Predicting Non-home Discharge After Lung Surgery: Analysis of the General Thoracic Surgery Database

Anticipating the need for non-home discharge (NHD) enables improved patient counseling and expedites placement, potentially reducing length of stay and hospital readmission. The objective of this study was to create a simple, preoperative, clinical prediction tool for NHD using The Society of Thoracic Surgeons General Thoracic Surgery Database (STS GTSD).

The STS GTSD was queried for patients who underwent elective anatomic lung cancer resection between 2009 and 2019. Exclusion criteria included age <18 years, percentage predicted diffusion capacity of the lung for carbon monoxide <20% or >150%, N3 or M1 disease, incomplete datasets, and mortality. The primary outcome was defined as discharge to an extended care, transitional care, rehabilitation center, or another hospital. Multivariable logistic regression was used to select risk factors and a nomogram for predicting risk of NHD was developed. The approach was cross-validated in 100 replications of a training set consisting of randomly selected two-thirds of the cohort and a validation set of remaining patients.

A total of 35 948 patients from the STS GTSD met inclusion criteria. Final model variables used to derive the nomogram for NHD risk prediction included age (P < .001), percentage predicted diffusion capacity of the lung for carbon monoxide (P < .001), open surgery (P < .001), cerebrovascular history (P < .001), and Zubrod score (P < .001). The receiver operating characteristic curve, using sensitivities and specificities of the model, yielded area under the curve of 0.74. In 100 replicated cross-validations, out-of-sample area under the curve ranged from 0.72-0.76.

Using readily available preoperative variables, our nomogram prognosticates the risk of NHD after anatomic lung resection with good discriminatory ability. Such risk stratification can enable improved patient counseling and facilitate better planning of patients' postoperative needs.