Observer error in systolic blood pressure measurement in the elderly. A case for automatic recorders?

Measurement of systolic blood pressure (BP) in the elderly can be inaccurate due to observer errors such as terminal digit preference and expectation bias. Efforts to reduce these errors include use of trained observers, random-zero sphygmomanometers, and automatic BP recorders. To evaluate the value of an infrasonic recorder, the infrasonde, we compared simultaneous systolic BP determinations obtained directly by intra-arterial measurements, and indirectly by a standard cuff-mercury sphygmomanometer using clinic nurses (casual and serial measurements), a trained physician assistant (PA), and the Infrasonde in 36 elderly hypertensive men. All the indirect measurements correlated positively with the direct measurement. A terminal digit preference for zero occurred more frequently in the casual and serial readings compared with the Infrasonde and PA cuff readings: 45% and 51% vs 21% and 22%, respectively. Our data suggest that the Infrasonde is a reliable, alternative device for systolic BP measurement that eliminates observer biases associated with standard cuff BP recordings.