Data were entered through the REDCap platform, hosted by the University of Witwatersrand.11 (link),12 (link) Data were exported directly from REDCap to STATA V.16.1 for analysis. Categorical variables collected were coded numerically and continuous variables were cleaned and standardised in terms of units measured in STATA V.16.1 in preparation for analysis.
Descriptive statistics were used to summarise the characteristics of the data set. A bivariate analysis (an analysis conducted to examine the relationship between two variables) was used to examine the relationship between each individual and workplace variable with each of the three burnout subscales of Emotional Exhaustion, Depersonalisation and Personal Accomplishment. The burnout subscales were the dependent variable. Statistically significant associations were examined using the crude estimate of 0.25. Multiple linear regression was used to examine any factors associated with burnout outcomes. A multiple linear regression model was built to account for confounding, using stepwise variable selection. A significance of 5% was used for statistical significance testing.
Descriptive statistics were used to summarise the characteristics of the data set. A bivariate analysis (an analysis conducted to examine the relationship between two variables) was used to examine the relationship between each individual and workplace variable with each of the three burnout subscales of Emotional Exhaustion, Depersonalisation and Personal Accomplishment. The burnout subscales were the dependent variable. Statistically significant associations were examined using the crude estimate of 0.25. Multiple linear regression was used to examine any factors associated with burnout outcomes. A multiple linear regression model was built to account for confounding, using stepwise variable selection. A significance of 5% was used for statistical significance testing.
Free full text: Click here
