JMIR Publications

ABSTRACT

Background:

For decades, the question has been studied how artificial intelligence can support diagnostics. Symptom checkers are designed for lay people and promise to provide a preliminary diagnosis, a sense of urgency and a suggested course of action.

Objective:

We use the international symptom checker app (SCA) Ada App as an example to answer the question: How do general practitioners experience the SCA in their daily work, and how does this interact with work-related psychosocial resources and demands?

Methods:

We conducted n=8 semi-structured interviews with GPs in Germany between December 2020 and February 2022. We analysed the data using the integrative basic method, an interpretative-reconstructive method, to identify core themes and modes of thematisation.

Results:

Although most GPs in our study were open to digitisation in health care and in their own practice, only one was familiar with the SCA. GPs considered the SCA as part of the "unorganised stage" of patients' searching about their conditions. Some preferred it to popular search engines. They considered it relevant to their work as soon as the SCA would influence patients' decisions to see a doctor. Some wanted to see the results of the SCA in advance in order to decide on the patient's next steps. GPs described the diagnostic process as guided shared decision making, with the GP taking the lead and the patient deciding. They saw diagnosis as an act of making sense of data, which the SCA would not be able to do, despite the huge amounts of data.

Conclusions:

GPs took a techno-pragmatic view of SCA. They operate in a health care system of increasing scarcity. They saw the SCA as a potential work-related resource if it helped them to reduce administrative tasks and unnecessary patient contacts. The SCA was seen as a potential work-related demand if it increased workload, e.g. if it increased patients' anxiety, was too risk-averse or made patients more insistent on their own opinion.