Abstract

Number preference for tens digits 0 and 5 by subjects when reading the peak expiratory flow (PEF) from mechanical meters may reduce the diagnosis of asthma and especially occupational asthma from serial PEF records. The records of 274 subjects (set A) recorded using mini-Wright PEF meters and 35 subjects (set B using electronic logging turbine spirometers (Micromedical) were examined for the ratio of tens digits 0 and 5 to the remaining tens digits (NP ratio; NPR). 24 records from set A and no records from set B were classified as having NP by visual inspection. No record from set B had a NPR >0.42 (mean 0.26, SD 0.08). NPR from set B were positively skewed (mean 1.81, SD 12.33). We chose a cut of 0.75 between no NP and possible NP and 1 between possible and definite NP. A further set (C1) of records recorded on mini-Wright peak flow meters was taken all with NPR <0.75. NP was induced in these records by rounding to the nearest 50l/min (set C2). The mean diurnal variation (DV, amp % predicted) did not change (paired t test p=0.87). Maximum change in amplitude was 23.4l, DV 5.7% and % of record with DV >15% was 47.6%. For OASYS-3 scored records the whole record score tended to decrease (paired t test p<0.001) but not for OASYS-NN scored records (paired t test p=0.93). Changes in classification of records is shown.

OASYS-NN OASYS-3
Work effect ? Work effect No Work effect Work effect No work effect
Work effect 42 6 1 29 4
? Work effect 4 16 10
No Work effect 0 10 67 5 112

NP can cause records with a possible work effect to be erroneously classified as no work effect. DV is not greatly affected but large changes can occur in amplitude of variation and % of record with a DV >15%.

Full Text

Comments