The association of medical mistrust (MM) to medical decision-making (MDM) preferences in cancer (CA) patients (pts).

Abstract

e13790 Background: Medical mistrust (MM) is associated with adverse health outcomes, but relatively few studies have assessed MM among CA pts. CA pts face frequent challenging decisions about diagnostic tests, treatment, supportive care, finances, and end-of-life. The Control Preferences Scale (CPS) measures MDM preference between a pt dominant (PD) versus clinician dominant (CD) MDM relationship. How medical mistrust (MM) impacts MDM preferences has not been previously explored in a racially and socioeconomically diverse cohort of CA pts. In the setting of high MM, PD-MDM could make pts vulnerable to poorly informed decisions about their CA care. Methods: A cross-sectional survey was conducted at an urban safety net hospital (n = 95) and suburban cancer center (n = 105) in Philadelphia PA. Recruitment occurred in outpatient clinics using a purposeful sampling approach to maximize age, sex and race/ethnicity variability. Primary outcome was MDM preference by CPS [patient-dominant (PD-MDM) vs clinician/shared dominant (CD-MDM)]. MM was measured with the Group-Based Medical Mistrust Scale (GBMMS) measuring race-based MM (Likert 1-5) and the Medical Mistrust Index (MMI) measuring institutional MM (Likert 0-10). Demographic and SES data, health literacy (HL, Chew screener), and medical information preferences (HINTS adapted) were assessed. Effects were assessed by logistic regression odds ratios (OR) for PD-MDM vs CD-MDM for a 1 point change in MM score, adjusted for covariates. Study was IRB approved FCCC #22-8003. Results: Of 200 participants, median age was 62 yrs; 62% female, 46% AA, 15% Hispanic, 47% ≤high school graduate, 52% had income < $50K/yr. PD-MDM preference was reported by 55% (110/200). No demographic/SES factors predicted PD-MDM preference. Race-based MM and institutional MM were higher in PD-MDM pts compared to CD-MDM (GBMMS OR = 1.73 (95%CI 1.19-2.51), p = 0.004; MMI OR = 1.28 (95%CI = 1.05 = 1.55), p = 0.013). 27% (54/200) sample reported HL impairment, with 22% for PD-MDM vs 33% CD-MDM, p = 0.07. Pts w/PD-MDM preference were less likely to prefer ‘a lot’ of information from healthcare providers (64% vs 81%, p = 0.023) and had low preference (‘not at all’) for information from family/friends (p = 0.019) or media sources (e.g., newspaper, p = 0.005). With separate adjustment for HL, race, education, and information preferences, the association of race-based MM (GBMMS) with PD-MDM remained significant, with the largest impact from adjustment for HINTS: healthcare providers (adjusted OR = 1.51, p = 0.046). For institutional based MM (MMI), the association with PD-MDM remained significant with covariate adjustment except for HINTS: healthcare providers (adjusted OR = 1.20, p = 0.077). Conclusions: PD-MDM preference was identified in > 50% of CA pts surveyed and was strongly associated with both institutional and race-based MM, even after adjusting for HL, race, and education.