In recent years there have been a number of projects and initiatives at local and national level to develop clinical information systems either for small-scale research or for wide-scale collation of information. The SHINDIG project undertook to provide a solution to the exchange of diabetes information between a general practice (GP) clinic and two hospitals. The consortium elected to base the development of the system on the information model developed for the EU-funded Good European Health Record (GEHR) project for which the development of a proof-of-concept prototype information system (known as PRISM) was already well underway. This paper looks at the aims, achievements and problems encountered during the SHINDIG project and draws a number of lessons from the development of one of the first systems to attempt to implement the GEHR architecture. Recommendations for consideration are also given to those embarking on similar projects and to inform the continued work in the healthcare record standards arena.