Developing Mixed-effects Models to Optimize Prediction of Postoperative Outcomes in a Modern Sample of Over 450,000 Patients Undergoing Elective Cervical Spine Fusion Surgery

Shahrestani, Shane; Brown, Nolan J.; Yue, John K.; Tan, Lee A

doi:10.1097/bsd.0000000000001512

Cited by 1 publication

(1 citation statement)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hospital identifiers were included as random effects. [26][27][28][29] Model performance was assessed using the area under the curve (AUC) of each ROC curve, acting as a proxy. An AUC of 0.50 indicates a random guess, whereas values exceeding 0.70 are considered acceptable.…”

Section: Discussionmentioning

confidence: 99%

A contemporary analysis of surgical ligation versus endovascular embolization in patients with intracranial dural arteriovenous fistulas: a propensity score–matched and mixed-effects model study

Shahrestani,

Michel,

Aguilera-Pena

et al. 2024

Neurosurgical Focus

View full text Add to dashboard Cite

OBJECTIVE Intracranial dural arteriovenous fistulas (dAVFs) are rare vascular lesions that can be asymptomatic or can lead to devastating hemorrhage based on the dAVF’s aggressiveness. Several approaches can be taken to treat dAVFs, such as endovascular embolization and surgical ligation. However, very few studies have evaluated the influence of surgery compared to endovascular approaches on patient outcomes. This study was performed to analyze the clinical characteristics and outcomes of patients who underwent treatment for intracranial dAVF in which either endovascular embolization or microsurgical ligation was used. METHODS The Nationwide Readmissions Database was reviewed for all patients who underwent treatment for dAVFs (n = 18,152) between 2016 and 2019. Patients who received only surgical ligation or endovascular embolization (i.e., not both) were included. Variables regarding demographics, clinical outcomes, and healthcare utilization were queried. Primary outcome measures were nonroutine discharge, 1-year readmission, top quartile length of stay (LOS), and top quartile of inpatient all-payer cost. Propensity score matching was performed to evaluate the influence of either surgery or embolization on patient outcomes. Receiver operating characteristic (ROC) curves were created for each outcome measure. The area under the curve (AUC) of each ROC was used to estimate mixed-effects model performance. RESULTS Following propensity score matching, 127 and 113 patients made up the surgical ligation and endovascular embolization cohort, respectively. There were no differences found in age (p = 0.16), sex (p = 0.57), or average Elixhauser Comorbidity Index (p = 0.32). Patients receiving surgical ligation had lower odds of readmission (OR 0.37, p = 0.028) and greater odds of nonroutine discharge (OR 2.21, p = 0.03) compared to patients who underwent endovascular embolization. The authors found no differences in the top quartile of LOS (p = 0.84), top quartile of cost (p = 0.38), or mortality (p > 0.99) between cohorts. ROC curves revealed that the mixed-effects models inclusive of approach outperformed models agnostic to approach with respect to nonroutine discharge (AUC with approach, 0.871; AUC without approach, 0.850; p = 0.018) and readmission (AUC with approach, 0.686; AUC without approach, 0.651; p = 0.019), but no differences were observed regarding top quartile of LOS (p = 0.17) and top quartile of cost (p = 0.40). CONCLUSIONS Surgical approach may influence perioperative outcomes in patients treated for intracranial dAVF—most significantly discharge disposition and 1-year readmission. Future longitudinal prospective studies with more clinical detail will be required to fully capture the predictive utility of surgical approach in patients treated for intracranial dAVF, particularly for various dAVF subtypes.

show abstract

Section: Discussionmentioning

confidence: 99%