Table 3. Long-term Benzodiazepine Use Among Simultaneous New Users by Baseline Patient Characteristic and Initial Prescription Detailsa.
| Variable | Non–Long-term Benzodiazepine Usersb (n = 54 900) |
Long-term Benzodiazepine Usersb (n = 6582) |
Long-term Benzodiazepine Use, RR (95% CI) | |
|---|---|---|---|---|
| Crude | Multivariablec | |||
| Sex | ||||
| Male | 17 091 (31.1) | 2610 (39.7) | 1 [Reference] | 1 [Reference] |
| Female | 37 809 (68.9) | 3972 (60.3) | 0.72 | 0.78 (0.74-0.81) |
| Age, median (IQR), y | 39 (31-49) | 42 (32-51) | NA | NA |
| Age group, y | ||||
| 18-24 | 6806 (12.4) | 706 (10.7) | 1 [Reference] | 1 [Reference] |
| 25-34 | 12 740 (23.2) | 1270 (19.3) | 0.96 | 1.00 (0.92-1.09) |
| 35-49 | 22 642 (41.2) | 2753 (41.8) | 1.15 | 1.15 (1.07-1.25) |
| 50-64 | 12 712 (23.2) | 1853 (28.2) | 1.35 | 1.27 (1.17-1.38) |
| Depression diagnosis (index diagnosis)d | ||||
| Not otherwise specified | 22 134 (40.3) | 2203 (33.5) | 1 [Reference] | 1 [Reference] |
| Major depressive disorder | 15 411 (28.1) | 2744 (41.7) | 1.67 | 1.19 (1.12-1.26) |
| Other | 17 355 (31.6) | 1635 (24.8) | 0.95 | 0.95 (0.90-1.01) |
| Specialty of physician who provided index depression diagnosis | ||||
| Family practice | 18 655 (34.0) | 1675 (25.4) | 1 [Reference] | 1 [Reference] |
| Psychiatry | 8661 (15.8) | 1884 (28.6) | 2.17 | 1.37 (1.27-1.47) |
| Other specialty | 18 185 (33.1) | 1968 (29.9) | 1.19 | 1.08 (1.02-1.15) |
| Psychologist, therapist | 4826 (8.8) | 524 (8.0) | 1.19 | 1.03 (0.94-1.14) |
| Unknown or multiple physicians | 4573 (8.3) | 531 (8.1) | 1.26 | 1.08 (0.98-1.18) |
| Short- vs long-acting benzodiazepine | ||||
| Short or intermediate acting | 41 820 (76.2) | 4283 (65.1) | 1 [Reference] | 1 [Reference] |
| Long actinge | 13 080 (23.8) | 2299 (34.9) | 1.61 | 1.28 (1.22-1.34) |
| Benzodiazepine index days’ supply,e d | 15 (10-30) | 30 (15-30) | ||
| 1-7 | 7948 (14.5) | 373 (5.7) | 1 [Reference] | 1 [Reference] |
| 8-21 | 24 154 (44.0) | 1790 (27.2) | 1.54 | 1.46 (1.30-1.64) |
| 22-35 | 22 041 (40.1) | 4103 (62.3) | 3.50 | 2.91 (2.54-3.33) |
| >35 | 756 (1.4) | 316 (4.8) | 6.58 | 4.90 (4.15-5.79) |
| Initial benzodiazepine quantity dispensed,e pills | 30 (30-50) | 30 (30-60) | ||
| 1-24 | 13 171 (24.0) | 767 (11.7) | 1 [Reference] | 1 [Reference] |
| 25-50 | 28 306 (51.6) | 3252 (49.4) | 1.87 | 1.18 (1.08-1.29) |
| >50 | 13 422 (24.4) | 2563 (38.9) | 2.91 | 1.14 (1.01-1.27) |
| Anxiety disorder diagnosis, recent | ||||
| Generalized anxiety disorder | 3443 (6.3) | 529 (8.0) | 1.27 | 1.16 (1.07-1.25) |
| Panic disorder | 1735 (3.2) | 249 (3.8) | 1.18 | 1.15 (1.02-1.29) |
| PTSD | 496 (0.9) | 75 (1.1) | 1.23 | 1.08 (0.83-1.39) |
| OCD, phobic disorder, or other anxiety | 893 (1.6) | 171 (2.6) | 1.51 | 1.27 (1.10-1.45) |
| Nonrecent insomnia or sleep disturbance diagnosis | 1939 (3.5) | 329 (5.0) | 1.37 | 1.23 (1.11-1.37) |
| Acute reaction to stress | 1307 (2.4) | 102 (1.5) | 0.67 | 0.76 (0.63-0.92) |
| Outpatient problem-oriented visits | ||||
| 0-1 | 10 592 (19.3) | 1548 (23.5) | 1 [Reference] | 1 [Reference] |
| 2-6 | 32 327 (58.9) | 3531 (53.6) | 0.77 | 0.86 (0.81-0.92) |
| >6 | 11 981 (21.8) | 1503 (22.8) | 0.87 | 0.83 (0.76-0.90) |
| Opioid prescription, recent | 3723 (6.8) | 773 (11.7) | 1.69 | 1.50 (1.39-1.61) |
Abbreviations: IQR, interquartile range; NA, not applicable; OCD, obsessive-compulsive disorder; PTSD, posttraumatic stress disorder; RR, risk ratio.
Characteristics were defined using inpatient and outpatient International Classification of Diseases, Ninth Revision, Clinical Modification diagnostic codes; Current Procedural Terminology, fourth edition, procedure codes; and outpatient dispensed prescriptions. A full list of characteristics is given in eTable 3 in the Supplement.
Data are presented as number (percentage) of patients unless otherwise indicated.
Multivariable RRs are adjusted for all the variables in eTable 3 in the Supplement and include, in addition to the variables in Table 3, the following: antidepressant class, inpatient depression diagnosis, benzodiazepine dose per day, psychiatric comorbidities (attention-deficit/hyperactivity disorder, adjustment disorder, or other episodic mood disorder), treated self-harm, nonrecent anxiety diagnosis, recent insomnia or sleep disturbance diagnosis, fall or fracture risk factures (fracture, fall injury, cataracts or glaucoma, or visual loss or disturbance), poisoning, nonpsychiatric comorbidities (cerebrovascular disease, myocardial infarction, chronic kidney disease, osteoarthritis and allied disorders, Crohn disease or gastroenteritis, hearing problem, overweight or obese, or urinary incontinence), health care use measures (inpatient admission, medication use), region, and year of treatment start. One person was excluded from multivariable analyses who started taking a benzodiazepine in solution form. The full model resulted in a C statistic of 0.700.
For index depression diagnosis, other includes adults diagnosed with dysthymic disorder, adjustment disorder with depressed mood, or prolonged depressive reaction; not otherwise specified includes adults without a specific diagnosis; and recent refers to measures within 30 days before antidepressant treatment initiation.
Days’ supply and quantity values of 0 or less were set to the median value (30), days’ supply values greater than 99 were truncated to 100, and quantity values greater than 200 were truncated to 200 (2% of all prescriptions). For patients with multiple agents (0.9%), if one agent was long acting, patients were classified as initiating treatment with a long-acting agent.