When a single diagnostic test is used to assess the disease state of a
patient, the classic univariate measures of sensitivity and specifici
ty are used. When combining two or more diagnostic tests, other method
s provide alternative methods for assessing the utility of combining t
hese tests. We have written and tested a program in FORTRAN that compu
tes the univariate statistics and bivariate concordant/discordant stat
istics analyses of all possible combinations of diagnostic tests as co
mpared with the 'gold standard'.