Reproducibility of non-specific bronchial challenge in adults: implications for design, analysis and interpretation of clinical and epidemiological studies.

BACKGROUND Poor reproducibility of an outcome measure reduces power and, in an independent variable, biases results. The intraclass correlation coefficient measures loss of power and degree of bias. Information is lacking on the intraclass correlation coefficient for bronchial responsiveness and factors affecting reproducibility. METHODS Papers containing information on reproducibility of bronchial responsiveness were identified using a Medline search and citations. Within and between person components of variance of PD20 or PC20 were expressed in doubling dose or concentration units, and the intraclass correlation coefficient calculated when not reported. RESULTS Results were extracted from 32 papers. Intraclass correlation coefficients were over 0.9 in short term studies of highly selected asthmatic patients, but larger and most long term studies had lower intraclass correlation coefficients, less than 0.5 in some cases, due to greater within person or lower between person variation. Reproducibility of dose or concentration-response slope was generally higher, but still less than that of forced expiratory volume in 1 second. CONCLUSIONS Information is available to calculate sample size for studies with bronchial responsiveness as the outcome, but results when bronchial responsiveness is an explanatory variable may be misleading.