Intensive care unit (ICU) readmission, also called ICU rebound, or unplanned ICU readmission, usually occurs when a patient suddenly deteriorates during the same hospitalization and needs to return to the ICU for treatment (Coughlin et al., 2018). Due to different follow-up times, ICU readmission rates vary widely, as reported in a systematic review ranging from 1.3 % to 13.7 % (Elliott et al., 2014). Patients who return to the ICU are associated with poor prognosis, including higher mortality, extended hospital stays, and higher adverse events rate (Grochla et al., 2019, Lee et al., 2022, Ponzoni et al., 2017, Zhong et al., 2021). Moreover, increasing hospitalization costs would significantly burden patients' families and healthcare systems (Brown et al., 2012, Markazi-Moghaddam et al., 2020). Many countries are concerned about this issue and have included the ICU readmission rate in their healthcare quality indicators (Zenghui et al., 2022, Nates et al., 2016, Rhodes et al., 2012).

Discharge decisions for critically ill patients usually rely on physicians' personal experience and judgment, which is highly subjective and easily affected by bed occupancy, resulting in premature discharge (Ponzoni et al., 2017, Zhou et al., 2022). Up to 42 % of patients return to the ICU due to it, and 11 % of readmissions are preventable, representing an opportunity to improve patient care (Al-Jaghbeer et al., 2016, Sauro et al., 2020, van Sluisveld et al., 2013). Early identification of readmission risk may help healthcare professionals to make targeted interventions and appropriate decisions (Cholack et al., 2021, Ponzoni et al., 2017).

Existing clinical predictive scoring systems, such as the Acute Physiologic and Chronic Health Evaluation (APACHE) system, the updated Simplified Acute Physiologic Score (SAPS), and the Mortality Prediction Model (MPM0), mainly focus on disease severity, in-hospital mortality, and do not determine individual treatment options or outcomes (Mark A Kelley, 2021). Recently, many risk assessment tools have been developed, but their type and reliability are unclear. Therefore, the purpose of this study is to review and evaluate existing risk assessment tools for ICU readmission.