0
    164
    views

    ICJR ABSTRACTS: Reconstructing Risk Scores in Primary Joint Reconstruction

    At the 7th Annual ICJR South Hip & Knee Course, 6 awards were given to orthopaedic residents and fellows who submitted the best research abstracts in hip and knee replacement. The abstract Reconstructing Risk Scores in Primary Joint Reconstruction received the Best Question Award: Knee Arthroplasty and was presented at the meeting by Chris Hoedt, MD.

    RELATED: Register for the 8th Annual ICJR South Hip & Knee Course

    Authors

    Chris Hoedt, MD; Alec Kellish, BS; Michael Lucian,i BS; Alisina Shahi, MD; Jack Shilling, MD; Tae Won Kim, MD; and Christina Gutowski, MD

    Introduction

    Identifying patients at highest risk for complications after total joint reconstruction (TJR) is essential to reduce costs, improve outcomes, and properly inform patients. The Charlson Comorbidity Index (CCI) & Elixhauser Comorbidity Measure (ECM) are often used in arthroplasty studies, but their predictive ability is only fair. There are no known studies of the predictive ability of their aggregate variables.

    Methods

    We reviewed the 2010-2015 National Inpatient Sample (NIS) for patients who underwent primary TJR. Comorbidities from CCI & ECM were analyzed for their predictive ability to identify patients who experienced myocardial infarction (MI), pneumonia (PNA), sepsis, pulmonary embolism (PE), death, mechanical complications (MC), prosthetic joint infection (PJI), higher charges, longer length of stay, and discharge to a facility. Backward elimination regression (BER) & Least Absolute Shrinkage and Selection Operator (LASSO) regression were performed on these variables and compared to the predictive ability using the area under the curve analysis of CCI & ECM.

    Results

    CCI was equivocal to BER & LASSO as good predictors of MI, a fair predictor (AUC 0.7-0.8) for death after THA, but very poor-poor for all other outcomes. ECM’s predictive ability was poor-fair for all outcomes, except for MC & PJI after TKA it was very poor. BER & LASSO were excellent predictors for PE & death after THA, good predictors for sepsis, death after TKA, and PJI & PNA after THA. All models failed at predicting total charges.

    Conclusions

    Aggregating the variables of CCI & ECM allowed for enhanced identification of patients at higher risk for adverse outcomes after primary TJR.

    Click the image above to watch Dr. Hoedt’s presentation of this Best Question Award-winning abstract.

    Author Information

    Chris Hoedt, MD; Alec Kellish, BS; Michael Lucian,i BS; Alisina Shahi, MD; Jack Shilling, MD; Tae Won Kim, MD; and Christina Gutowski, MD, are from the Cooper Bone & Joint Institute at the Cooper Medical School of Rowan University in Camden, New Jersey.