Background: Length of hospital stay (LOS) greatly influences costs of total hip arthroplasty (THA). LOS is, among others, dependent on patient comorbidity. American Society of Anesthetists (ASA) scores are traditionally used as comorbidity measure, but simple medication-based measures may suffice. Therefore, the objective of the study is to determine if medication-based measures (count of medication and the medication-based index Chronic Disease Score [CDS]) have a surplus value compared to a diagnosis-based measure for comorbidity (ASA score) to predict LOS in THA.
Methods: A retrospective cohort study. THA patients from 3 hospitals were included for 2009-2013, and demographic data, comorbidity (ASA score), and medication use (count of medication and CDS) were collected. A binomial regression model was used to compare the predictive ability of the medication-based and diagnosis-based measures.
Results: A total of 2,282 THA patients were included. Number of medications proved a statistical significant predictor of LOS. LOS increased 6% (incidence rate ratios: 10.06 [confidence interval: 1.03-1.09]) with each medication. Neither the CDS nor the ASA score had a statistically significant relation to LOS.
Conclusion: The medication-based measure count of medication is a better predictor for LOS in THA than the CDS and ASA score. This simple measure also appears to have more clinical relevance, as it has a higher range in scores than diagnosis-based indexes and is easily applicable.
Keywords: ASA classification; comorbidity; complexity; length of stay; medication; total hip arthroplasty.
Copyright © 2016 Elsevier Inc. All rights reserved.