Performance testing new peak flow meters.

Abstract

, a key target group will be reac-hed in an effective way and this may help to keep users ofpeak flow meters and monitoring devices adequately infor-med about the performance and quality of new equipment.Several important questions arise: what is a satisfactoryperformance, how is it best documented, and who shouldundertake the tests? Both the American Thoracic Society(ATS) and the European Respiratory Society (ERS) havedeveloped guidelines [2, 3] defining performance criteriafor the evaluation of spirometers. The following aspectsmight enter into a comprehensive evaluation: performanceassessment using mechanical test rigs, clinical testing,tests of patient compliance and ease of use, hygiene andmaintenance, and long-term reliability. Eventually, the re-sponsibility for the function of the meter should be placedwith the manufacturer, the distributor, the doctor and theusers in such a way that ensures optimal utilization in aclinical setting. Can the manufacturer be relied upon toproduce all of these test results, or should they be obtainedby an independent institution?With regard to the accuracy of measurements, there arethree main elements, which can all be assessed using a testrig. The first is calibration for steady flows, which maymean that the meter's scale would not necessarily be equi-distant [4] in order to give the best accuracy. The next ele-ment is the device's frequency response, which determineswhether or not the meter can react quickly enough tomeasure the fastest components of the expiration. Rapidlyoccurring events, such as PEF, may be underestimated ifthe device's response is overdamped, or may be overesti-mated if the device is underdamped. The third element isthe resistance of the meter, which has to be low so that itdoes not cause any gas compression in the lungs such thatPEF is lowered significantly [5]. The calibration of adevice and the determination of its internal resistance arestraight-forward. The difficult part is the determination ofthe frequency response. The ATS [2] has approached thisempirically by defining an adequate frequency response ifthe meter reacts satisfactorily to each one of a panel of 26waveforms which have arbitrarily been deemed to reflectthe common range of flow profiles. The ERS [3] has pro-posed another approach by instead considering that thefrequency response of PEF meters be tested with flow-time profiles having rise times (time from 10–90% PEF)and dwell times (duration of flow in excess of 95% or90% PEF) that are representative of the range found innormal subjects and those with airflow limitation.Considerable technical problems must be overcome inundertaking these tests. Using a servo-controlled calibra-tion pump to generate flow profiles with very short risetimes leads to considerable distortion of the output profilecompared with the input profile [6, 7]. Thus it is difficultfor the true output flow to be determined accurately. Thisproblem with profile distortion is not the case if the cali-bration flow is produced by explosive decompression froma pressure tank [8]. The disadvantage now is that this out-put flow has to be calibrated in advance, because it is notuniquely determined in the same way that the input flowfrom a servo-controlled pump is derived from its dimen-sions and the movement of the piston. No clinical testingshould be necessary, unless a meter has certain featuresthat suggest that a patient's interaction with the metermight be different from that for other types of meter. Thereis a good chance that two different meters will give thesame result if the calibration is correct, the frequency res-ponse is adequate and the internal resistance is not too high.In the paper by R