External validity and clinical usefulness of a risk prediction model for 30 day unplanned hospitalization in patients receiving outpatient parenteral antimicrobial therapy.

Abstract

OBJECTIVES: Outpatient parenteral antimicrobial therapy (OPAT) is increasingly used to treat a variety of infections. However, hospital readmissions remain relatively common. We examined the external validity and clinical usefulness of a previously derived risk prediction model for 30 day unplanned hospitalization in patients receiving OPAT. METHODS: A retrospective cohort study was conducted at two large teaching hospitals in the UK. The design comprised quasi-external temporal validation on patients from the same OPAT setting as the model development, and broader external validation on patients from a different setting. The model predictors were age, prior hospitalizations in the preceding 12 months, Charlson comorbidity score, concurrent IV antimicrobial therapy, type of infection and mode of OPAT treatment. Discriminative ability, calibration and clinical usefulness were assessed. RESULTS: Data from 2578 OPAT patients were analysed. The rates of 30 day unplanned hospitalization were 11.5% (123/1073), 12.9% (140/1087) and 25.4% (106/418) in the model derivation, temporal validation and broader external validation cohorts, respectively. The discriminative ability of the prediction model was adequate on temporal validation (c-statistic 0.75; 95% CI: 0.71-0.79) and acceptable on broader validation (c-statistic 0.67; 95% CI: 0.61-0.73). In both external cohorts, the model displayed excellent calibration between observed and predicted probabilities. Decision curve analysis showed increased net benefit across a range of meaningful risk thresholds. CONCLUSIONS: A simple risk prediction model for unplanned readmission in OPAT patients demonstrated reproducible predictive performance, broad clinical transportability and clinical usefulness. This model may help improve OPAT outcomes through better identification of high-risk patients and provision of tailored care.