Understanding Public Perception of Chatbots in Healthcare: A Focus on Diabetes
Since the introduction of ChatGPT in 2022, large language models have captured significant public interest, particularly in the realm of healthcare. A recent study sheds light on how people, both with and without diabetes, perceive the use of chatbots within this critical field.
Methods and Analysis
In 2024, an extensive online survey was conducted within the Health in Central Denmark cohort, targeting 136,229 individuals aged between 18 and 89. The survey aimed to gauge perceptions of artificial intelligence (AI) and chatbots, specifically regarding trust in various healthcare scenarios. These scenarios ranged from lifestyle advice and diagnostic support to interactions with primary care doctors and emergency situations. A unique aspect of the study was the random presentation of either a serious (emergency) or less serious (family doctor contact) scenario to participants. The relationship between diabetes status, demographic factors, and trust in chatbots was analyzed using multinomial logistic regression.
Results
Out of the respondents, 39,109 completed the questionnaire. A notable 94% were aware of AI, but familiarity with ChatGPT was slightly lower at 76%, and only 21% had actively used it. The data revealed that while a majority (49-55%) expressed trust in chatbots when healthcare professionals (HCPs) were involved, trust significantly dropped (3-6%) when HCPs were absent. Trust levels were further influenced by the severity of healthcare scenarios. For instance, there was a 0.63 odds ratio (OR, 95% CI 0.60 to 0.66) indicating reduced trust in chatbots during emergencies compared to general practitioner interactions. Interestingly, the study found that individuals with type 2 diabetes exhibited less trust in chatbots than those without diabetes, whereas type 1 diabetes did not show a similar effect. Factors such as age, gender, training length, and experience with ChatGPT also played a role in shaping trust.
Conclusion
The study underscores that while chatbots are generally seen as valuable aids when guided by healthcare providers, they are met with increased skepticism in high-stakes scenarios. To ensure equitable and effective adoption of this technology, it is crucial to address digital exclusion risks and consider demographic differences, including age, gender, and specific health conditions like type 2 diabetes.
For a more detailed exploration of the study, please refer to the original publication Here.
“`
