Co-creation of decision-relevant information with patients to support shared decision-making about adjuvant treatment in breast cancer care (Oral presentation)

Introduction
Patient Decision Aids (PDAs) usually provide benefit/harm information about treatment options and value clarification methods
(VCMs). Personalized risk information from prediction models is also increasingly integrated. We aimed to design this total package
of decision-relevant information about adjuvant breast cancer treatment in co-creation with patients, and to present it in an
understandable way.
Methods
Three co-creation sessions with breast cancer patients (N=7-10; of whom N=5 low health literate). Participants completed creative
assignments and evaluated prototypes of benefit/harm information and VCMs. Prototypes were further tested in a new group of
patients (N=10) and professionals (N=10). Notes, homework assignments, photos and audio-recordings were summarized and main
themes were identified. User test sessions were transcribed literally and analyzed using ATLAS.ti to assess key themes related to
comprehension. 
Results
Important information needs identified were: (a) need for overview/structure of information directly after diagnosis; (b) need for
transparent benefit/harm information for all adjuvant treatment options, including detailed side-effects/late effects information.
As for VCMs, patients stressed the importance of a summary or conclusion. A classical bar graph seemed the most appropriate way
of displaying personalized survival rates; the impact of most other formats was experienced as too distressful. The concept of
‘personalized information’ was not understood by multiple patients. 
Discussion
Patients’ need for numerical information about side-effects/late effects is in line with previous literature. This numerical
information is not always easily available, complicating SDM. The need for conclusions in VCMs also corresponds to previous
studies, but is still not included in many PDAs. Although we experimented with alternative visualizations of survival rates, patients
seemed to be supported better with a classical bar graph. 
Conclusion
This co-creation study provided concrete perspectives to meet patient information needs through a PDA. Further quantitative
studies are needed to establish comprehension of generated prototypes and the concept of ‘personalized information’.