Skip to main content
. 2016 Feb 1;19(2):174–182. doi: 10.1089/jpm.2015.0314

Table 2.

Multivariable Logistic Regression Analyses: Family Members' Characteristics and Family Members' Quality of Communication Scores (General and End-of-Life) as Predictors of Family Members' Decisional Conflict Scores

    Multivariable  
Outcome Covariate B SE p Exp (B)
Decisional conflict Quality of general communication −1.47 0.68 0.030 0.23
  Family member age −0.35 0.03 0.242 0.97
  Family member gender        
  Family member income        
  Family member education        
  Family member religious values        
    R2 0.23    
    n 49    
Decisional conflict Quality of end-of-life communication −0.34 0.14 0.014 0.71
  Family member age −0.03 0.03 0.286 0.97
  Family member gender        
  Family member income        
  Family member education        
  Family member religious values        
    R2 0.29    
    n 49    

The dependent variable in this analysis is decisional conflict coded so that 0 = “Little to no conflict” and 1 = “Moderate to high conflict. “The healthcare team didn't do this” was coded as a rating of 0. “I don't know” was coded as missing. The subscales: the General Communication Subscale (items 1–6) and the End-of-life Communication Subscale (items 7–13) had response options that ranged from 0 = “poor” to 10 = “absolutely perfect.” The middle of the scale with the value of 5 = “very good.”27