Surnames as markers of pathologies--two statistical techniques and their applications.

The objective of this research is to propose and to validate two different statistical techniques to test the hypothesis of an association between surnames and pathologies, in a population participating in a screening procedure for a given pathology. We propose two statistical methods: a first technique is based on the rarefaction method, and second one is based on the principle of resampling, and it can be considered a special case of a randomisation test. Both the techniques are applied to a data set of babies screened for congenital hypothyroidism (CH), and they gave similar results. The large overlapping of the results seems to suggest a substantial validity of the proposed techniques.