Dodgy data: national patent records

The world is defined by standards. Thanks to technical standards we have matching bolts and nuts, power plugs that fit and many other daily things we don’t even think about. Next to these strict and well documented standards, there are industry standards, which are defined by generally accepted requirements or norms in a particular industry.

In the patent information world, the standard is set by DOCDB/INPADOC (DI), a dataset gathered from many different patent authorities, and brought together by the European Patent Office (EPO). This complex dataset offers content from 106 different countries and contains more than 100 million patent records. The DOCDB part covers bibliographic information and patent family information, the INPADOC (International Patent Documentation) part provides data on the status of a patent.

With DI, the patent information industry has its standard for access to the world’s patent publications and it is used widely.

Gaps in completeness

When a closer look at this dataset is taken, it becomes clear that it is far from perfect. Of the 106 listed authorities, six no longer exist although they are still relevant for the state of the art; this leaves the dynamic corpus of DI covering only 100 countries. Of the 100 countries covered, only around 50 supply more or less regular updates, implying that information from the other 50 countries is not regularly updated, is incomplete, or is not updated at all.

Hence, the global patent information standard of 106 countries actually covers just 50 countries. For all patent information users, this should be of serious concern, because one cannot afford to miss out on any relevant patent out there.