gms | German Medical Science

Introduction: A clinical decision support system (CDSS) in the primary care domain for patients with unclear disease based on different artificial intelligence (AI) methods [1] can additionally support finding a diagnosis through the exchange of information with affected persons as experts of their disease [2]. In a user-centred approach, we want to identify how these stakeholder groups of affected persons (patients and relatives) imagine an ideal smart physician portal for patients with unclear disease and what role they can adopt in it.

Methods: To answer the research question a 90-minute online workshop was conducted to collect first ideas and needs for a new CDSS (target situation) and dealt with the study research question. Participants were invited by email via patient organizations across Germany. Participants were included with the following criteria: Affected person (Patient or relative) of a patient with an unclear diagnosis (= cause of a patient's current symptoms was not found by the available examinations in general practitioners practice, as well as by at least one specialist doctor; own definition based on definitions of centres for rare diseases) or a rare disease (disease affecting fewer than 5 people in 10000 [3]). A brainstorming session was conducted via video conference and recorded. After the brainstorming session, participants were asked to prioritise their brainstorming ideas in a given template ('Top five booklet '[4]). Following the method of Baxter et al. [4] we finalized the reports.

Results: Seven affected persons followed the invitation to the workshop (57% female, 43% male). In the workshop, 21 brainstorming items were gathered. 57% of participants sent us their Tops. The two items ‘Receive information about a confirmed diagnosis’ and ‘Enter missing data after receiving hint’ was named the most. In addition, participants were also interested in viewing further information such as recommendations of a family screening, examinations performed and findings, and being able to view everything in an accessible manner. Other ideas and desirable items ranged from the correction and, if necessary, entry of further data such as symptoms and medication by the patient for the CDSS to the transfer of health data from apps and the provision of further information.

Discussion: The discussion led to 21 ideas for further developing of an AI-based CDSS. The early inclusion of user feedback facilitated the design of a user-friendly system. As the focus of the planned CDSS is on diagnostic support and recommendation for further procedures, all therapeutic issues will take a secondary place but should be considered and discussed with all stakeholders in the future. As other studies have shown, patients with a rare disease have a high need for health information, higher than that of other patient groups [2], [5]. We were also able to confirm this. The patient approach should be pursued further in the project to go beyond initial ideas.

Conclusion: Involving all stakeholders early in the design of a CDSS provides many opportunities and chances to create a user-centred design. Further iterations should be conducted to include more patient groups and improve the user experience.

The authors declare that they have no competing interests.

The authors declare that a positive ethics committee vote has been obtained.