CSA Home
Central Surgical Association

Back to 2020 Abstracts


Defining The Relative Contribution Of Socioeconomic Characteristics To Patient Outcomes In The Model Of 30-day Readmissions After CABG
*Haroon Janjua, *Tara M Barry, *Evelena Cousin-Peterson, Paul C Kuo
USF, Tampa, FL

OBJECTIVE(S): Patient outcomes following healthcare interventions may be dependent on a variety of factors: patient, surgeon, hospital, information technology, temporal, cultural and socio-economic factors, among others. To date, studies have tended to focus only on individual components rather than the entirety of the global healthcare macroenvironment. In this study, we characterize the relative contribution of each of these factors using a model of 30-day readmission following CABG. METHODS: The HCUP, AHA, HIMMS and DCI from 2010-2013 were linked to reflect patient/temporal, hospital, hospital IT and socioeconomic characteristics for FL, IA, MA, MD, NY and WA. A stepwise logit model was run for factor selection using a combination of backwards elimination and forward selection; the final set contained 16 patient (HCUP), 5 hospital (AHA), 13 hospital IT (HIMMS) and 3 socioeconomic/cultural variables (HCUP+DCI ). Logistic regression, random forest, decision tree, gradient boosting, KNN classification, and XgBoost tree models were implemented. Modeling results were compared on the basis of predictive accuracy, sensitivity, specificity and AUC. Decision tree performed best: AUC =0.66, accuracy=0.87, sensitivity=0.89, specificity=0.68 and was selected for further analysis. GBM class plots were used to quantify factor contribution. RESULTS: The model had 45352 pts, 54096 admissions and 16.2% 30-day readmission rate after CABG. The top 10 predictors were: total diagnoses on d/c record, disposition at discharge, number of chronic conditions, total procedures, adults without high school diploma, primary payor method, median household income, patient location (urban-rural), admission type and AHRQ comorbidity-renal failure. The top three socioeconomic predictors were: Adults Without High School Diploma, Patient Location (urban vs rural designation), Estimated State Median Household Income. The relative contribution of patient/temporal, socioeconomic, hospital IT, and hospital factors to readmission is: 85%, 9%, 4% and 2%, respectively. CONCLUSIONS: In this model of 30 day readmission after CABG, socioeconomic factors' contribution is substantive but lags significantly behind patient/temporal factors. In contrast, hospital and hospital IT factors contributed even less. With ever increasing availability of data, identification of contributors to patient outcomes within the overall healthcare macroenvironment will allow prioritization of interventions.


Back to 2020 Abstracts
Milwaukee
Milwaukee
Milwaukee
Milwaukee