Fig. 1. This figure depicts our heuristic model for the association between postoperative delirium and opioid use. In this model, we describe that delirium status was represented by the four criteria used to measure the presence of delirium. Our hypotheses were (1) rest pain on postoperative day one affected delirium status on day one, (2) opioid dose on day one affected delirium status on day one, and (3) delirium status on day one did not affect opioid dose on day two.