Table 1. Characteristics of Study Cohorta.
| Characteristic | Commercial Insurance | Medicare Advantage | ||
|---|---|---|---|---|
| Long-term Opioid Episode | Short-term Opioid Episode | Long-term Opioid Episode | Short-term Opioid Episode | |
| Categorical variablesb | ||||
| Femaleb | 52.7 (0.25) | 55.7 (0.03) | 62.4 (0.22) | 61.9 (0.06) |
| Race/ethnicityb | ||||
| Non-Hispanic white | 77.9 (0.20) | 71.5 (0.03) | 70.5 (0.21) | 67.0 (0.06) |
| Non-Hispanic black | 10.2 (0.15) | 10.7 (0.02) | 17.1 (0.17) | 17.0 (0.05) |
| Hispanic | 7.96 (0.13) | 11.0 (0.02) | 7.88 (0.12) | 9.65 (0.04) |
| Non-Hispanic Asian | 1.14 (0.05) | 3.0 (0.01) | 1.09 (0.04) | 1.85 (0.02) |
| Unknown or other | 2.82 (0.08) | 3.7 (0.01) | 3.50 (0.08) | 4.59 (0.02) |
| Selected comorbiditiesc | ||||
| Hypertension without complications | 19.7 (0.14) | 9.94 (0.02) | 42.2 (0.17) | 38.6 (0.04) |
| Diabetes without complications | 7.03 (0.10) | 4.04 (0.01) | 17.50 (0.14) | 16.4 (0.04) |
| Chronic pulmonary disease | 4.48 (0.07) | 2.37 (0.01) | 14.1 (0.12) | 10.0 (0.03) |
| Depression | 7.39 (0.09) | 3.21 (0.01) | 11.6 (0.11) | 6.19 (0.02) |
| Diabetes with complications | 3.57 (0.07) | 1.64 (0.01) | 11.0 (0.11) | 9.56 (0.03) |
| Concurrent opioid and benzodiazepines prescriptions, mob | 23.00 (0.18) | 8.99 (0.02) | 25.70 (0.18) | 14.00 (0.04) |
| Proportion of opioid prescription days overlapping with benzodiazepinesb | ||||
| All person-months | 18.7 (0.17) | 9.27 (0.03) | 22.0 (0.18) | 13.6 (0.05) |
| Person-months with any overlap days | 79.3 (0.16) | 78.5 (0.06) | 83.9 (0.13) | 83.4 (0.06) |
| Person-months, No. | 168 810 | 583 020 | 228 195 | 454 771 |
| Same physician prescribing opioids and benzodiazepines (of person-months with any overlap)b | ||||
| Yes | 59.1 (0.12) | 51.8 (0.07) | 60.3 (0.10) | 50.4 (0.07) |
| No | 37.9 (0.12) | 39.5 (0.06) | 30.0 (0.10) | 39.3 (0.07) |
| Unknown | 2.99 (0.04) | 8.8 (0.04) | 9.7 (0.06) | 10.4 (0.05) |
| Person-months, No. | 168 810 | 583 020 | 228 195 | 454 771 |
| Opioid dose per opioid use day, morphine milligram equivalents | ||||
| <20 | 28.37 (0.19) | 18.94 (0.02) | 39.52 (0.19) | 33.87 (0.05) |
| 20 to <50 | 49.04 (0.20) | 59.11 (0.03) | 42.52 (0.19) | 53.01 (0.05) |
| 50 to <90 | 13.17 (0.12) | 18.36 (0.02) | 10.04 (0.11) | 10.60 (0.02) |
| 90 to <120 | 3.70 (0.06) | 2.36 (0.01) | 2.95 (0.06) | 1.59 (0.01) |
| ≥120 | 5.73 (0.10) | 1.23 (0.01) | 4.97 (0.10) | 0.94 (0.01) |
| Continuous variables | ||||
| Age, median (interquartile range), y | 51 (41-58) | 46 (34-56) | 70 (61-77) | 72 (67-78) |
| Opioid use, median (interquartile range), mo | ||||
| Opioid prescription days | 28 (19-30) | 5 (3-8) | 28 (20-30) | 7 (4-15) |
| Morphine milligram equivalents opioids | 630 (340-1200) | 150 (80-240) | 520 (300-990) | 160 (90-300) |
| Observations, No. | ||||
| Person-months | 734 545 | 6 485 891 | 889 459 | 3 255 263 |
| Unique beneficiariesd | 59 423 | 3 564 678 | 69 153 | 1 262 727 |
Except where noted otherwise, data are percentage of person-months (SE).
The SEs for proportions were clustered on beneficiary identification to account for multiple observations of the same beneficiaries.
Elixhauser comorbidities were calculated on a rolling 6-month basis using International Classification of Diseases, Ninth Revision and Tenth Revision codes from Quan et al.21
Some beneficiaries had both short-term and long-term opioid episodes while observed. Each is counted as a unique beneficiary in both categories of opioid use. The total number of unique commercial beneficiaries was 3 598 322; the total number of unique Medicare beneficiaries was 1 299 142. A small number of people appeared both as commercially insured and Medicare beneficiaries at different times (3708 [0.08%]).