Validation of the Oklahoma predictor model for discharge destination in patients with major traumatic injury: a cross-sectional study.

Abstract

Purpose: Protracted in-patient stays affect trauma patient costs and hospital resource utilizations. Proper discharge placement stratification may help with early discharge planning in this group of patients. No standardized discharge destination prediction model exists. A scoring model has been developed after examining Oklahoma Trauma Database discharge destination predictors. This study's goal was patient data-driven model validation.

Methods: Level II trauma center patient data over three months, including comorbidities, injuries, and demographics were analyzed. We compared the scoring model discharge destination prediction with actual destinations.

Results: The study included 459 patients, with 108 facility discharges. The scoring model demonstrated significant facility placement prediction (Scores ≤ 7: 11.94% or Negative Predictive Value of 88.1%; Score 8-14: 47.22% as Positive Predictive Value: and Score ≥ 15: 60.00%). Scoring 8-14 showed a 6.60-fold (95%CI: 4.11, 10.61) increase compared to ≤ 7. Scoring ≥ 15 was 11.07 times (95%CI: 1.79, 68.42) more likely than ≤ 7.

Conclusion: The Oklahoma Trauma Discharge Predictive Scoring Model demonstrated significant facility discharge prediction and may assist with decreasing delay of anticipated patient discharge destination.