Judicious use of critical care resources by predicting the need for routine ICU admission following esophagectomy

With advances made in the care for esophagectomy patients, the need for routine postoperative intensive care unit (ICU) admission needs reassessment. We developed a preoperative prediction model to distinguish patients requiring routine ICU admission from those who can be cared for in a post-anesthesia care unit (PACU). This retrospective cohort study included consecutive adults undergoing elective esophagectomy between January 2011 and June 2021 in the UMC Utrecht. Firth's corrected multivariable logistic regression was used for model development and internal validation using bootstrapping was performed to obtain optimism-corrected performance metrics. Among the 619 patients included, 380 (61%) received critical care support beyond the first morning following surgery: 83 (13%) were on invasive mechanical ventilation and 338 (55%) needed cardiovascular support (with 174 [28%] receiving only low-dose norepinephrine). Predictors retained in the final model included age, diabetes mellitus, hemoglobin level, kidney function, forced expiratory volume in 1 second, tumor stadium, type of neoadjuvant therapy and surgical approach. Discrimination was acceptable (adjusted c-statistic 0.67, 95% CI 0.62-0.71) with good calibration (O:E ratio 1.00). Using the model, approximately 50% of ICU beds could be conserved, at the cost of misallocating 22% of patients to a PACU (with only 12% of PACU-allocated patients requiring mechanical ventilation). Between one- and two-thirds of elective esophagectomy patients do not need routine ICU admission, depending on whether hemodynamic support can be provided in another high-dependency unit. Our model can help rationalize perioperative patient allocation and reduce ICU bed claims by roughly half.