Multiple-stressor research is very diverse, so doing comprehensive reviews with traditional tools is almost impossible. We used a machine-learning approach to screen one quarter of a million abstracts and found >2300 studies – an order of magnitude higher than previous reviews!