Table 5.
Changes in pain from baseline to 6 months as predictors of opioid use at 6 months for opioid users and opioid naïve on day of surgery separately.
| On opioids day of surgery | Odds Ratio | Relative Risk | Standard Error | AUC | p-value | 95% | CI |
|---|---|---|---|---|---|---|---|
| Change in overall body pain (BPI) | 0.62 | 0.78 | 0.09 | 0.43 | 0.001 | 0.46 | 0.82 |
| Change in surgical site pain (WOMAC) | 1.00 | 1.01 | 0.07 | 0.50 | 0.963 | 0.88 | 1.14 |
| Preoperative overall pain (BPI) | 1.24 | 1.11 | 0.24 | 0.53 | 0.269 | 0.85 | 1.82 |
| Preoperative overall pain (WOMAC) | 1.09 | 1.04 | 0.11 | 0.51 | 0.418 | 0.89 | 1.33 |
| Caucasian | 0.95 | 0.90 | 0.76 | 0.49 | 0.950 | 0.20 | 4.54 |
| Age | 0.97 | 0.99 | 0.02 | 0.50 | 0.203 | 0.93 | 1.02 |
| Female | 1.67 | 1.30 | 0.77 | 0.58 | 0.265 | 0.68 | 4.13 |
| Knee surgery (vs. Hip surgery) | 1.23 | 1.09 | 0.59 | 0.53 | 0.669 | 0.48 | 3.12 |
| Not on opioids day of surgery | Odds Ratio | Relative Risk | Standard Error | AUC | p-value | 95% | CI |
|---|---|---|---|---|---|---|---|
| Change in overall body pain (BPI) | 0.72 | 0.78 | 0.12 | 0.45 | 0.050 | 0.52 | 1.00 |
| Change in surgical site pain (WOMAC) | 0.99 | 0.99 | 0.11 | 0.50 | 0.939 | 0.80 | 1.23 |
| Preoperative overall pain (BPI) | 2.76 | 2.31 | 0.74 | 0.65 | <0.001 | 1.64 | 4.65 |
| Preoperative overall pain (WOMAC) | 0.89 | 0.92 | 0.14 | 0.48 | 0.459 | 0.65 | 1.21 |
| Caucasiana | – | – | – | – | – | – | |
| Age | 1.06 | 1.05 | 0.04 | 0.51 | 0.096 | 0.99 | 1.13 |
| Female | 0.37 | 0.42 | 0.22 | 0.35 | 0.091 | 0.11 | 1.18 |
| Knee surgery (vs. Hip surgery) | 3.89 | 3.36 | 2.62 | 0.70 | 0.044 | 1.04 | 14.6 |
Note. Overall model AUC for people on opioids day of surgery = 0.75. Overall model AUC for people not on opioids day of surgery = 0.853. AUC for each predictor is included as a measure of effect size. BPI = Brief Pain Inventory; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index. Change in overall body pain calculated by subtracting mean BPI Overall Body Pain score at 6 months from mean BPI Overall Body Pain score at day of surgery. Thus, positive values indicate decreases in overall body pain. Change in surgical site pain calculated by subtracting mean WOMAC Pain score at 6 months from WOMAC Pain score at day of surgery. Positive scores indicate decreases in surgical site pain from baseline to 6 months.
All participants were Caucasian in this sample so this variable was excluded from the model.