Table 2.
Modeling interrupted time series to evaluate the impact of the Narcotics Information Management System (NIMS) and COVID-19 on opioid use among outpatients with musculoskeletal and connective tissue disorders in South Korea.
| Outcome variables and estimator (ARa) | Segmented regression coefficient | P value | D-Wb | ||
| Proportion of patients on high-dose opioid treatment | 1.9426 | ||||
|
|
Intercept (β0) | 0.7947 | <.001 |
|
|
|
|
Baseline trend (β1) | 0.0040 | .48 |
|
|
|
|
Level change after NIMS (β2) | 0.1743 | .20 |
|
|
|
|
Trend change after NIMS (β3) | 0.0271 | .01 |
|
|
|
|
Level change after COVID-19 (β4) | 0.0484 | .78 |
|
|
|
|
Trend change after COVID-19 (β5) | 0.0309 | .19 |
|
|
| Proportion of patients receiving opioid prescriptions from multiple providers | 2.0825 | ||||
|
|
Intercept (β0) | 0.8141 | <.001 |
|
|
|
|
Baseline trend (β1) | –0.0076 | .31 |
|
|
|
|
Level change after NIMS (β2) | 0.6252 | .004 |
|
|
|
|
Trend change after NIMS (β3) | –0.0067 | .59 |
|
|
|
|
Level change after COVID-19 (β4) | 0.3969 | .09 |
|
|
|
|
Trend change after COVID-19 (β5) | 0.0323 | .28 |
|
|
| Overlap rate of opioid prescriptions per patient | 2.0046 | ||||
|
|
Intercept (β0) | 1.4828 | <.001 |
|
|
|
|
Baseline trend (β1) | –0.0113 | .16 |
|
|
|
|
Level change after NIMS (β2) | 0.3349 | .08 |
|
|
|
|
Trend change after NIMS (β3) | 0.0101 | .46 |
|
|
|
|
Level change after COVID-19 (β4) | 0.3709 | .12 |
|
|
|
|
Trend change after COVID-19 (β5) | –0.0442 | .16 |
|
|
| Naloxone use rate among opioid users | 1.9854 | ||||
|
|
Intercept (β0) | 0.3685 | .001 |
|
|
|
|
Baseline trend (β1) | 0.0156 | .01 |
|
|
|
|
Level change after NIMS (β2) | –0.2968 | .04 |
|
|
|
|
Trend change after NIMS (β3) | –0.0117 | .26 |
|
|
|
|
Level change after COVID-19 (β4) | 0.0652 | .72 |
|
|
|
|
Trend change after COVID-19 (β5) | –0.0324 | .18 |
|
|
aAR: first-order autocorrelation maximum likelihood estimate.
bD-W: Durbin-Watson test.