Quality control for blood pressure measurement in population studies: Shibata Children's Heart Study

To investigate the relation between observer performance for blood pressure measurement in a training process and in field conditions, measurement values were studied under training and field conditions among 21 blood pressure observers of 1434 subjects aged 6-15 years in Japan. The observers received training by a videotape, which included six audiovisual presentations of a falling mercury column in a standard sphygmomanometer with Korotkoff sounds. Observer bias was measured for each trainee as the mean difference between the observed and the standard values for each blood pressure reading, including systolic (SBP), fourth-phase diastolic (K4), and fifth-phase diastolic (K5) values. In multiple linear regression analyses, each 1 mmHg increment in observer bias was equivalent to 1.27, 0.88, and 1.25 mmHg difference in actual readings of SBP, K4, and K5, respectively, in the field. This finding indicates that observer performance in videotape training is predictive of measurement behavior in the field.