3080 — Validating Readmissions in the VA Risk-Standardized Readmission Rate

Horstman MJ, HSR&D/OAA Post-Doctoral Fellow, Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center, Houston TX; Petersen LA, Director, Center for Innovations in Quality, Effectiveness, and Safety, Michael E. DeBakey VA Medical Center, Houston, TX;

Objectives:
Hospital readmissions are used as an indicator of hospital quality by VA and other entities, yet the accuracy of data used in this metric has not been validated. We sought to validate the readmission data included in the 30-day risk-standardized readmission rate.

Methods:
We used the VA Inpatient Evaluation Center (IPEC) Drill Down Report to identify patients included in the Houston VA readmission metric from Q4 FY2011 to Q3 FY2014. Patients had an index admission with a coded principal diagnosis of acute myocardial infarction, heart failure, or pneumonia and were readmitted to the hospital within 30 days of discharge. We selected a simple random sample of 120 patients from 286 patients. Charts were reviewed to determine if the index admission met the inclusion criteria based on principal diagnosis and if any exclusion criteria were met. The data definition for the risk-standardized readmission rate was obtained from the VHA Service Support Center (VSSC).

Results:
We found that 37% (44/120) of patients included in the risk standardized readmission metric either did not meet inclusion criteria or met one of the exclusion criteria. The majority of misclassified patients were misclassified due to errors in coding with chart review identifying 27 patients who had an incorrect coded principal diagnosis. Three patients met exclusion criteria for leaving against medical advice, but were labeled as regular discharges in the PTF file. Two additional patients met exclusion criteria with less than 24-hour hospital stays and 12 patients met exclusion criteria for death within 30 days of discharge. Six patients had planned readmissions for chemotherapy or scheduled procedures.

Implications:
Our findings call into question the accuracy of data used in reporting 30-day readmission rates in VHA. From a systems perspective, several factors contributed to the inaccurate inclusion and exclusion of readmitted patients. Planned readmissions are not excluded from the VA readmission metric and thus, currently count against hospitals. Additional consideration should be given to the appropriateness of including planned readmissions in the risk-standardized readmission rate.

Impacts:
The questionable validity of IPEC readmission data limits the utility of the risk-standardized readmission rate as a measure of hospital quality.