Incidence of disorders tested by systematic screening: confidence limits and comparison of programmes.

Using the data from screening done in Switzerland since 1965, we showed that the probability of finding cases with increased phenylalanine or leucine concentrations is compatible with a Poisson distribution. If there are no trends present from year to year and if the Poisson distribution is an accurate fit further statistical treatment is possible. The advantage of using the Poisson distribution is exemplified by the calculation of confidence limits for the incidence of hyperphenylalaninaemia, maple syrup urine disease and hypothyroidism. Furthermore, statistical comparisons between different screening programs are easily done.