Researchers' and users' perceptions of the relative standing of biomedical papers in different journals

Journal citation impact factors, which are frequently used as a surrogate measure of research quality, do not correlate well with UK researchers' subjective views of the relative importance of journals as media for communicating important biomedical research results. The correlation varies with the sub-field: it is almost zero in nursing research but is moderate in more "scientific" subfields such as multiple sclerosis research, characterised by many authors per paper and appreciable foreign co-authorship. If research evaluation is to be based on journal-specific indicators, then these must cover different aspects of the process whereby research impacts on other researchers and on healthcare improvement.