Two dimensions of measurement error: Classical and Berkson error in residential radon exposure assessment

Abstract

Measurement error in exposure assessment is unavoidable. Statistical methods to correct for such errors rely upon a valid error model, particularly regarding the classification of classical and Berkson error, the structure and the size of the error. We provide a detailed list of sources of error in residential radon exposure assessment, stressing the importance of (a) the differentiation between classical and Berkson error and (b) the clear definitions of predictors and operationally defined predictors using the example of two German case-control studies on lung cancer and residential radon exposure. We give intuitive measures of error size and present evidence on both the error size and the multiplicative structure of the error from three data sets with repeated measurements of radon concentration. We conclude that modern exposure assessment should not only aim to be as accurate and precise as possible, but should also provide a model of the remaining measurement errors with clear differentiation of classical and Berkson components.

Air Pollution from Traffic and Risk for Lung Cancer in Three Danish Cohorts

“…We cannot see how such uncertainty could differ for patients and the subcohort and would therefore expect the misclassification to be nondifferential. A previous comparison between NO 2 concentrations measured and calculated by the Danish dispersion models (20,31) showed that the misclassification was primarily of the Berkson type as might be expected when exposure is predicted from a model (32,33). Berkson error is expected not to bias the risk estimates (34,35), although it decreases the precision (36), such that the confidence intervals reported in the present study are probably too narrow.…”

Radon and Lung Cancer in the American Cancer Society Cohort

“…Estimates of radon concentrations in individual homes are subject to a number of sources of uncertainty, including detector measurement error, variation due to detector placement, and changes in radon concentrations over time (seasonal and year-to-year variability; refs. 4,[38][39][40][41][42][43]. However, these measurement errors are most likely to be nonsystematic.…”

“…Although it is difficult to predict the total potential cumulative impact of such errors on the results observed in the current study, the observed relative risk estimates may be subject to some degree of downward bias (38,46,47). Mallick and colleagues (48) examined the impact of adjusting for plausible levels of exposure measurement error associated with ecological measures of ambient air pollution under a cohort design and found the relative risk estimates were subject to downward bias.…”

Effect of Soil Moisture on Seasonal Variation in Indoor Radon Concentration: Modelling and Measurements in 326 Finnish Houses

“…Based on large data from countries involved in the European epidemiological study (13) , the country-specific year-to-year variation expressed as the COV in the annual average radon activity concentration was 17-62 %. On the basis of the German epidemiological analysis, Heid et al (14) conclude in a corresponding error estimate of 0.55. High year-to-year variation indicates that variation between e.g.…”