Validation Studies using an Alloyed Gold Standard

Abstract

A key assumption made when using a validation study to correct an estimate of relative risk for bias due to misclassification or measurement error is that the available measure, known to have error but nonetheless used routinely in the main study, is compared to a gold standard measured without error. In most epidemiologic applications, the putative gold standard is in fact measured with error. The effect of the violation of the assumption on the corrected estimate depends on the magnitudes of the errors in the two measures and on their correlation. In particular, when the errors are negatively correlated, independent, or weakly positively correlated, the corrected estimate will tend to overcorrect beyond the true value.

Calibration of the dietary data obtained from the Brazilian center of the Natural History of HPV Infection in Men study: the HIM Study

“…However, after correction of QFFQ data based on linear regression analyses in which the reference method is as much based on reports as the QFFQ is, and is subject to similar errors, the relative risks may be biased 48,49 . Biomarkers can be used to make consumption measurements more accurate and consequently also the relative risk estimates.…”

Challenges to retrospective exposure assessment

“…If the estimation method is evaluated before the estimates are developed for a study, the measurement data can first be used to evaluate the method and then be incorporated into the study (5). The determination of validity is, however, problematic when the gold standard is "alloyed" (6), such as when historical air measurement data are nonrandom. (See the section Improving the Use of Exposure Data.)…”

Alguns pilares para a apreciação da validade de estudos epidemiológicos

“…Grande parte da literatura sobre as estraté- . Recentemente, no entanto, alguns autores têm apontado que, na presença de dependência entre os erros de mensuração e os verdadeiros valores das variáveis, a má-classificação pode até aumentar ou mesmo reverter a medida de efeito de interesse [44][45][46] .…”

“…Trata-se de uma literatura em constante crescimento e cujo aprofundamento transcende o escopo do artigo. O leitor poderá encontrar alguns exemplos interessantes em Clayton 51 (modelos lineares generalizados com efeitos aleatórios); Wacholder 45 (modelos que incorporam a correlação entre os erros e o verdadeiro valor das variáveis); Richardson 52 e Spiegelhalter e col. 53 (abordagem Bayesiana); e Wacholder e col. 44 , Brenner 54 e Spiegelman e col. 55 (estratégias de correção para estudos que utilizam um padrão "dourado" -alloyed gold standard).…”