Health Literacy Not Race Predicts End-of-life Care Preferences

Background: Several studies have reported that African Americans are more likely than whites to prefer aggressive treatments at the end of life.

Objective: Since the medical information presented to subjects is frequently complex, we hypothesized that apparent differences in end-of-life preferences and decision making may be due to disparities in health literacy. A video of a patient with advanced dementia may overcome communication barriers associated with low health literacy.

Design: Before and after oral survey.

Participants: Subjects presenting to their primary care doctors.

Methods: Subjects were asked their preferences for end-of-life care after they heard a verbal description of advanced dementia. Subjects then viewed a 2-minute video of a patient with advanced dementia and were asked again about their preferences. For the analysis, preferences were dichotomized into comfort care and aggressive care. Health literacy was measured using the Rapid Estimate of Adult Literacy in Medicine (REALM) and subjects were divided into three literacy categories: low (0-45, sixth grade and below), marginal (46-60, seventh to eighth grade) and adequate (61-66, ninth grade and above). Unadjusted and adjusted logistic regression models were fit using stepwise algorithms to examine factors related to initial preferences before the video.

Results: A total of 80 African Americans and 64 whites completed the interview. In unadjusted analyses, African Americans were more likely than whites to have preferences for aggressive care after the verbal description, odds ratio (OR) 4.8 (95% confidence interval [CI] 2.1-10.9). Subjects with low or marginal health literacy were also more likely than subjects with adequate health literacy to have preferences for aggressive care after the verbal description, OR 17.3 (95% CI 6.0-49.9) and OR 11.3 (95% CI 4.2-30.8) respectively. In adjusted analyses, health literacy (low health literacy: OR 7.1, 95% CI 2.1-24.2; marginal health literacy OR 5.1, 95% CI 1.6-16.3) but not race (OR 1.1, 95% CI 0.3-3.2) was an independent predictor of preferences after the verbal description. After watching a video of advanced dementia, there were no significant differences in the distribution of preferences by race or health literacy.

Conclusions: Health literacy and not race was an independent predictor of end-of-life preferences after hearing a verbal description of advanced dementia. In addition, after viewing a video of a patient with advanced dementia there were no longer any differences in the distribution of preferences according to race and health literacy. These findings suggest that clinical practice and research relating to end-of-life preferences may need to focus on a patient education model incorporating the use of decision aids such as video to ensure informed decision-making.