Readmission costs related to cardiac surgery. Analysis of risk factors and costs within six months after discharge using an administrative registry

AIM OF THE STUDY Analysis of readmission rate and costs of hospitalization after cardiac surgery using administrative data. BACKGROUND Medical procedures that require high technology have a higher rate of readmission over the time than other hospital interventions. Cardiac surgery patients are often complex and present high preoperative risk for subsequent disability or organ disfunction. These conditions are related to high risk of readmission for infection or organ failure. Clinical evidence and available literature suggest that readmission after cardiac interventions is strictly related to surgical complications, infections or poor medical status before surgery. The average reimbursement costs for cardiac surgery range from 18.000-21.000 euros in Italy, and fees for comorbidities or complications during the first episode of hospital stay. Subpopulation of cardiac patients at high risk of increase of costs may require further additional medical assistance in the following months due to poor health status and long-term complications. We assess the hypothesis that costs associated with initial treatment might be related to incremental costs for readmissions which occurred within six months after surgery. METHODS We analyzed the hospital administrative registry of Regione Piemonte, in northern Italy, and selected all patients who received cardiac surgery between January 1st and June 30th, 2009. The registry contains information related to each admission, including individual demographic characteristics, medical conditions and procedures according to ICD-9CM definitions. We followed hospital history of the selected patients for six months after discharge from the first intervention. We analyzed costs for all hospital stays and identified the ICD-9CM codes for morbidities and procedures other than cardiac surgery that are clinical relevant as known risk factors for complications and readmission. Hazard models were performed to identify predictors for recurrent hospitalization and regression analysis were used to evaluate initial hospitalization and readmission costs related to patients’ conditions. RESULTS Preliminary results suggest that the use of administrative registries compiled with ICD-9CM codes may in fact indicate the risk of readmission after cardiac surgery thus confirming the data from clinical literature. The diagnosis and procedure codes used for administrative purposes might be a reliable indicator for the actual risk according to clinical evidence. The initial length of stay is also related to an increase in readmission rate for any causes within six months from discharge, and known medical risk factors can be surrogated by ICD-9CM codes to predict an increase of hospitalization and costs. The implementation of our model with a broad registry analysis could yield more robust results, that could eventually be validated as a useful tool to predict the financial medium-term need for cardiac surgery in community patients.