Evaluation of visual displays to share expected quality of life changes with patients with breast cancer

Background

For patients deciding between mastectomy and breast conserving therapy, the incorporation of expected postoperative patient-reported outcomes may facilitate shared decision-making. We aim to understand how patients interact with a variety of patient-reported outcome display styles to support the development of decision aids for breast surgery.

Methods

A multidisciplinary team developed 6 visual displays of patient-reported outcome data in select BREAST-Q domains. Patients with stage 0–III breast cancer treated at an academic institution received questionnaires assessing their preferences for the patient-reported outcome displays. Demographic variables were extracted from the electronic health record. Display rankings were compared using means, standard deviations, and the Plackett-Luce model for ranked data with Bonferroni corrections to account for the pairwise comparisons. Differences in rankings by demographic and cancer variables were assessed by general linear models. Open-ended responses were assessed by thematic analysis.

Results

A total of 59 participants completed the questionnaire (Response rate: 17%). Three displays had significantly higher (more preferred) mean ranks without statistically significant differences among them; 2 of the 6 displays had significantly lower (less preferred) ranks. No differences were found between preference for different patient-reported outcome displays and patient demographic or disease-related variables. In open-ended comments, patients reported a preference for simple displays with longitudinal patient-reported outcome data; patients disliked displays that either had too little or too much information.

Conclusion

Displays with longitudinal patient-reported outcomes were preferred by patients, whereas displays with multiple BREAST-Q domains were overwhelming and therefore were less preferred. Patients generally preferred 3 to 4 data values per figure. There is no “one size fits all” approach, because several patient-reported outcome displays ranked highly. Further research on how patients interpret multivariable figures is warranted to understand the needs of diverse patient populations.