Fig. 2.  The numeric rating scale (NRS) pain score after discharge (by repeated-measures analysis of variance). Plots show mean NRS pain scores (mean ± SD) at different time points after discharge. A between-subject factor group was significant at level P = 0.02 (partial η squared = 0.115), and a within-subject factor time was significant at level P  less than 0.0005 (partial η squared = 0.432). An interaction (group × time) was also significant at level P = 0.001 (partial η squared = 0.130). The post hoc  comparisons with Bonferroni correction on factor group was found to be significant between the additive and LA control groups (P = 0.005) and between the additive and intramuscular control groups (P = 0.009). The post hoc  comparisons with Bonferroni correction on factor time was significant between adjacent time points (6 and 12 h: P = 0.004, partial η squared = 0.113; 12 and 24 h: P = 0.012, partial η squared = 0.086). Because the interaction (group × time) was significant, we also analyzed differences within groups across times with Bonferroni correction. In the LA control group, we found differences between 6 and 12 h (P = 0.036, partial η squared = 0.190). In the intramuscular control group, we found differences between 12 and 24 h (P = 0.001, partial η squared = 0.324). In the additive group, we found differences between 24 and 36 h (P = 0.028, partial η squared = 0.202). * Statistically significant difference within groups between certain times (P  less than 0.05). Group 1: LA control = local anesthetic block + intramuscular placebo; group 2: intramuscular control = local anesthetic block + intramuscular buprenorphine; group 3: additive = local anesthetic/buprenorphine block + intramuscular placebo.