[Estimation of initial interobserver variability of forced spirometry in the multicenter epidemiological study...

To estimate the interobserver variability and degree of agreement for basic spirometric parameters before beginning field work for the IBERPOC Project. Study of agreement between 7 observers (pneumologists) and a gold standard, using a scheme of incomplete balanced and randomized blocks with an equal number of spirometric measurements (n = 3) per patient and an equal number of measurements (n = 9) per observer performed at the same session. The study population consisted of 14 patients with different degrees of air flow obstruction and 7 normal volunteers. Statistically significant differences attributable to subjects (inter-patient variability) were found for the three variables analyzed. Variability attributable to the observer was found for FVC and FEV1 but not for FEV1/FVC. The greatest interobserver differences were found for FEV1, such that 4 of the 7 observers recorded values that were significantly different from the mean (p < 0.05). The differences were less marked for FVC and for the FEV1/FVC ratio, with only 2 observers recording significantly different values for each variable. The high degree of reproducibility as well as the excellent interobserver agreement found in this standardization session provide an a priori guarantee of validity for spirometric measurements and rule out the existence of differential bias in data recorded at the various geographic areas involved in the study.