Predictive model of readmission to internal medicine wards

Prevention of readmission to hospital is considered an outstanding example of a cost-effective practice. Our aim was to derive and validate a clinically useful index to quantify the risk of readmission among patients discharged from Internal Medicine departments.

We analysed hospital Basic Minimum Data Sets (BMDS) recorded between 2006 and 2008 to determine patterns of rehospitalization. Multivariate statistical analysis of routinely collected data was used to develop an algorithm ('SEMI INDEX') to identify patients predicted to have the highest risk of readmission in the 30 days following discharge. The algorithm was developed by using data from admissions in 2006-2007, for four age subgroups. Coefficients for the most powerful and statistically significant variables were applied against episodes recorded in 2008 to validate the findings of the algorithm developed from the first sample.

Of the 999,089 internal medicine admissions in Spain during 2006-2007, 12.4% were rehospitalized within 30 days. The key factors that predicted subsequent admission included male sex, length of stay, comorbidity of the patient, and some clinical conditions. There were small but relevant differences among the different age subgroups.

Readmissions to Internal Medicine departments are prevalent (12.4%). The SEMI INDEX can be used to assess accurately the risk of readmission within 30 days after discharge.