Abstract
This cohort study evaluates factors associated with the transition to long-term benzodiazepine use among older adults newly prescribed this drug by a nonpsychiatric clinician.
Benzodiazepine use among older adults is common despite evidence for many potential risks. While treatment guidelines recommend short-term use of benzodiazepines, up to one-third of use is long term, which is most common among older adults.1 To reduce benzodiazepine prescribing to older adults, one potential point for intervention is at the transition from new to long-term use, yet little is known about the factors that predict conversion to long-term use.2,3
Methods
We evaluated transition to long-term benzodiazepine use among older adults newly prescribed this drug (n = 576) by a nonpsychiatric clinician from 2008 to 2016 within Pennsylvania’s prescription assistance program for low-income older adults. Patients newly prescribed a benzodiazepine (ie, those with no use in the past year) were contacted for clinical assessment by a telephone-based behavioral health service (Supporting Seniors Receiving Treatment and Intervention [SUSTAIN]).4 The initial assessment included screening instruments assessing for depression, anxiety, sleep quality, and pain. Long-term benzodiazepine use was defined as a medication possession ratio (MPR) greater than 30% in the year following the initial prescription. Logistic regression was used to determine patient sociodemographic and clinical characteristics associated with long-term benzodiazepine use. After completing the clinical interview, participants provided verbal consent to allow the use of clinical data for research purposes in accordance with a University of Pennsylvania Institutional Review Board–approved protocol. Analyses were conducted using SAS statistical software, version 9.4 (SAS Institute Inc).
Results
The mean (SD) age of adults newly prescribed a benzodiazepine was 78.4 (7.0) years (Table 1). One year after the index prescription, 152 (26.4%) met the definition of long-term use; this group was prescribed a mean (SD) of 232.7 (82.6) benzodiazepine days. In adjusted analyses, white race (odds ratio [OR], 4.19; 95% CI, 1.51-11.59), days supplied in the index prescription (OR, 1.94; 95% CI, 1.52-2.47), and poor sleep quality (OR, 4.05; 95% CI, 1.44-11.43) were associated with increased long-term benzodiazepine use (Table 2). High anxiety and depression did not predict long-term benzodiazepine use in either unadjusted or adjusted analyses.
Table 1. Baseline Characteristics of SUSTAIN Enrollees With a Prescription for a Benzodiazepine and Those With Long-term Use, Percentage of Group, and Mean Scores.
| Characteristic | No. (%) | |
|---|---|---|
| Any Benzodiazepine Prescription (N = 576) | Long-term Benzodiazepine Use (n = 152) | |
| Age, mean (SD), y | 78.4 (7.0) | 77.4 (6.7) |
| Male sex | 95 (16.5) | 32 (21.1) |
| White race | 524 (91.0) | 147 (96.7) |
| Married | 148 (25.7) | 43 (28.3) |
| Financial statusa | 511 (88.7) | 136 (89.5) |
| PHQ-9 score, mean (SD) | 6.1 (5.2) | 6.9 (5.6) |
| Depression diagnosisb | ||
| Absent | 437 (75.9) | 110 (72.4) |
| Other depression | 81 (14.1) | 22 (14.5) |
| Major depression | 58 (10.1) | 20 (13.2) |
| High anxietyc | 31 (5.4) | 10 (6.6) |
| Pain interference | ||
| Quite a bit/extremely | 102 (17.7) | 28 (18.4) |
| A little bit/moderately | 242 (42.0) | 73 (48.0) |
| Not at all | 232 (40.3) | 51 (33.6) |
| Overall sleep qualityd | ||
| Very bad | 43 (7.5) | 22 (14.5) |
| Fairly bad | 102 (17.7) | 28 (18.4) |
| Fairly good | 320 (55.6) | 78 (51.3) |
| Very good | 111 (19.3) | 24 (15.8) |
| Usual hours of sleep at night, mean (SD) | 6.9 (1.9) | 6.6 (2.0) |
| Any psychosocial visit within past 2 ye | 36 (6.3) | 10 (6.6) |
| Psychotropic and opioid medications, mean (SD), total No.f | 1.6 (0.8) | 1.8 (0.9) |
| Days supplied in index dispensing, mean (SD) | 23.8 (9.7) | 27.9 (8.7) |
| Days supplied during the follow-up, mean (SD) | 88.5 (98.5) | 232.7 (82.6) |
| Index prescription was long-acting benzodiazepineg | 24 (4.2) | 5 (3.29) |
Abbreviations: PHQ-9, Patient Health Questionnaire–9; SUSTAIN, Supporting Seniors Receiving Treatment and Intervention.
Participant responded “Yes” to the following question: “Do you have enough money to get along?”
Depression diagnosis classified by response to PHQ-9 items: major depression if at least 5 positive items; other depression if 2 to 4 positive items, absent if no more than 1 item.
High anxiety is defined as either having the 7-item Generalized Anxiety Disorder (GAD-7) score of at least 15 or being diagnosed as having generalized anxiety disorder based on Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (Text Revision) criteria (≥3 or more symptoms endorsed).
Pittsburg Sleep Quality Index: modified version assessing quality, latency, and duration of sleep. For overall sleep quality, participants were asked to rate their overall sleep quality as very good, fairly good, fairly bad, or very bad.
Participants responded “Yes” to the following question: “Did you have any visit with a counselor, psychiatrist, therapist, or social worker within the past 2 years?”
Includes psychotropic medications, cognitive enhancers (cholinesterase inhibitors and memantine), and prescription opioids.
Classified as long- and short-acting based on previous studies.1 The most commonly prescribed benzodiazepines included lorazepam (n = 294 of 576 patients [51.0%]), alprazolam (n = 208 of 576 patients [36.1%]), and temazepam (n = 44 of 576 patients [7.63%]).
Table 2. Factors Associated With Benzodiazepine Medication Possession Ratio (MPR) Greater Than 30% During the 1 Year Following the Index Prescription.
| Characteristic | Unadjusted OR (95% CI) | P Value | Adjusted OR (95% CI)a | P Valueb |
|---|---|---|---|---|
| Age | 0.83 (0.69-1.00) | .046 | 0.81 (0.66-1.00) | .046 |
| Male sex | 1.53 (0.95-2.45) | .08 | 1.28 (0.75-2.18) | .37 |
| White race | 3.67 (1.43-9.40) | .007 | 4.19 (1.51-11.59) | .006 |
| Finance | 1.11 (0.61-2.02) | .73 | 1.45 (0.73-2.89) | .29 |
| Married | 1.20 (0.79-1.82) | .39 | 1.08 (0.67-1.72) | .76 |
| PHQ-9 score | 1.23 (1.03-1.47) | .03 | 1.04 (0.81-1.34) | .77 |
| High anxiety | 1.35 (0.62-2.94) | .45 | 0.84 (0.33-2.16) | .72 |
| Pain interference | ||||
| Quite a bit/extremely | 1.34 (0.79-2.29) | .28 | 0.88 (0.46-1.69) | .70 |
| A little bit/moderately | 1.53 (1.01-2.32) | .04 | 1.64 (1.05-2.58) | .03 |
| Not at all | 1 [Reference] | 1 [Reference] | ||
| Overall sleep quality | ||||
| Very bad | 3.80 (1.79-8.04) | <.001 | 4.05 (1.44-11.43) | .008 |
| Fairly bad | 1.37 (0.73-2.57) | .32 | 1.27 (0.59-2.73) | .55 |
| Fairly good | 1.17 (0.70-1.96) | .56 | 1.23 (0.70-2.18) | .47 |
| Very good | 1 [Reference] | 1 [Reference] | ||
| Usual hours of sleep at night | 0.82 (0.68-0.99) | .03 | 0.91 (0.71-1.16) | .44 |
| Days supplied in index dispensingc | 1.94 (1.52-2.46) | <.001 | 1.94 (1.52-2.47) | <.001 |
| Index prescription was long-acting benzodiazepine | 0.73 (0.27-1.98) | .53 | 0.76 (0.24-2.45) | .65 |
Abbreviations: OR, odds ratio; PHQ-9, Patient Health Questionnaire–9.
Adjusted model included all characteristics reported in the table.
Given the multiple comparisons, P < .01 was the threshold for statistical significance.
The unit for this predictor is 1 additional SD (ie, 9.7 days). So for each additional 9.7 days of benzodiazepine prescribed in the index prescription, the odds of long-term were 1.98-fold higher.
For sensitivity analyses, we completed adjusted models predicting 3 alternative long-term benzodiazepine use outcomes at 1 year, whether the patient (1) was prescribed at least 1 benzodiazepine refill (white race and days supplied were significant), (2) had a benzodiazepine prescription 1 year after index date (ie, on day 366; days supplied were significant); and (3) MPR as a continuous measure (white race and days supplied were significant).
Discussion
In this sample of older adults newly prescribed a benzodiazepine by a nonpsychiatric clinician, nearly one-third of patients went on to long-term use. While treatment guidelines recommend only short-term prescribing, if any, these long-term patients were prescribed nearly 8 months’ worth of benzodiazepine. Of the clinical measures evaluated, only poor sleep quality was associated with the likelihood of continued benzodiazepine use in adjusted analyses. White patient race and a larger initial prescription were also associated with conversion to long-term use. It is a cause for concern that these nonclinical factors are associated with benzodiazepine prescribing, which suggests that approaches to reduce prescribing of this drug that focus on specific clinical populations may have limited success.
Among our study’s limitations, it does not account for as-needed medication use, which may have an effect on our calculation of long-term use. The analysis is limited to low-income older adults from Pennsylvania, which may limit generalizability. Definitions of long-term benzodiazepine use vary,5 and it is possible that a different classification of long-term use might have yielded different results. However, we used 3 alternative definitions with no significant variation in our findings.
In conclusion, for new benzodiazepine users, prescribers should “begin with the end in mind” and immediately engage patients in discussion regarding the expected (brief) length of treatment, particularly when prescribed for insomnia. In light of the continued growth of psychotropic prescribing to older adults by nonpsychiatric clinicians, it is critical to improve access to and education regarding nonpharmacologic treatment so clinicians feel they have treatment alternatives to offer.6
References
- 1.Olfson M, King M, Schoenbaum M. Benzodiazepine use in the United States. JAMA Psychiatry. 2015;72(2):136-142. doi: 10.1001/jamapsychiatry.2014.1763 [DOI] [PubMed] [Google Scholar]
- 2.Gray SL, Eggen AE, Blough D, Buchner D, LaCroix AZ. Benzodiazepine use in older adults enrolled in a health maintenance organization. Am J Geriatr Psychiatry. 2003;11(5):568-576. doi: 10.1097/00019442-200309000-00012 [DOI] [PubMed] [Google Scholar]
- 3.Simon GE, Ludman EJ. Outcome of new benzodiazepine prescriptions to older adults in primary care. Gen Hosp Psychiatry. 2006;28(5):374-378. doi: 10.1016/j.genhosppsych.2006.05.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mavandadi S, Benson A, DiFilippo S, Streim JE, Oslin D. A telephone-based program to provide symptom monitoring alone vs symptom monitoring plus care management for late-life depression and anxiety: a randomized clinical trial. JAMA Psychiatry. 2015;72(12):1211-1218. doi: 10.1001/jamapsychiatry.2015.2157 [DOI] [PubMed] [Google Scholar]
- 5.Kurko TA, Saastamoinen LK, Tähkäpää S, et al. Long-term use of benzodiazepines: definitions, prevalence and usage patterns—a systematic review of register-based studies. Eur Psychiatry. 2015;30(8):1037-1047. doi: 10.1016/j.eurpsy.2015.09.003 [DOI] [PubMed] [Google Scholar]
- 6.Tannenbaum C, Martin P, Tamblyn R, Benedetti A, Ahmed S. Reduction of inappropriate benzodiazepine prescriptions among older adults through direct patient education: the EMPOWER cluster randomized trial. JAMA Intern Med. 2014;174(6):890-898. doi: 10.1001/jamainternmed.2014.949 [DOI] [PubMed] [Google Scholar]