Tracing the indirect societal impacts of biomedical research: development and piloting of a technique based on citations.

There is growing interest in assessing the societal impacts of research such as informing health policies and clinical practice, and contributing to improved health. Bibliometric approaches have long been used to assess knowledge outputs, but can they also help evaluate societal impacts? We aimed to see how far the societal impacts could be traced by identifying key research articles in the psychiatry/neuroscience area and exploring their societal impact through analysing several generations of citing papers. Informed by a literature review of citation categorisation, we developed a prototype template to qualitatively assess a reference's importance to the citing paper and tested it on 96 papers. We refined the template for a pilot study to assess the importance of citations, including self-cites, to four key research articles. We then similarly assessed citations to those citing papers for which the key article was Central i.e. it was very important to the message of the citing article. We applied a filter of three or more citation occasions in order to focus on the citing articles where the reference was most likely to be Central. We found the reference was Central for 4.4 % of citing research articles overall and ten times more frequently if the article contained three or more citation occasions. We created a citation stream of influence for each key paper across up to five generations of citations. We searched the Web of Science for citations to all Central papers and identified societal impacts, including international clinical guidelines citing papers across the generations.