Some observations on the variation in height coefficients in prediction equations for forced vital capacity.
J Clin Epidemiol 1992 Sep; 45(9):951-957
The effect of height on prediction equations for forced vital capacity (FVC) was examined. Estimation of the coefficients of the standard equation for the determination of FVC using regression methods has resulted in a high correlation between the estimated height coefficient and the estimated intercept. Twenty six published prediction equations for FVC and vital capacity were identified from the literature. An analysis was performed to examine the extent to which random variation accounted for the range between the coefficients in the equations. The coefficients were grouped based on number of subjects in the study and 95% confidence ellipses were generated for each group based on the mean number of subjects using statistics on the variability and covariability of coefficients obtained from regression of FVC on age, height, smoking pack years, and years underground in a group of coal miners. The relationship between the plots of observed coefficients after being centered at the grand mean of the 26 coefficients of these ellipses were examined. The relationship predicted by the ellipse for the intercept and height coefficient applicable to the group of coal miners was also seen in the plots from the 26 equations obtained from the independent data. This finding indicated that the variation between the studies was truly random. Another analysis grouping the studies by size of study also demonstrated that random variability could account for all of the variation between the height coefficients. The author concludes that caution is necessary when using equations for predicting the FVC of extremely tall people.
NIOSH-Author; Statistical-analysis; Mathematical-models; Height-factors; Pulmonary-function-tests; Mine-workers; Humans; Cigarette-smoking;
Author Keywords: Prediction formulas; Forced vital capacity
Journal of Clinical Epidemiology