A logistic regression model showed that implementation of the care bundle had a significant effect on prevention of postoperative pneumonia (odds ratio, 0.16; 95% confidence interval: 0.01-0.94) after controlling the following confounding factors: sex, blood urea nitrogen, amount of blood loss, recurrent laryngeal nerve palsy, and preoperative hospital stay.