We are a team of statisticians with a tool that would enable valid and “safe” causal claims to be made about the effects of COVID-19 treatments being attempted outside of RCTs, such as hydroxychloroquine, Remdesivir, and an increasing number of other agents. This tool may provide wider estimates than hoped but has the benefit of tolerating the confounding/selection that would bias ordinary analyses. We urgently require data from more hospitals where new treatments are being tried. Fortunately we do not need personally identifiable data, just a few summary statistics we can describe in further conversation. Please let us know if you have such data, and we can collaborate or help you with the analysis.