Statistical Analyses
MBI results were analyzed by descriptive statistics (Table 1). ANOVAs and Tukey’s post hoc tests determined significant differences in job demand, involvement in PSEs, and burnout in inpatient nurses, ambulatory nurses, MDs and APPs. Statistical analyses were performed using SAS, version 9.4 (SAS Institute).
To evaluate associations among MBI dimensions and NASA-TLX job demands, relative weights analysis (RWA) was conducted using RWA-Web.24,25 Relative weight refers to the proportionate contribution of each predictor variable to the variance accounted for in a regression model while considering its unique contribution and that when combined with other variables.24 RWA is superior to traditional regression analysis as it generates a more accurate ranking of each predictor’s unique contribution to an outcome variable. RWA provides accurate estimates of the relative importance of predictors in cases of high levels of collinearity.26 Our study conducted 9 RWA (3 cohorts×3 burnout dimensions). In all analyses, 95% confidence intervals for each individual relative weight and all corresponding significance tests were based on bootstrapping with 10,000 replications.