Background: Chatbots are promising tools that intuitively engage users and provide solutions autonomously and efficiently. Currently, chatbots already serve significant roles- from general symptom checkers to therapy for people with depression. Their benefits range from freeing up healthcare resources, to allowing patients greater accessibility to healthcare resources. The objectives of this study are to map the types of chatbots and their features, including their purpose, conversational attributes, precautions, and usability; define systematic methods to the novel approach of searching and assessing chatbot apps; and provide direction and framework for future research in this field.
Method: A systematic assessment of chatbots available on iOS and Android smartphones was performed, involving the search, selection and standardised assessment of included chatbots.
Results: The search identified 2719 apps, of which 48 were relevant as healthcare chatbots and were included in assessment. Mental wellness and symptom checker formed the majority of chatbot purposes. User input predominantly included combinations of text-based natural language input and constrained options, with few allowing free speech. Chatbot output was predominantly text-based, with some allowing additional speech output. Chatbots generally lacked personality and engagement. Clinical safety and information security features were incomplete in most chatbots, and usability was generally poor and compromised by lack of user-friendly features.
Discussion: App-based chatbots are strongly heterogeneous in features and serve a variety of healthcare functions with potential for greater efficiency and accessibility. As we eagerly adapt technology into our lives, it is crucial for us to consistently update our understanding of the available tools, their purposes, features, strengths and shortcomings. Doing so, we gain clearer insights and direction into how best to leverage tools like chatbots, how to improve future developments, and what precautions to be aware of.
Implications for practice: With increasing smartphone ownership and acceptability of health-related apps among the public, smartphone app-based chatbots are poised to improve healthcare.