Study objective: Vital sign measurements, specifically heart rate, respiratory rate, and blood pressure, play a fundamental role in many medical evaluations, yet little is known about the reliability of noninvasive vital sign measurements. We sought to determine whether trained observers can reproducibly assess vital signs in the clinical setting.
Methods: Two trained observers independently measured vital signs on 140 patients presenting to an urban emergency department with acute medical complaints. Heart rate and respiratory rate were each measured by auscultation of heart and breath sounds for 1 minute. Systolic and diastolic blood pressures were determined by auscultating Korotkoff sounds while viewing pressure measurements from a standard cuff and mercury manometer. The mean value of each vital sign and Bland-Altman statistics (mean difference between observers [MDO], expected range of agreement [ERA]) were used to provide absolute and relative indices of reliability.
Results: The observers found a mean heart rate of 78.5 beats/min, with an MDO of 0.02 beats/min (0.03%), and an ERA of +/- 10.6 beats/min (+/- 13.5%). Respiratory rate exhibited a mean of 17.5 breaths/min, an MDO of 0.04 breaths/min (0.2%), and an ERA of +/- 6.2 breaths/min (+/- 35.5%). The mean systolic blood pressure of 127.1 mm Hg was associated with an MDO of 1.3 mm Hg (1.0%), and an ERA of +/- 24.2 mm Hg (+/- 19.0%). Diastolic blood pressure exhibited a mean of 77.4 mm Hg, an MDO of 0.3 mm Hg (0.4%) with an ERA of +/- 19.9 mm Hg (+/- 25.7%).
Conclusion: The reproducibility of vital sign measurements may be limited by significant interobserver variability. Clinicians should recognize this inherent variability and interpret vital signs with caution.