Comparison of 5 experts and computer analysis in rule-based fetal heart rate interpretation

Objective The purpose of this study was to measure agreement among 5 expert clinicians and a computerized method with the use of a strict fetal heart rate classification method. Study Design Five providers independently scored 769 8-minute segments from the last 3 hours of 30 tracings with the use of a 5-tier color-coded framework that contains pattern descriptions and proposals for management. Computer analysis was performed with PeriCALM Patterns (PeriGen, Princeton, NJ) to detect and classify patterns. Results The clinicians agreed exactly with the majority opinion in 57% (95% confidence interval [CI], 49–64%) of the segments and were within 1 color code in 89% (95% CI, 81–96%). The average proportion of agreement was 0.83 (95% CI, 0.73–0.94). Weighted Kappa scores averaged 0.58 (range, 0.48–0.68). The computer-based results were not statistically different: 0.87 and 0.52, respectively. Conclusion These 5 clinicians achieved moderate-to-substantial levels of agreement overall using a strictly defined method to classify fetal heart rate tracings. The result of the computerized method was similar to the conclusions of these clinicians.