare providers (41%).
At the same time, though, 70% of physicians who responded had some reservations about using this technology. These included concerns that healthcare chatbots pose a risk to patients if they use them to self-diagnose too often (74%), if patients don’t accurately understand their diagnoses (74%), and if chatbots can’t provide detailed clarifications about patient assessments (71%).
In addition, many respondents believe that chatbots can’t care for all of a patient’s needs effectively (76%), cannot display human emotion (72%) and can’t provide detailed diagnosis and treatment as they don’t know all of a patient’s personal details (71%).
The research concludes that at least for now, healthcare chatbots could have a meaningful but limited part to play in the patient care process. They can be used to “support, motivate, and coach patients, as well as for streamlining organizational tasks,” the paper suggests. “In essence, chatbots could become a surrogate for non-medical caregivers.”
I’d argue that this is a pretty favorable set of results, though it’s possible the physicians involved will be less optimistic about healthcare chatbots when they use them more heavily and the “new tech” glow has faded. It’ll be interesting to see what future surveys conclude about chatbots after most doctors have had a chance to take them for a spin.