The American Journal on Addictions, 24: 265–270, 2015 Copyright © American Academy of Addiction Psychiatry ISSN: 1055-0496 print / 1521-0391 online DOI: 10.1111/ajad12182
Effects of Injectable Extended-Release Naltrexone (XR-NTX) for Opioid Dependence on Residential Rehabilitation Outcomes and Early Follow-up Douglas L. Leslie, PhD,1,2 William Milchak, LCSW,2 David R. Gastfriend, MD,3† Philip L. Herschman, PhD,4 Edward O. Bixler, PhD,2 Diana L. Velott, MPA, MS,1 Roger E. Meyer, MD2 1
Department of Public Health Sciences, Center for Applied Studies in Health Economics, Penn State College of Medicine, Hershey, Pennsylvania 2 Department of Psychiatry, Penn State College of Medicine, Hershey, Pennsylvania 3 Treatment Research Institute, Philadelphia, PA 4 CRC Health Group, Inc., Cupertino, California
Background and Objectives: Little is known about the use of extended-release naltrexone (XR-NTX) during residential rehabilitation, and its effects on early outcomes and rates of follow-up treatment. This study examined patient characteristics and rates of treatment completion and engagement in post-residential care of opioid dependent patients who received XR-NTX during residential rehabilitation, compared with patients who did not receive this medication. Methods: Electronic records for opioid dependent patients from three Pennsylvania residential detoxification and treatment facilities (N ¼ 7,687) were retrospectively analyzed. We determined the proportion of patients who received XR-NTX (INJ), and compared rates of treatment completion and engagement in follow-up care relative to a naturalistic control group of patients recommended for, but not administered, XR-NTX (Non-INJ). Data on whether the patient initiated follow-up care were available from one site (N ¼ 3,724). Results: Overall, 598 (7.8%) patients were recommended for XR-NTX and of these, 168 (28.1%) received injections. Compared to non-INJ patients, INJ patients were less likely to leave against medical advice (4.8% vs. 30.2%, p < .001) and more likely to initiate follow-up care (37.7% vs. 19.7%, p < .001). These differences remained significant after controlling for demographic covariates using regression analysis. Conclusions: XR-NTX was associated with higher rates of residential and early post-residential care engagement in patients with opioid dependence.
Received March 19, 2014; revised September 16, 2014; accepted October 31, 2014. † Dr. D. Gastfriend was an employee of Alkermes, Inc., Waltham, Massachusetts, at the time that this research was conducted. He is currently Chief Executive Officer of the Treatment Research Institute, Philadelphia, Pennsylvania. Address correspondence to Douglas L. Leslie, Penn State College of Medicine 90 Hope Drive, Hershey, PA 17033. E-mail: [email protected]
Scientific Significance: XR-NTX may be an effective adjunct in the residential treatment and aftercare of patients with opioid dependence. (Am J Addict 2015;24:265–270)
INTRODUCTION Opioid dependence causes great personal and societal costs.1 Successful treatment is associated with abstinence, reduced criminal activity, and an overall decrease in healthcare costs.2–13 Patient retention in treatment is a critical predictor of success and can be greater in medication-assisted programs.14 Approved medications fall into three categories: agonist (methadone), partial agonist (buprenorphine), or antagonist (naltrexone). In 1976, the National Institute on Drug Abuse postulated that an injectable formulation of extended-release naltrexone (XR-NTX) would improve the efficacy of narcotic antagonist treatment by enhancing patient adherence.15 By blocking the pharmacological effects of narcotics for a month, it might also promote engagement and enhance the efficacy of non-pharmacological aspects of treatment.16 1 XR-NTX (VIVITROL , Alkermes, Inc.) was approved by the Food and Drug Administration in 2006 for the treatment of alcohol dependence, and in 2010 for the treatment of opioid dependence. Its use in opioid dependence requires opioid detoxification, which makes inpatient or residential care a convenient setting for initiating treatment. Although the safety and effectiveness of XR-NTX has been established in research settings,17,18 there is almost no available information about the use of XR-NTX for opioid addiction in real-world community residential treatment settings and the interaction with patient outcomes. We conducted a retrospective study of XR-NTX 265
use among opioid dependent patients in residential rehabilitation programs. The objectives of the study were: (1) to determine the proportion of patients who were recommended for this treatment; and (2) to compare the early stage outcomes of patients who were injected with XR-NTX (INJ) with those who were recommended for XR-NTX but not injected (non-INJ).
METHODS Cohort Data for the study came from CRC Health Group, Inc., a large private network of specialized behavioral healthcare services in the U.S. Retrospective data that had been abstracted from CRC’s electronic health records were obtained for 11,146 patients attending one of three residential rehabilitation facilities in Pennsylvania during 2011, of whom 7,687 had a primary or secondary diagnosis of opioid dependence. Because not all opioid dependent patients would be good candidates for XR-NTX treatment, we limited our comparisons of early stage outcomes to patients who were identified in the data by CRC as having been recommended by clinical staff to receive XR-NTX and participated in an educational program about the drug. Patients who received XR-NTX (N ¼ 168) were identified, along with a comparison group of patients who were recommended for but failed to receive XR-NTX (N ¼ 430). No data were available that would identify the reasons why the largest group of patients in the cohort (N ¼ 7,089) were not considered for XR-NTX treatment (non-REC). Quality improvement staff at one site (Site 1) systematically made routine follow-up telephone calls to each of the follow-up facilities 7–10 days post-discharge to determine whether the patients referred by this residential facility attended their follow-up care. Data on the first followup visit were not routinely collected at the other two sites. Since the vast majority of patients were referred to non-CRC owned sites, HIPAA regulations prevented us from obtaining additional outcome information beyond the initial follow-up visit. All data were de-identified before being shared with researchers. The study was approved by the Institutional Review Board of the Penn State College of Medicine. Data Analyses The count and proportion of patients that were discharged prematurely against medical advice (AMA) was determined. In addition, for the facility with systematic follow-up data (Site 1 only), the proportion of patients who attended the postdischarge ambulatory level of care was determined. Chi-square tests were used to determine whether differences between the treatment groups were statistically significant using a two-tailed test with p < .05 as the threshold and were adjusted for multiple comparisons using Tukey’s test. Next, multivariate logistic regression models were developed to identify patient demographic and clinical 266
characteristics associated with receiving XR-NTX and the outcomes among those patients for whom XR-NTX was recommended but not administered. Three separate models were estimated, predicting receipt of XR-NTX, leaving AMA and, among patients from site 1 only, whether the patient attended their first post-discharge visit. Independent variables included medication group (INJ vs. non-INJ), site, age, gender, race/ethnicity, primary substance, type of insurance and referral source. In addition, interactions between medication group and each of the other independent variables were tested to examine patient characteristics that might predict outcomes. Mental health co-morbidities were not reliably available in the EHR data and were excluded from the model.
RESULTS Group Characteristics Of the 598 patients who were recommended for XR-NTX (8% of the total sample of patients with opioid dependence), 168 patients (28.1%) ultimately received the injection and 430 (71.9%) were not injected. Various reasons were offered for the latter group’s failure to receive their first injection including: the patient changed her/his mind (N ¼ 136; 31.6%); the patient was discharged before the medication arrived (N ¼ 121; 28.1%); inability to pay for the medication, usually due to insurance denial (N ¼ 89; 20.7%); the patient left against medical advice before being injected (N ¼ 67; 15.6%); or for unspecified reasons (N ¼ 17; 4%). Demographic characteristics were similar between the groups (Table 1). Although statistically significant, age differences were small, with the INJ patients slightly younger than the non-INJ and non-REC groups. The distribution of gender was similar in the INJ and non-INJ groups, while those in the non-REC group were more likely to be male. Most patients in the sample were Caucasian. Differences in the proportions of patients referred to treatment by State, County or Social Services were statistically significant, and were higher in the non-REC group (21.4%) than in either the INJ group (13.7%) or the non-INJ group (11.4%). The non-REC group was also less likely to have been referred by alumni (4.3%) than those in the non-INJ group (7.9%) or the INJ group (5.4%). INJ patients were much less likely to have commercial insurance (13.1%) than both the non-INJ patients (26.5%) and non-REC patients (22.4%), and conversely, were more likely to be covered by Medicaid (84.5%) than the non-INJ (71.2%) and non-REC (73.2%) groups. There were no group differences in primary substance dependence across the three groups. Rate of AMA discharge and length of stay are presented in Table 2. INJ patients were significantly less likely to leave AMA than non-INJ patients (4.8% vs. 30.2%, p < .001) or non-REC patients (4.8% vs. 30.2%, p < .001). On average, INJ patients stayed in rehab a week longer than both non-INJ and non-REC patients (p < .001 for both comparisons).
Injectable XR-NTX for Opioid Dependence in Rehab
April 2015
TABLE 1. Group characteristics
Injected N ¼ 168 Characteristics
N
Age (Mean SD) Gender Male Female Race White Black Hispanic Other Referral source Self State/county/social services CRC facility Alumni Drug alcohol provider Family Hospital/IP facility Insurance company Legal system OP facility Other Primary payer type Commercial insurance Medicaid Special payer Missing Primary substance Alcohol Opiates Cannabis Benzodiazepines Other
Not injected N ¼ 430
%
N
29.6 8.2
Not recommended N ¼ 7,089
%
N
28.3 8.2
p-values*
% 29.7 9.1
.006
Effects of Injectable Extended-Release Naltrexone (XR-NTX) for Opioid Dependence on Residential Rehabilitation Outcomes and Early Follow-up Douglas L. Leslie, PhD,1,2 William Milchak, LCSW,2 David R. Gastfriend, MD,3† Philip L. Herschman, PhD,4 Edward O. Bixler, PhD,2 Diana L. Velott, MPA, MS,1 Roger E. Meyer, MD2 1
Department of Public Health Sciences, Center for Applied Studies in Health Economics, Penn State College of Medicine, Hershey, Pennsylvania 2 Department of Psychiatry, Penn State College of Medicine, Hershey, Pennsylvania 3 Treatment Research Institute, Philadelphia, PA 4 CRC Health Group, Inc., Cupertino, California
Background and Objectives: Little is known about the use of extended-release naltrexone (XR-NTX) during residential rehabilitation, and its effects on early outcomes and rates of follow-up treatment. This study examined patient characteristics and rates of treatment completion and engagement in post-residential care of opioid dependent patients who received XR-NTX during residential rehabilitation, compared with patients who did not receive this medication. Methods: Electronic records for opioid dependent patients from three Pennsylvania residential detoxification and treatment facilities (N ¼ 7,687) were retrospectively analyzed. We determined the proportion of patients who received XR-NTX (INJ), and compared rates of treatment completion and engagement in follow-up care relative to a naturalistic control group of patients recommended for, but not administered, XR-NTX (Non-INJ). Data on whether the patient initiated follow-up care were available from one site (N ¼ 3,724). Results: Overall, 598 (7.8%) patients were recommended for XR-NTX and of these, 168 (28.1%) received injections. Compared to non-INJ patients, INJ patients were less likely to leave against medical advice (4.8% vs. 30.2%, p < .001) and more likely to initiate follow-up care (37.7% vs. 19.7%, p < .001). These differences remained significant after controlling for demographic covariates using regression analysis. Conclusions: XR-NTX was associated with higher rates of residential and early post-residential care engagement in patients with opioid dependence.
Received March 19, 2014; revised September 16, 2014; accepted October 31, 2014. † Dr. D. Gastfriend was an employee of Alkermes, Inc., Waltham, Massachusetts, at the time that this research was conducted. He is currently Chief Executive Officer of the Treatment Research Institute, Philadelphia, Pennsylvania. Address correspondence to Douglas L. Leslie, Penn State College of Medicine 90 Hope Drive, Hershey, PA 17033. E-mail: [email protected]
Scientific Significance: XR-NTX may be an effective adjunct in the residential treatment and aftercare of patients with opioid dependence. (Am J Addict 2015;24:265–270)
INTRODUCTION Opioid dependence causes great personal and societal costs.1 Successful treatment is associated with abstinence, reduced criminal activity, and an overall decrease in healthcare costs.2–13 Patient retention in treatment is a critical predictor of success and can be greater in medication-assisted programs.14 Approved medications fall into three categories: agonist (methadone), partial agonist (buprenorphine), or antagonist (naltrexone). In 1976, the National Institute on Drug Abuse postulated that an injectable formulation of extended-release naltrexone (XR-NTX) would improve the efficacy of narcotic antagonist treatment by enhancing patient adherence.15 By blocking the pharmacological effects of narcotics for a month, it might also promote engagement and enhance the efficacy of non-pharmacological aspects of treatment.16 1 XR-NTX (VIVITROL , Alkermes, Inc.) was approved by the Food and Drug Administration in 2006 for the treatment of alcohol dependence, and in 2010 for the treatment of opioid dependence. Its use in opioid dependence requires opioid detoxification, which makes inpatient or residential care a convenient setting for initiating treatment. Although the safety and effectiveness of XR-NTX has been established in research settings,17,18 there is almost no available information about the use of XR-NTX for opioid addiction in real-world community residential treatment settings and the interaction with patient outcomes. We conducted a retrospective study of XR-NTX 265
use among opioid dependent patients in residential rehabilitation programs. The objectives of the study were: (1) to determine the proportion of patients who were recommended for this treatment; and (2) to compare the early stage outcomes of patients who were injected with XR-NTX (INJ) with those who were recommended for XR-NTX but not injected (non-INJ).
METHODS Cohort Data for the study came from CRC Health Group, Inc., a large private network of specialized behavioral healthcare services in the U.S. Retrospective data that had been abstracted from CRC’s electronic health records were obtained for 11,146 patients attending one of three residential rehabilitation facilities in Pennsylvania during 2011, of whom 7,687 had a primary or secondary diagnosis of opioid dependence. Because not all opioid dependent patients would be good candidates for XR-NTX treatment, we limited our comparisons of early stage outcomes to patients who were identified in the data by CRC as having been recommended by clinical staff to receive XR-NTX and participated in an educational program about the drug. Patients who received XR-NTX (N ¼ 168) were identified, along with a comparison group of patients who were recommended for but failed to receive XR-NTX (N ¼ 430). No data were available that would identify the reasons why the largest group of patients in the cohort (N ¼ 7,089) were not considered for XR-NTX treatment (non-REC). Quality improvement staff at one site (Site 1) systematically made routine follow-up telephone calls to each of the follow-up facilities 7–10 days post-discharge to determine whether the patients referred by this residential facility attended their follow-up care. Data on the first followup visit were not routinely collected at the other two sites. Since the vast majority of patients were referred to non-CRC owned sites, HIPAA regulations prevented us from obtaining additional outcome information beyond the initial follow-up visit. All data were de-identified before being shared with researchers. The study was approved by the Institutional Review Board of the Penn State College of Medicine. Data Analyses The count and proportion of patients that were discharged prematurely against medical advice (AMA) was determined. In addition, for the facility with systematic follow-up data (Site 1 only), the proportion of patients who attended the postdischarge ambulatory level of care was determined. Chi-square tests were used to determine whether differences between the treatment groups were statistically significant using a two-tailed test with p < .05 as the threshold and were adjusted for multiple comparisons using Tukey’s test. Next, multivariate logistic regression models were developed to identify patient demographic and clinical 266
characteristics associated with receiving XR-NTX and the outcomes among those patients for whom XR-NTX was recommended but not administered. Three separate models were estimated, predicting receipt of XR-NTX, leaving AMA and, among patients from site 1 only, whether the patient attended their first post-discharge visit. Independent variables included medication group (INJ vs. non-INJ), site, age, gender, race/ethnicity, primary substance, type of insurance and referral source. In addition, interactions between medication group and each of the other independent variables were tested to examine patient characteristics that might predict outcomes. Mental health co-morbidities were not reliably available in the EHR data and were excluded from the model.
RESULTS Group Characteristics Of the 598 patients who were recommended for XR-NTX (8% of the total sample of patients with opioid dependence), 168 patients (28.1%) ultimately received the injection and 430 (71.9%) were not injected. Various reasons were offered for the latter group’s failure to receive their first injection including: the patient changed her/his mind (N ¼ 136; 31.6%); the patient was discharged before the medication arrived (N ¼ 121; 28.1%); inability to pay for the medication, usually due to insurance denial (N ¼ 89; 20.7%); the patient left against medical advice before being injected (N ¼ 67; 15.6%); or for unspecified reasons (N ¼ 17; 4%). Demographic characteristics were similar between the groups (Table 1). Although statistically significant, age differences were small, with the INJ patients slightly younger than the non-INJ and non-REC groups. The distribution of gender was similar in the INJ and non-INJ groups, while those in the non-REC group were more likely to be male. Most patients in the sample were Caucasian. Differences in the proportions of patients referred to treatment by State, County or Social Services were statistically significant, and were higher in the non-REC group (21.4%) than in either the INJ group (13.7%) or the non-INJ group (11.4%). The non-REC group was also less likely to have been referred by alumni (4.3%) than those in the non-INJ group (7.9%) or the INJ group (5.4%). INJ patients were much less likely to have commercial insurance (13.1%) than both the non-INJ patients (26.5%) and non-REC patients (22.4%), and conversely, were more likely to be covered by Medicaid (84.5%) than the non-INJ (71.2%) and non-REC (73.2%) groups. There were no group differences in primary substance dependence across the three groups. Rate of AMA discharge and length of stay are presented in Table 2. INJ patients were significantly less likely to leave AMA than non-INJ patients (4.8% vs. 30.2%, p < .001) or non-REC patients (4.8% vs. 30.2%, p < .001). On average, INJ patients stayed in rehab a week longer than both non-INJ and non-REC patients (p < .001 for both comparisons).
Injectable XR-NTX for Opioid Dependence in Rehab
April 2015
TABLE 1. Group characteristics
Injected N ¼ 168 Characteristics
N
Age (Mean SD) Gender Male Female Race White Black Hispanic Other Referral source Self State/county/social services CRC facility Alumni Drug alcohol provider Family Hospital/IP facility Insurance company Legal system OP facility Other Primary payer type Commercial insurance Medicaid Special payer Missing Primary substance Alcohol Opiates Cannabis Benzodiazepines Other
Not injected N ¼ 430
%
N
29.6 8.2
Not recommended N ¼ 7,089
%
N
28.3 8.2
p-values*
% 29.7 9.1
.006
