Hi, my name is Sanket Mehta, and I'm coming from the Department of Orthopedic Surgery at the Columbia University Medical Center, here to talk to you about the discriminative ability of different comorbidity indices to predict adverse outcomes in traumatic ankle fracture. Our only disclosure is that this project was funded by a grant from the Orthopedic Science Research Fund. Traumatic ankle fractures are the most common lower extremity fracture, and there's over such fractures evaluated in the U.S. emergency rooms annually. There's been a lot of literature that's identified individual comorbid medical conditions such as COPD, vascular disease, and smoking as risk factors for adverse outcomes after ankle fracture surgery, but capturing the cumulative impact of these individual medical conditions has continued to be a challenge. Comorbidity indices such as the modified Charlson, the Oegshauser, and the modified five-factor frailty index are cumulative measures of comorbidities that have been shown to predict adverse events in procedures such as hip fracture surgery and posterior lumbar fusion. Although these indices have been validated individually in orthopedic procedures, very few studies have examined the clinical predictive performance power of these measures in a comparative context. Our project looked to examine the clinical relevance of modified Charlson, the Oegshauser, and the modified five-factor frailty indices in terms of their discriminative ability to predict in-hospital mortality and adverse events after surgical management of traumatic ankle fracture. We looked at all adult patients registered in the National Trauma Data Bank experiencing malleolar fracture and undergoing surgical management from the years 2011 to 2016. The modified Charlson, Oegshauser, and frailty indices were calculated using ICD-9 and ICD-10 coding algorithms, and we looked at in-hospital mortality, severe adverse events, minor adverse events, infectious events, and any adverse events using area under the curve analyses. Our findings showed that the modified Charlson indice was the most discriminative for in-hospital mortality, whereas the Oegshauser comorbidity measure was most discriminative for any adverse event, major adverse event, minor adverse event, and any infectious events during hospitalization. In terms of demographic variables, we examined age, gender, and race. We saw that age was most discriminative for in-hospital mortality as well as minor adverse events, whereas gender was most predictive of any adverse event, major adverse events, and infectious events. Looking at a combination analysis of the most predictive comorbidity measure in conjunction with the most predictive demographic variable, we found that for all five outcomes, a combination of the comorbidity measure and the demographic measure outperformed the individual comorbidity measure versus the demographic measure. In addition, we've displayed a few selective curves here for death and adverse events, showing that the modified Charlson plus the age is indeed more discriminative than the individual parts. Similarly, for any adverse event, the Oegshauser plus gender is more discriminative than the sum of its parts. In terms of limitations, we have several. This study was a retrospective approach, and the National Trauma Data Bank is quite dependent on coding information by various centers across the United States, which can result in selection bias and variations in data quality. In addition, we only account for short-term events during initial hospitalization and do not account for later hospitalizations relating to the same traumatic ankle fracture or outpatient follow-up. In conclusion, patients with traumatic ankle fracture undergoing surgical management, the modified Charlson demonstrates robust discriminative ability for mortality, and the Oegshauser demonstrates robust discriminative ability for multiple adverse events, both major, minor, and infectious. The use of these indices, along with easily accessible demographics such as age, can result in further improvements in clinical predictive power. The use of these easily calculable indices may assist in the identification of patients at risk, helping in further optimizing stratification algorithms, informing expectations and guiding reimbursement. Thank you for your time.

comorbidity indices

adverse outcomes

traumatic ankle fracture surgery

predictive variables

identifying at-risk patients