Reliability and Validity Study of a Simple Assessment Tool for Early Detection of Perioperative Stroke

Abstract

Background: Stroke is a feared postoperative complication. The diagnosis is usually delayed resulting in a reduced possibility for therapeutic intervention. Objective: To assess the reliability and validity of a simple assessment tool for early detection of perioperative stroke. Materials and Methods: The authors set up a short course and workshop for all service nurses in neurological screening method for the assessment tool for early detection of perioperative stroke including observing eye deviation or extraocular muscle limitation, asymmetrical limb weakness, and conscious changes in 24 surgical wards in Siriraj Hospital. The protocol was applied during routine postoperative vital sign measurement to evaluate new neurological deficits within 14 days after surgery. The postoperative patients were selected under the permission. The authors used content validity, sensitivity, specificity, and analysis to measure content and criterion validity, respectively. The authors also tested inter-rater reliability by using Kappa and Fleiss’s Kappa statistics. Results: Four hundred twenty-five postoperative patients between June and December 2018 were assessed by the protocol. The sensitivity was 85.71% and specificity was 98.02%. Positive predictive value (PPV) and Negative predictive value (NPV) was 69.23% and 99.25%, respectively. The accuracy of the tool was 97.41%. The interrater reliability was substantial agreement. Conclusion: This simple assessment tool is a valid and reliable clinical tool used for perioperative stroke screening. The tool is simple and quick to perform among nurses and physicians. The tool was designed to detect only moderate to severe stroke that required treatment intervention due to postoperative treatment limitation.