Ultrasound fetal biometry: number needed to audit

Area of Science:

• Medical Imaging
• Obstetrics
• Statistical Modeling

Background:

• Accurate fetal measurements are crucial for optimal pregnancy care and outcomes.
• Regular audits are essential for maintaining and improving the accuracy of these measurements.
• The impact of audit sample size on the reliability of fetal measurement data is not well-defined.

Purpose of the Study:

• To model the potential outcomes of fetal measurement audits with varying numbers of images and measurements.
• To explore the implications of different audit sizes on the improvement and maintenance of clinical standards.
• To provide guidance on interpreting audit results to ensure effective targeted interventions.

Main Methods:

• Statistical modeling using binomial distribution for pass/fail image scoring.
• Sample size calculations for estimated fetal weight (EFW) error analysis.
• Modeling of audit outcomes for sample sizes of 3, 10, 25, and 100 patients across various pass rates and error margins.

Main Results:

• Confidence intervals for pass rates narrowed significantly with increased sample size (±32% to ±56% for 3 images vs. ±6% to ±10% for 100 images).
• Significant differences in mean EFW errors decreased with larger sample sizes (16.8% for 3 samples vs. 2.9% for 100 samples).
• Small audits (e.g., 3-10 images) yield imprecise results with wide confidence intervals.

Conclusions:

• Results from small fetal measurement audits lack precision and can lead to misdirected interventions.
• Careful interpretation of audit data is necessary to avoid inappropriate targeting of training and support.
• Larger audit sample sizes are recommended for more reliable assessment of fetal measurement accuracy and effective quality improvement.