Advertisement

Patterns of patient discontinuation from buprenorphine/naloxone treatment for opioid use disorder: A study of a commercially insured population in Massachusetts

Published:April 19, 2021DOI:https://doi.org/10.1016/j.jsat.2021.108416

      Highlights

      • Little evidence exists on the buprenorphine/naloxone treatment discontinuation for commercially insured patients.
      • Patient and provider level factors of buprenorphine/naloxone treatment discontinuation are examined.
      • Substantial percentage of commercially insured patients discontinue treatment within one year of initiation.
      • High volume prescribers' patients have higher associated risk for treatment discontinuation.

      Abstract

      Background

      Research has shown buprenorphine/naloxone to be an effective medication for treating individuals with opioid use disorder. At the same time, treatment discontinuation rates are reportedly high though much of the extant evidence comes from studies of the Medicaid population.

      Objectives

      To examine the pattern and determinants of buprenorphine/naloxone treatment discontinuation in a population of commercially insured individuals.

      Research design

      We performed a retrospective observational analysis of Massachusetts All Payer Claims Data (MA APCD) covering years 2013 through 2017. We defined treatment discontinuation as a gap of 60 consecutive days without a prescription for buprenorphine/naloxone within a time frame of 24 months from the initiation of treatment. A mixed-effect Cox proportional hazard model examined the associated risk of discontinuing treatment with baseline predictors.

      Subjects

      A total of 5134 individuals who were commercially insured during the study period.

      Measures

      Buprenorphine/naloxone treatment discontinuation.

      Results

      Overall 75% of individuals had discontinued treatment within two years of initiating treatment, and median time to discontinuation was 300 days. Patients aged between 18 and 24 years (HR = 1.436, 95%, CI = 1.240–1.663) and receiving treatment from prescribers with high panel-size (HR = 1.278, 95% CI = 1.112–1.468) had higher risk of discontinuing treatment. On the contrary, patients receiving treatment from multiple prescribers had lower associated risk of treatment discontinuation.

      Conclusions

      A substantial percentage of patients discontinue treatment well before they can typically meet criteria for sustained remission. Further investigations should assess the clinical outcomes following premature discontinuation and identify strategies for retaining patients in treatment.

      Keywords

      1. Introduction

      Opioid use disorder (OUD) affects more than 15 million individuals across the globe, with an estimated 2.1 million individuals affected in the United States as of 2017 (
      • Degenhardt L.
      • Charlson F.
      • Mathers B.
      • Hall W.D.
      • Flaxman A.D.
      • Johns N.
      • Vos T.J.A.
      The global epidemiology and burden of opioid dependence: Results from the global burden of disease 2010 study.
      ;
      • Welty L.
      • Harrison A.
      • Abram K.
      • Olson N.
      • Aaby D.
      • McCoy K.
      Substance Abuse and Mental Health Services Administration (2017). Key substance use and mental health indicators in the United States: Results from the 2016 National Survey on Drug Use and Health (HHS Publication no. SMA 17-5044, NSDUH series H-52). Rockville, MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. Retrieved.
      ). Alongside the substantial death toll from this condition, OUD is associated with blood-borne infectious diseases (e.g., hepatitis and HIV), emergency department visits, productivity loss, and worse social-emotional outcomes (
      • Birnbaum H.G.
      • White A.G.
      • Schiller M.
      • Waldman T.
      • Cleveland J.M.
      • Roland C.L.
      Societal costs of prescription opioid abuse, dependence, and misuse in the United States.
      ;
      • Florence C.
      • Luo F.
      • Xu L.
      • Zhou C.
      The economic burden of prescription opioid overdose, abuse and dependence in the United States, 2013.
      ;
      • Hansen R.N.
      • Oster G.
      • Edelsberg J.
      • Woody G.E.
      • Sullivan S.D.
      Economic costs of nonmedical use of prescription opioids.
      ;
      • McAdam-Marx C.
      • Roland C.L.
      • Cleveland J.
      • Oderda G.M.
      Costs of opioid abuse and misuse determined from a Medicaid database.
      ;
      • Oderda G.M.
      • Lake J.
      • Rüdell K.
      • Roland C.L.
      • Masters E.T.
      Economic burden of prescription opioid misuse and abuse: A systematic review.
      ;
      • Strassels S.
      Economic burden of prescription opioid misuse and abuse.
      ;
      • White A.G.
      • Birnbaum H.G.
      • Mareva M.N.
      • Daher M.
      • Vallow S.
      • Schein J.
      • Katz N.
      Direct costs of opioid abuse in an insured population in the United States.
      ). As such, the prevention and sustained treatment of OUD is an important health policy issue.
      Buprenorphine/naloxone is one of three FDA approved medications for OUD. Unlike methadone, which providers administer in opioid treatment program settings, qualified clinicians can prescribe buprenorphine/naloxone in office settings to treat OUD. Research has shown that successful treatment with this medication is associated with reduced mortality (
      • Chang H.-Y.
      • Daubresse M.
      • Saloner B.
      • Alexander G.C.
      Chronic disease medication adherence after initiation of buprenorphine for opioid use disorder.
      ;
      • Larochelle M.R.
      • Bernson D.
      • Land T.
      • Stopka T.J.
      • Wang N.
      • Xuan Z.
      • Walley A.Y.
      Medication for opioid use disorder after nonfatal opioid overdose and association with mortality: A cohort study.
      ;
      • Sordo L.
      • Barrio G.
      • Bravo M.J.
      • Indave B.I.
      • Degenhardt L.
      • Wiessing L.
      • Pastor-Barriuso R.
      Mortality risk during and after opioid substitution treatment: Systematic review and meta-analysis of cohort studies.
      ;
      • Tkacz J.
      • Severt J.
      • Kassed C.
      • Ruetsch C.
      Clinical differences between opioid abuse classes ameliorated after 1 year of buprenorphine-medication assisted treatment.
      ).
      However, retaining patients in treatment with this medication can be challenging. Guidelines recommend the minimum duration for buprenorphine/naloxone treatment to be a year, and some evidence suggests that lifetime use of buprenorphine/naloxone should be the standard of care to avoid recurrence of non–prescribed opioid use (
      • Kraus M.L.
      • Alford D.P.
      • Kotz M.M.
      • Levounis P.
      • Mandell T.W.
      • Meyer M.
      • Wyatt S.A.
      Statement of the American Society of Addiction Medicine Consensus Panel on the use of buprenorphine in office-based treatment of opioid addiction.
      ). Yet studies indicate that 50% or more of patients who begin treatment with buprenorphine/naloxone do not remain in treatment for even 12 months (
      • Hser Y.I.
      • Saxon A.J.
      • Huang D.
      • Hasson A.
      • Thomas C.
      • Hillhouse M.
      • Wiest K.
      Treatment retention among patients randomized to buprenorphine/naloxone compared to methadone in a multi-site trial.
      ;
      • Morgan J.R.
      • Schackman B.R.
      • Leff J.A.
      • Linas B.P.
      • Walley A.Y.
      Injectable naltrexone, oral naltrexone, and buprenorphine utilization and discontinuation among individuals treated for opioid use disorder in a United States commercially insured population.
      ;
      • Saloner B.
      • Daubresse M.
      • Alexander G.C.
      Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multi-state population.
      ;
      • Shcherbakova N.
      • Tereso G.
      • Spain J.
      • Roose R.J.
      Treatment persistence among insured patients newly starting buprenorphine/naloxone for opioid use disorder.
      ). One randomized control trial found that approximately 25% of patients discontinued treatment within the first month and 50% within 6 months (
      • Hser Y.I.
      • Saxon A.J.
      • Huang D.
      • Hasson A.
      • Thomas C.
      • Hillhouse M.
      • Wiest K.
      Treatment retention among patients randomized to buprenorphine/naloxone compared to methadone in a multi-site trial.
      ). Another study reported that 28% discontinued treatment within the first month and 65% discontinued within the first 6 months (
      • Saloner B.
      • Daubresse M.
      • Alexander G.C.
      Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multi-state population.
      ). The consequences of premature discontinuation are severe; evidence suggests that as many as 50% of the patients who discontinue treatment resume use of non–prescribed opioids within one month (
      • Bentzley B.S.
      • Barth K.S.
      • Back S.E.
      • Book S.W.
      Discontinuation of buprenorphine maintenance therapy: Perspectives and outcomes.
      ). Such premature discontinuation also increases the risk of mortality (
      • Manhapra A.
      • Rosenheck R.
      • Fiellin D.A.
      Opioid substitution treatment is linked to reduced risk of death in opioid use disorder.
      ;
      • Sordo L.
      • Barrio G.
      • Bravo M.J.
      • Indave B.I.
      • Degenhardt L.
      • Wiessing L.
      • Pastor-Barriuso R.
      Mortality risk during and after opioid substitution treatment: Systematic review and meta-analysis of cohort studies.
      ).
      While many patients may discontinue treatment with buprenorphine/naloxone prematurely, extant studies largely focus on vulnerable populations such as individuals enrolled in Medicaid programs or the Veterans Health Administration (
      • Lo-Ciganic W.H.
      • Gellad W.F.
      • Gordon A.J.
      • Cochran G.
      • Zemaitis M.A.
      • Cathers T.
      • Donohue J.M.J.A.
      Association between trajectories of buprenorphine treatment and emergency department and in-patient utilization.
      ;
      • Lo-Ciganic W.H.
      • Donohue J.M.
      • Kim J.Y.
      • Krans E.E.
      • Jones B.L.
      • Kelley D.
      • safety, d.
      Adherence trajectories of buprenorphine therapy among pregnant women in a large state Medicaid program in the United States.
      ;
      • Samples H.
      • Williams A.R.
      • Olfson M.
      • Crystal S.
      Risk factors for discontinuation of buprenorphine treatment for opioid use disorders in a multi-state sample of Medicaid enrollees.
      ;
      • Shcherbakova N.
      • Tereso G.
      • Spain J.
      • Roose R.J.
      Treatment persistence among insured patients newly starting buprenorphine/naloxone for opioid use disorder.
      ). Much less is known about patterns of medication discontinuation for other types of patients even though OUD is widespread in the U.S. population. Moreover, little information exists about the rate of medication discontinuation beyond 12 months because most studies are confined to timelines of 12 months or less from the time treatment was initiated (
      • Meinhofer A.
      • Williams A.R.
      • Johnson P.
      • Schackman B.R.
      • Bao Y.
      Prescribing decisions at buprenorphine treatment initiation: Do they matter for treatment discontinuation and adverse opioid-related events?.
      ;
      • Morgan J.R.
      • Schackman B.R.
      • Leff J.A.
      • Linas B.P.
      • Walley A.Y.
      Injectable naltrexone, oral naltrexone, and buprenorphine utilization and discontinuation among individuals treated for opioid use disorder in a United States commercially insured population.
      ;
      • Ronquest N.A.
      • Willson T.M.
      • Montejano L.B.
      • Nadipelli V.R.
      • Wollschlaeger B.A.
      Relationship between buprenorphine adherence and relapse, health care utilization and costs in privately and publicly insured patients with opioid use disorder.
      ;
      • Saloner B.
      • Daubresse M.
      • Alexander G.C.
      Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multi-state population.
      ).
      In this paper, we report results for an investigation of buprenorphine/naloxone treatment discontinuation for a large cohort of commercially insured individuals in Massachusetts, covering a 2-year post-treatment period that used the most recent data that were available for Massachusetts during the time of this study. Our study extends a previous study of a commercially insured population that included a post-treatment period of more than 12 months (
      • Manhapra A.
      • Agbese E.
      • Leslie D.L.
      • Rosenheck R.A.
      Three-year retention in buprenorphine treatment for opioid use disorder among privately insured adults.
      ) by using more current data and also by examining prescriber- and patient-level characteristics for their association with treatment discontinuation. Massachusetts is among those states that the opioid epidemic has most severely affected, with an estimated prevalence of OUD approaching 5% of the population as of 2015 (
      • Barocas J.A.
      • White L.F.
      • Wang J.
      • Walley A.Y.
      • LaRochelle M.R.
      • Bernson D.
      • Linas B.P.
      Estimated prevalence of opioid use disorder in Massachusetts, 2011–2015: A capture–recapture analysis.
      ). Our findings are intended to help address important clinical and policy issues regarding the treatment of OUD with buprenorphine/naloxone.

      2. Data and methods

      2.1 Data sources

      In this retrospective, longitudinal study we used administrative data from the Massachusetts All Payer Claims Database (MA APCD). These de-identified data are collected and maintained by the Center for Health Information and Analysis (CHIA). The APCD is a source of health claims data, including pharmacy, medical, and dental claims, for the majority of Massachusetts residents covered by commercial insurance carriers (including Medicare Advantage and Medicare supplement policies) and also Medicaid. The APCD does not include claims covered by workers' compensation, TRICARE, Veterans Health Administration, and Federal Employees Health Benefit Plan. In addition, as self-insured health plans operate under federal rather than state regulatory authority, they are not required to submit claims to CHIA and a subset do opt out. However, because this study focused on individuals who were commercially insured throughout a defined time period, these omissions from the APCD did not impact our ability to reliably assess patterns of treatment discontinuation. The corresponding author's university Institutional Review Board (IRB) approved the study before the authors conducted any analyses.
      For the study, we used APCD pharmacy claims for commercial health plans for the years 2013 through 2017. This dataset contains records of prescriptions reimbursed by private insurers, including information about the drug name and associated national drug codes (NDC), quantity and days for which the drugs were dispensed, information about the patient (age, gender), and the prescribing physician (name, national provider identifier). To collect information about prescriber specialty, we linked the National Provider Identification (NPI) dataset to the APCD. To gather information on the annual volume of patients that prescribers treated using buprenorphine/naloxone, we also linked the APCD to the Massachusetts Prescription Monitoring Program (PMP) database based on prescribers' NPI. We used member eligibility data from APCD to ensure that the individuals included in the study sample were continuously insured throughout the study period.

      2.2 Study sample

      The study sample included patients who met three criteria: (1) 18 to 64 years of age, (2) initiated buprenorphine/naloxone treatment for the first time between January 2014 and December 2014, and (3) Massachusetts resident who was continuously enrolled in a health plan within the commercial insurance market from January 2013 to December 2017. The third criterion was adopted to ensure that discontinuation was not the result of individuals losing their health insurance or moving out of Massachusetts. We used NDCs to extract claims for buprenorphine/naloxone. To identify individuals who appeared to be initiating buprenorphine/naloxone treatment during the relevant time frame, we searched the database for individuals who had a claim for this medication at some point between January and December 2014 but did not have any previous claims for at least 12 months. The date at which the first claim appeared during the relevant time frame served as the index date for treatment. We then followed each patient's medication claims profile for up to two years from the index date to determine if (and if so when) treatment discontinuation occurred.

      2.3 Outcomes and measures

      The primary outcome measure was the discontinuation of buprenorphine/naloxone treatment. Consistent with previous studies (
      • Khemiri A.
      • Kharitonova E.
      • Zah V.
      • Ruby J.
      • Toumi M.
      Analysis of buprenorphine/naloxone dosing impact on treatment duration, resource use and costs in the treatment of opioid-dependent adults: A retrospective study of US public and private health care claims.
      ;
      • Meinhofer A.
      • Williams A.R.
      • Johnson P.
      • Schackman B.R.
      • Bao Y.
      Prescribing decisions at buprenorphine treatment initiation: Do they matter for treatment discontinuation and adverse opioid-related events?.
      ;
      • Saloner B.
      • Daubresse M.
      • Alexander G.C.
      Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multi-state population.
      ;
      • Samples H.
      • Williams A.R.
      • Olfson M.
      • Crystal S.
      Risk factors for discontinuation of buprenorphine treatment for opioid use disorders in a multi-state sample of Medicaid enrollees.
      ;
      • Shcherbakova N.
      • Tereso G.
      • Spain J.
      • Roose R.J.
      Treatment persistence among insured patients newly starting buprenorphine/naloxone for opioid use disorder.
      ;
      • Weinstein Z.M.
      • Kim H.W.
      • Cheng D.M.
      • Quinn E.
      • Hui D.
      • Labelle C.T.
      • Samet J.H.
      Long-term retention in office based opioid treatment with buprenorphine.
      ), we defined treatment discontinuation as an apparent gap in the patient's medication regimen. Specifically, we considered a patient to have discontinued treatment when the patient did not fill a buprenorphine/naloxone prescription for 60 consecutive days. We counted the number of days that elapsed from treatment initiation to treatment discontinuation (i.e., last day of the 60-day gap) and defined this as time to treatment discontinuation. We considered observations for patients remaining on buprenorphine/naloxone without a gap of 60 consecutive days right censored. We also compared medication adherence between patients who discontinued treatment (up to the point of discontinuation) and patients who remained in treatment for the entire two-year follow up period. We measured adherence using the proportion of medication days covered (PDC) for the period in which the patient was in treatment (i.e., 730 days for patients who did not discontinue).
      We investigated whether selected characteristics of a patient and a patient's prescriber were associated with treatment discontinuation. For patient-level characteristics, we included age, gender, and insurance types. We stratified age into four categories: 18 to 24, 25 to 39, 40 to 54, and 55 to 64. We also classified patients based on type of commercial insurance coverage: Health Maintenance Organization (HMO), Preferred Provider Organization (PPO), and other (which primarily comprises patients with indemnity coverage). We also constructed a variable for the number of prescribers from whom a patient received buprenorphine/naloxone treatment for the two-year treatment period. We assigned patients into one of three categories: one prescriber, between two and four prescribers, and more than four prescribers.
      For prescriber-level variables, we included specialty training (i.e., psychiatry, family medicine, internal medicine, and others), and whether a prescriber was certified in addiction medicine. We also accounted for the prescriber's panel size regarding the number of patients for whom they had prescribed buprenorphine/naloxone medication during the year in which the focal patient began treatment. For patients with multiple prescribers, we used the prescriber who was responsible for most of the patient's prescriptions during the treatment period. We included three categories for prescriber panel size: fewer than 50 patients, 51 to 199 patients, and 200 patients or more.

      2.4 Statistical analysis

      We conducted several types of analyses. First, we computed descriptive statistics to compare the two groups of patients—those who discontinued and those who did not—based on the preselected patient- and prescriber-level characteristics. Because all the covariates were categorical, we reported data as proportions.
      Second, we conducted a multi-variate survival analysis to model time-to-discontinuation of buprenorphine/naloxone treatment. We constructed a Kaplan-Meier survival curve depicting the probability of buprenorphine/naloxone treatment continuation over the 2 years of study time-frame and determined the median time to discontinue treatment for the entire study sample. In addition, we calculated median time to treatment discontinuation for all the covariates and performed Chi-squared tests to evaluate the statistical significance of observed differences in median discontinuation times.
      As a third analysis, we also estimated a mixed-effect Cox proportional hazard model to evaluate the statistical significance of the association of different predictors of treatment discontinuation. The mixed effect model allowed us to capture the random effects due to the clustering of patients under an individual prescriber. We also computed intra-class correlation coefficients (ICC) to assess the amount of variance for treatment discontinuation accounted for at the level of the prescriber.
      The descriptive bivariate comparisons and Cox-regression estimates were conducted in R statistical computing software (version 3). We determined statistical significance in both bivariate and multi-variate analyses using a two-tailed test with a p-value set at <0.05.

      3. Results

      The study sample comprised 5134 patients. Table 1 presents the baseline characteristics for the entire study sample. The average age was 36.1 ± 10.76 and approximately one-half of the patients (52.5%) were between the ages of 25 and 39. Approximately 72% of the patients had health insurance coverage through an HMO. While almost 36% of the patients received buprenorphine/naloxone prescriptions from one prescriber only, approximately 20% received such prescriptions from more than four prescribers during the two-year study period. More than 60% of patients received prescriptions from prescribers whose panel size for buprenorphine/naloxone prescriptions was between 51 and 199 patients, with remaining patients divided between prescribers with relatively low (fewer than 50 patients) and high-volume panels (200 or more patients). A little more than one-third of the patients received treatment from prescribers who were trained in psychiatry and approximately 7% received treatment by prescribers who were certified in addiction medicine.
      Table 1Characteristics of Massachusetts commercially insured individuals with buprenorphine/naloxone prescriptions and by treatment discontinuation
      Treatment discontinuation was defined as a gap of 60 consecutive days without a buprenorphine/naloxone prescription.
      pattern.
      TotalContinuedDiscontinued
      Number of patients513412833851
      Patient characteristics
       Patient age
      18–24696 (13.5)95 (7.4)601 (15.6)
      25–392693 (52.5)671 (52.3)2022 (52.5)
      40–541394 (27.2)409 (31.9)985 (25.6)
      55–64351 (6.8)108 (8.4)243 (6.3)
       Patient gender
      Male3130 (61.0)544 (57.6)1460 (62.1)
      Female2004 (39.0)739 (42.4)2391 (37.9)
       Patient insurance type
      HMO3707 (72.2)936 (73.1)2771 (72.0)
      PPO1295 (25.2)329 (25.6)966 (25.1)
      Other132 (2.6)17 (1.3)112 (2.9)
       Number of unique prescribers issuing buprenorphine/naloxone prescription
      1 prescriber1831 (35.7)360 (28.1)1471 (38.2)
      2–4 prescribers2251 (43.8)601 (46.8)1650 (42.8)
      More than 4 prescribers1052 (20.3)322 (25.1)730 (19.0)
      Prescriber characteristics
       Prescriber panel size
      Low (1–50 patients)946 (18.4)250 (19.5)696 (18.1)
      Medium (51–199 patients)3164 (61.6)843 (65.7)2321 (60.3)
      High (200 and above)1024 (20.0)190 (14.8)834 (21.7)
       Prescriber specialty
      Psychiatry1789 (34.9)449 (35.0)1340 (34.8)
      Family medicine1132 (22.0)281 (21.9)851 (22.1)
      Internal medicine1046 (20.4)271 (21.1)775 (20.1)
      Others1167 (22.7)282 (22.0)885 (23.0)
       Addiction medicine certification
      Without certificate4782 (93.1)1195 (93.1)3587 (93.1)
      With certificate352 (6.9)88 (6.9)264 (6.9)
      Numbers in parentheses indicate (%).
      a Treatment discontinuation was defined as a gap of 60 consecutive days without a buprenorphine/naloxone prescription.
      Table 1 also presents descriptive statistics for patients stratified by treatment continuation/discontinuation status. We identified a total of 3851 patients (75%) who discontinued treatment based on a gap of 60 consecutive days without a buprenorphine/naloxone prescription during the two-year treatment period. The PDC of individuals who discontinued treatment (mean PDC = 51.94) was significantly lower than those who continued (mean PDC = 89.61) (not shown in Table 1).
      Fig. 1 depicts the Kaplan-Meier survival curve showing the probability of treatment continuation over the 2-year treatment period. The figure reveals that treatment continuity declines substantially initially and then somewhat flattens out. The median time to treatment discontinuation for the entire study cohort was 300 days (95% CI = 300–330). Table 2 presents the median treatment discontinuation time according to baseline patient- and prescriber-level characteristics. The median time differed significantly according to age, gender, insurance types, number of prescribers, and prescriber's patient volume.
      Fig. 1
      Fig. 1Risk of buprenorphine/naloxone treatment discontinuation (with a gap of 60 consecutive days).
      Table 2Median time (in days) to discontinuation
      Treatment discontinuation was defined as a gap of 60 consecutive days without a buprenorphine/naloxone prescription.
      from buprenorphine/naloxone treatment according to patent- and prescriber-level characteristics.
      Median time (days) to discontinuation (95% CI)p-Value
      p value was calculated using chi-squared test.
      Patient characteristics
       Patient age
      18–24180 (150–210)<0.001
      25–39330 (300–360)
      40–54390 (330–450)
      55–64390 (300–480)
       Patient gender
      Male330 (300–360)0.005
      Female300 (270–330)
       Patient insurance type
      HMO300 (300–360)
      PPO300 (270–360)<0.001
      Other180 (150–300)
       Number of unique prescribers issuing buprenorphine/naloxone prescription
      1 prescriber240 (210–270)
      2–4 prescribers330 (300–360)<0.001
      More than 4 prescribers450 (390–510)
      Prescriber characteristics
       Prescriber panel size
      Low (1–50 patients)330 (270–390)<0.001
      Medium (51–199 patients)360 (330–390)
      High (200 and above)210 (180–240)
       Prescriber specialty
      Psychiatry330 (270–360)0.3
      Family medicine300 (270–360)
      Internal medicine330 (270–390)
      Others270 (240–300)
       Addiction medicine certification
      Without certificate300 (300−330)0.9
      With certificate330 (240–390)
      a Treatment discontinuation was defined as a gap of 60 consecutive days without a buprenorphine/naloxone prescription.
      b p value was calculated using chi-squared test.
      Table 3 presents results from the mixed-effect Cox regression model in the form of hazard ratios (HR). Patients between the ages of 18 and 24 had a 43.6% higher risk (HR = 1.436, 95% CI = 1.240–1.663) of discontinuing treatment compared with those between 55 and 64 years of age. Patients with health insurance coverage other than HMO or PPO had somewhat higher associated risk (HR = 1.252, 95% CI = 1.03–1.522) of discontinuing treatment. In addition, patients with more than one prescriber had a significantly lower risk of experiencing treatment discontinuation. Specifically, those with between 2 and 4 prescribers had approximately 18.7% lower risk (HR = 0.813; 95% CI = 0.753–0.877) and those with more than 4 prescribers had approximately a 34% lower risk (HR = 0.665, 95% CI = 0.602–0.734). Compared with individuals receiving treatment from low-volume prescribers, those who received their treatment from high-volume prescribers experienced a 27.8% higher risk of discontinuing treatment (HR = 1.278, 95% CI = 1.112–1.468). We did not observe any statistically significant association between treatment discontinuation and the specialty or certification status of the prescriber. The intra-class correlation coefficient value was 0.10, indicating some degree of clustering effect at the prescriber level.
      Table 3Estimates from mixed effect cox-proportional hazard model predicting buprenorphine/naloxone treatment discontinuation.
      Treatment discontinuation was defined as a gap of 60 consecutive days without a buprenorphine/naloxone prescription.
      Variables/predictorsHazard ratios95% CI on hazard ratiosp-Value
      Patient characteristics
       Patient age
      55–64ReferenceReferenceReference
      18–241.4361.240–1.663<0.000
      25–391.0790.949–1.2260.246
      40–540.9550.835–1.0930.505
       Patient gender
      FemaleReferenceReferenceReference
      Male1.0550.986–1.1290.121
       Patient insurance type
      HMOReferenceReferenceReference
      PPO1.0120.939–1.0920.751
      Other1.2521.03–1.5220.024
       Unique prescribers issuing buprenorphine/naloxone prescription
      1 prescriberReferenceReferenceReference
      2–4 prescribers0.8130.753–0.877<0.000
      More than 4 prescribers0.6650.602–0.734<0.000
      Prescriber characteristics
       Prescriber panel size
      Low (1–50 patients)ReferenceReferenceReference
      Medium (51–199 patients)0.9640.868–1.070.492
      High (200 and above)1.2781.112–1.4680.001
       Prescriber specialty
      OthersReferenceReferenceReference
      Psychiatry0.9720.875–1.0790.591
      Family medicine1.0120.901–1.1370.839
      Internal medicine0.9540.848–1.0730.429
       Addiction medicine certification
      Without certificateReferenceReferenceReference
      With certificate0.9180.790–1.0660.263
      Intra-class correlation coefficient (ICC)0.10
      a Treatment discontinuation was defined as a gap of 60 consecutive days without a buprenorphine/naloxone prescription.
      We conducted several sensitivity analyses regarding the way the medication gap (i.e., a gap of 60 consecutive days) and treatment period (i.e., 2 years) were specified for the investigation. For one analysis, we re-defined treatment discontinuation as a gap of 30 consecutive days without a buprenorphine/naloxone prescription and maintained a 2-year treatment period. We found a total of 4392 patients had discontinued treatment (Appendix Table A, Characteristics of Massachusetts commercially insured individuals with buprenorphine/naloxone prescriptions and by treatment discontinuation pattern). Thus, the percentage of discontinuing patients increased from 75% to 85.5% and the median treatment discontinuation time declined from 300 days to 210 days (Appendix Fig. A, Risk of buprenorphine/naloxone treatment discontinuation [with a gap of 30 consecutive days]). However, differences in median discontinuation time for baseline patient-level and prescriber-level characteristics did not change in a material way (Appendix Table B, Median time to experience buprenorphine/naloxone treatment discontinuation according to patent- and prescriber-level characteristics). Nor did results change from the Cox model (Appendix Table C, Estimates from mixed effect cox-proportional hazard model predicting buprenorphine/naloxone treatment discontinuation) that we re-estimated for the 30-days medication gap.
      For the other analysis, we reduced the treatment period to 1 year and used a medication gap of 60 days. The percentage of discontinuing patients declined from 75% to 61%. Otherwise, the results from this analysis were not materially different from those previously presented, including median discontinuation time and the association between covariates and discontinuation in terms of magnitude and statistical significance.
      As an additional analysis, we also investigated how many patients re-engaged in buprenorphine/naloxone treatment after discontinuation. To conduct this analysis, we longitudinally tracked patients' buprenorphine/naloxone prescription filling patterns for six months following treatment discontinuation. We considered patients to have re-initiated treatment if they had refill prescriptions for buprenorphine/naloxone in two consecutive months within the six months following treatment discontinuation. Based on these parameters, we found that approximately 20% of patients re-engaged in treatment within six months of initial discontinuation.

      4. Discussion

      Treatment providers are beginning to use buprenorphine/naloxone widely for OUD; however, high rates of medication discontinuation are a major challenge. The data indicate that a substantial percentage of patients discontinue treatment less than one year after initiating therapy, well before they can typically meet criteria for sustained remission. While the observed discontinuation rates are in line with what has been reported in previous studies (
      • Meinhofer A.
      • Williams A.R.
      • Johnson P.
      • Schackman B.R.
      • Bao Y.
      Prescribing decisions at buprenorphine treatment initiation: Do they matter for treatment discontinuation and adverse opioid-related events?.
      ;
      • Ronquest N.A.
      • Willson T.M.
      • Montejano L.B.
      • Nadipelli V.R.
      • Wollschlaeger B.A.
      Relationship between buprenorphine adherence and relapse, health care utilization and costs in privately and publicly insured patients with opioid use disorder.
      ;
      • Saloner B.
      • Daubresse M.
      • Alexander G.C.
      Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multi-state population.
      ;
      • Samples H.
      • Williams A.R.
      • Olfson M.
      • Crystal S.
      Risk factors for discontinuation of buprenorphine treatment for opioid use disorders in a multi-state sample of Medicaid enrollees.
      ), as noted, our investigation is one of the very few studies to examine a medication window longer than one year for a commercially insured population. Our observed one-year retention rate of approximately 40% is also in line with what
      • Manhapra A.
      • Agbese E.
      • Leslie D.L.
      • Rosenheck R.A.
      Three-year retention in buprenorphine treatment for opioid use disorder among privately insured adults.
      reported (i.e., 45%) for their study, which did include a follow-up period of more than 12 months for a commercially insured population. As such, the evidence is quite strong that patients with OUD have a low likelihood of remaining on buprenorphine/naloxone treatment beyond one year, which raises serious concerns about the prognosis for such patients even after they begin treatment with buprenorphine/naloxone. Many patients with OUD require medications for extended periods of time, if not for life. While our research reveals that some patients do return to treatment, whether success rates improve during subsequent treatment cycles requires further investigation.
      Several of the covariates that we examined were associated with treatment discontinuation. In particular, younger patients were at higher risk for discontinuation. This finding is consistent with previous studies of discontinuation rates for buprenorphine/naloxone (
      • Saloner B.
      • Daubresse M.
      • Alexander G.C.
      Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multi-state population.
      ;
      • Samples H.
      • Williams A.R.
      • Olfson M.
      • Crystal S.
      Risk factors for discontinuation of buprenorphine treatment for opioid use disorders in a multi-state sample of Medicaid enrollees.
      ) and also aligns with evidence indicating that the younger a person is when he/she begins to use substances the more serious the addiction, possibly because drug use has a more pronounced effect on the functional structure of developing brains (
      • Lynskey M.T.
      • Heath A.C.
      • Bucholz K.K.
      • Slutske W.S.
      • Madden P.A.
      • Nelson E.C.
      • Martin N.G.
      Escalation of drug use in early-onset cannabis users vs co-twin controls.
      ;
      • Squeglia L.M.
      • Jacobus J.
      • Tapert S.F.
      The influence of substance use on adolescent brain development.
      ). Younger patients may also be less likely to engage in treatment possibly due to immaturity and insulation from the full negative consequences of addiction that older adults are likely to experience, such as issues with employment, geographic stability, financial security, and family unity.
      Two additional findings offer potentially interesting considerations regarding the role of prescribers in managing patients who are undergoing treatment with buprenorphine/naloxone. One of our findings indicated that patients under the care of prescribers with a relatively large panel size for buprenorphine/naloxone were at higher risk for treatment discontinuation. On one hand, these prescribers are likely to be highly experienced and may assume treatment responsibilities for patients with very challenging addiction issues. On the other hand, prescribers dealing with high caseloads may have less time to closely monitor the ongoing care of patients and, consequently, are more likely to miss early signs of potential treatment discontinuation.
      The other finding is that patients with more prescribers during the treatment period were at lower risk of treatment discontinuation. To some degree this association may reflect that patients who discontinue at an early point in their treatment will simply not have an opportunity to see more than one prescriber. Moreover, some patients have multiple prescribers because they undergo treatment in group practice settings where more than one prescriber participates in their care. We lacked the data to determine the extent to which this was the case for patients with multiple prescribers, but it is common. Why this type of arrangement might be associated with a lower risk of treatment discontinuation is not readily apparent, though perhaps these settings offer more accessible care to such patients when needed or employ practitioners with diverse training backgrounds who can offer a multidisciplinary treatment approach (i.e., physician, nurse practitioner, physician assistant, clinical pharmacist, etc.). Some patients who are engaging in treatment may change prescribers to find one with whom they feel most comfortable for managing their treatment based on such considerations as practice style and accessibility, which contributes to long-term adherence. Given the uncertainties for interpreting these two provider-level variables, we re-estimated the models without either one, but this had no impact on the results for the remaining variables.
      While buprenorphine/naloxone offers patients with OUD an effective outpatient treatment option, the high rate of discontinued treatment undermines the true value of the medication. In preparing for this investigation, we conducted interviews with fifteen experienced prescribers of buprenorphine/naloxone where we inquired about their experiences in managing treatment for OUD. The prescribers represented a mix of clinical settings—community-based primary care, specialized drug treatment programs, and academically affiliated outpatient centers. All the prescribers emphasized the difficulty of avoiding medication discontinuation during the early phase of treatment, a time during which many patients are still using non–prescribed opioids. Some of the prescribers with whom we spoke seemingly lacked well-developed patient registries and other systems for monitoring their patients' treatment progress, which does align with the previously noted point that some prescribers with large panel sizes may be more likely to miss early signs of treatment discontinuation. At the same time, the prescribers with whom we spoke emphasized that reducing rates of discontinuation for buprenorphine/naloxone treatment requires more widely available social support systems for patients who are undergoing treatment. As many patients receive buprenorphine/naloxone treatment in settings that do not offer integrated social supports such as individual counseling and family therapy, a recognized need exists to help patients access such services and coordinate them with their clinical treatment.
      Our study has several limitations worth noting that may also offer opportunities for future research. First, we based our measure of medication discontinuation entirely on information from prescription claims. Despite the gaps in their medication fill history, some patients may have remained on buprenorphine/naloxone, possibly obtaining the medication from family members, friends, or other contacts. We were also unable to determine whether a patient who discontinued filling their prescription for buprenorphine/naloxone did so to begin treatment in a methadone clinic, which is an alternative treatment for OUD. Second, we lacked information for examining associations between patients' socioeconomic status and treatment discontinuation. The prescription claims data that we obtained for the study only included 3-digit zip codes for patients' residence to ensure confidentiality of their information, thus limiting our ability to link location of residence with census data for socioeconomic characteristics. In addition, we did not have patient-level information about race/ethnicity, which is a potentially important socioeconomic characteristic for investigating patterns of treatment discontinuation particularly within the context of health disparities. As claims data have often lacked such information, future research will hopefully benefit from datasets in which race/ethnicity has been collected reliably for all patients. Third, as noted, the APCD does not include all Massachusetts' commercial claims as some self-insured plans choose to opt out of this claims submission program. While this omission did not affect our ability to reliably assess treatment discontinuation, it is a study limitation because the cohort comprised only individuals who were able to maintain their commercial health insurance during the study period and these individuals will be better able to remain in treatment than those who lose health insurance even if only temporarily. As such, this limitation may conservatively bias the rates of discontinuation that we report as they may be lower than what would be observed for a cohort that included individuals who lost health insurance. Additionally, our study did not examine patients' outcomes following treatment discontinuation. Guidelines exist for how long patients should be in treatment. A few studies have examined outcomes reporting an association between discontinuation rates and higher rates of emergency room visits and health service utilization (
      • Lo-Ciganic W.H.
      • Gellad W.F.
      • Gordon A.J.
      • Cochran G.
      • Zemaitis M.A.
      • Cathers T.
      • Donohue J.M.J.A.
      Association between trajectories of buprenorphine treatment and emergency department and in-patient utilization.
      ;
      • Ronquest N.A.
      • Willson T.M.
      • Montejano L.B.
      • Nadipelli V.R.
      • Wollschlaeger B.A.
      Relationship between buprenorphine adherence and relapse, health care utilization and costs in privately and publicly insured patients with opioid use disorder.
      ;
      • Tkacz J.
      • Volpicelli J.
      • Un H.
      • Ruetsch C.
      Relationship between buprenorphine adherence and health service utilization and costs among opioid dependent patients.
      ), but these studies have relied on medication windows of 1 year or less. As such, an important opportunity exists for studying patient outcomes in relation to longer term patterns of medication discontinuation for buprenorphine/naloxone.
      In conclusion, we believe that our study contributes to the literature on OUD by reporting high rates of treatment discontinuation for buprenorphine/naloxone within a commercially insured population and by identifying determinants of such discontinuation. At the same time, our study as well as much of the extant research on this topic does not point to patient- or provider-level factors that can readily be put into action to reduce discontinuation rates. As such, an important opportunity exists for future research to offer specific steps that providers and programs can take to improve treatment retention for patients who are struggling with OUD, a challenge that is likely complicated by the fact that many such patients are reluctant to seek treatment in the first place. We hope that our study is a positive step in this important endeavor.

      CRediT authorship contribution statement

      • Md Mahmudul Hasan: Conceptualization; Formal analysis; Investigation; Methodology; Validation; Visualization; Roles/Writing - original draft; Writing - review & editing.
      • Md. Noor-E-Alam: Conceptualization; Methodology; Project administration; Resources; Supervision; Writing - review & editing.
      • Prathamesh Mohite: Conceptualization; Methodology; Formal analysis; Visualization.
      • Md Saiful Islam: Conceptualization; Visualization; Writing - review & editing.
      • Alicia Sasser Modestino: Conceptualization; Methodology; Investigation; Validation; Writing - review & editing.
      • Alyssa Peckham: Conceptualization; Validation; Writing - review & editing.
      • Leonard D. Young: Data curation; Conceptualization; Validation; Writing - review & editing.
      • Gary J. Young: Data curation; Conceptualization; Methodology; Investigation; Validation; Roles/Writing - original draft; Writing - review & editing.

      Disclosure of funding received

      Northeastern University was supported by funding from the Centers for Disease Control and Prevention's (CDC's) Overdose to Action (OD2A) grant received by the Massachusetts Department of Public Health (MDPH).

      Declaration of competing interest

      None.

      Acknowledgement

      We would like to acknowledge Jiesheng Shi, Graduate Student in Data Analytics Engineering, Department of Mechanical and Industrial Engineering, Northeastern University for his help to prepare the data.

      Appendix A

      Appendix Table ACharacteristics of Massachusetts commercially insured individuals with buprenorphine/naloxone prescriptions and by treatment discontinuation
      Treatment discontinuation was defined as a gap of 30 consecutive days without a buprenorphine/naloxone prescription.
      pattern.
      TotalContinuedDiscontinued
      Number of patients51347424392
      Patient characteristics
       Patient age
      18–24696 (13.5)49 (6.6)647 (14.7)
      25–392693 (52.5)382 (51.5)2311 (52.6)
      40–541394 (27.2)250 (33.7)1144 (26.1)
      55–64351 (6.8)61 (8.2)290 (6.6)
       Patient gender
      Male3130 (61.0)399 (53.8)2731 (62.2)
      Female2004 (39.0)343 (46.2)1661 (37.8)
       Patient insurance type
      HMO3707 (72.2)542 (73.1)3165 (72.1)
      PPO1295 (25.2)191 (25.7)1104 (25.2)
      Other132 (2.6)9 (1.2)120 (2.7)
       Number of unique prescribers issuing buprenorphine/naloxone prescription
      1 prescriber1831 (35.7)203 (27.4)1628 (37.1)
      2–4 prescribers2251 (43.8)354 (47.7)1897 (43.2)
      More than 4 prescribers1052 (20.3)185 (24.9)867 (19.7)
      Prescriber characteristics
       Prescriber panel size
      Low (1–50 patients)946 (18.4)141 (19)805 (18.4)
      Medium (51–199 patients)3164 (61.6)492 (65.3)2672 (60.8)
      High (200 and above)1024 (20.0)109 (14.7)915 (20.8)
       Prescriber specialty
      Psychiatry1789 (34.9)242 (32.6)1547 (35.2)
      Family medicine1132 (22.0)167 (22.5)965 (22)
      Internal medicine1046 (20.4)157 (21.2)889 (20.2)
      Others1167 (22.7)176 (23.7)991 (22.6)
       Addiction medicine certification
      Without certificate4782 (93.1)697 (94.0)4085 (93.0)
      With certificate352 (6.9)45 (6.0)307 (7.0)
      Numbers in parentheses indicate (%).
      a Treatment discontinuation was defined as a gap of 30 consecutive days without a buprenorphine/naloxone prescription.
      Appendix Table BMedian time to experience buprenorphine/naloxone treatment discontinuation
      Treatment discontinuation was defined as a gap of 30 consecutive days without a buprenorphine/naloxone prescription.
      according to patent- and prescriber-level characteristics.
      Median time (days) to discontinuation (95% CI)p-Value
      p value was calculated using chi-squared test.
      Patient characteristics
       Patient age
      18–24120 (120–150)<0.001
      25–39210 (180–240)
      40–54210 (210–270)
      55–64210 (180–270)
       Patient gender
      Male210 (210–240)<0.001
      Female180 (180–210)
       Patient insurance type
      HMO210 (180–210)<0.001
      PPO210 (180–210)
      Other150 (120–180)
       Number of unique prescribers issuing buprenorphine/naloxone prescription
      1 prescriber150 (150–180)<0.001
      2–4 prescribers210 (210–240)
      More than 4 prescribers270 (240–300)
      Prescriber characteristics
       Prescriber panel size
      Low (1–50 patients)210 (180–210)<0.001
      Medium (51–199 patients)210 (210–240)
      High (200 and above)150 (120–180)
       Prescriber specialty
      Psychiatry210 (180–240)0.9
      Family medicine180 (180–210)
      Internal medicine210 (180–240)
      Others180 (150–210)
       Addiction medicine certification
      Without certificate210 (180–210)0.5
      With certificate210 (180–240)
      a Treatment discontinuation was defined as a gap of 30 consecutive days without a buprenorphine/naloxone prescription.
      b p value was calculated using chi-squared test.
      Appendix Table CEstimates from mixed effect cox-proportional hazard model predicting buprenorphine/naloxone treatment discontinuation.
      Treatment discontinuation was defined as a gap of 30 consecutive days without a buprenorphine/naloxone prescription.
      Variables/predictorsHazard ratios95% CI on hazard ratiosp-Value
      Patient characteristics
       Patient age
      55–64ReferenceReferenceReference
      18–241.3951.215–1.601<0.000
      25–391.0540.936–1.1860.389
      40–540.9440.833–1.0690.363
       Patient gender
      FemaleReferenceReferenceReference
      Male1.0821.016–1.1520.015
       Patient insurance type
      HMOReferenceReferenceReference
      PPO1.0080.940–1.0820.820
      Other1.2501.036–1.5090.020
       Unique prescribers issuing buprenorphine/naloxone prescription
      1 prescriberReferenceReferenceReference
      2–4 prescribers0.8270.770–0.888<0.000
      More than 4 prescribers0.7050.643–0.773<0.000
      Prescriber characteristics
       Prescriber panel size
      Low (1–50 patients)ReferenceReferenceReference
      Medium (51–199 patients)0.9280.843–1.0220.129
      High (200 and above)1.1521.014–1.3090.030
       Prescriber specialty
      OthersReferenceReferenceReference
      Psychiatry1.0040.911–1.1060.935
      Family medicine1.0080.905–1.1230.885
      Internal medicine0.9760.875–1.0890.668
       Addiction medicine certification
      Without certificateReferenceReferenceReference
      With certificate0.9790.854–1.1230.763
      Intra-class correlation coefficient (ICC)0.10
      a Treatment discontinuation was defined as a gap of 30 consecutive days without a buprenorphine/naloxone prescription.
      Appendix Fig. A
      Appendix Fig. ARisk of buprenorphine/naloxone treatment discontinuation (with a gap of 30 consecutive days).

      References

        • Barocas J.A.
        • White L.F.
        • Wang J.
        • Walley A.Y.
        • LaRochelle M.R.
        • Bernson D.
        • Linas B.P.
        Estimated prevalence of opioid use disorder in Massachusetts, 2011–2015: A capture–recapture analysis.
        American Journal of Public Health. 2018; 108: 1675-1681
        • Bentzley B.S.
        • Barth K.S.
        • Back S.E.
        • Book S.W.
        Discontinuation of buprenorphine maintenance therapy: Perspectives and outcomes.
        Journal of Substance Abuse Treatment. 2015; 52: 48-57
        • Birnbaum H.G.
        • White A.G.
        • Schiller M.
        • Waldman T.
        • Cleveland J.M.
        • Roland C.L.
        Societal costs of prescription opioid abuse, dependence, and misuse in the United States.
        Pain Medicine. 2011; 12: 657-667
        • Chang H.-Y.
        • Daubresse M.
        • Saloner B.
        • Alexander G.C.
        Chronic disease medication adherence after initiation of buprenorphine for opioid use disorder.
        Medical Care. 2019; 57: 667-672
        • Degenhardt L.
        • Charlson F.
        • Mathers B.
        • Hall W.D.
        • Flaxman A.D.
        • Johns N.
        • Vos T.J.A.
        The global epidemiology and burden of opioid dependence: Results from the global burden of disease 2010 study.
        . 2014; 109: 1320-1333
        • Florence C.
        • Luo F.
        • Xu L.
        • Zhou C.
        The economic burden of prescription opioid overdose, abuse and dependence in the United States, 2013.
        Medical Care. 2016; 54: 901
        • Hansen R.N.
        • Oster G.
        • Edelsberg J.
        • Woody G.E.
        • Sullivan S.D.
        Economic costs of nonmedical use of prescription opioids.
        The Clinical Journal of Pain. 2011; 27: 194-202
        • Hser Y.I.
        • Saxon A.J.
        • Huang D.
        • Hasson A.
        • Thomas C.
        • Hillhouse M.
        • Wiest K.
        Treatment retention among patients randomized to buprenorphine/naloxone compared to methadone in a multi-site trial.
        Addiction. 2014; 109: 79-87
        • Khemiri A.
        • Kharitonova E.
        • Zah V.
        • Ruby J.
        • Toumi M.
        Analysis of buprenorphine/naloxone dosing impact on treatment duration, resource use and costs in the treatment of opioid-dependent adults: A retrospective study of US public and private health care claims.
        Postgraduate Medicine. 2014; 126: 113-120
        • Kraus M.L.
        • Alford D.P.
        • Kotz M.M.
        • Levounis P.
        • Mandell T.W.
        • Meyer M.
        • Wyatt S.A.
        Statement of the American Society of Addiction Medicine Consensus Panel on the use of buprenorphine in office-based treatment of opioid addiction.
        Journal of Addiction Medicine. 2011; 5: 254-263
        • Larochelle M.R.
        • Bernson D.
        • Land T.
        • Stopka T.J.
        • Wang N.
        • Xuan Z.
        • Walley A.Y.
        Medication for opioid use disorder after nonfatal opioid overdose and association with mortality: A cohort study.
        Annals of Internal Medicine. 2018; 169: 137-145
        • Lo-Ciganic W.H.
        • Donohue J.M.
        • Kim J.Y.
        • Krans E.E.
        • Jones B.L.
        • Kelley D.
        • safety, d.
        Adherence trajectories of buprenorphine therapy among pregnant women in a large state Medicaid program in the United States.
        Pharmacoepidemiology and Drug Safety. 2019; 28: 80-89
        • Lo-Ciganic W.H.
        • Gellad W.F.
        • Gordon A.J.
        • Cochran G.
        • Zemaitis M.A.
        • Cathers T.
        • Donohue J.M.J.A.
        Association between trajectories of buprenorphine treatment and emergency department and in-patient utilization.
        Addiction. 2016; 111: 892-902
        • Lynskey M.T.
        • Heath A.C.
        • Bucholz K.K.
        • Slutske W.S.
        • Madden P.A.
        • Nelson E.C.
        • Martin N.G.
        Escalation of drug use in early-onset cannabis users vs co-twin controls.
        Jama. 2003; 289: 427-433
        • Manhapra A.
        • Agbese E.
        • Leslie D.L.
        • Rosenheck R.A.
        Three-year retention in buprenorphine treatment for opioid use disorder among privately insured adults.
        Psychiatric Services. 2018; 69: 768-776
        • Manhapra A.
        • Rosenheck R.
        • Fiellin D.A.
        Opioid substitution treatment is linked to reduced risk of death in opioid use disorder.
        British Medical Journal Publishing Group, 2017
        • McAdam-Marx C.
        • Roland C.L.
        • Cleveland J.
        • Oderda G.M.
        Costs of opioid abuse and misuse determined from a Medicaid database.
        Journal of Pain & Palliative Care Pharmacotherapy. 2010; 24: 5-18
        • Meinhofer A.
        • Williams A.R.
        • Johnson P.
        • Schackman B.R.
        • Bao Y.
        Prescribing decisions at buprenorphine treatment initiation: Do they matter for treatment discontinuation and adverse opioid-related events?.
        Journal of Substance Abuse Treatment. 2019; 105: 37-43
        • Morgan J.R.
        • Schackman B.R.
        • Leff J.A.
        • Linas B.P.
        • Walley A.Y.
        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. 2018; 85: 90-96
        • Oderda G.M.
        • Lake J.
        • Rüdell K.
        • Roland C.L.
        • Masters E.T.
        Economic burden of prescription opioid misuse and abuse: A systematic review.
        Journal of Pain & Palliative Care Pharmacotherapy. 2015; 29: 388-400
        • Ronquest N.A.
        • Willson T.M.
        • Montejano L.B.
        • Nadipelli V.R.
        • Wollschlaeger B.A.
        Relationship between buprenorphine adherence and relapse, health care utilization and costs in privately and publicly insured patients with opioid use disorder.
        Substance Abuse and Rehabilitation. 2018; 9: 59
        • Saloner B.
        • Daubresse M.
        • Alexander G.C.
        Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multi-state population.
        Medical Care. 2017; 55: 669
        • Samples H.
        • Williams A.R.
        • Olfson M.
        • Crystal S.
        Risk factors for discontinuation of buprenorphine treatment for opioid use disorders in a multi-state sample of Medicaid enrollees.
        Journal of Substance Abuse Treatment. 2018; 95: 9-17
        • Shcherbakova N.
        • Tereso G.
        • Spain J.
        • Roose R.J.
        Treatment persistence among insured patients newly starting buprenorphine/naloxone for opioid use disorder.
        Annals of Pharmacotherapy. 2018; 52: 405-414
        • Sordo L.
        • Barrio G.
        • Bravo M.J.
        • Indave B.I.
        • Degenhardt L.
        • Wiessing L.
        • Pastor-Barriuso R.
        Mortality risk during and after opioid substitution treatment: Systematic review and meta-analysis of cohort studies.
        Bmj. 2017; 357
        • Squeglia L.M.
        • Jacobus J.
        • Tapert S.F.
        The influence of substance use on adolescent brain development.
        Clinical EEG and Neuroscience. 2009; 40: 31-38
        • Strassels S.
        Economic burden of prescription opioid misuse and abuse.
        Journal of Managed Care Pharmacy. 2009; 15: 556-562
        • Tkacz J.
        • Severt J.
        • Kassed C.
        • Ruetsch C.
        Clinical differences between opioid abuse classes ameliorated after 1 year of buprenorphine-medication assisted treatment.
        Journal of Addictive Diseases. 2012; 31: 100-111
        • Tkacz J.
        • Volpicelli J.
        • Un H.
        • Ruetsch C.
        Relationship between buprenorphine adherence and health service utilization and costs among opioid dependent patients.
        Journal of Substance Abuse Treatment. 2014; 46: 456-462
        • Weinstein Z.M.
        • Kim H.W.
        • Cheng D.M.
        • Quinn E.
        • Hui D.
        • Labelle C.T.
        • Samet J.H.
        Long-term retention in office based opioid treatment with buprenorphine.
        Journal of Substance Abuse Treatment. 2017; 74: 65-70
        • Welty L.
        • Harrison A.
        • Abram K.
        • Olson N.
        • Aaby D.
        • McCoy K.
        Substance Abuse and Mental Health Services Administration (2017). Key substance use and mental health indicators in the United States: Results from the 2016 National Survey on Drug Use and Health (HHS Publication no. SMA 17-5044, NSDUH series H-52). Rockville, MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. Retrieved.
        College of Health Sciences. 2019; 106: 128
        • White A.G.
        • Birnbaum H.G.
        • Mareva M.N.
        • Daher M.
        • Vallow S.
        • Schein J.
        • Katz N.
        Direct costs of opioid abuse in an insured population in the United States.
        Journal of Managed Care Pharmacy. 2005; 11: 469-479