An Assessment of Three Portable Peak Flow Meters* - Chest ...

More documents

Recommendations

Info

Table 1-.Mini-Wright Meter Flow (Lpm) 120 180 300 480 Overall 5 D%= -2.7 Percentage of Discrepancy (D%) Instrument Mean D% 1 18.1 24.8 11.8 #{149} 1.4 14.0 2 8.1 16.5 7.7 1.9 8.6 3 24.2 30.7 13.8 8.9 19.4 4 26.4 38.3 Malfunctioned 32.4 5 17.3 22.4 11.7 5.2 14.2 Models 18.8 26.5 11.3 4.4 Overall D%= 15.3 Percentage of Discrepancy (D%) Table 2-Pulmonary Monitor Flow (Lpm) 120 180 300 480 Instrument Mean D% 1 -27.5 8.2 0.4 5.1 -3.7 2 -45.4 -12.3 13.4 8.2 -9.0 3 -55.7 -6.9 7.7 -2.2 -14.3 4 -54:5 -7.4 2.4 -4.4 -16.0 5 -55.3 -8.1 4.0 1.3 -145 Model -47.7 -5.3 5.4 1.6 Overall D%=-11.5 Percentage of Discrepancy (D%) Table 3-Peak Flow Meter Flow (Lpm) 120 180 300 480 Instrument Mean D% 1 -1.9 -5.1 -1.9 -4.9 -3.5 2 0.5 -1.1 0.0 0.7 0.0 3 -6.4 -2.1 -4.2 -0.6 -3.3 4 -7.4 0.6 1.3 0.4 -1.3 5 -2.9 -4.1 -6.2 -8.3 -5.4 Model -3.6 -2.4 -2.2 -2.5 converted to percentages (percent discrepancy, D%). Under-readings, thus, were negative values compared to actual, while over-readings were positive. The mean D% for each of the five measurements at each flow rate for each instrument was then calculated. The mean D% for each flow range between instruments (inter-instrument variability) of similar manufacture, as well as the mean D% for all flows on a single instrument (intra-instrument 20- 10. w 30. 0 U, -10- -20- -30- -40- -50- #{149}a- I I I I I -ARM 10 240 360 480 Flow (I/rn) Ficunx 1. Inter-instrument variability, D%. RESULTS The percentage of discrepancy for each of the Armstrong, Vitalograph, and HealthScan peak flow meters are depicted in Tables 1-3. Intra-instrument, as well as inter-instrument means are also displayed. Graphic depictions of results are displayed for inter- instrument D% in Figure 1, and for intra-instrument D% in Figure 2. As can be seen, only the Health- Scan peak flow meter had acceptable mean interand intra-instrument D% of less than 5 percent. In testing the Armstrong Industries’ mini-Wright peak flow meter, one malfunctioned during the testing. The diaphragm became displaced from the cylindrical housing, and could not be manually replaced. Though the widest recorded ranges were >- C.) z aw C.) Cl) 30- 20’ 10 -ARM -- VIT -.- HS-O variability) were then determined. Ficuas 2. Intra-instrument variability, D%. Downloaded From: http://journal.publications.chestnet.org/ on 01/02/2013 1 __.-1#{149}-..,#{149} -I;_._44.. I
noted on this particular instrument prior to mal- function, elimination of all data from the defective instrument would not significantly alter the net results for the model as a whole. The Vitalograph pulmonary monitor data are ob- viously biased by extremely poor performance at the 120 Lpm flow range. The instrument scale begins at 150 Lpm, and hence, results were extrapolated back to include the 120 Lpm range, assuming the same linearity of scale to apply at flows below, as well as above the scale minimum. This may not, in fact, be the case. The length-tension relationship of the spring within the cannister may be such that quite a large initial force must be directed to overcoming inertia. Hence, all flow measurements at 120 Lpm may have little validity. Ignoring the flows at 120 Lpm, the Vitalograph pulmonary monitor’s performance improves significantly. The instrument tends to underestimate flows at 180 Lpm, and overestimate flows above this range. Though inter-instrument and intra-instrument mean percentages of discrepancy meet ACCP and ATS standards (0.57 and 2.2, respectively), this was achieved only through wide variation in D% at the 180 Lpm flow range. No single percentage of discrepancy met the established standards at this range, though there were sufficient positive and negative values so that the mean that resulted was acceptable. Likewise, though intra-instrument van- ability was acceptable, this was often the result of averaging significantly varied individual results (instruments 2 and 3 especially). This factor limits model acceptability. The HealthScan peak flow meter performed more uniformly. Though some units tended to vary from actual readings at either high or low flow ranges, there were no wide swings in results evident as was the case for the Vitalograph product. HealthScan instruments generally tended toward slight underestimation of true values. Overall, both inter- and intra-instrument accuracy was quite good. DiscussioN Results of peak flow measurements are now widely accepted for following the course of bronchial asthma.7 Alterations in peak flow provide an ob- jective means of documenting patient status, and with the development of portable instrumentation to achieve this purpose, close monitoring of pulmonary function at home can be made. Three light- weight, inexpensive devices for peak flow measure- ment are currently widely marketed as described. Ideally, these devices should meet ATS and ACCP standards. The results of our studies performed at varying flow rates for multiple instruments support the accuracy and reproducibility of only the Health- Scan instrument, though only the Armstrong unit meets the established criteria for flow range. Prior documentation of performance of both the Armstrong and Vitalograph products have been previously described. In the original description of the Armstrong mini-Wright device, documentation of an underestimation of peak flow of 17 percent was made, though allowance for this was pro- vided in calibration.’#{176} Subsequent studies showed correlation between the standard Wright and mini- Wright peak flow meters, but the study design in- volved the comparison of separate maximal efforts by a trained subject into each device, rather than recording flow in series into both, as done in our study.” This may have introduced some variability in the actual flow delivered, since the studies as- sumed the subject produced the exact peak flow each time. Likewise, significant variation between instruments for presumably similar flows was also documented. Perks et al’2 also found significant inter-instrument variability for the mini-Wright unit, but felt that this difference remained constant for any given machine so that deviations in measurements were valid, though exact accuracy was not necessarily so. Similar studies are also available for the Vitalo- graph pulmonary monitor. When initially described, arbitrary units of measurements were used, rather than liters per minute, which probably limited clinical utilization.13 When revised to provide read- lug in Lpm, Perks et al’ found fair correlation with a standard Wright peak flow meter ( again on suc- cessive peak flow maneuvers, rather than in series), but again found significant inter-instrument variability. They also stated that the inability to measure flows below 150 Lpm would limit the clinical utility of the unit. This is particularly apt to be true in management of asthma.3 Burns15 compared the Armstrong and Vitalograph products and found that significant errors of estimate existed for both devices, being larger for the pulmonary monitor. He concluded that though the accuracy of both devices was low, they could be of use in determining discrete peak flow changes from 308 Assessment of Three Portable Peak Flow Meters (E!chenhorn .t a!) Downloaded From: http://journal.publications.chestnet.org/ on 01/02/2013 baseline. Our studies support the inter-instrument van- ability of both the Armstrong and Vitalograph devices when compared to a pneumotachograph. Likewise, significant intra-instrument variability ex- ists for both devices. Much of the problem in accuracy and reproducibility may stem from the design of both which utilizes a single expanding spring to measure flow. Variability of the length-tension relationship for any particular spring might well be
Page 1: An Assessment of Three Portable Pea

An Assessment of Three Portable Peak Flow Meters* - Chest ...

Create successful ePaper yourself

Delete template?

Save as template?