Using bibliometric analyses of patent literature for predicting the clinical fates of developing drugs.
Certain bibliometric features of the early literatures of developing drugs can be used to predict their ultimate clinical fates. The chronological sequence of publications is expressed as a binary vector with 1 for a patent and 0 for a nonpatent. The decimal equivalents for standardized vector lengths provide scalar values for comparing one drug with another. In order to incorporate concordant patents, fuzzy subsets are employed, with the number of attempts required to achieve transitive closure being the values for comparison. The methods involved are described using minoxidil as an example.