Results
There were 83 participants, 49 in the PRN narcotic group and 34 assigned to ERAS. Sixty-six percent were male and the mean age was 53 years (Table 1 ). There were no differences in patient characteristics between PRN and ERAS groups; this includes no significant different in age (54.37±13.56 PRN vs 51.76±14.83; p=0.41), history of IV drug use (n=4 PRN vs. n=2 ERA; p=0.69) or prevalence of prior pain medication prescriptions (n=11 PRN vs n=10 ERA; p=0.47) The type of operations included was heterogeneous and included coronary artery bypass grafting (CABG) alone (n=27; 32%), valve surgery alone (n=15; 18%), CABG + valve (n=3; 4%), CABG + other (n=2;2%), Valve + other (n=13;16%), and other procedures (n=23; 28%), with the vast majority of these procedures consisting of thoracic aortic repair. All procedures were performed via median sternotomy. Surgical characteristics are described in Table 2 . There was no difference in hospital mortality (n=1 PRN vs n=1ERA; p=0.79), hospital length of stay (16.8 days PRN vs. 15.35 days ERAS; p=0.66), or ICU length of stay (6.88 days PRN vs 6.88 days ERAS; p=0.99). There was a trend toward significant for operative cardiopulmonary bypass times (145.41 minutes PRN vs. 114.85 minutes ERAS; p=0.06) but no difference in aortic cross clamp times (97 minutes PRN vs 75.82 minutes ERAS; p=0.11).
There were no supported differences in patient-reported outcome and pain control between the ERAS protocol and traditional narcotic PRN-only pain medications (Table 3 ). As an illustrative example of patient-reported trajectories at the 4 time points, Figure 3shows percentage of patients reporting headache either none of the time (green) or some of the time (orange) for both ERAS (solid lines) and PRN (dashed lines). For this particular outcome, headache remained stable across time, with those on the ERAS protocol displaying a higher proportion of having no headache. Trajectory relationships for other outcomes are shown in Figures 4 and 5 . Low-level presence of almost all patient reported adverse outcomes and pain was observed across the full battery of symptoms reported (Figures 3-5 ), with the exception of moderate pain being reported in equal numbers across all time points.