Andy C Leon was a mensch. One of the highest praises that can be given in Yiddish is to call someone a mensch. Literally, the word simply means ‘human being’. Figuratively, the word has deep harmonics and resonances; a mensch is someone who is a real human being, someone of great character, someone with integrity, decency and goodness, someone who knows right from wrong and chooses right, someone who is responsible with dignity and fortitude, and someone to admire and aspire to be like. Andy was a mensch.

Andy gave without reservation and without expectation of anything in return. He gave for the pleasure and satisfaction of giving. He spent years training new investigators and was generous with his time and wise advice. His intellectual powers were impressive. Yet, when he helped young investigators, who were struggling to design a study or interpret statistics, he would gently correct them, never criticizing or humiliating them. He would say, ‘Well, I have a slightly different take on that and you may want to consider approaching the problem this way’.

As a colleague, Andy was the best. He was open, collaborative, inventive, and innovative. He loved the give and take, and was delighted when he helped find a solution to a problem. He was a lover of jazz and this spilled over to his work, able to improvise with deep knowledge when no clear solution was available. He adapted a metric, ‘intent to attend’, that allowed an analysis of the assumption of ignorable attrition for mixed models. When faced with the problem of what to do with a hybrid comparative effectiveness trial, he figured out that if, on top of randomization to the study treatments, participants could receive other treatments needed for their treatment, then any differences in clinical outcomes between the study treatments would be blunted. He determined that it would make sense to have, in addition to the symptom outcomes, a co-primary outcome measure that would quantify the need for additional treatments. He called these ‘Necessary Clinical Adjustments’ or NCAs (the average number of changes in medications per unit of time). In this way, if treatment A and treatment B had similar outcomes, but treatment A required no NCAs and treatment B required 100 NCAs, then A was the winner.

Andy was a good friend. We shared music, laughs, and railed against injustice and narrow mindedness together. He was intolerant of intolerance. I will miss his good company and wise counsel. The world is a better place because he was here.