<< Previous
Next >>

Paul Brennan, George Davey-Smith, Debby Lawlor,
Rodolfo Saracci, Duncan Thomas

Discrepancies between results from observational studies and randomized trials, as for example on the beneficial and adverse effects of hormonal replacement therapy, have raised fresh questions on the extent to which confounding can be effectively controlled in observational studies. The course will offer in the first part a detailed critical presentation of methods for confounding control while the second part will be entirely devoted to the novel approach of Mendelian randomization in its principles and first applications to epidemiological studies. The course is addressed to students already conversant with epidemiological methods and the practice of data analysis.