BACKGROUND: To evaluate unobserved heterogeneity in trauma patients' desire for autonomy in medical decision making (DAD). MATERIAL/ METHODS: This cross-sectional study at an inner-city emergency department screened 1009 patients for DAD (using the Decision Making Preference Scale of the Autonomy Preference Index), education level, and substance use. To investigate unobserved heterogeneity, a covariate adjusted finite mixture model was established. Model fit was evaluated with the Bayesian Information Criterion. RESULTS: The median age of participants was 32 years (range, 18-84 years) and 62% were male. Unobserved heterogeneity explained more variance in DAD than did sex, age, or substance use, but less variance than level of education. The best overall model fit was found with 3 latent subpopulations: 53.3% of patients with low DAD, 35.6% of patients with medium DAD, and 11.1% of patients with high DAD. Female sex and level of education showed a positive association; higher age and substance use showed a negative association with patients' DAD. CONCLUSIONS: Apart from a negative association with substance use and the known associations with sex and level of education, trauma patients' DAD showed substantial variability between individuals, and this variability could not be explained by these factors.
BACKGROUND: To evaluate unobserved heterogeneity in traumapatients' desire for autonomy in medical decision making (DAD). MATERIAL/ METHODS: This cross-sectional study at an inner-city emergency department screened 1009 patients for DAD (using the Decision Making Preference Scale of the Autonomy Preference Index), education level, and substance use. To investigate unobserved heterogeneity, a covariate adjusted finite mixture model was established. Model fit was evaluated with the Bayesian Information Criterion. RESULTS: The median age of participants was 32 years (range, 18-84 years) and 62% were male. Unobserved heterogeneity explained more variance in DAD than did sex, age, or substance use, but less variance than level of education. The best overall model fit was found with 3 latent subpopulations: 53.3% of patients with low DAD, 35.6% of patients with medium DAD, and 11.1% of patients with high DAD. Female sex and level of education showed a positive association; higher age and substance use showed a negative association with patients' DAD. CONCLUSIONS: Apart from a negative association with substance use and the known associations with sex and level of education, traumapatients' DAD showed substantial variability between individuals, and this variability could not be explained by these factors.