LAUSR

Introduction: Unplanned readmissions are associated with a high cost to health insurances and the incidence of preventable readmissions could be considered as a quality indicator for the initial hospital admission. We aimed to assess the predictive value for unplanned readmission of higher pain scores at discharge of the initial admission as well as of other pain and demographic characteristics. The documentation of significant associations would provide further support for a structured pain management policy. Methods: A retrospective analysis of a large single university hospital data-set of 33.122 admissions within a 13-month period allowed for the assessment of the predictive relationship of pain toward unplanned readmission at 7 and at 30 days after discharge through logistic regression, and of other characteristics through linear regression. Results: Pain scores at discharge of the initial admission were not significantly different (p > 0.05) with or without unplanned readmission and hence have no predictive value on the individual patient level. The prediction of the number of patients for each group, for example the number of patients that will be readmitted (size of the group), shows significance for pain at the moment of discharge (p_initial = 0.000), pain medication (p = 0.0044), and age (p = 0.0017). Pathology (p = 0.6151) and gender (p = 0.7029) have no significant predictive value. Conclusion: Pain as dichotomous variable upon discharge cannot be used as single risk predictor for unplanned readmission. However, the pain score at discharge in combination with the use of pain medication and age is a risk factor for the number of short-term unplanned readmissions.