If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The use of medications for treating OUD was examined in a 12 step treatment setting.
•
Nearly all participants successfully completed residential treatment.
•
Patients who were compliant with medications had positive outcomes after treatment.
•
Patients who were not compliant with naltrexone had especially unfavorable outcomes.
Abstract
Rationale
Opioid overdose deaths and healthcare costs associated with opioid use disorder (OUD) continue to escalate while the majority of addiction treatment providers in the United States do not use medication-assisted treatment (MAT) in spite of proven efficacy. The primary resistance to the use of MAT has been associated with the philosophical conflict many 12-step based treatment programs have with the use of these medications.
Objective
This study sought to determine whether patients self-selecting into a treatment program based upon the 12-step philosophy would elect to use MAT and, if so, what initial outcomes might result.
Methods
This naturalistic, prospective study of patients (N = 253) with OUD included a combination of OUD-specific group therapy and the use of buprenorphine-naloxone, oral naltrexone, injectable naltrexone, or no medication with standard 12-step treatment initiated in a residential or day treatment setting with outpatient follow-up. Baseline assessment of subjects with OUD included level of craving and opioid withdrawal symptom severity. Post-residential treatment outcomes at 1- and 6-months included craving, opioid withdrawal, residential treatment completion, continuing care compliance, medication compliance, substance use frequency and 12-step meeting attendance.
Results
Irrespective of medication condition, nearly all patients successfully completed residential treatment and the majority attended additional programming afterward. Among those who elected to take a medication (71%), differences were associated with medication compliance. Patients who reported compliance with their medication at 1 and 6 months following residential treatment had significantly higher abstinence rates than patients who reported noncompliance. Among those who relapsed post-discharge, neither medication use nor compliance was significantly related to a change in the frequency of alcohol use days or drug use days at 6 months.
Conclusion
These preliminary results suggest that it is feasible to administer medications, including partial opioid agonists like buprenorphine, within the context of 12-step based treatment and taking these medications as prescribed is associated with favorable outcomes.
The need for effective treatment for patients with opioid use disorder (OUD) is more urgent now than ever before. Overdose deaths in the United States from illicit and prescribed opioids have dramatically increased over the last several years and are currently at record highs (
Substance Abuse and Mental Health Services Administration
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).
Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration,
Rockville, MD2017
Patients with OUD are at increased risk for premature death, are more likely to exhibit poor social and economic functioning, and are more likely to be involved with crime than individuals who are involved with other substances (
). Individuals with OUD also pose a strong challenge to clinicians and other mental health professionals when presenting for treatment because they are highly likely to drop out of treatment prematurely and often resume opioid use (
Effectiveness of injectable extended-release naltrexone vs. daily buprenorphine-naloxone for opioid dependence: A randomized clinical noninferiority trial.
). An additional challenge for individuals with OUD is that many have a high degree of clinical severity, including misuse of substances other than opioids, severe drug use histories, and the presence of co-occurring psychological disorders like anxiety and depression (
A great deal of research has focused on identifying effective treatment for individuals with OUD. One modality of OUD treatment that has received a large amount of scientific support is medication-assisted treatment (MAT;
). Examples of psychosocial therapies commonly employed within a MAT approach include cognitive-behavioral therapy, motivational enhancement therapy, and contingency management (
). Increased dissemination of MAT across a variety of programs is critical given the marked gaps between the need for MAT and the capacity of MAT providers in the United States (
explores explanations for the slow incorporation of MAT within traditional 12-step based treatment centers, including a philosophical conflict between administering partial opioid agonists such as buprenorphine and the abstinence-based goals of 12-step based treatment. Galanter also points out that OUD medications are used in clinical settings that do not typically utilize a 12-step approach, and 12-step based treatment settings focusing on OUD do not typically utilize medications, especially buprenorphine. A study involving interviews of over 300 administrators of privately funded substance use disorder treatment organizations found that only 34% of patients with OUD received naltrexone, buprenorphine or methadone as part of treatment. These rates were dramatically lower than the rates of medications for treatment of co-occurring psychiatric disorders (
). As a result of underutilization, very little is known about how patients attending 12-step based treatment for OUD would benefit from receiving medications like naltrexone and buprenorphine-naloxone.
In January 2013, the Hazelden Betty Ford Foundation began using naltrexone and buprenorphine-naloxone with OUD patients attending 12-step based residential or day treatment (for simplicity, referred to as residential treatment hereafter). This approach, combined with OUD-specific group therapy sessions, is called Comprehensive Opioid Response with the 12-steps (COR-12). This paper provides initial data on the feasibility and effectiveness of the combined approach. Given strict definitions of “sobriety” and “recovery” espoused among traditional 12-step mutual-support communities (
), the study sought to determine whether patients self-selecting into a treatment program based upon the 12-step philosophy would elect to use MAT and, if so, what initial outcomes might result.
2. Material and methods
2.1 Study design
This study was a prospective, naturalistic examination of the feasibility and initial impact of combining MAT for OUD within the context of a 12-step based professional treatment model. Prior to conducting this study, it was approved by the Institutional Review Board at Schlmann Associates IRB, an independent review board. The study was monitored throughout by this IRB, and all participants provided informed consent and signed informed consent documents.
2.2 Participants
Two hundred fifty-nine patients with OUD who attended residential or day treatment in the period spanning from June 2013 to June 2017 were recruited into the study, with 98% (n = 253) subsequently enrolled in the COR-12 program. Patients in both programs participated in the same treatment activities during the day: those in residential stayed on campus continuously, while those in day treatment were allowed to leave campus after programming concluded in order to attend work, school, or other activities. All patients met DSM criteria for OUD and identified an opioid as their drug of choice at the time of treatment admission. Professional, licensed addiction counselors provided the substance use disorder (SUD) diagnostic evaluations. The initial 18 months of the study used DSM-IV criteria to define substance dependence, followed by a transition to DSM-V criteria for the later period of the study. Four or more DSM-V criteria were required for diagnosis of a moderate to severe OUD, thus appropriate for inclusion in the study. An active OUD was the only requirement for inclusion in the study. The only exclusionary criterion was an active diagnosis of schizophrenia or other severe psychosis.
2.3 Description of the COR-12 program
In January 2013, the Hazelden Betty Ford Foundation implemented the COR-12 program. The program was a supplement to the existing residential treatment program, which focuses on 12-step based, professionally delivered treatment. The residential treatment model is based on the principles and philosophy of Alcoholics Anonymous and stresses the importance of the 12-steps in recovery from substance use disorders using 12-step facilitation therapy. The model also includes the use of motivational interviewing techniques, cognitive-behavioral therapy, and contingency management. While the primary treatment method is group therapy, individual sessions regularly occur with licensed addiction counselors. Co-occurring mental health disorders are treated by psychology and psychiatry staff in residential and outpatient settings. SUDs are seen as brain diseases that are primary, chronic, and progressive and a goal of treatment is to help the individual achieve long-term abstinence from all alcohol, illicit drugs, and/or misused prescription drugs.
Two primary elements of treatment specific to the COR-12 program for OUD include: 1) the administration of buprenorphine-naloxone or naltrexone if appropriate for the patient, and 2) regular attendance at OUD-specific group therapy sessions throughout treatment; these were provided at no extra cost to patients.
The OUD-specific group therapy sessions focus on challenges unique to OUD issues, including the experience of taking these medications as part of treatment, the risks associated with opioid relapse, shame associated with maladaptive behaviors related to opioid use, the use of naloxone in case of overdose, and addressing stigma associated with the use of medications that may be encountered in 12-step meetings. These tailored groups continue in the outpatient setting if patients transfer within the same system of care after completion of residential treatment.
Level of care decision-making was based on American Society of Addiction Medicine (ASAM) criteria, with regular interaction with insurance reviewers. Level of care decisions also included assessment of and attention to major risk factors, emotional and cognitive conditions and establishment of continuing care planning. Decisions to transfer patients into lower levels of care were made by an interdisciplinary team. Subsequent levels of care included intensive outpatient, outpatient, sober living residences, and other services (e.g., outpatient mental health care).
2.4 Measures
2.4.1 Baseline clinical characteristics
Information relevant to patients' baseline characteristics was drawn from each patient's clinical chart, including basic demographics, the number of previous residential treatment episodes, mental health diagnoses, drug and alcohol use disorder diagnoses, and opioid of choice. In addition, information regarding the current residential episode was pulled from patients' charts, including the length of stay, discharge status, and if the patient stepped down to a lower level of care following discharge from residential treatment.
) measures a patient's current level of craving (i.e., at the present moment) through 14 items that cover three domains: desire and intention to drug use, negative reinforcement, and perceived control over drug use. For the purposes of this study, the wording of questions was slightly modified to refer to the patient's opioid of choice if other than heroin. Each item is a statement, such as: “my desire to use this drug now seems overwhelming,” and the patient is asked to indicate agreement on a 7-point scale (1 = strongly disagree to 7 = strongly agree). The two items on the control over use scale: “I could easily limit how much of this drug I would use if I used now,” and “if I started using this drug now I would be able to stop,” were reverse-scored, such that higher scores on all items represented a higher level of craving or desire. The heroin version of the DDQ has been found to have high test-retest reliability and concurrent validity (
), and has been used in several studies of individuals with OUD.
2.4.2.2 Single-item craving
Two additional items were added to the DDQ to capture self-reported opioid craving at treatment admission and at the time of the initial study assessment: “on the first day that you were admitted to treatment here, to what extent were you experiencing craving for this drug,” and “at this moment, to what extent are you currently experiencing craving for this drug?” Both were answered using a response format of frequency (1 = no craving at all to 10 = extremely strong craving). These measurements allowed us to track changes in craving early in the treatment episode as well as compare craving during early treatment with craving after treatment (the same item was used to measure craving in the post-treatment follow-up assessments).
) is a self-report 10-item scale that assesses several physical aspects of opioid withdrawal. Participants rate withdrawal symptoms experienced in the past 24-h on a 4-point scale of severity (0 = none to 3 = severe). Sample items include: “feeling sick,” “muscle spasms/twitching,” and “insomnia/problems sleeping.” The SOWS is a psychometrically validated instrument and has been utilized in studies of patients with OUD (
At roughly 1- and 6-months after discharge from the residential program, participants completed an outcomes survey over the phone. The survey at 1 month referenced the timeframe since discharge from residential treatment and the 6-month survey referenced the timeframe since the last survey (or since discharge if the 1-month survey was missed).
Post-residential treatment outcomes included self-reported craving for opioid drug of choice (1 = no craving at all to 10 = extremely strong craving), number of alcohol and illicit drug use days, and number of 12-step meetings attended. Because the intended result of participation in a 12-step program is sobriety from all substances, continuous abstinence from both drugs and alcohol was the primary outcome of interest (0 = relapsed, 1 = continuous abstinence).
In addition, patients were asked a series of questions to ascertain if they took an anti-craving medication during follow-up. If they answered “yes” to the question, “do you have a prescription for anti-craving medication,” they were asked, “which of the following anti-craving medications have you been prescribed,” and “have you taken your anti-craving medication as prescribed?” The answer to the last question was then used to determine a 1- and 6-month medication compliance score for each patient (0 = noncompliant, 1 = compliant).
Because recommendations about stepping down to a lower level of care are often included in patients' continuing care plans following residential treatment, study participants were asked at 1- and 6-month follow-up if they were currently participating in any substance use disorder treatment programs or services, and if so, what type (e.g., outpatient program, sober living).
2.5 Procedure
Research staff examined the treatment provider's electronic health record (EHR) to identify patients with OUD who admitted to the residential program and subsequently enrolled in the COR-12 program. Eligible participants were approached as early in the episode as possible and those who agreed to be in the study provided written voluntary informed consent. At the time of the study's onset, clinical staff provided patients with the option of participating in the OUD-specific support groups, a process that could take up to several weeks (median = 25 days). Approximately half-way through the study's recruitment period, clinical staff began automatically enrolling all OUD patients in COR-12 at treatment onset, thus shortening the time between admission and recruitment to an average of 7.6 days (SD = 3.32).
Because this was an observational study, research staff did not decide which OUD patients were inducted on buprenorphine-naloxone or naltrexone. The no medication group was comprised primarily of those who refused medication. Medical staff (e.g., physicians, nurse practitioners) conducted the initial medication conversation with patients and made recommendations based on contraindications (e.g., liver function, pregnancy) and patient preferences. Limited predictors exist distinguishing responsiveness to these two medications. One subgroup was excluded from use of buprenorphine-naloxone; those who had experience with buprenorphine as their first intoxicating opioid and could not comprehend it as a useful medication. Unless medications were deemed medically inappropriate, all OUD patients were encouraged to begin taking either buprenorphine-naloxone or naltrexone. Among patients interested in naltrexone, oral naltrexone was provided when patients refused injectable extended-release naltrexone (hereafter referred to as injectable naltrexone), most often due to cost. Due to onerous regulatory issues, and the relative ease of use of buprenorphine-naloxone and naltrexone post-discharge, methadone was not offered to patients.
COR-12 counselors reinforced the importance of using MAT as a primary component of treatment. While direct assessment of staff attitudes toward MAT was not conducted, anecdotal reports consistently indicated buy-in from the majority of staff. At the time COR-12 was implemented within the treatment system, many staff members had been personally impacted by an opioid overdose death of either a former patient, family member, or member of their social network and, as a result, most supported the use of these medications to improve outcomes.
Research staff tracked recruited patients throughout the study and examined charts to determine: 1) which recruited participants were successfully inducted on either form of naltrexone or buprenorphine-naloxone during the residential episode, and 2) among patients who took buprenorphine-naloxone, which patients continued taking it as a maintenance medication (i.e., beyond the withdrawal period). Because patients are commonly prescribed buprenorphine-naloxone for detoxification/withdrawal purposes, only those who elected to continue it for maintenance purposes were assigned to the buprenorphine-naloxone condition for the purposes of this study.
After recruitment and consent, participants completed a brief baseline assessment session wherein they completed the DDQ, the single item craving questions, and the SOWS. They were then told they would be contacted by research staff at several points after discharge from the residential treatment program.
Outcome surveys are conducted as a part of routine healthcare operations at the Hazelden Betty Ford Foundation. Throughout the year following discharge from residential treatment, research staff contact patients at systematic intervals in order to track patient outcomes. The research staff conducting the calls have no clinical contact with patients and are trained extensively to emphasize confidentiality of survey responses, remain neutral in regards to alcohol and drug use disclosures, and maintain a professional demeanor in order to minimize under-reporting of drug and/or alcohol use or over-reporting of estimated medication compliance. For the current study, research staff contacted patients by phone roughly 1- and 6-months after discharge and asked them to complete a follow-up outcomes survey. Because these patients were included in a research study, they were contacted more frequently (once per week during their survey window) in order to reduce attrition. If outcome callers were unable to reach a patient, they called patients' designated “locators,” consisting of friends or family members who a patient identified at treatment admission as someone who could help locate them following discharge.
2.6 Analysis plan
Successful induction on either buprenorphine-naloxone or naltrexone (oral or injectable) was determined via a review of the patient's medical record. For purposes of analysis, patients successfully inducted on one of the medications prior to their discharge from residential were assigned to either the buprenorphine-naloxone condition (for those electing to continue it for maintenance purposes) or one of the naltrexone conditions. Because of established differences in the efficacy of oral and injectable naltrexone for reducing craving in OUD patients (e.g.,
), naltrexone patients were split into two separate groups depending on the route of drug administration selected. In a few cases, patients did not take a medication during their residential stay, but reported taking a medication for all or a portion of the follow-up (post-treatment) period. These patients were placed in the level of medication condition corresponding to the medication they were taking during follow-up. Patients who did not take one of the medications during treatment or during the follow-up period were assigned to the no medication condition. This group consisted of patients who opted not to go on medications while in treatment or who were deemed inappropriate for medications for medical reasons. In addition, six patients in the sample left treatment against staff advice before a medication decision was made by medical staff. Because this was considered the naturalistic study corollary to the randomization point, these individuals were not included in the analysis. To ensure this did not bias the results, analyses were repeated including these subjects in the no medication group, which did not result in any changes in statistical significance (
). In sum, the medication condition variable consisted of four levels: 1) patients taking buprenorphine-naloxone; 2) patients taking oral naltrexone; 3) patients taking injectable naltrexone; 4) and patients taking no medications.
Because patients in each of the medication conditions may have differed in both demographic and baseline clinical characteristics, the first set of analyses examined whether medication condition was related to each of these variables. Chi-square tests were used for qualitative variables, analysis of variance for quantitative variables, and nonparametric tests for variables with a high degree of skewness. An alpha level of 0.05 was used for all analyses.
All variables that significantly differed across medication groups were treated as covariates in subsequent outcomes analyses. For example, because buprenorphine-naloxone patients had significantly higher craving for their drug of choice early in treatment, as well as a higher SOWS score compared to the other three groups, craving and SOWS scores were treated as covariates in substance use outcomes analyses.
In addition, because medication compliance is likely to have an impact on substance use (
), all analyses examining post-residential outcomes (including abstinence status and substance use days) were conducted with a revised medication condition variable. The medication condition/compliance variable incorporated whether patients reported on the 1- and 6-month follow-up surveys being compliant with the medication during the follow-up period (since discharge from residential). As a result, outcomes analyses consisted of a comparison of seven groups: 1) buprenorphine-naloxone/compliant, 2) buprenorphine-naloxone/noncompliant, 3) oral naltrexone/compliant, 4) oral naltrexone/noncompliant, 5) injectable naltrexone/compliant, 6) injectable naltrexone/noncompliant, and 7) no medications. Analyses also compared the buprenorphine-naloxone groups to the two naltrexone groups to determine which medication and/or route of drug administration produced better outcomes.
The impact of the seven-level medication condition/compliance variable on continuous abstinence from both drugs and alcohol during each post-residential follow-up period (1 and 6 months) was examined via binomial logistic regression. Because different factors may predict the presence (i.e., abstinence vs. relapse) and level (i.e., total number of days using) of alcohol or drug use, two zero-inflated negative binomial models were used to examine the impact of the medication condition/compliance variable on the number of alcohol use days and the number of drug use days. This model was ideal because both variables were highly skewed (due to an excess of zeroes) and overdispersed (i.e., the standard deviation far exceeded the mean;
Inference in regression models of heavily skewed alcohol use data: A comparison of ordinary least squares, generalized linear models, and bootstrap resampling.
Table 1 summarizes the demographic and baseline clinical characteristics of the study sample. Sixty-eight percent of participants were male and the average age was 30.24 years. Fifty-two percent of the sample was 26 years or younger (the youngest participants were 19 years old). Ninety-six percent were Caucasian. Fifty-eight percent of patients identified heroin as their drug of choice and 30% identified oxycodone. In addition to an OUD, 48% of the sample had an alcohol use disorder and 42% had a cannabis use disorder. Other co-occurring substance use disorder diagnoses are shown in Table 1. Over 20% of patients had three or more substance use disorders in addition to OUD. Regarding mental health, 90% had at least one DSM mental health disorder, typically a depressive (33%) or anxiety disorder (42%). Fifty-three percent had at least one prior treatment for a substance use disorder; 25% had three or more prior episodes. Approximately two-thirds of the sample (62.6%) reported previous SUD treatments (M = 1.30, SD = 1.60).
Table 1Demographic and clinical characteristics at baseline.
The average length of stay in the residential treatment program was 39.73 days (SD = 25.89) with 92% of patients successfully completing treatment. No differences in treatment completion were found among those enrolled in COR-12 later in their stay versus those enrolled automatically at treatment admission (χ2 = 6.53, p > .05). Completion of treatment was defined as being discharged with a status of “with staff approval” or “conditional with staff approval.” Conditional with staff approval typically meant that the patient completed all aspects of treatment but that clinical staff had at least one concern about the patient, such as hesitation or ambivalence in following through with continuing care recommendations. Seventy-three percent of the sample also attended at least one other program in the same system of care (e.g., intensive outpatient, sober housing) after their discharge from residential, while 5.4% reported attending another program outside the Hazelden Betty Ford system.
Four of the 253 participants (1.6%) died after leaving treatment. Two were in the injectable naltrexone group, one was in the oral naltrexone group, and one was in the buprenorphine-naloxone group. No patients from the no medication group died. Adverse events were reported to the organizational quality review team and the IRB. No significant medication related adverse events occurred during the course of the study.
3.2 Medication condition
Based on review of patient charts, the medication administration record, and follow-up data, we identified 87 patients taking buprenorphine-naloxone (34%), 27 patients taking oral naltrexone (11%), 65 patients taking injectable naltrexone (26%), and 74 patients taking no medications (29%).
3.3 Demographic and baseline clinical characteristics as a function of medication condition
Full results are shown in Table 2. No significant group differences emerged for gender, χ2(3) = 5.18, p = .16, or age, F(3,249) = 1.79, p = .15. No group differences were found for total number of co-occurring SUD diagnoses in addition to OUD, F(3,249) = 1.26, p = .29. A significant difference emerged for the total number of co-occurring mental health disorders, F(3,249) = 2.87, p < .05: buprenorphine-naloxone patients and oral naltrexone patients had a significantly higher number of mental health disorders than the no medication patients.
Table 2Demographic and baseline clinical characteristics as a function of medication condition.
Medication Condition
Buprenorphine-naloxone (N = 87)
Oral naltrexone (N = 27)
Injectable naltrexone (N = 65)
No medication (N = 74)
Demographics
Age (SD)
30.98 (11.6)
29.63 (8.80)
27.89 (8.04)
31.68 (11.44)
Gender (% male)
59
74
75
67
Baseline clinical characteristics
Diagnoses
No. mental health (SD)
1.18a (1.09)
1.30b (1.32)
0.84 (1.08)
0.78a,b (0.94)
No. SUD (SD)
1.66 (1.20)
1.93 (1.24)
1.68 (1.06)
1.45 (1.22)
% alcohol use disorder
44
56
60
42
% cannabis use disorder
42
44
49
38
% cocaine use disorder
16
22
17
14
% amphetamine use disorder
30
33
14
22
% depressive disorder
39
48
28
25
% anxiety disorder
51
55
38
33
No. of past alcohol/drug treatments (SD)
0.93 (1.16)
0.88 (1.12)
1.03 (1.15)
0.58 (0.94)
Opioid drug of choice
% Heroin
53
37
78
51
% Oxycodone
27
44
14
26
DDQ scores (craving)
Desire/intention to drug use scale (SD)
3.03a,b (1.40)
2.61 (1.34)
2.32a (1.02)
2.25b (1.25)
Negative reinforcement scale (SD)
4.00a,b (1.78)
3.51 (1.81)
3.36a (1.89)
3.23b (1.68)
Control over use scale (SD)
6.02 (1.47)
6.06 (1.49)
6.36 (1.12)
5.81 (1.75)
Craving scores
Treatment admission (SD)
8.39a,b,c (2.47)
6.28a (3.22)
6.65b,d (2.90)
5.51c,d (3.43)
Baseline (SD)
4.22a,b,c (2.47)
3.08a (2.27)
2.92b (2.02)
2.69c (2.07)
Change – admission to baseline (SD)
−4.17 (3.12)
−3.20 (4.50)
−3.72 (3.12)
−2.82 (3.41)
Opioid withdrawal – SOWS (SD)
8.53a,b (6.58)
6.88c (4.95)
4.23a,c (3.96)
5.44b (5.78)
Treatment engagement
Length of stay in residential program (SD)
42.94 (28.93)
37.93 (22.53)
42.10 (25.94)
36.91 (22.47)
% completed residential
93
89
97
95
% stepped down to another program within the same system of care
74a
67b
88a,b,c
64c
Note. Groups with the same superscript letter next to thema are statistically different at p < .05.
Additional analyses compared medication condition to the three subscale scores for the DDQ. Buprenorphine-naloxone patients had significantly higher scores on the desire and intention to drug use scale than injectable naltrexone patients and no medication patients, F(3,249) = 6.83, p < .001. The same result emerged for the negative reinforcement scale, F(3,242) = 2.86, p < .05, with buprenorphine-naloxone patients reporting significantly higher scores than injectable naltrexone patients and no medication patients. No differences between groups were observed for the control over use scale, F(3,242) = 1.52, p = .21.
Ratings on the single-item opioid craving scale varied by medication condition at admission to treatment and at the time of the baseline interview. Patients who ultimately chose to maintain on buprenorphine-naloxone reported significantly higher levels of craving than both naltrexone groups and the no medication group at treatment admission, F(3,241) = 13.14, p < .001, and at baseline, F(3,241) = 7.38, p < .001. An analysis of the change in craving from treatment admission to baseline interview showed no significant differences between med conditions, F(3,241) = 25.59, p = .08. Mean ratings are shown in Table 2.
Further differences were observed for baseline opioid withdrawal symptoms, F(3,242) = 8.24, p < .001. Patients choosing to maintain on buprenorphine-naloxone had a significantly higher SOWS score than injectable naltrexone and no medication patients, and those choosing to maintain on oral naltrexone had significantly higher SOWS scores relative to those who chose injectable naltrexone.
3.5 Engagement with treatment services
The four groups of patients did not differ in the number of days they attended the residential program, F(3,249) = 0.91, p = .44, but did differ regarding the percentage who stepped down to another program in the same system of care, χ2(3) = 10.57, p < .05. Injectable naltrexone patients were significantly more likely to step down to another treatment program within Hazelden Betty Ford (88%) than patients taking buprenorphine-naloxone (74%), patients taking oral naltrexone (67%), and patients not taking medications (64%). The four groups did not differ in discharge status, χ2(3) = 2.40, p = .49, indicating no differences by medication condition in the successful completion of residential treatment.
3.6 One-month substance use outcomes
One-hundred ninety patients (73% of the sample) were successfully contacted by research staff for the 1 month outcomes survey, and medication condition was related to 1-month survey follow-up rates, χ2(3) = 18.44, p < .001. Injectable naltrexone patients had a higher follow-up rate (92%) than the buprenorphine-naloxone patients (76%), the oral naltrexone patients (67%), and the no medication patients (62% p < .001). A series of analyses comparing contacted versus noncontacted patients revealed no significant differences in demographic and baseline clinical characteristics.
Among patients who completed a 1-month survey and who left residential taking either form of naltrexone or buprenorphine-naloxone, 118 (81%) reported compliance with their medication during the follow-up period, defined as continuing to take an oral medication or having received another injection since the time of discharge. Eighty-seven percent of buprenorphine-naloxone patients reported compliance, 65% of oral naltrexone patients reported compliance, and 80% of injectable naltrexone patients reported compliance, a difference that was nonsignificant, χ2(3) = 2.99, p = .39.
Based on answers to the alcohol and drug use questions, we determined whether each patient was continuously abstinent from all alcohol, illicit drugs, and/or misused prescription drugs since leaving residential treatment. A binomial logistic regression was used to ascertain the effect of medication condition/compliance (buprenorphine-naloxone/complaint, buprenorphine-naloxone/noncompliant, oral naltrexone/compliant, oral naltrexone/noncompliant, injectable naltrexone/compliant, injectable naltrexone/noncompliant, and no medication) on abstinence status (abstinent vs relapsed). Continuous abstinence rates at 1 month were significantly related to medication condition/compliance, F(12) = 30.06, p < .01 (see Fig. 1). The model explained 27.8% (Nagelkerke R2) of the variance in continuous abstinence at 1-month follow-up. Buprenorphine-naloxone/compliant patients were significantly more likely than buprenorphine-naloxone/noncompliant patients (B = 2.44, p < .01) and oral naltrexone/noncompliant patients to be abstinent at 1 month (B = 2.85, p < .01). Oral naltrexone/compliant patients were more likely to be abstinent than oral naltrexone/noncompliant patients (B = 3.20, p < .05) and buprenorphine-naloxone/noncompliant patients (B = 2.80, p < .05). Injectable naltrexone/compliant patients were more likely than buprenorphine-naloxone/noncompliant patients (B = 3.38, p < .01) and oral naltrexone/noncompliant patients (B = 3.79, p < .01) to be abstinent at 1-month follow-up. Finally, patients in the no medications group were significantly more likely to be continuously abstinent at 1 month than the oral naltrexone/noncompliant group (B = 2.19, p < .05). Of the covariates entered into the model, the desire and intention to use subscale of the DDQ (B = −0.77, p < .05) and total SOWS score (B = 0.17, p < .01) at baseline were significant. No other covariates or differences between medication condition/compliance groups were significant.
Fig. 1One month continuous abstinence as a function of medication condition and drug compliance (N = 186).
Note. Contrasts with superscripta–f are significantly different at p < .05. Contrasts with superscriptg are significantly different at p < .01. Buprenorphine-naloxone/compliant v. buprenorphine-naloxone/noncomplianta. Buprenorphine-naloxone/compliant v. oral naltrexone/noncompliantb. Injectable naltrexone/compliant v. buprenorphine-naloxone/noncompliantc. Injectable naltrexone/compliant v. oral naltrexone/noncompliantd. No medications v. oral naltrexone/noncompliante. Oral naltrexone compliant v. buprenorphine-naloxone.noncompliantf. Oral naltrexone/compliant v. oral naltrexone/noncompliantg. Continuous abstinence is based on subjects who completed the 1-month follow-up survey and are therefore not intent-to-treat. Error bars represent 95% confidence interval.
We were not able to conduct analyses of alcohol and drug use days during the 1-month follow-up period due to the highly skewed nature of these variables. The majority of participants had not relapsed at 1 month, and when use days were examined as a function of medication condition/compliance group, the buprenorphine-naloxone/compliant group had no variance (no one in that group had relapsed). All alcohol and drug use days reported during the 1-month follow-up period were taken into account for the 6-month outcomes analyses for substance use days, reported below.
3.7 One-month secondary outcomes
A one-way ANCOVA was run to determine the effect of medication condition/compliance on craving rating of patients' opioid drug of choice at 1-month follow-up. Craving ratings for the same scale taken at the baseline session as well as ratings on the three DDQ subscales were entered into the model as covariates (see Table 3). Craving ratings at 1-month post-residential were not significantly related to medication condition/compliance, F(6,170) = 1.66, p = .13.
Table 31- and 6-month outcomes as a function of medication condition and medication compliance.
1 month outcomes
6 month outcomes
# 12-step meetings
craving score
# 12-step meetings
craving score
Medication condition
Buprenorphine-Naloxone/compliant
16.35 (13.09)
3.02 (1.94)
65.77 (43.16)a
2.24 (1.48)
Buprenorphine-naloxone/noncompliant
10.88 (8.11)
3.89 (2.26)
65.07 (40.44)
3.07 (2.87)
Oral naltrexone/compliant
14.55 (18.48)
2.56 (2.30)
69.10 (25.06)
1.67 (1.15)
Oral naltrexone/noncompliant
13.17 (13.14)
2.00 (0.89)
46.00 (33.41)b
3.75 (3.49)
Injectable naltrexone/compliant
19.66 (11.34)
2.02 (1.24)
104.10 (77.29)a,b
1.53 (0.61)
Injectable naltrexone/noncompliant
24.33 (13.61)
2.63 (2.06)
47.00 (34.87)
2.76 (2.22)
No medication
16.00 (11.25)
2.11 (1.79)
72.18 (49.65)
2.03 (1.69)
Note: groups with the same superscript letter next to thema are statistically different at P < .05. Percent continuously abstinent is based on subjects who completed the 1- or 6-month follow-up survey and are therefore not intent-to-treat.
) are a foundation of Hazelden Betty Ford's treatment model and regular attendance at meetings is recommended to each patient as part of continuing care after residential treatment. A one-way ANCOVA examined the total number of meetings patients reported attending during the 1-month follow-up period as a function of medication condition/compliance, controlling for the significant covariates at baseline. The test was not significant, F(6,174) = 1.15, p = .34, indicating no group differences in meeting attendance (see Table 3).
3.8 Six-month substance use outcomes
One hundred forty-nine patients (57% of the sample) were successfully contacted by research staff for the 6-month outcomes survey. A series of analyses comparing contacted versus noncontacted patients revealed some differences in demographic and baseline clinical characteristics. Patients who were successfully contacted at 6 months were more likely than noncontacted patients to be older (M = 31.40 vs. 28.65; p < .05), have an alcohol use disorder (56% vs. 37%; p < .01), have a sedative-hypnotic use disorder (34% vs. 22%; p < .05), and have a higher number of co-occurring SUDs (1.76 vs. 1.38; p < .01). There were no significant differences in 6-month follow-up rates as a function of medication condition/compliance, (χ2(6) = 2.22, p = .89).
Similar to the 1-month analyses, we identified which patients in the medication groups reported complying with their medication during the entire 6-month follow-up period. Among patients who left residential treatment taking either form of naltrexone or buprenorphine-naloxone and who completed a 6-month survey, 66 (59%) reported compliance with their medications. Seventy-two percent of buprenorphine-naloxone patients reported compliance, 29% of oral naltrexone patients reported compliance, and 53% of injectable naltrexone users reported compliance, a significant difference (buprenorphine-naloxone vs. oral naltrexone compliance; χ2(2) = 9.83, p < .01).
Similar to 1-month results, medication condition/compliance continued to be significantly related to continuous abstinence from drugs and alcohol, F(12) = 32.54, p < .01 (see Fig. 2). The model explained 28% (Nagelkerke R2) of the variance in continuous abstinence at 6-month follow-up. Oral naltrexone/compliant patients were significantly more likely than oral naltrexone/noncompliant patients to be continuously abstinent at 6 months (B = 3.30, p < .05). Injectable naltrexone/compliant patients were significantly more likely to be abstinent at 6 months than buprenorphine-naloxone/compliant (B = 1.44, p < .05), buprenorphine-naloxone/noncompliant (B = 1.75, p < .05), and oral naltrexone/noncompliant (B = 2.59, p < .01) patients. Lastly, patients in the no medication group were more likely than oral naltrexone/noncompliant patients to be abstinent at 6 months (B = 2.01, p < .05). Baseline total number of mental health diagnoses (B = −0.44, p < .05), baseline DDQ desire and intention to use subscale (B = −0.68, p < .05), and craving rated at residential treatment admission (B = 0.17, p < .05) were significant covariates in the model.
Fig. 2Six month continuous abstinence as a function of medication condition and drug compliance (N = 144).
Note. Contrasts with superscripta–d are significantly different at p < .05. Contrasts with superscripte are significantly different at p < .01. Injectable naltrexone/compliant v. buprenorphine-naloxone/complianta. Injectable naltrexone/compliant v. buprenorphine-naloxone/noncompliantb. Oral naltrexone/compliant v. oral naltrexone/noncompliantc. No medications v. oral naltrexone/noncompliantd. Injectable naltrexone/compliant v. oral naltrexone/noncompliante. Continuous abstinence is based on subjects who completed the 6-month follow-up survey and are therefore not intent-to-treat. Error bars represent 95% confidence interval.
To examine whether medication compliance was related to alcohol use during the 6 months following residential treatment, we conducted a zero-inflated negative binomial (see Table 4 for full results). The table includes odd ratios and percent change in number of alcohol days for a 1-unit increase in each predictor for ease of interpretation. The zero-inflated model, which predicts the likelihood for an individual to remain abstinent from alcohol through the 6-month follow-up, showed that medication condition/compliance was not a significant predictor of the absence of alcohol days, nor were any of the covariates entered into the model. Similarly, medication condition/compliance was not a significant predictor in the count model, where variables were assessed to determine if they predicted change in the frequency of alcohol days through the 6-month follow-up. Of the covariates entered into the model, only craving at treatment admission was significantly related to a change in the frequency of alcohol days: among patients who used alcohol prior to the 6-month follow-up, a one-unit increase in craving for opioids as rated at treatment admission increased alcohol use days by 33.64%.
Table 4Estimates of zero-inflated negative binomial model to predict absence of alcohol use days and number of alcohol use days at 6-month follow-up (N = 143).
To examine whether medication condition/compliance was associated with drug use at 6 months, we conducted a second zero-inflated negative binomial (see Table 5 for full results). In the zero-inflated model predicting the likelihood for an individual to remain abstinent from drugs through the 6-month follow-up, being in the oral naltrexone/noncompliant condition was the only medication condition/compliance group that influenced the odds of remaining abstinent. Holding other factors constant, if a patient was noncompliant with their oral naltrexone prescription, the likelihood of staying abstinent from drugs through 6-month follow-up decreased by 87.63%. Otherwise, medication condition/compliance and the additional covariates were not a significant predictor of the absence of drug use days. In the count model where variables were assessed to determine if they predicted change in the frequency of drug use days through the 6-month follow-up, none of the medication condition/compliance variables or covariates entered into the model were found to be significant predictors of change in frequency of drug use days.
Table 5Estimates of zero-inflated negative binomial model to predict absence of drug use days and number of drug use days at 6-month follow-up (N = 143).
The single-item craving question also appeared on the 6-month survey. A one-way ANCOVA with medication condition/compliance as the independent variable and baseline craving variables entered as covariates did not reach significance, F(6,107) = 1.61, p = .15, indicating that medication condition/compliance was not significantly related to craving score (see Table 3). The total number of 12-step meetings attended between the 1- and 6-month follow-up differed as a function of medication condition/compliance, F(6,131) = 3.13, p < .01. Injectable naltrexone/compliant patients reported attending a significantly higher number of meetings than buprenorphine-naloxone/compliant patients (mean difference of 47.83 (95% CI, 6.24 to 89.42) meetings, p < .05) and oral naltrexone/noncompliant patients (mean difference of 64.48 (95% CI, 7.95 to 120.99) meetings, p < .05). None of the baseline covariates entered into the model were significant (see Table 3).
4. Discussion
This study examined the feasibility of using medications as part of a 12-step based treatment model for patients with OUD. Given the potential role of self-selection into a 12-step based treatment program, the degree to which patients would engage in MAT was unknown. Of 253 patients, 71% were successfully inducted on a medication during their residential treatment episode. Nearly all patients completed residential treatment and 73% attended at least one other program in the same system of care after residential. These results suggest that the use of medications, including buprenorphine-naloxone, within a 12-step abstinence-based treatment program is feasible and associated with a high level of patient engagement with psychosocial care.
The high retention observed here may be accounted for by factors independent of the MAT component. Study participants entered treatment at a private facility with a network of continuing care options including day treatment, intensive outpatient, outpatient, and sober housing, among others. As such, clinical staff members worked to transfer and engage patients in the next level of care as seamlessly as possible. Treatment settings with fewer resources may not see the same level of retention and outcome.
reviewed 55 studies and found retention rates in MAT generally ranged from 19 to 95% at 3 months, reflecting a high degree of variability among settings. While the present study lacked a randomized control condition, the residential facility in which it took place has seen substantial improvements in OUD patient retention since implementing the COR-12 program, with atypical discharge rates (e.g., leaving against staff advice) dropping from 22% to 6% in its residential program (
Consistent with clinical trials, medication compliance was a critical factor among this treatment sample. Overall, patients who were compliant with their medications were more likely to remain continuously abstinent at 1- and 6-months post-discharge regardless of which medication they chose. Different levels of drug compliance arose between the three medication groups: oral naltrexone patients were less likely to be compliant with their medication than buprenorphine-naloxone patients at six months. Given that previous studies have shown that substance use outcomes with naltrexone are strongly related to medication compliance (
), it is critical that clinical staff develop strategies for enhancing compliance if oral naltrexone is to continue being prescribed.
The only significant contrast that emerged between the three compliant drug conditions was that injectable naltrexone patients who took their medication as prescribed were more likely to be continuously abstinent at 6 months than buprenorphine-naloxone patients who were similarly compliant. This suggests injectable naltrexone may have the potential to be more efficacious than buprenorphine-naloxone in preventing relapse to drugs or alcohol in patients with OUD. However, this finding should be replicated in a larger trial with biochemical confirmation of drug compliance and abstinence in order to determine if there is a real difference in efficacy.
For the purposes of this study, compliance was determined based on a single item that asked a patient if they took their anti-craving medication as prescribed. It is much easier for an individual to answer such a question in the affirmative if they know they did not miss one of a small number of injections. In contrast, there is the potential for much more variation in an individual's understanding of taking an oral medication “as prescribed,” which could be taken to mean not missing a single dose, or only missing a few here and there, a difference which has been shown to influence the overall effectiveness of the medication. For example,
found that patients who took their buprenorphine fewer than 80% of days in a 3-month period were 10 times as likely to return to opioid use than compliant patients. Further research is needed to determine if this finding is an artifact of the way compliance was measured in the current study, or if it represents a true difference in drug efficacy.
Nearly three quarters of buprenorphine-naloxone patients reported compliance with their medication during the 6 months post-discharge from residential treatment. This rate is higher than that observed in other studies. For example, a study reporting analyses of a large, private United States insurance claims database found that 64% of patients prescribed sublingual film buprenorphine-naloxone and 58% of patients prescribed the oral (pill) form were still taking the medication at 6 months (
Persistence and healthcare utilization associated with the use of buprenorphine/naloxone film and tablet formulation therapy in adults with opioid dependence.
). Participants in the present study received ongoing individual and group therapy, which likely helped with buprenorphine-naloxone compliance.
Ninety-one percent of patients who were compliant with their buprenorphine-naloxone were abstinent at 1-month follow-up, but this dropped to 59% by 6 months post-discharge. Buprenorphine-naloxone compliant patients did not have fewer drug use days at 6-month follow-up than the no medication group (or any other groups). It is unclear why buprenorphine-naloxone did not have an impact on drug use in the present study. Follow-up rates are one potential factor. Another is the amount of psychosocial treatment completed among the sample, which may have played a role in bolstering outcomes among all groups. In addition, patients in the buprenorphine-naloxone condition had the highest levels of opioid craving both early in the residential treatment episode and at the 1-month follow-up, suggesting that this group struggled with craving more than participants taking naltrexone or no medications.
Challenges with medication compliance may stem from a number of factors, including altered activity in the medial prefrontal cortex from prolonged opioid use (
Patients' beliefs about medications are associated with stated preferences for methadone, buprenorphine, naltrexone, or no medication-assisted therapy following inpatient opioid detoxification.
). Future research is needed to examine how clinical and medical staff can work with patients to improve medication compliance both during and after OUD treatment.
Medication compliance was not a significant predictor of the absence of alcohol use days, and the only significant finding for the absence of drug use days was for the oral naltrexone/noncompliant group. For this group, medication noncompliance was found to be a significant predictor of the absence of drug use days, decreasing the likelihood that a patient would remain abstinent from drugs at 6-month follow-up. This is not surprising given the results from the outcome analyses of continuous abstinence, where patients in the oral naltrexone/noncompliant group had significantly worse continuous abstinence rates compared to the other medication/compliance conditions. In the present study, patients who were interested in using naltrexone were encouraged to use the injectable form of the drug and many patients only opted to take the oral form when insurance would not cover the cost of the injections. Future studies should endeavor to recruit a large enough sample of oral naltrexone users in order to compare outcomes of patients who chose that specific mode of delivery against those who took it when their first choice was not available.
Of note, patients who opted not to take a medication were more likely to report continuous abstinence than patients in the oral naltrexone/noncompliant group at both 1- and 6-months and had continuous abstinence rates that were not significantly different from any of the other compliant or noncompliant medication conditions. This suggests that perhaps not all patients with OUD need a craving medication to ensure success, and is a powerful reminder of the impact of patient choice on outcomes.
Overall, independent of medication condition, the outcomes for this sample of OUD patients are positive and suggest that the COR-12 program holds promise as a treatment for this population. However, some limitations should be noted. First, while some studies have verified opioid and other drug use with urine drug screens, our study relied on self-reports of substance use. Future studies should use drug screens to confirm abstinence when possible. Another limitation of this study is that over 40% of patients were not successfully reached for a 6-month outcomes survey. This is consistent with other studies of OUD patients receiving medications, some of which reported higher attrition rates than those found here. For example,
found attrition rates as high as 74% for some study conditions (e.g., people in the placebo condition). Still, caution must be exercised in interpreting the 1- and 6-month outcomes data because patients who are contacted at follow-up may be more likely than noncontacted patients to be doing well, which could potentially overinflate outcomes data. This possibility seems fairly unlikely, however, given that patients contacted at 6 months tended to be more clinically severe (i.e., more likely to have additional SUD diagnoses) than noncontacted patients.
Next, the number of patients in the oral naltrexone group was small to begin with (11% of the total sample), and made smaller at follow-up due to attrition. The small sample sizes in the oral naltrexone complaint and noncompliant groups led to large error terms, particularly in the 6-month continuous abstinence analysis, which reflect a higher degree of uncertainty of the significance of the observed point estimates for these groups (
). As a result, the significant contrasts involving oral naltrexone should be interpreted with caution. Future studies should endeavor to employ much larger samples to better power statistical analyses of 6-month outcomes data.
While we went to great lengths to ensure honest responses from patients at follow-up (e.g., emphasizing confidentiality of survey responses, remaining neutral during alcohol and drug use disclosures), having Hazelden Betty Ford staff perform the follow-up surveys instead of an unaffiliated firm could have led some participants to provide socially desirable responses, including under-reporting their alcohol or drug use or over-reporting their drug compliance.
Another caveat pertains to the generalizability of these results based on the characteristics of the sample: the majority of patients were under 30 years of age and nearly all were Caucasian. While socioeconomic status was not assessed in the current study, a prior published study on young adults within the same treatment system found the patients to be representative of the middle-class population and largely dependent upon insurance coverage for care (
). Because the sample is not typical of the broader population of individuals who have OUD, the results may not generalize to other individuals receiving medication-assisted treatment for OUD. At the same time, this paper adds useful knowledge to the field regarding the impact of MAT for an understudied segment of people with OUD, namely Caucasian Millennial heroin and oxycodone users having multiple co-occurring SUD and mental health diagnoses.
A final caveat is that this was a naturalistic observational study. It reflects how medications were prescribed by a group of providers in a 12-step residential addiction treatment setting. Provider bias combined with a lack of standards in the field for making medication decisions for the treatment of OUD may have played a role in the study results. The results reflect general clinical practice and are likely not as strong as those that would have been generated via a randomized controlled trial, where participants were randomly assigned to a medication condition. In addition, because we did not use random assignment, there were baseline differences between patients who took naltrexone, buprenorphine-naloxone, and opted not to take any medications. Buprenorphine-naloxone and oral naltrexone patients were more severe at baseline than the other groups on a number of characteristics, including having more co-occurring mental health diagnoses, very strong drug cravings and intention/desire to use opioids, and more pronounced opioid withdrawal early in treatment. Though we addressed these differences in our statistical analysis, randomization would have eliminated them.
The challenges of implementing MAT in a 12-step, abstinence-based treatment program were considerable, but were less problematic than expected by the implementation team, which included both authors. Hazelden Betty Ford Foundation is a non-profit and board approval was required to pursue such a “controversial” alteration of treatment protocols. The president/CEO was supportive, as was the board chair, which greatly facilitated the effort. The Foundation had been using MAT for alcohol use disorders, so a precedent had been set, but not for maintenance medication for OUD, which was the primary concern. We established a multidisciplinary team to review the literature, wrote a white paper outlining our findings and our plans for the organization, and then took a full year to establish a detailed plan and to train staff for implementation. Staff attitudes and resistance were not measured but were recognized as differing throughout this multi-state organization. When asked to be on the implementation team, one counselor asked, “Is this a good career move?” In addition, one author (MDS) received a four-page, single-spaced email outlining why the Foundation should not use medication for OUD. Many explanatory conversations took place and a small number of clinical staff left the organization. Public criticism from other 12-step, abstinence-based treatment programs was much more negative than our internal experience.
Staff training sessions revealed much more support than dissent. The passion of our clinical staff for improving outcomes and preventing patient death was paramount; data and information appeared to overcome resistance. Knowing counselor bias can undermine medication compliance, we provided a great deal of training and monitored patterns of low engagement and medication adherence, using further training and corrective action when issues arose. Medical provider bias may have played a role in medication decisions, which is consistent with a naturalistic observational study. Prescribing buprenorphine-naloxone more frequently for patients with higher craving, higher withdrawal symptoms and more co-occurring mental health diagnoses as well as prescribing naltrexone more frequently for those with co-occurring alcohol use disorders has strong clinical justification. The COR-12 program has evolved and is now an integral, essential aspect of Hazelden Betty Ford's treatment for OUD and clinical staff compete for open positions within this program.
In summary, the findings reported here demonstrate the feasibility of including MAT within a 12-step based professionally led treatment program. Despite concerns about philosophical disagreements or provider bias, the majority of patients elected to begin taking either buprenorphine-naloxone or naltrexone with the intention of maintaining on the medications after discharge. Some interesting differences emerged among medication groups, suggesting patients with higher craving and more severe withdrawal symptoms may be more likely to elect buprenorphine-naloxone maintenance over naltrexone options or no medications at all. These findings are important given the context within which MAT studies have traditionally been conducted, which has been primary care settings for the majority of studies (
). Furthermore, and for sound research purposes, patients with greater clinical severity (e.g., suicidality, co-occurring disorders, physical health conditions) are often excluded from clinical trials (
). The naturalistic and observational study reported here demonstrates that MAT use is feasible with clinically severe patients in 12-step residential care and may have a positive impact on initial outcomes.
Funding sources
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Declaration of Competing Interest
None.
Acknowledgements
The authors wish to acknowledge Kate Gliske, PhD for her significant role in reconfiguring statistical data and other contributions to this manuscript.
References
Bentzley B.S.
Barth K.S.
Back S.E.
Book S.W.
Discontinuation of buprenorphine maintenance therapy: Perspectives and outcomes.
Persistence and healthcare utilization associated with the use of buprenorphine/naloxone film and tablet formulation therapy in adults with opioid dependence.
Inference in regression models of heavily skewed alcohol use data: A comparison of ordinary least squares, generalized linear models, and bootstrap resampling.
Substance Abuse and Mental Health Services Administration
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).
Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration,
Rockville, MD2017 (Retrieved from)
Effectiveness of injectable extended-release naltrexone vs. daily buprenorphine-naloxone for opioid dependence: A randomized clinical noninferiority trial.
Patients' beliefs about medications are associated with stated preferences for methadone, buprenorphine, naltrexone, or no medication-assisted therapy following inpatient opioid detoxification.
p < .01.
30055-2&id=gr1.jpg){kind=link}
30055-2&id=gr2.jpg){kind=link}