| Keywords |
Problem drinking ; Workers ; Decision tree |
| Abstract |
This study was done to develop a predictive model on problem drinking that will provide a new paradigm of problem drinking prevention program for employees. The cross-sectional descriptive study was done with a self administered questionnaire survey method. Participants included in the final analysis were 296 workers. The questions were on demographic factors, job related factors, and AUDIT. Decision tree analysis using SPSS Statistics 21 program was applied to build an optimum and significant predictive model on problem drinking of workers. From the data analysis, the predictive model for factors related to workers' problem drinking presented with 4 pathways. The most important predictive factor was gender, and the following factors were on-duty hour, income and job type in order. In terms of the accuracy of the predictive model, overall predictive percent incorrect was 27.7%, and predictive percent corrects of normal drinking and problem drinking were 67.5% and 78.2% respectively. According to the results, it is necessary to provide a differentiated program depending on the predictive factors of workers' problem drinking. |