Unregistered Submission 24 Mar 2017 Erratum:The odds ratio for Insulin in table 3 should probably be 1.090 not 1090. Problem with the statistics: 22 drugs were tested for two outcomes and 9 groups of drugs with the same two outcomes...