Revisiting the accuracy of peak flow meters: a double-blind study using format methods of agreement

Background: There is widespread use of peak flow meters in both hospitals and general practice. Previous studies to assess peak flow meter accuracy have shown significant differences in the values obtained from different meters. However, many of these studies did not use human subjects for peak flow measurements and did not compare meters of varying usage. In this study human subjects have been used with meters of varying usage. Methods: Participants were tested using two new (meters A and C) and one old peak flow meter (meter B) in random order. The study was double-blinded. Participants were recruited from the university campus. Results: Four hundred and nine individuals participated. The difference between peak flow means of A and B was -9.93 l/min (95% Cl: -12.37 to -7.48, P <0.0001). The difference between peak flow means of B and C was 20.08 l/min (95% Cl: 17.85-22.29, P <0.0001). The difference between peak flow means of A and C was 10.15 l/min (95% Cl: 7.68-12.61, P <0.0001). Conclusion: There was a significant difference between the values obtained from the new and old peak flow meters and also between the two new peak flow meters. We conclude that there is need for caution in interchangeably using flow meters in clinical practice. (c) 2004 Elsevier Ltd. All rights reserved.