DSP and Patient characteristics within dyads
Baseline characteristics for the Latina and non-Latina patient-DSP dyads are summarized in Table 1. Among the dyads which include a Latina patient and their DSP (N=292), 78%, 17.5% and 4.5% of their DSPs self-identified as Latinx, White, and Asian/Black/Other, respectively. The Latinx DSPs within Latina/Latinx dyads were younger (p<0.0001), and had lower educational attainment (p<0.0001) and acculturation (p<0.0001) when compared to Latina patient/non Latinx dyads. Latinx DSPs mean age was 44 years, 80% were of Mexican origin, 41% had low acculturation, and only 26.6% had a college graduate degree or higher (Table 1). Overall, 46% and 55% of Latina patients in this cohort had high and low acculturation, whereas 59% and 41% of Latinx DSPs had high and low acculturation, respectively. While 40% of low acculturation Latina patients chose a highly acculturated Latinx DSP, 60% of Latina patients with low acculturation identified a key DSP who also had low acculturation (data not shown). Overall, high acculturated Latinas paired consistently with highly acculturated Latinx DSPs (82%) (data not shown).
Patient characteristics did not vary by marital/partnered status, surgery received or delivery of adjuvant radiation. There were statistically significant differences across age (p<0.0001), stage at diagnosis (p=0.009), and the proportion who received chemotherapy (p=0.01) (Table 1). While the proportion (68%) who were married or partnered was not statistically significantly different across dyads (p=0.332), the identified key DSP for Latinas within the Latina/Latinx dyads was more often a daughter over a husband/partner (37% vs 21%) (Table 1). This contrasts with all other dyads where husband/partner was preferred over daughter: Latina/White DSP (59% vs 16%), Latina/Other race/ethnicity DSP (46% vs 15%), and non-Latina patient/DSP dyads (47% vs 17%).
The identified key DSP within Latina/Latinx dyads was similar regardless of the patient marital status (Figure 2 and Table S1). Married/partnered and not married/partnered Latinas were still more likely to choose a daughter as their DSP over a husband, other family-member (i.e., mother, son, sister), or friend, when compared to all other Latina/DSPs (Latina/White, Latina/Other) and non-Latina patient/DSP dyads (Figure 2). Overall, among the married/partnered patients, the third most common DSP after a daughter or husband was a Friend/Non-family member (Table S1). For not married/partnered patients, daughters remained the preferred DSPs for the Latina/Latinx dyads as discussed, however for all other dyads a Friend/Non-family member was the most common DSP, followed by a daughter, and a sister (Figure 2).
DSP characteristics and level of engagement
Table 2 displays the multivariable-adjusted mean DSP-reported engagement scores across the 4 domains (being informed, extent of involvement, satisfaction, awareness), stratified by dyad type. Latinx DSPs reported being more informed (adjusted mean 4.26, p=0.058) compared to the other dyads after adjusting by DSP age, type, education, race, ethnicity, as well as patient age, stage, surgery type, and delivery radiation and/or chemotherapy (Table 2). No statistically significant differences were seen for the other DSP engagement domains (Involved extent, satisfaction, and awareness) (Table 2).
Association of DSP-reported engagement and knowledge with patient-reported SDQ and treatment deliberation
Table 3 displays the multivariable-adjusted mean differences in patient-reported SDQ and treatment deliberation for each of the 4 DSP-reported engagement measures, stratified by DSP type. A more informed Latinx DSP within the Latina/Latinx dyads was positively associated with higher patient reported SDQ (adjusted mean difference 0.176, p=0.034), despite no statistically significant difference in treatment deliberation (-0.024, p=0.787) (Table 3). No statistically significant associations were observed for the Latina/non-Latinx dyads. Within non-Latina patient/DSP dyads, having a highly informed and aware DSP was associated with higher patient reported SDQ (adjusted mean difference, 0.115, p=0.001) and treatment deliberation (adjusted mean difference, 0.138, p<0.001), respectively.
Correlation between DSP preferred treatment and patient received treatment
Within the entire sample of dyads (both Latina/DSP dyads [N=292] and non-Latina/DSP dyads [N=881]), 61% and 39% of patients underwent lumpectomy and mastectomy, respectively (Figure 3A). The proportion of mastectomy with and without reconstruction was 22% and 17%. Overall, there was high concordance between treatment received by patient and her DSPs’ preferred treatment (Figure 3A).
When stratified by Latina patient (Figure 3B) and non-Latina patient dyads (Figure 3C), the distribution of patients’ treatment received did not vary, but there were differences between their DSPs preferred treatment among Latinas (Figure 3B). Latinx DSPs within Latina/Latinx dyad had a notably higher preference for mastectomy (56%, Figure 3B) when compared to DSPs for non-Latina patients (non-Latina/DSPs) (43%, Figure 3C), specifically for mastectomy with reconstruction (40%) when compared to non-Latina/DSP dyads (28%) (Figure 3, panels B and C). Over a quarter (27%) of Latina patients (versus 13% of non-Latina patients), underwent lumpectomy despite their DSPs preference for mastectomy (See Table S2.1 and S2.2). Overall, DSPs preference for mastectomy was greater than the proportion of patients who underwent mastectomy with reconstruction (22%). Despite these differences seen in the proportions of DSP preferred treatment and patient received treatment, concordance analysis showed overall good agreement for both Latinas (82%, kappa=0.651) and non-Latina patient dyads (90%, kappa=0.790).