🔍 Key Findings
- Remote respiratory function grading (RFG) had poor to moderate reliability compared to in-person assessment.
- Expert graders had higher agreement with in-person scores than novices (Cohen’s kappa 0.37–0.48 vs. 0.21–0.47).
- Interobserver agreement was moderate among experts (Fleiss’ kappa = 0.59) and poor among novices (Fleiss’ kappa = 0.39).
- Remote recordings suffered from background noise, short clip durations, and technical limitations of electronic stethoscope recordings.
- Final RFG scores were based on the highest grade across categories (respiratory noise, inspiratory effort, dyspnea/cyanosis/syncope).
- Clinical impact: Only in-person grading reliably supports decisions for surgical intervention or breeding restrictions.
