Development of a tailored, computerized, breast cancer risk assessment and decision support tool: What do clinicians want?

Abstract

e20660 Background: End user input into the design of decision support tools is critical to enhance integration and future routine use in clinical practice. As part of the development of an evidence-based, tailored, computerized breast cancer (BC) risk assessment and management tool, we examined clinicians’ requirements. Methods: Australian breast surgeons (BSs) and primary care clinicians (PCCs) were recruited through local professional networks. Facilitated focus group discussions about current practice of assessing and managing BC risk and perceptions of the proposed tool were audiotaped transcribed and managed using QSR NVivo. A coding framework was developed based on the transcripts, data were coded and each code further analyzed to identify key themes. Results: Four focus groups, involving 12 BSs, 17 PCCs (12 doctors, 5 practice nurses) were conducted. 55 % were male, mean age 45yrs (range 25-67), mean of 14yrs in practice (range1-25+). Clinicians identified difficulties assessing and managing BC risk and lack of available tools to standardise their current inconsistent approach to risk assessment and management. Most felt confident identifying high risk women, but found differentiating women at population risk from those at moderately increased risk more difficult. They felt a tool would help them reassure anxious low to moderate risk women and avoid unnecessary onward referral or investigations and better identify and refer or manage high risk women. Desirable features in a tool included: evidence-based, accessible (web-based), visual, simple data entry process, displays absolute risk (rather than relative) and mutable risk and risk reduction estimates in multiple formats (words, pictographs, graphs) to improve comprehension, printable summary sheets. Clinicians felt that women could input risk factors before the clinic visit but that joint consumer and clinician data entry was preferable and women should only receive a risk estimate with a clinician present. Conclusions: Development of tools for BC risk assessment and management should address these clearly identified needs of end users in order to optimise translation of current and future knowledge into clinical practice.