Abstract
Introduction:
Buprenorphine is a highly effective medication treatment for opioid use disorder (OUD) that can be prescribed in multiple treatment settings. Treatment retention, however, remains a challenge. In this study, we examined the association of days of supply as well as daily dosage of the initial buprenorphine prescription with treatment discontinuation and adverse opioid-related events following buprenorphine initiation.
Methods:
2011 to 2015 Health Care Cost Institute commercial claims data were analyzed for individuals aged 18–64 years initiating buprenorphine treatment (N=17,158). Treatment discontinuation was defined as a gap of 30 days or more in buprenorphine use within 180 days of initiation. Adverse opioid-related events were defined as having at least one emergency department visit or inpatient admission involving opioid poisoning, dependence or abuse within 360 days of initiation. We conducted multivariate logistic regressions to estimate adjusted odds ratios of outcomes associated with daily dose (≤4 mg vs. >4mg) and days of supply (≤7, 8-15, 16-27, or ≥28 days) of the initial buprenorphine prescription.
Results:
Over one-half (55%) of individuals discontinued buprenorphine within 180 days and 13% experienced at least one adverse opioid-related event within 360 days of initiation. Both a lower initial dose [≤4 mg, OR=1.79, p < .01] and fewer initial days of supply [≤7 days vs. ≥28 days, OR=1.32, p < .01] [8-15 days vs. ≥28 days, OR=1.22, p < .01] were associated with increased odds of discontinuation. While a lower initial dose was not associated with adverse events, fewer initial days of supply were associated with a higher risk of adverse events, even after controlling for treatment discontinuation.
Conclusion:
In this population of commercially insured, non-elderly adults, we found that fewer initial days of supply as well as a lower initial dose were associated with increased likelihood of treatment discontinuation, highlighting the importance of prescribing decisions when initiating buprenorphine for OUD.
Keywords: buprenorphine, treatment discontinuation, opioid use disorder, Buprenorphine treatment
1. Introduction
In 2017, an estimated 2.1 million individuals aged 12 or older had an opioid use disorder (OUD) [1]. Medication treatment with opioid agonists methadone or buprenorphine are considered the most effective treatment for OUD [2, 3], more than doubling abstinence from illicit opioid use in randomized controlled trials and reducing the risk of mortality, HIV, and hepatitis C [4, 5, 6, 7, 8, 9, 10, 11, 12]. This association between medication treatment and improved outcomes is stronger with longer treatment durations [13, 14].
Buprenorphine treatment is increasingly being adopted in the United States because of its lower toxicity and availability in office-based settings [15]. Treatment retention, however, remains a challenge since a large proportion of patients discontinue treatment within the first few months [16, 17, 18, 19]. In a study of patients receiving buprenorphine at opioid treatment programs, approximately 25% discontinued treatment within the first month and over 50% discontinued within the first 6 months since treatment initiation [18]. In another study of Medicaid enrollees receiving buprenorphine prescriptions, approximately 28% discontinued treatment within the first month and 65% discontinued within the first 6 months since treatment initiation [17]. Similarly, a study based on an all-payer database found that 59% of patients discontinued treatment within the first 6 months [20].
Common risk factors for buprenorphine treatment discontinuation relating to patient characteristics or circumstances include a younger age, male gender, minority race/ethnicity, payer type (i.e. Medicaid, Medicare), work or school conflicts, and a lack of transportation [17, 20, 21]. Risk factors relating to provider characteristics include prescriber specialty other than psychiatry and primary care, inpatient setting, and strict program requirements [15, 19, 20, 21, 22]. Buprenorphine treatment characteristics such as buprenorphine taper, product type (i.e. sublingual buprenorphine without naloxone) and low dosage have also been identified as risk factors for discontinuation [13, 16, 17, 18, 19]. A study found that a lower maximum dose of buprenorphine during an episode predicted treatment discontinuation at 6 months [18]. Another study found that a low dose of the initial buprenorphine prescription (<4 mg) predicted treatment discontinuation at 6 months among Medicaid patients [17]. Understanding and overcoming barriers to buprenorphine treatment retention is important as patients who discontinue treatment are at an elevated risk of relapse, overdose, and death [15, 23, 24, 25].
In this study, we examined the association between initial buprenorphine treatment characteristics and treatment discontinuation within 180 days of initiation and adverse opioid-related events within 360 days of initiation. Specifically, we focused on days of supply as well as daily dosage of the first buprenorphine prescription, decisions usually within the control of prescribing clinicians. We consider our examination of days of supply innovative and hypothesize that a longer duration of the initial prescription will be associated with a lower risk of discontinuation and adverse opioid-related events by providing greater flexibility and reducing patient burden of recurrent visits. We focused on non-elderly, commercially insured patients, a population that accounts for nearly 40% of individuals with OUD [26] but that remains understudied.
2. Material and Methods
2.1. Data Source
We used data from the 2011 to 2015 Health Care Cost Institute (HCCI) claims database, including claims for about fifty million individuals per year enrolled in a commercial health insurance plan (including Medicare Advantage plans) offered or administered by Aetna, Humana, or UnitedHealthcare [27]. The data included beneficiary enrollment information and inpatient, outpatient and pharmacy claims. Pharmacy claims provided information on National Drug Code (NDC), days of supply, quantity, strength, and other information for each prescription.
2.2. Study Sample
Our study sample contained 17,158 individuals who initiated buprenorphine treatment for OUD, who were 18–64 years old at the time of treatment initiation, and who were continuously enrolled in a health insurance plan contributing data to HCCI for at least 19 months surrounding the initial buprenorphine prescription, 6 months before (baseline period) and 12 months after (follow-up period) the month in which buprenorphine was initiated. To determine buprenorphine initiation, we identified the first buprenorphine prescription filled during study years. Although this may not be the very first buprenorphine prescription a patient ever received, the continuous enrollment requirement ensured that the initial buprenorphine represented a clinically new episode of treatment after at least 6 months without documented buprenorphine treatment. This approach is consistent with previous studies [14, 17, 22].
We used NDCs provided by the Centers for Disease Control and Prevention (CDC) to identify buprenorphine prescriptions for OUD treatment in pharmacy claims [28]. We excluded buprenorphine products used to treat pain such as Butrans, Buprenex, and Belbuca. The under documentation of OUD in our data was substantial, with only 6.7% of patients who initiated buprenorphine having an OUD diagnosis at baseline. Therefore, we did not require an OUD diagnosis in the 6 months prior to treatment initiation.
2.3. Measures
2.3.1. Buprenorphine Treatment
Buprenorphine treatment discontinuation was defined as a gap of 30 days or more in buprenorphine prescriptions in the first 180 days following the initial prescription. This treatment continuation threshold is based on National Quality Forum performance measurement criteria [29] and is consistent with previous studies [17, 22]. To account for early refills and patients having buprenorphine in stock even when claims indicate depletion of medication, we carried overlapping days of buprenorphine prescriptions forward to the end of the last prescription when determining whether a gap occurred.
We considered two characteristics of the initial buprenorphine prescription: daily dosage and days of supply. Daily dosage was calculated by first deriving total dose for the initial prescription (per-unit strength x units) and then dividing by days of supply. We generated a dichotomous indicator of low (≤ 4 mg) or high (>4 mg) initial daily dosage based on clinical guidelines for buprenorphine treatment, which recommend a first day dosage of at most 8 mg, and of at least 2 mg for those with long-acting opioid addiction or of 4 mg for those with short-acting opioid addiction [30]. A sensitivity analysis using finer categories of daily dosage (≤4, >4 and ≤8, >8 and ≤15, >15 and ≤23, >23 mg) found no statistically or clinically meaningful differences in the rate of discontinuation between the last four categories. We categorized initial days of supply into ≤7 days, 8-15 days, 16-27 days, ≥28 days based on the empirical distribution of the sample as well as the common clinical practice of prescribing quantities in increments of weeks or 30 days.
2.3.2. Adverse Opioid-Related Events
We generated a dichotomous indicator of adverse opioid-related events at follow-up. Adverse events were defined as having at least one emergency department (ED) or inpatient admission with a first or secondary diagnosis of opioid poisoning, dependence or abuse in the 360 days following treatment initiation. We included diagnostic codes pertaining to opioid abuse and dependence (in addition to poisoning) since previous research has shown that opioid overdoses are severely under-coded in hospital settings and that this broader approach can improve sensitivity without compromising specificity [31]. In sensitivity analyses, we considered adverse events in the 180 days following treatment initiation (see Table A5). Appendix Table A1 includes diagnostic codes from the International Classification of Diseases, 9th revisions (ICD-9) used to identify these events.
2.3.3. Control Variables
We controlled for patient gender and age group at treatment initiation (18-24, 25-34, 35-44, 45-54 and 55-64 years old) to account for demographic characteristics. We also controlled for baseline comorbid conditions with dichotomous indicators of chronic pain, cancer, mental health conditions, tobacco use conditions, drug use conditions and alcohol use conditions captured in the 6 months prior to the month of the initial buprenorphine prescription based on primary and secondary diagnoses in claims. Adverse opioid-related events as defined in section 2.3.2 but captured at baseline, were also included. Finally, we controlled for indicators of opioid analgesics and benzodiazepine prescriptions at baseline, using NDCs provided by the [28]. Appendix Table A1 provides diagnostic codes used to define baseline controls.
2.4. Analysis
We estimated two sets of multivariate logistic models to examine whether characteristics of the initial buprenorphine prescription predicted outcomes, adjusting for patient demographics and baseline comorbid conditions. In the first set of regressions, we estimated the adjusted odds ratios (AORs) of treatment discontinuation associated with daily dosage and days of supply of the initial prescription. In the second set, we estimated the AORs of adverse opioid-related events associated with dosage and days of supply of the initial prescription. We first estimated the model without including the indicator of treatment discontinuation. We then included treatment discontinuation to see how the associations changed after accounting for possible mediation through treatment discontinuation. We estimated robust standard errors in all models. In sensitivity analyses, we excluded patient demographics and baseline comorbid conditions. Adjusted odds ratios associated with dosage or days of the first buprenorphine did not change in statistically or clinically meaningful ways (Appendix Table A3). Analyses were conducted using Stata version 15. This study was approved by the Institutional Review Board of the Weill Cornell Medical College.
3. Results
3.1. Sample Characteristics
Our sample included 17,158 individuals who initiated buprenorphine treatment between July 1, 2011 and December 31, 2014 (Table 1). Individuals in our sample were mostly males (64%) and roughly one-half were between ages 18 and 34 (52%). A significant proportion of individuals had a diagnosed drug use condition (37%), a mental health condition (42%), and/or chronic pain (55%) at baseline. Moreover, 51% and 32% of individuals received at least one prescription for opioid analgesics and benzodiazepines at baseline, respectively. Approximately 10% of initial buprenorphine prescriptions in our sample had a daily dose of 4mg or less, while 31%, 27%, 9%, and 33% were for ≤7, 8-15, 16-27, and ≥28 days of supply, respectively. Over one-half (55%) of individuals in our sample discontinued buprenorphine within 180 days of treatment initiation and 18% discontinued within 30 days. During the 180 days before and after treatment initiation, 17% and 8% of individuals experienced an adverse opioid-related event, respectively. Additionally, 13% experienced an adverse opioid-related event within 360 days of treatment initiation, 68% of which occurred after treatment discontinuation.
Table 1.
Sample characteristics (N=17,158)
| N | % | |
|---|---|---|
| Gender | ||
| Male | 11,053 | 64.4 |
| Female | 6,105 | 35.6 |
| Age | ||
| 18-24 | 4,726 | 27.5 |
| 25-34 | 4,121 | 24 |
| 35-44 | 3,267 | 19 |
| 45-54 | 3,167 | 18.5 |
| 55-64 | 1,877 | 10.9 |
| Baseline comorbidities | ||
| Cancer | 1,015 | 5.9 |
| Chronic pain | 9,410 | 54.8 |
| Mental health conditions | 7,352 | 42.8 |
| Alcohol use conditions | 1,058 | 6.2 |
| Drug use conditions | 6,269 | 36.5 |
| Tobacco use conditions | 1,955 | 11.4 |
| Baseline prescription drugs | ||
| Benzodiazepine Rx | 5,494 | 32 |
| Opioid Rx | 8,732 | 50.9 |
| Days supply of initial Bup Rx | ||
| 0-7 days | 5,318 | 31 |
| 8-15 days | 4,696 | 27.4 |
| 16-27 days | 1,501 | 8.7 |
| 28+ days | 5,643 | 32.9 |
| Daily dose of initial Bup Rx | ||
| 0< to 4 mg | 1,675 | 9.8 |
| 4< to 8 mg | 4,070 | 23.7 |
| 8< to 15 mg | 2,002 | 11.7 |
| 15< to 23 mg | 6,649 | 38.8 |
| 23< mg | 2,762 | 16.1 |
| Treatment discontinuation | ||
| Within 30 days post initiation | 3,012 | 17.6 |
| Within 60 days post initiation | 5,620 | 32.8 |
| Within 90 days post initiation | 7,081 | 41.3 |
| Within 180 days post initiation | 9,479 | 55.2 |
| Adverse opioid-related events | ||
| Within 180 days pre initiation | 2,880 | 16.8 |
| Within 180 days post initiation | 1,406 | 8.2 |
| Within 360 days post initiation | 2,265 | 13.2 |
| Event post Tx discontinuation | 1,528 | 67.5 |
Notes: Data are from the 2011 to 2015 Health Care Cost Institute claims database and capture non-elderly adults initiating buprenorphine treatment.
Cross tabulations between days of supply and dose show that patients receiving a longer supply at the initial prescription only had a slightly greater likelihood of receiving a daily dose of 4mg or less (≤7 days: 7.9%; 8-15 days: 8.7%, 16-27 days: 10.3%; ≥28 days: 12.3%) (see Appendix Table A2).
3.2. Treatment discontinuation
Both daily dosage and days of supply of the index buprenorphine prescription were associated with treatment discontinuation within 180 days of initiation (Table 2, Column 1). A daily dose of ≤4 mg [vs. >4 mg] was associated with an AOR of 1.79 of treatment discontinuation [p < .01]. Similarly, receiving an initial prescription of 15 days or less was associated with increased odds of treatment discontinuation [≤7 days vs. ≥28 days, AOR=1.32, p < .01] [8-15 days vs. ≥28 days, AOR=1.22, p < .01].
Table 2.
Characteristics of the initial buprenorphine prescription and treatment outcomes
| (1) Tx Discontinuation | (2) Adverse Event | (3) Adverse Event | |
|---|---|---|---|
| Tx Discontinuation | 3.200*** (0.187) | ||
| Dose = 4 mg or less | 1.790*** (0.0991) | 0.956 (0.0774) | 0.838** (0.0687) |
| Days of supply = 1, 0-7 | 1.323*** (0.0524) | 1.392*** (0.0870) | 1.295*** (0.0824) |
| Days of supply = 2, 8-15 | 1.221*** (0.0496) | 1.375*** (0.0883) | 1.299*** (0.0848) |
| Days of supply = 3, 16-27 | 1.038 (0.0618) | 1.020 (0.102) | 1.012 (0.104) |
| Age = 3, 25-34 | 0.460*** (0.0212) | 0.300*** (0.0201) | 0.344*** (0.0235) |
| Age = 4, 35-44 | 0.458*** (0.0230) | 0.171*** (0.0148) | 0.193*** (0.0170) |
| Age = 5, 45-54 | 0.499*** (0.0259) | 0.205*** (0.0171) | 0.230*** (0.0195) |
| Age = 6, 55-64 | 0.647*** (0.0400) | 0.202*** (0.0208) | 0.214*** (0.0224) |
| Gender = 1, Female | 1.092*** (0.0370) | 1.088 (0.0571) | 1.062 (0.0567) |
| Benzodiazepine Rx | 1.059 (0.0396) | 1.119* (0.0647) | 1.118* (0.0656) |
| Opioid Rx | 1.227*** (0.0485) | 0.999 (0.0599) | 0.980 (0.0604) |
| Alcohol use conditions | 0.939 (0.0648) | 1.228** (0.109) | 1.245** (0.114) |
| Drug use conditions | 0.846*** (0.0308) | 1.364*** (0.0765) | 1.413*** (0.0803) |
| Tobacco use conditions | 1.009 (0.0513) | 1.176** (0.0849) | 1.170** (0.0857) |
| Mental health conditions | 1.027 (0.0368) | 1.233*** (0.0684) | 1.251*** (0.0703) |
| Cancer | 1.258*** (0.0862) | 1.200* (0.130) | 1.152 (0.125) |
| Chronic pain | 1.063 (0.0412) | 1.070 (0.0615) | 1.055 (0.0623) |
| Adverse event at baseline | 1.351*** (0.0637) | 2.418*** (0.145) | 2.331*** (0.142) |
| Constant | 1.396*** (0.0639) | 0.167*** (0.0112) | 0.0766*** (0.00621) |
| Observations | 17,158 | 17,158 | 17,158 |
Notes: Data are from the 2011 to 2015 Health Care Cost Institute claims database and capture non-elderly adults initiating buprenorphine treatment. Outcome variables include treatment discontinuation, defined as a gap of 30 days or more in buprenorphine use within 180 days of initiation, and adverse opioid-related events, defined as having at least one emergency department visit or inpatient admission involving opioid poisoning, dependence or abuse within 360 days of initiation. Coefficients are adjusted odds ratios based on multivariable logistic regressions. Robust standard errors are in parenthesis.
p<0.01,
p<0.05,
p<0.1.
3.3. Adverse opioid-related events
In the model that did not control for treatment discontinuation, a daily dose of 4 mg or less was not associated with a higher AOR of adverse opioid-related events [AOR=0.95, p =0.57] (Table 2, Column 2). However, fewer days of supply were associated with increased odds of adverse events [≤7 days vs. ≥28 days, AOR=1.39, p < .01] [8-15 days vs. ≥28 days, OR=1.37, p < .01] (Table 2, Column 2). In the model that controlled for treatment discontinuation along with daily dose and days of supply, treatment discontinuation was associated with an AOR of 3.2 [p < .01] of having at least one adverse event (Table 2, Column 3). Notably, even after controlling for treatment discontinuation, fewer days of supply were still associated with increased odds of adverse events [≤7 days vs. ≥28 days, AOR=1.29, p < .01] [8-15 days vs. ≥28 days, OR=1.29, p < .05], suggesting that treatment discontinuation only partially mediated the association between days of supply and adverse opioid-related events (Table 2, Column 3). In contrast, after controlling for treatment discontinuation, a low daily dose was associated with reduced odds of adverse events [≤4 mg vs. >4 mg, AOR=0.83, p < .01].
4. Discussion
In this population of commercially insured, non-elderly adults who initiated buprenorphine treatment for OUD, we found that about 55% discontinued buprenorphine within 180 days of initiation and 18% discontinued within 30 days. We also found that 13% of patients experienced an adverse opioid-related event suggesting relapse or overdose within 360 days of initiation, nearly 70% of which occurred after treatment discontinuation. The high percentage of early discontinuation and subsequent adverse events underscores the challenges of retaining patients with OUD in buprenorphine treatment and the potentially detrimental consequences of failing to do so. These patterns of discontinuation among the commercially insured are generally comparable to those in previous studies based on other populations and research designs [17, 18, 20].
Characteristics of the initial buprenorphine prescription were associated with treatment discontinuation and adverse opioid-related events. In particular, fewer initial days of supply as well as a lower initial daily dose were associated with increased odds of treatment discontinuation. Moreover, fewer days of supply was associated with increased odds of adverse opioid-related events even after controlling for treatment discontinuation. Our findings related to initial daily dose are in line with those from a previous population study of Medicaid enrollees [17], suggesting better stabilization and control of withdrawal symptoms in patients who are prescribed a higher dose at the initial prescription. While a low daily dose did not appear to be associated with adverse events, in a model that controlled for treatment discontinuation, it was associated with reduced odds of adverse events. This suggests multiple pathways through which a low daily dose may be associated with increases or decreases in the likelihood of adverse events. On the one hand, a low daily dose may be associated with increased odds of adverse events through treatment discontinuation. On the other, a low daily dose may reflect greater clinical stability for some patients and therefore be associated with a lower likelihood of adverse events once discontinuation is controlled for.
Our findings related to initial days of supply are novel and reflect another important dimension of clinical decision-making with potential to improve buprenorphine treatment retention and prevent relapse or overdose. By prescribing buprenorphine for fewer days (e.g., 7 days), the prescribing clinician plausibly intends to reevaluate the patient upon depletion of the prescription and assess the need for adjusting the medication at that point. However, having to return to the clinician for reevaluation in a short period of time may constitute a major barrier to treatment continuation for patients lacking transportation, child care or having work conflicts. Thus, longer initial days of supply might offer protection through reductions in patient burden associated with recurrent visits. Nonetheless, the fact that the association between days of supply and adverse opioid-related events remained strong even after controlling for treatment discontinuation suggests there might be other mechanisms at play. One possible mechanism is reductions in the likelihood of treatment gaps that are shorter than 30 days. Another possible mechanism is that, among those who discontinued, earlier discontinuation was more detrimental (associated with a greater likelihood of adverse events) than discontinuation later in the process and that fewer days of supply of the initial buprenorphine prescription was positively associated with earlier discontinuation. Indeed, our data showed that fewer days of supply were more strongly associated with the odds of earlier treatment discontinuation (Appendix Table A4). Finally, it is possible that having longer days of supply in hand allows patients to better stabilize their cravings and withdrawal symptoms by taking extra buprenorphine ad lib. We know from the literature that early response to treatment can be an important predictor of later outcomes [32]. As such, some patients might benefit from more flexible, higher buprenorphine dosing early on. A more thorough investigation of these potential mechanisms is beyond the scope of this study and should be examined in future research.
It is worth noting that our findings capture average associations seen in a largely privately insured, non-elderly population. The appropriate initial daily dose and days of supply will likely vary across clinical cases and should be the outcome of a shared decision-making process between the treating provider and the patient, while reflecting best practices.
4.1. Limitations
This study has several limitations. First, because of significant under documentation of OUD in our data, we did not restrict our sample to individuals with an OUD diagnosis. By doing so, we could have inadvertently included an unknown number of individuals without an OUD who were prescribed buprenorphine intended for OUD only off-label. Second, patients who discontinued treatment after the first buprenorphine prescription might have been prescribed buprenorphine for detoxification without being inducted to maintenance treatment. While this possibility does not invalidate the association between characteristics of the initial buprenorphine prescription and adverse opioid-related events, it has implications for how findings should be interpreted. Third, we cannot ascertain if the first buprenorphine prescription observed in our data was truly the very first buprenorphine prescription received by the patient. It is possible that some patients were inducted in a different setting (e.g., inpatient or an outpatient Opioid Treatment Program) where buprenorphine or methadone was initially dispensed without generating a pharmacy claim. However, to the extent that the first buprenorphine prescription observed in our data reflected the early phase of treatment, our findings regarding how dosage and days of supply in the early phases of treatment relate to treatment discontinuation and opioid-related adverse events should still be of salience to clinicians providing medications for OUD in office-based settings. Fourth, it is possible that our 19-month continuous enrollment requirement introduced some selection bias. In particular, people who discontinued buprenorphine and/or relapsed might be at a greater risk of losing or switching their jobs and even of drug overdose mortality, and in turn, of being excluded from our sample because of discontinued enrollment. As such, our findings may not generalize to all privately insured adults initiating buprenorphine treatment. Finally, to an unknown extent, our findings may be explained by underlying disease severity. This issue of confounding by indication might occur if less stable patients or those with more severe OUD were purposefully prescribed fewer initial days of supply or a lower dose. As a result, the documented associations between these initial treatment characteristics and the likelihood of treatment discontinuation and adverse events might partially reflect this selection bias. While our models controlled for common risk factors for relapse and comorbid conditions at baseline (e.g. mental health, medical and substance use conditions, adverse opioid-related events, and opioid/benzodiazepine prescriptions), and our estimates were robust to the inclusion or exclusion of these controls, claims-based measures may not sufficiently capture differences in underlying disease severity. Our claims data also do not capture patient race or ethnicity. We were thus not able to examine potentially different treatment characteristics and outcomes associated with patient race/ethnicity.
5. Conclusion
Using a national claims database of commercially insured, non-elderly adults who initiated buprenorphine treatment for OUD, we found that characteristics of the initial buprenorphine prescription were associated with treatment discontinuation and adverse opioid-related events suggesting relapse or overdose. Specifically, fewer initial days of supply as well as a lower initial dose were associated with increased likelihood of treatment discontinuation. In addition, having at least 16 days of supply was associated with reduced odds of adverse events even after controlling for treatment discontinuation. Our findings highlight the importance of prescribing decisions in the early phase of buprenorphine treatment and provide support to policies, clinical guidelines, and health care system interventions targeting the quality of prescribing when or shortly after buprenorphine is initiated.
Highlights.
Patients initiating buprenorphine for opioid use disorder often discontinue early
A lower initial dose of buprenorphine was associated with discontinuation
Fewer initial days of supply (1-15 days) was associated with discontinuation
Longer days of supply was associated with reductions in adverse opioid events
Acknowledgments
Funding Sources: This work was supported by the National Institute of Mental Health T32MH073553 and by the National Institute on Drug Abuse P30DA040500.
Appendix
Table A1.
Measure Definitions
| Measures | Code | Source | Definition |
|---|---|---|---|
| Adverse opioid-related events | ICD-9 | IP claims, OP claims with a mention of ED | 305.5x, 304.0x, 304.7x, 965.0x, E850.x, E935.x, E950.0, E980.0 |
| Buprenorphine | NDC | Pharmacy claims | Buprenorphine Hydrochloride, Buprenorphine/naloxone, and their brand name eq. |
| Mental health conditions | ICD-9 | IP/OP claims | 290.xx, 293.xx, 294.xx, 295.xx, 296.xx, 297.x, 298.x, 300.xx, 301.xx, 302.xx, 306.xx, 307.xx, 308.x, 309.xx, 311, 312.xx, 313.xx 314.xx, 316 |
| Any Pain | |||
| Back pain | ICD-9 | IP/OP claims | 722.30, 722.32, 722.33, 722.70, 722.72, 722.73, 722.80, 722.82, 722.83, 722.90, 722.92, 722.93, 737.1, 737.3, 738.4, 738.5, 739.2, 739.3, 739.4, 756.10, 756.11, 756.12, 756.19, 805.4, 805.8, 839.2, 839.42, 846, 846.0, 847.1, 847.2, 847.3, 847.9, 721.3x-721.9x, 722.2x, 724.xx, 756.13 |
| Neck Pain | 721.0x, 721.1x, 722.0x, 722.31, 722.71, 722.81, 722.91, 723.xx, 839.0, 839.1, 847.0 | ||
| Arthritis and joint pain | 711.xx, 712.xx, 713.x, 714.xx, 715.xx, 716.xx, 717.xx, 718.xx, 719.xx, 725, 726.xx, 727.xx, 728.xx, 729.3x, 729.7x, 729.8x, 729.9x, 730.xx, 731.x, 732.x, 733.xx, 734, 735.x, 736.xx, 737.2x, 737.4x, 738.1x, 710, 710.1, 710.3, 710.4, 710.5, 710.8, 710.9, 729, 729.2, 729.4, 729.5, 729.6, 737, 737.8, 737.9, 738, 738.2, 738.3, 738.6, 738.7, 738.8, 738.9, 739, 739.1, 739.5, 739.6, 739.7, 739.8, 739.9 | ||
| Other pain | 524.60, 524.61, 524.62, 524.63, 524.64, 524.69, 307.81, 784.0, 723.8, 375.15, 729.1, 786.59, 530.5, 564.xx, 595.1, 625.7x, 617.x, 536.8, 710.2, 388.3, 780.71, 346.00, 346.01, 346.02, 346.03, 346.10, 346.11, 346.12, 346.13, 346.20, 346.21, 346.22, 346.23, 346.30, 346.31, 346.32, 346.33, 346.40, 346.41, 346.42, 346.43, 346.50, 346.51, 346.52, 346.53, 346.60, 346.61, 346.62, 346.63, 346.70, 346.71, 346.72, 346.73, 346.80, 346.81, 346.82, 346.83, 346.90, 346.91, 346.92, 346.93, 780.52 | ||
| Substance use conditions | |||
| Tobacco use conditions | ICD-9 | IP/OP claims | 305.1x |
| Alcohol use conditions | ICD-9 | IP/OP claims | 291.xx, 303.xx, 305.0x, 571.0, 571.1, 571.2, 571.3, 535.3x |
| Drug use conditions | ICD-9 | IP/OP claims | 292.xx, 304.xx, 305.2x, 305.3x, 305.4x, 305.5x, 305.6x, 305.7x, 305.8x, 305.9x |
Table A2.
Initial buprenornhine Rx days of supply and dose cross tabulations
| Days of supply | 0-7 days | 8-15 days | 16-27 days | 28+ days | Total | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Daily dose | N | % | N | % | N | % | N | % | N | % |
| 0< to 4 mg | 418 | 7.9 | 408 | 8.7 | 155 | 10.3 | 694 | 12.3 | 1,675 | 9.8 |
| 4< to 8 mg | 1,230 | 23.1 | 1,148 | 24.4 | 340 | 22.7 | 1,352 | 24.0 | 4,070 | 23.7 |
| 8< to 15 mg | 631 | 11.9 | 643 | 13.7 | 210 | 14.0 | 518 | 9.2 | 2,002 | 11.7 |
| 15< to 23 mg | 2,156 | 40.5 | 1,823 | 38.8 | 539 | 35.9 | 2,131 | 37.8 | 6,649 | 38.8 |
| 23< mg | 883 | 16.6 | 674 | 14.4 | 257 | 17.1 | 948 | 16.8 | 2,762 | 16.1 |
| Total | 5,318 | 100.0 | 4,696 | 100.0 | 1,501 | 100.0 | 5,643 | 100.0 | 17,158 | 100.0 |
Table A3.
Unadjusted logistic regressions
| (1) Tx Discontinuation | (2) Adverse Event | (3) Adverse Event | |
|---|---|---|---|
| Tx Discontinuation | 3.670*** (0.202) | ||
| Dose = 4 mg or less | 1.841*** (0.101) | 1.115 (0.0836) | 0.949 (0.0723) |
| Days of supply = 1, 0-7 | 1.388*** (0.0538) | 1.640*** (0.0955) | 1.520*** (0.0900) |
| Days of supply = 2, 8-15 | 1.262*** (0.0504) | 1.539*** (0.0928) | 1.461*** (0.0895) |
| Days of supply = 3, 16-27 | 1.002 (0.0585) | 0.997 (0.0957) | 0.996 (0.0975) |
| Constant | 0.989 (0.0272) | 0.113*** (0.00513) | 0.0500*** (0.00315) |
| Controls | No | No | No |
| Observations | 17,158 | 17,158 | 17,158 |
Notes: Data are from the 2011 to 2015 Health Care Cost Institute claims database and capture non-elderly adults initiating buprenorphine treatment. Coefficients are odds ratios based on multivariable logistic regressions. Robust standard errors are in parenthesis.
p<0.01,
p<0.05,
p<0.1.
Table A4.
Treatment discontinuation within 180, 120, 90, and 60 days of initiation
| 180 days | 120 days | 90 days | 60 days | |
|---|---|---|---|---|
| Dose = 4 mg or less | 1.790*** (0.0991) | 1.750*** (0.0930) | 1.725*** (0.0905) | 1.739*** (0.0926) |
| Days of supply = 1, 0-7 | 1.323*** (0.0524) | 1.378*** (0.0543) | 1.487*** (0.0595) | 1.841*** (0.0780) |
| Days of supply = 2, 8-15 | 1.221*** (0.0496) | 1.235*** (0.0501) | 1.314*** (0.0543) | 1.588*** (0.0697) |
| Days of supply = 3, 16-27 | 1.038 (0.0618) | 1.012 (0.0602) | 1.100 (0.0670) | 1.252 (0.0818) |
| Controls | Yes | Yes | Yes | Yes |
| Observations | 17,158 | 17,158 | 17,158 | 17,158 |
Notes: Data are from the 2011 to 2015 Health Care Cost Institute claims database and capture non-elderly adults initiating buprenorphine treatment. Coefficients are adjusted odds ratios based on multivariable logistic regressions. Robust standard errors are in parenthesis.
p<0.01,
p<0.05,
p<0.1.
Table A5.
Adverse Opioid-Related Events within 180 Days
| Adverse Events | Adverse Events | |
|---|---|---|
| Tx Discontinuation | 4.808*** (0.393) | |
| Dose = 4 mg or less | 1.083 (0.102) | 0.930 (0.0884) |
| Days of supply = 1, 0-7 | 1.420*** (0.109) | 1.301*** (0.102) |
| Days of supply = 2, 8-15 | 1.493*** (0.117) | 1.394*** (0.111) |
| Days pf supply = 3, 16-27 | 0.993 (0.126) | 0.976 (0.127) |
| Controls | Yes | Yes |
| Observations | 17,158 | 17,158 |
Notes: Data are from the 2011 to 2015 Health Care Cost Institute claims database and capture non-elderly adults initiating buprenorphine treatment. Coefficients are adjusted odds ratios based on multivariable logistic regressions. Robust standard errors are in parenthesis.
p<0.01,
p<0.05,
p<0.1.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declarations of Interest: None.
References
- 1.Substance Abuse and Mental Health Services Administration. (2018). Key substance use and mental health indicators in the United States: Results from the 2017 National Survey on Drug Use and Health (HHS Publication No. SMA 18-5068, NSDUH Series H-53). Rockville, MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration; https://www.samhsa.gov/data/report/2017-nsduh-annual-national-report. [Google Scholar]
- 2.Schuckit MA, 2016. Treatment of opioid-use disorders. N. Engl. J. Med 375 (4), 357–368. [DOI] [PubMed] [Google Scholar]
- 3.Volkow ND, Frieden TR, Hyde PS, Cha SS, 2014. Medication-assisted therapies: tackling the opioid-overdose epidemic. N. Engl. J. Med 370 (22), 2063–2066. [DOI] [PubMed] [Google Scholar]
- 4.Connock M, Juarez-Garcia A, Jowett S, Frew E, Liu Z, Taylor RJ, … & Burls A (2007). Methadone and buprenorphine for the management of opioid dependence: a systematic review and economic evaluation. In NIHR Health Technology Assessment programme: Executive Summaries. NIHR Journals Library. [DOI] [PubMed] [Google Scholar]
- 5.Connery HS, 2015. Medication-assisted treatment of opioid use disorder: review of the evidence and future directions. Harv. Rev. Psychiatry 23 (2), 63–75. [DOI] [PubMed] [Google Scholar]
- 6.Whelan PJ, Remski K, 2012. Buprenorphine vs methadone treatment: a review of evidence in both developed and developing worlds. J. Neurosci. Rural Pract 3 (1), 45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Mattick RP, Breen C, Kimber J, & Davoli M (2014). Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane database of systematic reviews, (2). [Google Scholar]
- 8.Bell J, Trinh L, Butler B, Randall D, & Rubin G (2009). Comparing retention in treatment and mortality in people after initial entry to methadone and buprenorphine treatment. Addiction, 104(7), 1193–1200. [DOI] [PubMed] [Google Scholar]
- 9.Schwartz RP, Gryczynski J, O’grady KE, Sharfstein JM, Warren G, Olsen Y, … & Jaffe JH (2013). Opioid agonist treatments and heroin overdose deaths in Baltimore, Maryland, 1995–2009. American journal of public health, 103(5), 917–922. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tsui JI, Evans JL, Lum PJ, Hahn JA, & Page K (2014). Association of opioid agonist therapy with lower incidence of hepatitis C virus infection in young adult injection drug users. JAMA internal medicine, 174(12), 1974–1981. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Sullivan LE, Moore BA, Chawarski MC, Pantalon MV, Barry D, O’Connor PG, … & Fiellin DA (2008). Buprenorphine/naloxone treatment in primary care is associated with decreased human immunodeficiency virus risk behaviors. Journal of substance abuse treatment, 55(1), 87–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Murphy SM, & Polsky D (2016). Economic evaluations of opioid use disorder interventions. PharmacoEconomics, 34(9), 863–887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Fiellin DA, Schottenfeld RS, Cutter CJ, Moore BA, Barry DT, & O’connor PG (2014). Primary care–based buprenorphine taper vs maintenance therapy for prescription opioid dependence: a randomized clinical trial. JAMA internal medicine, 174(12), 1947–1954. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lo-Ciganic WH, Gellad WF, Gordon AJ, Cochran G, Zemaitis MA, Cathers T, … & Donohue JM (2016). Association between trajectories of buprenorphine treatment and emergency department and in- patient utilization. Addiction, 111(5), 892–902. [DOI] [PubMed] [Google Scholar]
- 15.Bentzley BS, Barth KS, Back SE, & Book SW (2015). Discontinuation of buprenorphine maintenance therapy: perspectives and outcomes. Journal of substance abuse treatment, 52, 48–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Timko C, Schultz NR, Cucciare MA, Vittorio L, & Garrison-Diehn C (2016). Retention in medication-assisted treatment for opiate dependence: A systematic review. Journal of addictive diseases, 35(1), 22–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Samples H, Williams AR, Olfson M, & Crystal S (2018). Risk factors for discontinuation of buprenorphine treatment for opioid use disorders in a multi-state sample of Medicaid enrollees. Journal of Substance Abuse Treatment, 95, 9–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hser YI, Saxon AJ, Huang D, Hasson A, Thomas C, Hillhouse M, … & Cohen A (2014). Treatment retention among patients randomized to buprenorphine/naloxone compared to methadone in a multi- site trial. Addiction, 109(1), 79–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Morgan JR, Schackman BR, Leff JA, Linas BP, & Walley AY (2018). Injectable naltrexone, oral naltrexone, and buprenorphine utilization and discontinuation among individuals treated for opioid use disorder in a United States commercially insured population. Journal of substance abuse treatment, 85, 90–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Saloner B, Daubresse M, & Caleb Alexander G (2017). Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multistate population. Medical care, 55(1), 669–676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gryczynski J, Mitchell SG, Jaffe JH, O’Grady KE, Olsen YK, & Schwartz RP (2014). Leaving buprenorphine treatment: Patients’ reasons for cessation of care. Journal of substance abuse treatment, 46(3), 356–361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Shcherbakova N, Tereso G, Spain J, & Roose RJ (2018). Treatment persistence among insured patients newly starting buprenorphine/naloxone for opioid use disorder. Annals of Pharmacotherapy. 10.1177/1060028017751913. [DOI] [PubMed] [Google Scholar]
- 23.Sordo L, Barrio G, Bravo MJ et al. Mortality risk during and after opioid substitution treatment: systematic review and meta-analysis of cohort studies. BMJ. 2017;357:j 1550:1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.LaRochelle MR, Bernson D, Land T, et al. (2018). Medication for Opioid Use Disorder After Nonfatal Opioid Overdose and Association With Mortality: A Cohort Study. Ann Intern Med. doi: 10.7326/M17-3107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Degenhardt L, Bucello C, Mathers B, et al. Mortality among regular or dependent users of heroin and other opioids: a systematic review and meta-analysis of cohort studies. Addiction. 2010;106:32–51. [DOI] [PubMed] [Google Scholar]
- 26.Zur J & Tolbert J (2018). The Opioid Epidemic and Medicaid’s Role in Facilitating Access to Treatment. https://www.kff.org/medicaid/issue-brief/the-opioid-epidemic-and-medicaids-role-in-facilitating-access-to-treatment/. Accessed February 19, 2019.
- 27.Health Care Cost Institute. (2019). http://www.healthcostinstitute.org/. Accessed February 19, 2019.
- 28.Centers for Disease Control and Prevention. CDC compilation of benzodiazepines, muscle relaxants, stimulants, zolpidem, and opioid analgesics with oral morphine milligram equivalent conversion factors, 2016 version. [Google Scholar]
- 29.National Quality Forum (2017). Behavioral health 2016–2017: Technical report. Department of Health and Human Services; https://www.qualityforum.org/Publications/2017/08/Behavioral_Health_2016-2017_Final_Report.aspx. [Google Scholar]
- 30.Center for Substance Abuse Treatment. (2004). Clinical guidelines for the use of buprenorphine in the treatment of opioid addiction. [Google Scholar]
- 31.Rowe C, Vittinghoff E, Santos GM, et al. Performance Measures of Diagnostic Codes for Detecting Opioid Overdose in the Emergency Department. Acad Emerg Med. 2017;24(4):475–483. [DOI] [PubMed] [Google Scholar]
- 32.Weiss RD, & Rao V (2017). The prescription opioid addiction treatment study: what have we learned. Drug and alcohol dependence, 173, S48–S54. [DOI] [PMC free article] [PubMed] [Google Scholar]