R. N. Mubang1,3, J. C. Stoltzfus6, B. A. Hoey3, C. D. Stehly2,3, D. C. Evans4, C. Jones4, T. J. Papadimos5, M. S. Cohen1, J. Grell1, W. S. Hoff1,3, P. Thomas1,3, J. Cipolla1,3, S. P. Stawicki2 1St. Luke’s University Health Network,Department Of Surgery,Bethlehem, PA, USA 2St. Luke’s University Health Network,Department Of Research & Innovation,Bethlehem, PA, USA 3St. Luke’s University Health Network,Regional Level I Trauma Center,Bethlehem, PA, USA 4The Ohio State University College Of Medicine,Department Of Surgery,Columbus, OH, USA 5The Ohio State University College Of Medicine,Department Of Anesthesiology,Columbus, OH, USA 6St. Luke’s University Health Network,The Research Institute,Bethlehem, PA, USA
Introduction: Traditional injury severity assessment is insufficient in estimating the morbidity and mortality risk for older (≥45 years) trauma patients. Commonly used tools involve complex calculations or tables, do not consider comorbidities, and often rely upon data that is not available early in the trauma patient’s hospitalization. The Comorbidity-Polypharmacy Score (CPS), a simple sum of all pre-injury medications and comorbidities, has been found to independently predict morbidity and mortality in previously published reports. These studies, however, have been limited by relatively small sample sizes. We hypothesized that CPS is indeed associated with outcomes in older trauma patients and sought to validate this association using a large administrative dataset.
Methods: After IRB approval, a retrospective study of patients aged 45 years and older was performed using the administrative database from a Level I Trauma Center. The study period was Jan 1, 2008 to Dec 31, 2013. Abstracted data included patient demographics, injury characteristics and severity, Glasgow Coma Scale (GCS), hospital (HLOS) and intensive care unit lengths of stay (ILOS), morbidity, post-discharge destination, and in-hospital mortality. Univariate analyses were conducted with mortality, all-cause morbidity, and discharge destination as primary end-points. Variables trending toward statistical significance (p<0.20) were subsequently included in multivariate logistic regression. Data are presented as adjusted odds ratios (AOR), with p<0.05 denoting statistical significance.
Results: A total of 5,839 complete patient records were analyzed. Average patient age was 68.5±15.3 years (52% male, 89% blunt mechanism), with mean GCS 14.3. Mean HLOS increased from 4.27 days for patients with CPS≤15 to 5.55 days for those with CPS≥16 (p<0.01). Likewise, ILOS increased from a mean of 0.67 days for patients with CPS≤7 group to 1.33 days for those with CPS≥22 (p<0.01). Independent predictors of mortality included age (AOR 1.05, p<0.01), CPS (per-unit AOR 1.08, p<0.02), GCS ≤8 (AOR 48.16, p<0.01), and ISS (per-unit AOR 1.08, p<0.01). Independent predictors of all-cause morbidity included age (AOR 1.02, p<0.01), GCS ≤8 (AOR 5.4, p<0.01), ISS (per-unit AOR 1.09, p<0.01), and CPS (per-unit AOR 1.04, p<0.01). CPS did not independently predict need for discharge to a facility.
Conclusion: This study confirms that CPS is an independent predictor of morbidity and mortality in older trauma patients. However, CPS was not found to be independently associated with need for discharge to a facility in the current dataset. Prospective multicenter studies are warranted to evaluate the use of CPS as a predictive and interventional tool, with special focus on correlations between specific pre-existing conditions, potential pharmacologic interactions, and morbidity/mortality patterns.