Purpose: Most women newly diagnosed with early stage invasive breast cancer are surgically treated with breast conserving surgery (BCS) or mastectomy. In this preference-sensitive decision, the quality of the decision can be gauged by the extent to which the choice reflects the preferences of an informed patient. The study purpose was to collect descriptive data about changes in patients' knowledge and values across the decision making process, and to field test a decision quality instrument (DQI).

Methods: Newly diagnosed breast cancer patients viewed a video-based decision aid and then completed questionnaires: (1) immediately after watching the video; (2) after the initial visit with the surgeon; and (3) four weeks after surgery. The DQI included items about patients' knowledge (multiple-choice responses), values (responses on a 10-point scale), preferred treatment (times 1 and 2), and treatment received (time 3). Analysis focused on a) the within-sample and across-time distributions of knowledge levels and value scores, and b) the association between reported values and treatment received. We hypothesized that knowledge would decrease over time, and that values would become more strongly associated with treatment received over time.

Results: 43 patients completed all 3 assessments. 30 (70%) had BCS and 13 (30%) had mastectomy. DQI knowledge scores were high and did not differ significantly across the 3 time points (92%, 89%, 88% correct; F = 0.98; p = 0.3808). Importance scores for keeping the breast (KB) and removing the breast for ‘peace of mind' (PoM) were associated with treatment choice at the decision point (time 2). With each point increase in importance: for KB, patients were less likely to have mastectomy (OR 0.69: 95% CI 0.53-0.9); for PoM, they were more likely to have mastectomy (OR 1.87: 95% CI 1.24-2.61). Between assessments, patients' scores on each of the 3 values items changed 1-2 points in the direction of increased concordance with actual choice (e.g., for PoM, ORs 1.63, 1.87, 1.91 at times 1, 2 and 3, respectively); these changes were non-significant.

Conclusion: These results imply that key aspects of Decision Quality (the level of comprehension and the constellation of personal values) remain consistent across this time, which, in turn, suggests that a DQI audit may be valid from the point of actual therapeutic choice until four weeks after treatment.