Using a Validated Predictive Algorithm to Identify Patients at High Risk for Hospital Readmission

Purpose:
The purpose of the study was two-fold. First, to identify the utility, sensitivity, and specificity of the LACE index in identification of patients at high risk for hospital readmission and secondly, to determine if additional patient-level risk factors enhanced readmission predictability.

Background/Significance:
Hospital readmission is an adverse patient outcome that is serious, common, costly, and considered a marker of poor quality of care. Hospital readmissions add an estimated cost of $15 to $17.4 billion a year to Medicare spending. For hospitals, identifying patients at high risk for readmission is a priority to improve quality of care and reduce costs.

Methods:
A retrospective study was conducted in a community hospital located in the Southeast. Patients(N = 598)discharged from 2010 through 2011 were included. Data were collected from the organization's database and manually abstracted from the electronic medical record at discharge date using a structured tool. Logistic regression models were fit for the probability of readmission within 30 days after discharge with two separate models. The first model used the LACE index as the predictor variable, and the second model used the LACE index with additional risk factors. The two models were compared to determine if the additional patient-level risk factors increased the model's predictive ability.

Results:
The results indicate both models have reasonable prognostic capability. The LACE index with additional risk factors did little to improve prognostication, while adding to the model's complexity. Findings indicated a one-point increase in the LACE index yielded an increase of 33.3% in the odds of a patient being readmitted, and a 49.4% increase in the odds of readmission with the LACE index with additional clinical factors. Findings support the use of the LACE index as a practical tool to identify patients at high risk for hospital readmission.

Conclusions and Implications for Practice:
Findings may be useful for developing interventions to reduce these readmission events, identifying risk factors of populations at readmission risk, developing disease specific registries for chronic disease management, and incorporating readmission risk assessments into existing clinical processes.