Hospital use following anterior cervical discectomy and fusion

Using Machine Learning and Deep Learning Algorithms to Predict Postoperative Outcomes Following Anterior Cervical Discectomy and Fusion

STUDY DESIGN

A retrospective cohort study from a multisite academic medical center.

OBJECTIVE

To construct, evaluate, and interpret a series of machine learning models to predict outcomes related to inpatient health care resource utilization for patients undergoing anterior cervical discectomy and fusion (ACDF).

SUMMARY OF BACKGROUND DATA

Reducing postoperative health care utilization is an important goal for improving the delivery of surgical care and serves as a metric for quality assessment. Recent data has shown marked hospital resource utilization after ACDF surgery, including readmissions, and ED visits. The burden of postoperative health care use presents a potential application of machine learning techniques, which may be capable of accurately identifying at-risk patients using patient-specific predictors.

METHODS

Patients 18-88 years old who underwent ACDF from 2011 to 2021 at a multisite academic center and had preoperative lab values within 3 months of surgery were included. Outcomes analyzed included 90-day readmissions, postoperative length of stay, and nonhome discharge. Four machine learning models-Extreme Gradient Boosted Trees, Balanced Random Forest, Elastic-Net Penalized Logistic Regression, and a Neural Network-were trained and evaluated through the Area Under the Curve estimates. Feature importance scores were computed for the highest-performing model per outcome through model-specific metrics.

RESULTS

A total of 1026 cases were included in the analysis cohort. All machine learning models were predictive for outcomes of interest, with the Random Forest algorithm consistently demonstrating the strongest average area under the curve performance, with a peak performance of 0.84 for nonhome discharge. Important features varied per outcome, though age, body mass index, American Society of Anesthesiologists classification >2, and medical comorbidities were highly weighted in the studied outcomes.

CONCLUSIONS

Machine learning models were successfully applied and predictive of postoperative health utilization after ACDF. Deployment of these tools can assist clinicians in determining high-risk patients.

LEVEL OF EVIDENCE

III.

IDENTIFICATION OF RISK FACTORS ASSOCIATED WITH 30-DAY READMISSION OF PATIENTS SUBMITTED TO ANTERIOR OR POSTERIOR ACCESS CERVICAL SPINE SURGERY

ABSTRACT Aim: To conduct a systematic review of the literature to identify risk factors associated with 30-day readmission of patients submitted to anterior or posterior access cervical spine surgery. Methods: The databases used to select the papers were PubMed, Web of Science, and Cochrane, using the following search strategy: patient AND readmission AND (30 day OR “thirty day” OR 30-day OR thirty-day) AND (spine AND cervical). Results: Initially, 179 papers that satisfied the established search stringwere selected. After reading the titles and abstracts, 46 were excluded from the sample for not effectively discussing the theme proposed for this review. Of the 133 remaining papers, 109 were also excluded after a detailed reading of their content, leaving 24 that were included in the sample for the meta-analysis. Conclusions: The average readmission rate in the studies evaluated was 4.85%. Only the occurrence of infections, as well as the presence of patients classified by the American Society of Anesthesiology (ASA) assessment system with scores greater than III, were causal factors that influenced the readmission of patients. No significant differences were noted when comparing the anterior and posterior surgical access routes. Level of evidence II; Systematic Review of Level II or Level I Studies with discrepant results.