Introductive remarks on causal inference

Silvana A. Romio, Rino Bellocco, Giovanni Corrao

Abstract


One of the more challenging issues in epidemiological research is being able to provide an unbiased estimate of the causal exposure-disease effect, to assess the possible etiological mechanisms and the implication for public health. A major source of bias is confounding, which can spuriously create or mask the causal relationship. In the last ten years, methodological research has been developed to better de_ne the concept of causation in epidemiology and some important achievements have resulted in new statistical models. In this review, we aim to show how a technique the well known by statisticians, i.e. standardization, can be seen as a method to estimate causal e_ects, equivalent under certain conditions to the inverse probability treatment weight procedure.

Full Text:

PDF (English)

References


S. GREENLAND, (2004), Applied Bayesian Modeling and Causal Inference form Incomplete-Data Perspectives, Wiley & Sons.

S. GREENLAND, J. PEARL, (2006), Causal Diagrams, Computer Science Department. University of California, Technical Report R-332.

S. GREENLAND, J.M. ROBINS, J. PEARL, (1999), Confounding and Collapsability in Causal Inference, “Statistical Science”, 14:29-46.

M.A. HERNÀN, (2004), A definition of causal effect for epidemiological research, “Journal of Epidemiology and Community Health”, 58:265-271.

M.A. HERNÀN, J.M. ROBINS, (2006), Estimating causal effect for epidemiological data, “Journal of Epidemiology and Community Health”, 60:578-586.

D.G. HORVITZ, D.J. THOMPSON, (1952), A generalization of sampling without replacement from a finite population, “Journal of the American Statistical Association”, 647:663-685.

D. HUME, (1748), An Enquiry Concerning Human Understanding, section VII, part II.

O. MIETTINEN, E.F. COOK, (1981), Confounding: Essence and Detection, “American Journal of Epidemiology”, 114:593-603.

A. STEPTOE, L. BRYDON, (2005), Association between acute lipid stress responses and fasting lipid levels 3 years later, “Health Psychology”, 24:601-607.




DOI: 10.6092/issn.1973-2201/3591