J. Mark Ansermino

Background: counting respiratory rate over 60 seconds can be impractical in a busy clinical setting.

Methods: 870 respiratory rates of 272 acutely ill medical patients estimated from observations over 15 seconds and those calculated by a computer algorithm were compared.

Results: The bias of 15 seconds of observations was 1.85 breaths per minute and 0.11 breaths per minute for the algorithm derived rate, which took 16.2 SD 8.1 seconds. The algorithm assigned 88% of respiratory rates their correct National Early Warning Score points, compared with 80% for rates from 15 seconds of observation.

Conclusion: The respiratory rates of acutely ill patients are measured nearly as quickly and more reliably by a computer algorithm than by observations over 15 seconds.