Analysis of intra-operative variables as predictors of 30-day readmission in patients undergoing glioma surgery at a single center

Iahn Cajigas, Anil K. Mahavadi, Ashish H. Shah, Veronica Borowy, Nathalie Abitbol, Michael E. Ivan, Ricardo J. Komotar, Richard H. Epstein

Research output: Contribution to journalArticle

Abstract

Purpose: Reducing the time from surgery to adjuvant chemoradiation, by decreasing unnecessary readmissions, is paramount for patients undergoing glioma surgery. The effects of intraoperative risk factors on 30-day readmission rates for such patients is currently unclear. We utilized a predictive model-driven approach to assess the impact of intraoperative factors on 30-day readmission rates for the cranial glioma patient. Methods: Retrospectively, the intraoperative records of 290 patients who underwent glioma surgery at a single institution by a single surgeon were assessed. Data on operative variables including anesthesia specific factors were analyzed via univariate and stepwise regression analysis for impact on 30-day readmission rates. A predictive model was built to assess the capability of these results to predict readmission and validated using leave-one-out cross-validation. Results: In multivariate analysis, end case hypothermia (OR 0.28, 95% CI [0.09, 0.84]), hypertensive time > 15 min (OR 2.85, 95% CI [1.21, 6.75]), and pre-operative Karnofsky performance status (KPS) (OR 0.63, 95% CI [0.41, 0.98] were identified as being significantly associated with 30-day readmission rates (chi-squared statistic vs. constant model 25.2, p < 0.001). Cross validation of the model resulted in an overall accuracy of 89.7%, a specificity of 99.6%, and area under the receiver operator curve (AUC) of 0.763. Conclusion: Intraoperative risk factors may help risk-stratify patients with a high degree of accuracy and improve postoperative patient follow-up. Attention should be paid to duration of hypertension and end-case final temperature as these represent potentially modifiable factors that appear to be highly associated with 30-day readmission rates. Prospective validation of our model is needed to assess its potential for implementation as a screening tool to identify patients undergoing glioma surgery who are at a higher risk of post-operative readmission within 30 days

Original languageEnglish (US)
JournalJournal of neuro-oncology
DOIs
StateAccepted/In press - Jan 1 2019

    Fingerprint

Keywords

  • Cranial
  • Glioma
  • Hospital readmission
  • Neurosurgery
  • Predictive modeling

ASJC Scopus subject areas

  • Oncology
  • Neurology
  • Clinical Neurology
  • Cancer Research

Cite this