Factors Influencing Satisfaction with Patient-Controlled Analgesia Among Postoperative Patients Using a Generalized Ordinal Logistic Regression Model

Abstract

Purpose: The purpose of this study was to identify the factors affecting the satisfaction with patient-controlled analgesia (PCA) of patients using a generalized ordinal logistic regression model and to evaluate the difference in results of the ordinal regression from those of binary regression. Methods: The study design involved secondary analysis of electronic medical records from a single tertiary care hospital in Seoul, Korea. It included 2,409 patients treated with PCA for postoperative pain management after open or laparoscopic abdominal surgery. Binary logistic regression and generalized ordinal logistic regression were used to identify factors affecting satisfaction. Results: Binary logistic regression analysis showed that there was insufficient information for analysis. Generalized ordinal logistic regression revealed that sex, age, pain, PCA usage, and side-effects were common factors affecting PCA satisfaction. However, the effect of some factors affecting PCA satisfaction differed with the level of satisfaction. In open surgery patients, the effect of pain at 6 hours after surgery was significantly greater in the group with lower satisfaction. While, in the laparoscopic surgery patients, the effect of pain at 6-24 hours after surgery was significantly greater in the group with lower satisfaction. Conclusion: Generalized logistic regression may be an appropriate statistical method for analyzing ordinal data. Degree of postoperative pain and assessment interval are the most important factors associated with PCA satisfaction. Because the factors affecting PCA satisfaction were different for the two types of abdominal surgeries, customizing PCA to individual patients may potentially improve pain management and consequently increase PCA satisfaction. (C) 2020 Korean Society of Nursing Science. Published by Elsevier BV.