Risk factors and their identification second part: study designs for identification of risk factors.

This is the second a series of three articles which reviews the identification of risk factors of a disease, here: diabetes or complications of diabetes. In the first of the series [1], we gave the definition of a risk factor, along with measures of its force-relative risk and odds ratio, followed by the epidemiological definitions of the diseases: diabetes, coronary heart disease and hypertension. Risk factors were further discussed and we completed the discussion by some observations on the bias which can arise from a study or from its analysis, which can lead the researcher to the wrong conclusion. In this second article we define the three types of epidemiological studies which are used to determine whether factors are associated with a disease: observational or cross-sectional studies, cohort studies and casecohort studies. Examples are provided of each of these study types; their advantages and disadvantages are discussed. The final paper will provide some examples of the identification of risk factors from the literature. The first example involves diabetes and pancreatic cancer, the second birth weight and non-insulin dependent diabetes. Having found an association between a risk factor and diabetes, we will discuss whether it can be considered to be a risk factor, and if so whether it is likely to be a cause of the disease.