A rebuild of a landmark randomized study of health insurance (NEJM, 2013) in Python and R, plus a check of how well its evidence supports its headline claims.
