Predictors of 30-day readmission in patients with bloodstream infection: a population-based cohort study.

PURPOSE: To identify predictors of hospital readmissions within 30 days after discharge following a bloodstream infection (BSI).

METHODS

Population-based study of 30,843 patients surviving after their first BSI episode, 2007-2016. Using multivariate competing risk regression analyses, we estimated subhazard ratios (sHRs) and 95% confidence intervals (CIs) for all-cause, infection-related, and unplanned readmissions. Predictors included sex, age, comorbidities, causative microorganisms, length of stay (LOS), and prior hospital admissions.

RESULTS

Within 30 days of discharge, 9655 patients (31.3%) were readmitted of which 3333 (34.5%) were infection-related and 7725 (80.0%) unplanned. Higher risks of all-cause readmissions were observed for males (sHR [95% CI]: 1.07 [1.03-1.12]), those aged 0-15 years (1.41 [1.26-1.58]), and patients with several comorbidities, especially malignancies and liver diseases (sHR 1.26-2.09). BSIs caused by most microorganisms other than Escherichia coli also predicted an increased readmission risk (sHR 1.11-1.39). LOS in the first, third, and fourth quartiles also predicted a higher readmission risk (sHR 1.26, 1.14, and 1.25, respectively). Patients with prior hospitalization also had a higher readmission risk, within 31-365 (sHR 1.22 [1.16-1.28]) and 1-30 days (1.70 [1.61-1.79]). Results were similar for unplanned readmissions. For infection-related readmissions, only LOS in the first quartile (1-5 days) was associated with a higher readmission risk (1.67 [1.52-1.85]). Over the study period, the median LOS decreased, whereas readmission rates increased.

CONCLUSION

Nearly one-third of patients discharged alive after a BSI were readmitted within 30 days. Further studies are needed to determine whether some of these readmissions may be avoidable.