Key Points Drug discovery has traditionally responded to the 'pull' of unmet medical need and commercial potential. Here, we evaluate the scientific areas that are providing 'push' in terms of scientific innovation, as measured by publications and patents. The amount of funding from the US National Institutes of Health, the number of doctoral degrees awarded, the number of publications and the number of approvals for new molecular entities issued by the US Food and Drug Administration show a general upward trend from the 1950s to the present. Although explanations can be suggested for some of the short-term ups and downs, these trends are also affected by numerous hidden variables, and well as time lags, feedback loops and other complications. The numbers of publications in various disease areas correlate well with global disease burden.The correlation is greater in the developed world, where the impact of infectious and parasitic diseases, respiratory infections and infant mortality, is considerably lower than in the developing world. A few therapeutic areas stand out when analysing trends in publications, citations, publications in high-impact journals and patents. Oncology is the clearest outlier on almost every metric. Viruses, nutrition and metabolism, and the immune system also show increases in the majority of metrics. In terms of individual diseases, insulin resistance, orthomyxoviridae infections, depression, autism, macular degeneration, inflammation, obesity, cognitive disorders and ventricular dysfunction have strong publication growth in both 2- and 5-year periods. Publications related to the genes encoding forkhead box P3, leucine-rich repeat kinase 2, janus kinase 2, transcription factor 7-like 2, interleukin-17A, toll-like receptor 2 (TLR2), TLR4, FK506 binding protein 12-rapamycin associated protein 1and ADIPOQ (adiponectin, C1Q and collagen domain-containing) show the strongest publication growth in the same periods. Assessing scientific innovation is a complex endeavour; however, the 'bibliome' seems to offer many approaches that could enhance decision-making in drug discovery. Supplementary information The online version of this article (doi:10.1038/nrd2973) contains supplementary material, which is available to authorized users.