We should create an issue implementation that is based on a document, (which has those fields normalised) for a performance increase.