This is great. It also underscores a methodological opportunity - is there a way to group together a huge amount of studies, regardless of disease or treatment,

This is great. It also underscores a methodological opportunity - is there a way to group together a huge amount of studies, regardless of disease or treatment, and create a semantic association graph that takes things like climate, culture, geography, and the like into account? You could rank researchers and institutions and locations in order to weight contributions to meta-studies, but you would be also able to associate confounders like parasites almost immediately. I have to imagine it would be the ultimate tool in identifying novel uses for existing drugs. If this could be done with open source software, it could be a killer app for scihub. Medicine, nutrition, chemical processes, energy science, all sorts of things could benefit the world. Hell, what if there's a correlation with elevation? It looks like there's a lower death rate at higher elevation - how much air pressure is the threshold for viruses getting into lungs? Could a strong coughing fit at sea level be more dangerous than one at the summit of everest? Does less coughing occur in low humidity, low dust places?