Assessing Epidemiological Relations and the Role of Measurement Errors

Conclusion

As discussed in this chapter, epidemiologic studies coupled with careful, hypothesis-driven data analysis [45,46] can be useful in providing empirical evidence for the role of intermediate endpoints and cancer precursors in the genesis of cancer, and for the role of susceptibility traits and exposures in the development of cancer precursors. The specification of a priori models that depict the relations among variables and their directionality requires an eclectic consideration of all aspects of disease biology, histopathology, and genetics within the timetested epidemiologic framework for assessing causal mechanisms of disease [47]. However, even the most multidisciplinary molecular epidemiology studies must contend with the impact of measurement error in remote exposure, genetic susceptibility markers, intermediate endpoints, and precursor lesions. Although careful consideration of the possible effects of misclassification of key variables helps our understanding of the nature and degree of the ensuing biases on the measures of association, a preventive approach to minimizing misclassification is a preferred solution. This involves not only the development and validation of better survey instruments and laboratory assays, but also the design of epidemiologic studies that can properly measure the dynamic changes occurring as the early events in the natural history of cancer.