 |  |
Peter describes the work presented as follows:
“I showed the group's work on mapping Norwegian Varenummer to OMOP standard vocabularies. Across 339 million dispensations, automated mapping followed by manual curation raised full drug identity (ingredient, strength, form) from 64% to 89%, with 99.9% mapped overall.
The main lesson is that drug mapping and its validation are a project in their own right, needing yearly maintenance as product catalogues and vocabularies change.
The work improved PharmaSafe's contribution to federated research networks, where we have participated in more than 20 studies using OMOP and the ConcePTION CDM.
Many questions followed, especially on how AI could reduce the manual curation burden.
Thanks to the mapping team, Zeenat F. Chatha, Hedvig M. E. Nordeng, and Saeed Hayati, and to Sitra and NOR-OMOP.”