TY - JOUR T1 - Improving the evaluation of worldwide biomedical research output: classification method and standardised bibliometric indicators by disease JF - BMJ Open JO - BMJ Open DO - 10.1136/bmjopen-2017-020818 VL - 8 IS - 6 SP - e020818 AU - Lissy van de Laar AU - Thijs de Kruif AU - Ludo Waltman AU - Ingeborg Meijer AU - Anshu Gupta AU - Niels Hagenaars Y1 - 2018/06/01 UR - http://bmjopen.bmj.com/content/8/6/e020818.abstract N2 - Objective Since most biomedical research focuses on a specific disease, evaluation of research output requires disease-specific bibliometric indicators. Currently used methods are insufficient. The aim of this study is to develop a method that enables detailed analysis of worldwide biomedical research output by disease.Design We applied text mining techniques and analysis of author keywords to link publications to disease groups. Fractional counting was used to quantify disease-specific biomedical research output of an institution or country. We calculated global market shares of research output as a relative measure of publication volume. We defined ‘top publications’ as the top 10% most cited publications per disease group worldwide. We used the percentage of publications from an institution or country that were top publications as an indicator of research quality.Results We were able to classify 54% of all 6.5 million biomedical publications in our database (based on Web of Science) to a disease group. We could classify 78% of these publications to a specific institution. We show that between 2000 and 2012,‘other infectious diseases’ were the largest disease group with 337 485 publications. Lifestyle diseases, cancers and mental disorders have grown most in research output. The USA was responsible for the largest number of top 10% most cited publications per disease group, with a global share of 45%. Iran (+3500%) and China (+700%) have grown most in research volume.Conclusions The proposed method provides a tool to assess biomedical research output in new ways. It can be used for evaluation of historical research performance, to support decision-making in management of research portfolios, and to allocate research funding. Furthermore, using this method to link disease-specific research output to burden of disease can contribute to a better understanding of the societal impact of biomedical research. ER -