ISSUES IN PEAK EXPIRATORY FLOW RATE MEASUREMENTS: ANALYSIS AND INTERPRETATION OF RESULTS
I read with interest the paper by Al-Taweel et al entitled, “Peak expiratory flow rate in a sample of normal Saudi males at Riyadh, Saudi Arabia” published in the Journal of Family & Community Medicine (July 1999 issue).1 Because the paper was stimulating, it raised the following points:
1. The objective of the study was “to identify the normal peak expiratory flow rate (PEFR) for a sample of Saudi males and to compare the finding with British standards”. As the sample was drawn from Primary Health Care Centre ( PHCC) attendants, I am afraid that this sample may not be representative of the general population. This is important especially if such results will be used to compare with general population from elsewhere i.e. British population.
2.Even if it is assumed that, the objective was to compare PHCCattendants with the same from other populations, still the sample has to be representative. No information was available from the paper to suggest how the sample size (n=680) was calculated and how far it was quantitatively representative to all Riyadh PHCCattendants? (Representing which percentage?) And how it was selected from each PHCC?
3.As age was an important factor in PEFR measurement, the sample collected should have been selected in such a way to give equal number of subjects among different age strata. Failure to do that has resulted in 52.6% of the sample to be from one stratum (<25 years of age), which may represent PHCCattendants but not the Saudi Arabian male adults living in Riyadh. This has resulted in the following:
(a) Age was not normally distributed in the sample (severely skewed to the left, if represented in frequency distribution or histogram). Having had this and in such a case, parameteric methods of analysis may not apply and use of non-parametric methods may be inevitable. Moreover, describing age by mean and standard deviation (SD), of the raw skewed data may not be precise. Instead, geometric (logarithm) mean and SD may be more precise following transformation of raw data to their natural logs.
(b) The drop in the curve in figure 2 for Saudis beyond age of 25 years may be due to the fact that majority of the sample (52.6%) was below 25 years of age and only 12.8% were above age 45 years. This may have shifted the curve to the left with severe drop at older ages compared to the British curve. Similarly, figure 1 showed severe skewness to the left with peak at age 25 years followed with a drop.
The multiple regression equation showed R2 = 21% which is a model of a relatively weak predictive power (only 21% of the variation in PEF could be explained by the variation of variables included in the equation). The fact that its p-value was <0.0001 may only indicate that variables included in the model correlate significantly and lineary with PEFR.
In table 2, the equation contained two constant values (-338.5, which appeared both in the equation and table) and (+9, which appeared only in the equation but not in the table), which one is correct?
It is advisable not to include adolescents when designing or studying adults’ standards of any pulmonary function test, including peak expiratory flow rate. This is because the “difference relate mainly to the ratios of thorax to total body height”.2 Therefore, “adolescents should not be compared with adult standards until puberty and growth are complete.2 Consequently, inclusion of subjects at prepubertal age in this study was better avoided.
7.Based on 1, 2 and 3 above, I feel that a conclusion such as “this study has shown that Saudi Arabian male adults living in Riyadhhave lower peak expiratory flow rate than adults of different racial origin living in UK” was not adequately supported and lacks such generalizability.
1. Al-Taweel AA, Kalantan KA, Ghani HA. Peak expiratory flow rate in a sample of normal Saudi males at Riyadh, Saudi Arabia. Journal of Family & Community Medicine 1999; 6:23-27.
2. Mueller GA, Eigen H. Pediatric pulmonary function testing in asthma. Asthma 1992; 39:1243-57.
Dr. Kasim M. Al-Dawood, FFCM(KFU) MSc (Epidemiol), DLSHTM
Associate Professor and Consultant
Department of Family & Community
Medicine, College of Medicine
King Faisal University
P.O. Box 2290, Al-Khobar 31952
NB:The Journal has not yet received a reply from the corresponding author until the publication of this issue.