The final analytic sample size for this analysis was 925 patients with stage 0, I, II, or IIIa and no clinical contraindications to BCS. The mean time between the date of definitive surgical treatment and response to the survey was 9 months. There were no significant differences in tumor behavior (DCIS vs. invasive) or treatment received by race/ethnicity (). About 18% of all women received mastectomy. About two-thirds of women reported that they had completed radiation and chemotherapy treatment by the time they participated in the survey.
Characteristics of the patient sample by race/ethnicity (N=877)1
3.1. Decision Outcomes
Overall, 37% of women reported a shared surgical treatment decision, 36% reported a patient-based decision, and 27% reported a surgeon-based decision. The majority (93%) reported a match between actual and preferred involvement in the breast cancer surgery decision, while 3% reported having too much involvement and 4% reported having too little. Thirty three percent of the sample reported having high decision satisfaction, while 38% had moderate satisfaction and 30% had low decision satisfaction (i.e., high decision dissatisfaction). About thirty-nine percent had a lot of decision regret, 30% a moderate amount of regret and another 31% reported very little regret.
3.2. Patient Characteristics
The race/ethnic distribution of the sample was: 25.2% Latina-SP, 20.6% Latina-EP, 24% African American, 26.1% Caucasian and 4.1% other. For purposes of these analyses those indicating ‘other’ race/ethnic group were excluded (N=48). The mean age of patients was 59 years (range 29 -79). shows the proportion of women with different demographic and clinical factors by race/ethnicity. Latina women, both Spanish and English preferent, were significantly younger than Caucasian women (P<0.001). Latina-SP women were significantly less likely to have some college or more education compared to other groups (P<0.001). Latina-SP women were significantly more likely to have low health literacy as measured in our study than any of the other racial/ethnic groups (28.1%, 8.0%, 6.2%, and 4.5% for Latina-SP, Latina-EP, AA and Caucasian women, respectively, p<0.001). As expected, Latina-SP women used more translation services than other racial/ethnic groups. About three-quarters of all women reported having someone with them during treatment discussions.
3.3 Decision Outcomes by Race/ethnicity
provides the adjusted proportion of respondents reporting each of the three categories of actual involvement (surgeon-based, shared, patient-based) by racial/ethnic group, controlling for demographic and clinical factors. Latina-SP women were slightly less likely to report a patient-based decision than the other groups (P<0.10). We also found that Latina-SP women were slightly more likely to report too little involvement compared to other groups (results not shown). shows the proportion of women in each racial/ethnic group who reported high decision dissatisfaction and/or a lot of decision regret. Latina-SP women reported having both high dissatisfaction and a lot of regret significantly more often than women of other racial/ethnic groups (P<0.001).
Actual decision involvement by race/ethnicity1
Decision dissatisfaction and regret, by race/ethnicity1
3.4. Factors Associated with Decision Dissatisfaction and Regret
Tables and provide the logistic regression results for correlates of high decision dissatisfaction () and a lot of decision regret (). The first column of both tables provides the unadjusted results for race/ethnicity. The second column provides the results, controlling for other demographic and clinical factors, and the third column shows results controlling for the mechanistic variables.
Logistic regression of decision dissatisfaction (N=877)
Logistic regression of decision regret (N=877)
As seen in Tables and , without adjusting for other factors, Latina-SP women were 8.7 times as likely as Caucasian women to report high levels of decision dissatisfaction and 9.4 times as likely to report decision regret. The unadjusted column shows that Latina-EP and African American women were also significantly more likely than Caucasian women to report high decision dissatisfaction and regret (P<0.01). When other factors were controlled, Latina-SP women continued to be the most likely to report high decision dissatisfaction and regret though the odds were reduced in both cases (OR: 5.6; 95% CI 3.2-9.8 for dissatisfaction; OR: 5,6; 95%CI 3.2-9.9 for decision regret) compared to Caucasian women. The other racial/ethnic groups were also more likely than Caucasian women to report dissatisfaction and regret controlling for other factors (P<0.05 across racial/ethnic groups). While women with less than a high school education were more likely than those with some college to report decision dissatisfaction (OR: 1.8; 95% CI 1.1-2.8), there was no significant association between low education and decision regret. However including education in the model did reduce the difference in decision regret between Latina-SP and Caucasian women (Unadjusted OR: 9.4 vs. adjusted OR: 5.6) None of the clinical variables, including type of surgical treatment, were significantly associated with high decision dissatisfaction or regret.
The third column of both tables shows the results for race/ethnicity when controlling for the three mechanistic variables. Women with low health literacy were significantly more likely than those with high health literacy to report both decision dissatisfaction and regret (OR: 5.6; 95%CI 2.9-11.1 for dissatisfaction; OR 3.5; 95% CI 1.8-7.1 for regret). Those with moderate health literacy were also more likely to report dissatisfaction and regret (P<0.01) than those with high literacy. Having someone with you during treatment discussions was negatively associated with decision dissatisfaction (OR: 0.65; 95% CI 0.44-0.96); while it was also negatively associated with regret this association was not significant (). Translation was not associated with either dissatisfaction or regret. Although including the mechanistic variables, specifically health literacy, did not attenuate the association between race/ethnicity and decision outcomes, it did significantly reduce differences in decision regret between Latina-SP and Caucasian women (from an adjusted OR of 5.6 to an adjusted OR of 4.1 including mechanistic variables).