Background and Aims

The 30-day hospital re-admission rate is an outcome of growing attention due to its use as a metric of quality of care as well as its association with increased healthcare expenditures.

Patients with Acute Stroke (AS) are at high risk of hospital re-admission that is often associated with increased mortality rate, greater levels of disability and higher costs as compared with the initial stroke events.

The aim of our study was to assess the validity of a predictive model of unplanned 30-day hospital re-admissions and elaborate a 30-day re-admission risk score for patients with AS.

Methods

We conducted a retrospective study on adult patients with AS who were admitted to Fondazione Policlinico Universitario A. Gemelli and discharged alive between January 1st, 2014 and December 31th, 2019. Data collection included demographic features, clinical and laboratory parameters, diagnostic and invasive procedures as well as discharge and re-admission data.

Results

Of the 7599 screened patients with AS, 4561 patients met the inclusion criteria. Of these patients, 361 patients (7.91%) were readmitted within 30 days from discharge. After the identification of 7 predicting early readmission variables by machine learning model, the risk score was calculated (Figure 1A). Based on this risk score, our patients were stratified in low, medium and high risk groups (Figure 1B).

Conclusions

The identification of risk factors that contribute to early hospital re-admission after AS and the stratification of AS patients at discharge can provide clinicians with a useful tool to plan a personalized follow-up.