Articles Relapse to opioid use in opioid-dependent individuals released from compulsory drug detention centres compared with those from voluntary methadone treatment centres in Malaysia: a two-arm, prospective observational study Martin P Wegman, Frederick L Altice, Sangeeth Kaur, Vanesa Rajandaran, Sutayut Osornprasop, David Wilson, David P Wilson, Adeeba Kamarulzaman Summary Background Detention of people who use drugs into compulsory drug detention centres (CDDCs) is common Lancet Glob Health 2017; throughout East and Southeast Asia. Evidence-based pharmacological therapies for treating substance use disorders, 5: e198–207 such as opioid agonist treatments with methadone, are generally unavailable in these settings. We used a unique Published Online December 7, 2016 opportunity where CDDCs coexisted with voluntary drug treatment centres (VTCs) providing methadone in Malaysia http://dx.doi.org/10.1016/ to compare the timing and occurrence of opioid relapse (measured using urine drug testing) in individuals S2214-109X(16)30303-5 transitioning from CDDCs versus methadone maintenance in VTCs. See Comment page e123 Yale University School of Methods We did a parallel, two-arm, prospective observational study of opioid-dependent individuals aged 18 years and Medicine, New Haven, CT, USA older who were treated in Malaysia in the Klang Valley in two settings: CDDCs and VTCs. We used sequential sampling (M P Wegman BSc, to recruit individuals. Assessed individuals in CDDCs were required to participate in services such as counselling Prof F L Altice MD, A Kamarulzaman FRACP); Centre sessions and manual labour. Assessed individuals in VTCs could voluntarily access many of the components available of Excellence for Research in in CDDCs, in addition to methadone therapy. We undertook urinary drug tests and behavioural interviews to assess AIDS, Faculty of Medicine, individuals at baseline and at 1, 3, 6, 9, and 12 months post-release. The primary outcome was time to opioid relapse University of Malaya, Kuala post-release in the community confirmed by urinary drug testing in individuals who had undergone baseline Lumpur, Malaysia (M P Wegman, Prof F L Altice, interviewing and at least one urine drug test (our analytic sample). Relapse rates between the groups were compared V Rajandaran MPP, using time-to-event methods. This study is registered at ClinicalTrials.gov (NCT02698098). Prof A Kamarulzaman); University of Florida College of Medicine, Gainesville, FL, USA Findings Between July 17, 2012, and August 21, 2014, we screened 168 CDDC attendees and 113 VTC inpatients; of (M P Wegman); National these, 89 from CDDCs and 95 from VTCs were included in our analytic sample. The baseline characteristics of the two Antidrugs Agency, Selangor, groups were similar. In unadjusted analyses, CDDC participants had significantly more rapid relapse to opioid use Malaysia (S Kaur MBBS); post-release compared with VTC participants (median time to relapse 31 days [IQR 26–32] vs 352 days [256–unestimable], The World Bank Group, Washington, DC, USA log rank test, p<0·0001). VTC participants had an 84% (95% CI 75–90) decreased risk of opioid relapse after adjustment (S Osornprasop PhD, for control variables and inverse propensity of treatment weights. Time-varying effect modelling revealed the largest D Wilson PhD); and Burnet hazard ratio reduction, at 91% (95% CI 83–96), occurs during the first 50 days in the community. Institute, Melbourne, VIC, Australia (Prof D P Wilson PhD) Interpretation Opioid-dependent individuals in CDDCs are significantly more likely to relapse to opioid use after Correspondence to: Prof Frederick Altice, Yale release, and sooner, than those treated with evidence-based treatments such as methadone, suggesting that CDDCs University School of Medicine have no role in the treatment of opioid-use disorders. and School of Public Health, New Haven, CT 06510-2283, USA Funding The World Bank Group, Doris Duke Charitable Foundation, National Institute on Drug Abuse, Australian frederick.altice@yale.edu National Health & Medical Research Council, National Institute of Mental Health, and the University of Malaya-Malaysian Ministry of Higher Education High Impact Research Grant. Copyright © The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY-NC-ND license. Introduction drug control laws mandate 2 years of detention, Criminalisation of drug possession and use is common followed by community supervision for another worldwide, with many Asian countries confining 18 months after release.2 people who use drugs, or those suspected of using Although Malaysia introduced opioid-agonist therapies them, in specialized facilities called compulsory drug and needle and syringe programmes in 2005 when it detention centres (CDDCs).1 In Malaysia, CDDCs were failed to meet its political goal of reducing HIV infections,2 first introduced in 1978 in response to a growing heroin CDDCs remain central to drug control efforts.3 By 2010, epidemic and have been operated by the Malaysian Malaysia’s Ministry of Health had expanded opioid National Anti-Drug Agency (NADA). As of 2010, NADA agonist therapies in communities and prisons. The was operating 28 of these detention facilities housing perceived effectiveness of community-based opioid 7000 individuals. For those placed in CDDCs, national agonist therapies in contrast to the perceived high failure www.thelancet.com/lancetgh Vol 5 February 2017 e198 Articles Research in context Evidence before this study pharmacological treatments for treating substance use disorders We sought to compare the rates of relapse after mandatory are not provided in these settings. Globally, only 40 countries confinement in compulsory drug detention centres (CDDCs), provide treatment with methadone or buprenorphine in prison, where methadone was not available, with voluntary treatment albeit with low coverage rates. Many high-income countries programs providing methadone, for persons with opioid such as Australia, Canada, and most of the European Union have dependence. We reviewed the scientific literature by searching made methadone maintenance therapy available in criminal PubMed, EMBase, the Cochrane Database of Systematic Reviews, justice settings. In Asia, only six countries provide methadone and Google Scholar for any original articles published through maintenance therapy in prisons, including Indonesia and December, 2015 with no language restrictions, with the search Malaysia. In Asia where CDDCs exist, none provide methadone terms “methadone”, “opioid”, opiate”, “addiction”, “opioid maintenance therapy, and because of this absence of drug substitution therapy”, opioid agonist therapy”, “methadone”, treatments and the evidence of human rights abuses, many ‘‘substance abuse’’, “substance use”, “dependence”, ”detention”, international agencies have called for their systematic closure. “forced treatment”, “compulsory treatment”, “mandated Added value of this study treatment”, “mandatory treatment”, “addiction”, “addiction This is the first prospectively assessed study that directly treatment”, “involuntary treatment”, “involuntary addiction compared post-release drug use outcomes for people who treatment”, “detained”, “compulsory”, “prison”, “jail”, “correction”, completed so-called drug rehabilitation at CDDCs with those for “incarc”, “effective”, “relapse”, and “urine drug testing”. participants of voluntary drug treatment centres (VTCs) in From our search we concluded that peer-reviewed research Malaysia. By designing a study that simultaneously assessed comparing the effects of voluntary opioid agonist therapies two different, but coexisting, drug treatment programmes, we programmes with CDDCs on post-release opioid use outcomes provided robust, previously unavailable information about the is non-existent. The predominance of information on CDDC effectiveness of CDDCs and their role in relapse reduction effectiveness is at a high risk of bias, and has equivocal findings, compared with voluntary treatment with methadone. These as documented in a systematic review. By contrast, findings striking findings are also urgently needed to counter the from clinical trials and systematic reviews of community-based continued expansion of CDDCs across the country and recent methadone maintenance therapy are available in Asia and developments in the region where VTCs are being suspended or other regions, and confirm the effectiveness of methadone reverted to CDDCs, in the absence of documented evidence of maintenance therapy for treating opioid dependence in their benefit. reducing illicit opioid use compared with no pharmacological therapy. Similarly, findings from several trials have shown the Implications of all the available evidence value of methadone provided in confined settings, with The findings from our study showed that relapse to opioid use increased post-release treatment retention and decreased is more likely and faster after release from CDDCs compared likelihood of relapse. with VTCs suggesting that CDDCs have no role in the treatment of opioid use disorders. The sum of evidence strongly supports Although WHO recommends providing maintenance with international calls for all countries in Asia that support CDDCs opioid-agonist treatments such as methadone or to cease such human rights violations and scale-up buprenorphine as best practice for treating opioid dependence evidence-based treatments such as opioid agonist therapies in prisoners with opioid dependence, CDDCs are not subjected that can be accessed voluntarily and made potentially available to the same oversight, and such evidence-based to individuals as part of an alternative to incarceration strategy. rates of CDDCs resulted in NADA partially shifting its Up to now, no findings from studies have supported policy toward treating addiction from compulsory, any sustained rehabilitation benefits from CDDCs;7 institutional interventions to voluntary, evidence-based instead, they are associated with negative health treatment in line with that provided by the Ministry consequences, increased HIV risk-taking, compounded of Health.4 Several CDDCs were subsequently transitioned stigma and discrimination, human rights violations, and to VTCs, called Cure and Care centres, which provided absence of evidence-based practices in treating drug inpatient and outpatient methadone maintenance with a dependence.3,8,9 Despite many international agencies menu of voluntary psychosocial interventions, calling for all countries to close CDDCs over concerns of recreational programming, and vocational training.5,6 By their ineffectiveness and human-rights abuses,10 CDDCs contrast with CDDCs, patients could voluntarily present continue to operate, and in some settings proliferate, themselves for treatment at VTCs. After a thorough across east and southeast Asia. Approximately medical assessment (which was not available in the 600 000 people are mandatorily detained in more than CDDCs) patients at VTCs could receive 1 to 3 months of 1000 facilities annually.11–14 inpatient methadone treatment, followed by continued The pathways by which individuals enter CDDCs, the outpatient methadone maintenance upon release. duration of detention, and the services available in these e199 www.thelancet.com/lancetgh Vol 5 February 2017 Articles centres all vary substantially. In most countries, questionnaires, anonymised data, and analytic code are detention in CDDCs is predicated by a complex interplay deposited publically. of individual, social, and political factors.15 The main reasons for entering CDDCs include a positive urine Participants and setting drug test, suspicion of illicit drug use by police, or Eligibility criteria included being aged 18 years or older, insistence by family members.7,16 Proponents argue that the ability to provide informed consent, meeting criteria these centres are central policy components of a for opioid dependence,20 and intending to live in the Klang comprehensive response to opioid use, and serve to Valley. Assessed individuals in CDDCs were required to balance individuals’ needs for rehabilitation with the participate in non-evidence-based services, including right to safety for families and communities.17 individual, group and family counselling sessions, Individuals are held in these centres, however, which spiritual programmes, physical exercise, manual labour, often do not have trained healthcare personnel or and vocational training (eg, farming or electronics). In evidence-based drug treatments, without due process addition to methadone therapy, individuals enrolled in the protections or judicial oversight of detention. Opioid VTC arm could access many of the components of the agonist therapies such as methadone and CDDC programme, but did so voluntarily.5,6,21 Recruitment buprenorphine, which are included in the model list of in CDDCs occurred within 90 days before expected essential medicines by WHO for opioid dependence release. Common reasons for non-participation included treatment, are unavailable;18 and instead, educational not returning to Klang Valley, confidentiality concerns, and vocational training programmes, and hard labour and concerns over potential harassment by law are often mandated.7 enforcement because of study participation. Despite more than 30 years of experience, concerns We used sequential sampling for participant over CDDCs’ ineffectiveness and continued expansion, recruitment. During the recruitment periods, all facility few studies have empirically examined how CDDCs attendees meeting eligibility criteria at three VTCs affect drug use outcomes. A systematic review19 of providing methadone maintenance therapy in Greater compulsory inpatient and outpatient treatment strategies Kuala Lumpur and at six CDDCs were offered study showed little evidence that compulsory drug treatment is participation. Recruitment was halted early in September, effective in promoting abstention from drug use or in 2014, because of reversion of some VTCs to CDDCs; and reducing criminal recidivism. This review did not, because interim analyses revealed large differences however, compare the effectiveness of CDDCs relative to between study arms in the primary outcome. evidence-based treatment, such as voluntary medical After group informational sessions, interested clients treatment with opioid agonist therapies. met privately with trained researchers to complete For our study, we took advantage of a unique informed consent procedures. Everyone screened opportunity where CDDCs coexisted with voluntary drug received referral information for healthcare and drug treatment centres (VTCs) providing methadone in treatment. Consented participants were reimbursed Malaysia. This transition allowed contemporaneous RM50 (approximately US$15) for each visit and provided comparison of two divergent policies towards addressing mobile phones with phone credit. Additional RM50 problematic drug use in Malaysia with objective drug bonuses were provided for completing all of the first treatment outcomes in opioid-dependent individuals. In six and 12 follow-up interviews (one per month). our analysis, we compared the timing and occurrence of relapse with opioids and other illicit drugs confirmed by CDDC (n) VTC (n) urine drug testing between the two groups. Given the evidence of methadone’s effectiveness in reducing opioid 168 screened 113 screened use relative to treatment without opioid agonist therapies, we hypothesised that individuals transitioning to the 70 excluded 15 excluded community after release from VTCs would have fewer relapses and longer times to relapse than those released from CDDCs. 98 enrolled 98 enrolled Methods 9 ineligible 3 ineligible Study design We did a parallel, two-arm, prospective study of opioid- dependent individuals treated in two settings: Malaysian 89 analysed 95 analysed CDDCs and VTCs in the Klang Valley. We did not select a randomised design because the Malaysian judicial system Figure 1: Participant flow chart Reasons for screening failure were not systematically recorded, but common reasons For the protocol see https:// determined who entered CDDCs. The study was approved included not returning to Klang Valley, and concerns regarding confidentiality and dataverse.harvard.edu/dataset. by institutional review boards at the University of Malaya potential harassment by law enforcement due to study participation. xhtml?persistentId=doi:10.7910/ and Yale University, and by NADA, and the protocol, CDDC=compulsory drug detention centre. VTC=voluntary drug treatment centre. DVN/RO34OK www.thelancet.com/lancetgh Vol 5 February 2017 e200 Articles Compulsory drug Voluntary p value or detention history, lifetime and recent drug use, addiction detention centres treatment centres severity,22 opioid cravings using an 11-point Likert scale, (n=89) (n=95) motivation for drug treatment using SOCRATES,23 HIV Age (years) 39 (34–46); 25–56 37 (30–41); 21–70 0·0119* testing and treatment history, social support,24 and drug- Ethnic origin 0·9232† related and sex-related HIV-risk behaviours. The survey Malay 65 (73%) 67 (71%) was translated and back-translated to Bahasa Malaysia to Indian 15 (17%) 17 (18%) ensure the accuracy of intended meaning. Researchers Chinese and other 9 (10%) 11 (12%) undertook and recorded urine drug tests for Completed secondary school 0·0220† five metabolites: opioids, methamphetamines, benzo- No 58 (65%) 46 (48%) diazepines, methadone, and buprenorphine using a Yes 31 (35%) 49 (52%) custom RapiDip InstaTest (Cortez Diagnostics, CA, USA). Married 0·1822† The tests have good diagnostic accuracy, interoperator No 68 (76%) 80 (84%) reliability, and performance on interference testing with Yes 21 (24%) 15 (16%) high specificity when tested with other common Previous housing type 0·3841† metabolites. We also did HIV testing, but this is the subject Missing data 2 (2%) 0 of a companion analysis. Urine drug test assessments Permanent 28 (32%) 25 (26%) occurred on the day of release for CDDC participants and Temporary 59 (68%) 70 (74%) baseline surveys occurred within 90 days before, or 7 days Number of times imprisoned 3 (2–5); 0–16 3 (1–4); 0–10 0·2480* after, release from CDDC or inpatient treatment at VTCs. Number of times in lockup or jail 7 (3–10); 0–49 5 (3–10); 0–60 0·5085* Follow-up visits, especially those where urine drug testing Number of times detained in compulsory 1 (0–2); 0–8 1 (0–2); 0–10 0·7774* assessments were made, were scheduled (within a 2-week drug detention centres‡ target window) in person in a private setting. All interviews Age at first drug use (years) 18 (15–21); 9–40 18 (16–20); 12–48 0·5866* were conducted in Bahasa Malaysia. Drug of choice 0·1109† Missing data 2 (2%) 0 Outcomes Heroin 82 (94.3%) 83 (87%) The primary outcome, specified a priori, was occurrence Other 5 (6%) 12 (13%) and timing of urine drug test-confirmed opioid use in Duration of heroin use (years) 16 (10–21); 1–40 13 (8–20); 3–41 0·1430* the community, because all participants met criteria for Daily use of heroin before entering facility 0·4774† opioid dependence, the most frequently used illicit drug Missing data 3 (3%) 7 (7%) in Malaysia. Timing of relapse was based on free choice No 14 (16%) 11 (13%) in the community and not within a controlled setting. A Yes 72 (84%) 77 (88%) secondary outcome was urine drug test-confirmed use of Drug use severity 0·5167† any of three illicit drug types: opioids, amphetamine-type Missing data 0 2 (2%) substances, or benzodiazepines. Low or moderate 19 (21%) 14 (15%) In our adjusted, weighted analyses, we included Substantial 59 (66%) 65 (70%) variables which could explain potential differences in Severe 11 (12%) 14 (15%) treatment allocation or baseline risk of opioid relapse Opioid cravings (0–10) 3 (1–7); 0–10 3 (0–7); 0–10 0·4550* between the arms: receptiveness, ambivalence and taking Ever injected drugs 0·0944† steps towards change in drug use; daily heroin use before Missing data 1 (1%) 4 (4%) detention or inpatient VTC entry; age of first drug use No 60 (68%) 51 (56%) and years of heroin use; addiction severity; drug injection; Yes 28 (32%) 40 (44%) previous drug treatment; lifetime use of alcohol, stimulants, benzodiazepines and non-heroin opioids; Alcohol use (lifetime) 0·2439† age; ethnicity; marital, housing, and education status; No 17 (19%) 25 (26%) social support; and number of times imprisoned, jailed, Yes 72 (81%) 70 (74%) and previously detained in a CDDC (appendix). Non-heroin opioid use (lifetime) 0·7269† No 73 (82%) 76 (80%) Statistical analysis Yes 16 (18%) 19 (20%) We used five number summaries and Mann–Whitney (Table 1 continues on next page) U test statistics (continuous variables) or proportions and χ² test statistics (categorical variables) to compare group See Online for appendix Procedures characteristics. For the main analyses, we employed time- Results from urine drug tests were obtained at baseline to-event approaches, for which the time origin was the and at 1, 3, 6, 9, and 12 months post-release; baseline and first day not being in a controlled environment (release monthly behavioral surveys were also obtained through date from CDDCs or inpatient VTC units). The target 12 months post-release. We interviewed participants about event was the first positive urine drug test, assuming that their demographic and social characteristics, incarceration any missing intervening follow-up measurements were e201 www.thelancet.com/lancetgh Vol 5 February 2017 Articles negative (non-events). With this definition, we censored Compulsory drug Voluntary p value observations only at the latest non-missing negative urine detention centres treatment centres drug test, given that the event had not yet occurred. (n=89) (n=95) For each group, we estimated Kaplan-Meier curves for (Continued from previous page) time-to-relapse, cumulative relapse-free proportions at Benzodiazepine use (lifetime) 0·99832 selected intervals, and median relapse times, applying the No 74 (83%) 79 (83%) log-rank test of equality. In our prespecified primary Yes 15 (17%) 16 (17%) analysis, we employed Cox-regression with Efron’s Stimulant use (lifetime) 0·6526† method for ties handling. To account for potential No 28 (32%) 27 (28%) selection effects, a logistic regression model of the Yes 61 (69%) 68 (72%) propensity of seeking care at a VTC was developed using Use of more than one drug at the same 0·1390† control variables measuring characteristics of participants time (lifetime) before treatment allocation.25 Common support and Missing data 0 2 (2%) balance diagnostics suggested good model performance No 40 (45%) 52 (56%) (appendix).25 Inverse propensity of treatment scores were Yes 49 (55%) 41 (44%) then incorporated as stabilised weights in the final Cox Ever received buprenorphine treatment‡ 0·8846† regression with the remaining control variables (ie, opioid Missing data 0 11 (12%) cravings and the SOCRATES subscales) included as No 78 (88%) 73 (87%) explanatory variables with study arm as the main variable Yes 11 (12%) 11 (13%) of interest (see appendix for further description of this Recent buprenorphine treatment‡ 0·1670† approach and related robustness checks).26 Missing data 0 11 (12%) Detecting the presence of a time-varying effect of study No 87 (98%) 84 (100%) arm,27 we used Akaike Information Criteria to select a Yes 2 (2%) 0 piece-wise model which included a time-varying Readiness for change specification (appendix). We provided hazard ratio Recognition 40 (20–60); 10–70 50 (30–70); 10–70 0·1335* estimates for both the time-invariant and time-varying specifications, with the former interpretable as the Ambivalence 60 (40–70); 10–90 60 (40–70); 10–90 0·6663* averaged effect over the follow-up interval. We also Taking steps 70 (50–90); 10–90 90 (70–90); 40–90 0·0001* plotted adjusted survival curves. These were implemented Recent emergent or urgent care 0·8569† after a two-stage imputation approach to address partial Missing data 1 (1%) 0 missing information on length of inpatient stay for the No 83 (94%) 89 (94%) VTC arm, timing of the first urine drug test measurement Yes 5 (6%) 6 (6%) for the CDDC arm, and baseline control variables Ever tested for HIV 0·0508† (appendix contains description of missingness, Missing data 1 (1%) 5 (5%) imputation details, and related sensitivity analyses). No 7 (8%) 16 (18%) Because of attrition and missing follow-up Yes 81 (92%) 74 (82%) measurements, we did several additional robustness HIV-test result 0·0055† checks (sensitivity analyses) by redefining our missing Missing data 3 (3%) 6 (6%) follow-up measurement assumptions in several ways, HIV-negative 72 (84%) 61 (69%) such that all missing follow-up urine drug tests were HIV-positive 5 (6%) 2 (2%) equivalent to a non-event; all missing follow-up urine Unknown 9 (11%) 26 (29%) drug tests were equivalent to an event; and missing Social support follow-up urine drug tests were 25% likely to be events Significant partner 16 (12–20); 4–24 16 (12–22); 4–24 0·7011* for CDDC participants and 75% likely to be events for Family 22 (20–24); 11–24 23 (20–24); 10–24 0·2397* VTC participants (ie, a 50% absolute difference). Friends 16 (12–20); 4–24 19 (12–21); 8–24 0·3566* Analyses were done in SAS version 9.4 (SAS Institute Time as inpatient (days) ·· 80 (58–93); 15–100 ·· Inc., Cary, NC). This study is registered at ClinicalTrials. gov (NCT02698098). Data are median (IQR); range, or n (%). Denominators are different for the missing percentage calculations and the cariable percentage calculations ··=not applicable. *Kruskal-Wallis. †χ². ‡12% non-response in voluntary treatment arm. Role of the funding source Table 1: Baseline characteristics of the study population One of the funders of the study (the World Bank) had a role in study design and review of the manuscript but had no role in data collection, data analysis, data interpretation, Results or decision to publish the findings. All other funders had Between July 17, 2012, and August 21, 2014, we screened no role in these stages. The corresponding author had full 281 opioid-dependent individuals in Malaysia; 168 in access to all the data in the study and had final CDDCs and 113 in inpatient units of VTCs. 98 in both responsibility for the decision to submit for publication. groups completed baseline interviews and 89 (CDDC) www.thelancet.com/lancetgh Vol 5 February 2017 e202 Articles Days from No opioid use No illicit drug use 100 CDDC group release VTC group Probability of no opioid use (%) Voluntary Compulsory Voluntary Compulsory 75 treatment drug treatment drug centres detention centres detention centres centres 50 Hazard ratio 0·198 (95% CI 0·12–0·31) Month 1 30 0·90 0·51 0·89 0·46 25 (0·81– (0·39–0·61) (0·79– (0·35–0·57) 0·95) 0·94) Log-rank p<0·0001 0 3 90 0·80 0·23 0·78 0·19 0 30 90 180 270 365 (0·68– (0·14–0·33) (0·66– (0·11–0·28) Time (days) Number at risk 0·88) 0·87) CDDC 89 42 17 12 6 2 6 180 0·69 0·19 0·62 0·12 VTC 76 61 45 33 26 7 (0·55– (0·11–0·28) (0·48– (0·06–0·20) 0·79) 0·73) Figure 2: Unadjusted probability of no opioid use 9 270 0·62 0·12 0·53 0·07 CDDC=compulsory drug detention centre. VTC=voluntary drug treatment centre. (0·48– (0·05–0·21) (0·39– (0·02–0·16) 0·73) 0·66) 100 CDDC group 12 365 0·50 0·10 0·42 0·05 VTC group Probability of no opioid use (%) (0·34– (0·04–0·19) (0·27– (0·01–0·13) 0·64) 0·56) 75 Unadjusted Kaplan Meier estimates (95% CIs) reflecting the cumulative probability of no drug relapse for selected follow-up times. 50 Table 2: Unadjusted probability of no drug relapse using urine drug testing 25 and 95 (VTC) of these individuals had at least one 0 subsequent urine drug test, representing our analytic 0 30 90 180 270 365 sample (figure 1). Loss between recruitment and baseline Time (days) measurement was due to inability to locate participants Figure 3: Adjusted probability of no opioid use (including early release) and absence of communication CDDC=compulsory drug detention centre. VTC=voluntary drug treatment centre. with the study team. Delaying the origin of time from day of entry at the VTC to day of inpatient release reduced unadjusted, adjusted, and adjusted with time-varying the VTC arm sample by between 13 and 34 participants group effect hazard ratios in table 3). depending on the imputation model used, representing VTC participants had an 80% (95% CI 69–88) lower risk attrition before the discharge date (appendix). The of opioid relapse—an effect that was accentuated to 84% number of completed outcome measurements for each (75–90) after adjustment using control variables and group were similar, with 50% completed at month 3 and inverse propensity of treatment weights. Time-varying a quarter to a third completed at month 12 (appendix). effect modelling revealed the largest hazard ratio reduction, Participants were similarly matched for most baseline of 91% (83–96), occurred during the first 50 days of characteristics (table 1) except that CDDC participants observation. This hazard ratio reduction diminished over were older, had higher education levels, were incarcerated the post-release period to 61% (12–83) at 180 days and by more frequently, were less likely to have injected opioids, 270 days a large difference remained in the arms. Moreover, and were less likely to be taking steps towards changing increased craving for opioids at baseline corresponded to a their drug use. reduction in hazard of relapse; hazard ratios for recognition In unadjusted analyses, CDDC participants had of, ambivalence for, and taking steps towards changing significantly more rapid relapse to opioid use post-release drug use were not significantly different between the compared with VTC participants (median time to relapse groups. Similar adjusted hazard ratio estimates were 31 days [95% CI 26–32] vs 352 days [256 to unestimable], computed for any-illicit-drug use, including amphetamine- log rank test, p<0·0001; table 2, figure 2, appendix); type substances (appendix). Additionally, sensitivity additional analyses with relapse to any drug use were analyses affecting the imputation of missing dates or similar (30 days [95% CI 24–32] vs 317 days [177 to alternate event coding did not substantively change the unestimable]; table 2, appendix), favouring more rapid results for opioid or any-illicit-drug use (appendix). median time to relapse for CDDC participants (30 vs 317 days, log rank test, p<0·0001). Cox-regression Discussion modelling, including inverse propensity score weighting Our study showed opioid-dependent participants treated and adjustment for post-treatment-assignment variables with methadone in VTCs experienced a seven-fold revealed consistent results (adjusted curves in figure 3; decreased risk of relapse to opioids and any-illicit-drug e203 www.thelancet.com/lancetgh Vol 5 February 2017 Articles after release, compared to similarly matched individuals Unadjusted Adjusted Adjusted, time-varying released from CDDCs in Malaysia. Not only did we find group effect that relapse was markedly faster for those released from Hazard ratio 95% CI Hazard ratio 95% CI Hazard ratio 95% CI CDDCs compared to those treated in VTCs, but considered on its own, relapse to opioid use was rapid All variables in 0·198 (0·12–0·31) 0·155 (0·10–0·25) ·· ·· voluntary drug after CDDC release, suggesting CDDCs have no role in treatment centres treating opioid use disorders. This is one of the first peer- Time after release reviewed study comparing objective drug use outcomes from voluntary contemporaneously for opioid-dependent persons drug treatment centres released from CDDCs with similar participants receiving evidence-based methadone maintenance in community- >0–<50 days ·· ·· ·· ·· 0·087 (0·04–0·17) based VTCs. It contributes to a growing body of evidence 90 days ·· ·· ·· ·· 0·270 (0·14–0·51) of how drug policies negatively impact individual and 180 days ·· ·· ·· ·· 0·388 (0·17–0·88) public health.28 270 days ·· ·· ·· ·· 0·456 (0·18–1·13) The findings here strongly support international calls 365 days ·· ·· ·· ·· 0·509 (0·19–1·35) for all countries that support CDDCs to cease operations Opioid cravings ·· ·· 0·903 (0·85–0·96) 0·904 (0·85–0·96) in light of the ineffectiveness of these centres in treating Recognition of ·· ·· 1·015 (0·88–1·17) 1·015 (0·89–1·16) change drug dependence. Simultaneously, these countries Ambivalence ·· ·· 0·977 (0·86–1·12) 0·967 (0·85–1·11) should scale-up evidence-based opioid agonist therapies towards change such as methadone or buprenorphine maintenance in Taking steps ·· ·· 0·991 (0·86–1·14) 0·993 (0·87–1·14) communities, which should be encouraged and towards change voluntary. Promisingly, policy modifications are underway in southeast Asia where some CDDCs are Data compare the hazard of the specified post-release or discharge opioid relapse in the voluntary treatment centres group versus the compulsory drug detention centres group. The unadjusted hazard ratios were estimated using Cox regression transitioning to VTCs where opioid agonist therapies are models for which group was the only covariate and no additional weighting was used. The adjusted Cox models available.14 Yet, these findings are also urgently needed to additionally adjusted for measurements of opioid cravings and ambivalence, recognition, and taking steps towards change counter developments in Vietnam and Malaysia where and applied inverse propensity of treatment weights. The adjusted, time varying group effects modified the adjusted model by incorporating a time-varying (non-proportional) hazard for the group, implemented as an interaction between VTCs are being suspended or reverted to CDDCs, in the group and the logarithm of time beyond 50 days post-release. Stochastic regression for missing discharge and inpatient absence of clear evidence that they reduce drug use.29 times was used; missing outcome measurements were ignored. CDDCs share similarities with many prisons globally, Table 3: Cox regression hazard ratios for opioid relapse after release where people who use drugs are concentrated, and transitions to the community are marked by similar high rates of drug relapse and disruptions in social networks. we specifically compared two policy programmes to Findings from several countries empirically support address opioid dependence that coexist in several provision of opioid agonist therapies within prisons for countries in Asia. Accordingly, we have not explored nor reducing within-prison transmission of blood-borne discussed all potential policy options, such as provision viruses.12 In the post-release period, opioid agonist of opioid agonist therapies in compulsory settings or therapies substantially reduce drug use and HIV voluntary residential treatment without opioid agonist transmission risk30 and increases retention in care,31,32 therapies. especially if the opioid agonist therapy is optimally Second, this study was observational in nature, such dosed.33 Unlike prisons, however, CDDCs do not adhere that treatment exposures were allocated non-randomly. to international regulatory oversight because entry Participants in the CDDC arm were detained by police ignores judicial processes, and they do not provide an for suspected or real drug use. By contrast, VTC equivalence of treatment available in the community, participants probably sought treatment of their own including opioid agonist therapies and medical care.3,11,34 volition, including through support from family and Despite the new findings from this study, the data friends. This difference, however, is partly mitigated by should be interpreted in the context of several our eligibility criteria for which only those meeting considerations. First, the two comparison programmes opioid dependence criteria were enrolled. We differ not just by the presence and absence of methadone characterised latent dissimilarity between these two maintenance therapy, but also by other optional services populations by obtaining several measures associated available and the voluntariness of the two strategies. Our with drug relapse. These measures yielded only small study was not designed to isolate the precise differences, especially in addiction severity, between the component(s) responsible for the difference. However, two groups. We further incorporated these variables in given systematic reviews that document a substantial our modelling to adjust for the different propensities of difference in treatment outcomes between drug-free seeking treatment. treatment and patients prescribed opioid agonist Third, there was high attrition in our study. After therapies,35 methadone is likely to have played a recruitment, 53% participants in the CDDC arm prominent role in the observed differences. In this study, completed baseline interviews and had at least one urine www.thelancet.com/lancetgh Vol 5 February 2017 e204 Articles drug test, compared with 84% from the VTCs. relapse to use of amphetamine-type substances, and Nonetheless, we noted considerable similarity between should be closed even in regions where amphetamine groups retained and subsequently analysed. Thus, for use disorders are common. problematic bias to occur, we would have to believe that Although the individual and societal costs of the full sample of CDDC participants were substantially maintaining CDDCs are high and despite incontrovertible less likely to relapse than the participants recruited from evidence that Malaysia’s harm reduction programmes VTCs. Additionally, among those who had a baseline are cost-effective,42 government and public resistance to interview and at least one urine drug test (our analytic closure of CDDCs or even conversion to VTCs remains sample), attrition did not differ between groups. This high.14 Key factors sustaining CDDCs in Malaysia include suggests that although the analysed sample might have the country’s anachronistic culture of zero tolerance been predisposed to more favourable outcomes, both towards people who use drugs and abstinence-based groups were similarly affected. For this reason, estimates treatment. In addition, NADA’s performance metrics are of the between-group effects are still likely to be valid. focused on maintaining or increasing arrests and Furthermore, alternative event coding which assumed a detentions, rather than the societal goals of rehabilitation large (50%) absolute difference in risk of event for missing and public health that focus on reducing drug use, crime values between each group produced findings that or recidivism, and HIV transmission.14,43 remained significant. Taken together, these limitations Since 2010, several international agencies have provided suggest that the findings, while of strong internal validity, many regional consultations on CDDCs in Asia, from are likely to be most reflective of a subset of people who which an expert working group has been established to use drugs in these settings, and may exaggerate the overall formulate evidence-based recommendations to support effects experienced had all people who use drugs in the transition to a comprehensive system of voluntary, CDDCs been shifted to VTCs. Despite these limitations, community-based treatment, harm reduction and social this study provides clear evidence of the ineffectiveness of support services. Along with these recommendations, this CDDCs in addressing opioid dependence, and shows a group proposes a three-step strategy for transition that several-fold decrease in relapse to opioid use after release includes establishing a national, multisectoral decision- among those prescribed methadone in VTCs. making mechanism with responsibility for the transition; Understanding the extent to which the effectiveness of implementing reforms to develop and strengthen the methadone provided in VTCs is reflected in other various mechanisms responsible for addressing operations outcomes such as criminal activity, rearrest, HIV and treatment of substance use disorders across different transmission, mortality, and quality-of-life, is important sectors; and examining related drug policies, including and requires further assessment. Findings from previous laws, regulations, strategies and practices, and shifting studies36,37 have shown that drug-treatment effectiveness is away from criminalisation and punishment, to health- strongly associated with improvements in these indicators. based and rights-based drug policy measures.44 Furthermore, rapid relapse to drug use after release from Despite regional consultations and a 2012 joint statement controlled settings like prisons is associated with high by 12 UN agencies calling for immediate closure of CDDCs rates of HIV risk-behaviors,38 overdose, and death.39 and for implementation of voluntary, evidence-informed These results are likely to be generalisable beyond and rights-based health and social services in the Klang Valley to greater Malaysia, and more broadly, to community, CDDCs continue to operate in east and other regions of east and southeast Asia. This conclusion southeast Asia. Our study provides the first prospective, is supported by evidence suggesting that relapse is comparative evidence that CDDCs are ineffective in common among released detainees who are not provided preventing drug relapse, especially when compared with opioid agonist therapies in both Malaysian CDDCs3,40 and voluntary evidence-based treatments like methadone. In those elsewhere in Asia.7,41 This finding would also light of this, and numerous studies documenting the cost- probably hold for settings where amphetamine-type effectiveness of opioid agonist therapies in treating opioid substances are prevalent. For example, use of dependence, a renewed effort by governments to transition amphetamine-type substances was common in our to and expand a comprehensive system of voluntary, sample of individuals with opioid dependence (around community-based treatment is urgently needed, especially 70% with lifetime use). Although there are currently no in Asia. Ultimately, this effort should be situated within a evidence-based pharmacological treatments for people more comprehensive review of national and regional drug with amphetamine use disorders, the first quartile policies that continue to criminalise drug use and limit median time to relapse to use of amphetamine-type people who use drugs from accessing evidence-based substances in our sample was 33 versus 355 days for treatment, care, and support. CDDC and VTC participants, respectively. Even though it Contributors was beyond the scope of this study to examine outcomes MPW, FLA, VR, and AK did the literature search. FLA, SK, VR, SO, DW, for those using amphetamine-type substances but DPW, and AK contributed to the study design and conceptualisation. VR obtained the data. MPW did the data analysis and produced the without opioid dependence, findings here provide figures. MPW, FLA, DPW, and AK interpreted the data. MPW, FLA, SK, evidence that CDDCs are ineffective in preventing VR, SO, DW, DPW, and AK drafted and revised the report. e205 www.thelancet.com/lancetgh Vol 5 February 2017 Articles Declaration of interests 15 Human Rights Watch. Skin on the cable: the illegal arrest, arbitrary We declare no competing interests. detention and torture of people who use drugs in Cambodia, 2010. https://www.hrw.org/report/2010/01/25/skin-cable/illegal-arrest- Acknowledgments arbitrary-detention-and-torture-people-who-use-drugs https://www. Funding for this research was provided by the World Bank, Malaysian hrw.org/sites/default/files/reports/cambodia0110webwcover.pdf Ministry of Education High Impact Research Grant (HIRGA (accessed Sept 25, 2016). E000001-20001), National Institute of Mental Health for career 16 Human Rights Watch. Torture in the name of treatment: human development (F30MH105153), National Institute on Drug Abuse for rights abuses in Vietnam, China, Cambodia, and Lao PDR, 2012. research (R01 DA025943 and R01 DA041271) and career development https://www.hrw.org/report/2012/07/24/torture-name-treatment/ (K24 DA017072), and Doris Duke Charitable Foundation through a human-rights-abuses-vietnam-china-cambodia-and-lao-pdr grant supporting the Doris Duke International Clinical Research (accessed Sept 25, 2016). Fellows Program at Yale University School of Medicine. We thank the 17 Wu Z. Arguments in favour of compulsory treatment of opioid participants in this study; the staff at Malaysian PUSPEN and Cure & dependence. Bull World Health Organ 2013; 91: 142–45. Care facilities; our research team Muhammad Azri, Kamal Sapilo, and 18 WHO. 19th WHO model List of essential medicines. Geneva: Nuruljannah; Jeannia Fu for her contributions to the design and World Health Organization, 2015: p 40. initiation of the study and comments on the manuscript; the Malaysian 19 Werb D, Kamarulzaman A, Meacham MC, et al. The effectiveness National Anti-Drug Agency and its former Director-General Puan of compulsory drug treatment: a systematic review. Int J Drug Policy Seri Dato Zuraida; and support staff at the Centre of Excellence on 2015; 28: 1–9. Research In AIDS. The content of this report is solely the responsibility 20 Wickersham JA, Azar MM, Cannon CM, Altice FL, Springer SA. of the authors and does not necessarily represent the official views of Validation of a brief measure of opioid dependence: the rapid opioid the funders. dependence screen (RODS). J Correct Health Care 2015; 21: 12–26. 21 Krishnan A, Brown S-E, Ghani M, Khan F, Kamarulzaman A, References Altice FL. Pre-treatment drug use characteristics and experiences 1 UN Office on Drugs and Crime, Economic and Social Commission among patients in a voluntary substance abuse treatment center in for Asia and the Pacific, Joint United Nations Programme on Malaysia: a mixed methods approach. Subst Abus 2016; 37: 542–49. HIV/AIDS. Report of the second regional consultation on 22 Skinner HA. The drug abuse screening test. Addict Behav 1982; compulsory centres for drug users in Asia and the Pacific. 7: 363–71. Kuala Lumpur: UNODC, 2012. 23 Miller WR, Tonigan JS. Assessing drinkers’ motivation for change: 2 Kamarulzaman A. Impact of HIV prevention programs on drug the stages of change readiness and treatment eagerness scale users in Malaysia. J Acquir Immune Defic Syndr 2009; (SOCRATES). Psychol Addict Behav 1996; 10: 81. 52 (suppl 1): S17–19. 24 Zimet GD, Powell SS, Farley GK, Werkman S, Berkoff KA. 3 Fu JJ, Bazazi AR, Altice FL, Mohamed MN, Kamarulzaman A. Psychometric characteristics of the multidimensional scale of Absence of antiretroviral therapy and other risk factors for perceived social support. J Pers Assess 1990; 55: 610–17. morbidity and mortality in Malaysian compulsory drug detention 25 Austin PC. Balance diagnostics for comparing the distribution of and rehabilitation centers. PLoS One 2012; 7: e44249. baseline covariates between treatment groups in propensity-score 4 WHO Western Pacific Regional Office, WHO Office of the matched samples. Stat Med 2009; 28: 3083–107. Representative for Brunei, Darussalam, Malaysia, and Singapore. 26 Vaughan AS, Kelley CF, Luisi N, del Rio C, Sullivan PS, Good practices in Asia: effective paradigm shifts towards an Rosenberg ES. An application of propensity score weighting to improved national response to drugs and HIV/AIDS: scale-up of quantify the causal effect of rectal sexually transmitted infections on harm reduction in Malaysia. Geneva: World Health Organization, incident HIV among men who have sex with men. 2011. BMC Med Res Methodol 2015; 15: 25. 5 Ghani MA, Brown S-E, Khan F, et al. An exploratory qualitative 27 Kleinbaum DG, Klein M. Survival analysis: a self-learning text. assessment of self-reported treatment outcomes and satisfaction New York: Springer; 2012. among patients accessing an innovative voluntary drug treatment centre in Malaysia. Int J Drug Policy 2015; 26: 175–82. 28 Csete J, Kamarulzaman A, Kazatchkine M, et al. Public health and international drug policy. The Lancet 2016; 387: 1427–80. 6 Khan F, Krishnan A, Ghani M, et al. Assessment of an innovative voluntary substance abuse treatment program designed to 29 Tuoi Tre News. Voluntary drug detoxification a flop in Vietnam. 2014. transition from compulsory drug detention centers in Malaysia. http://tuoitrenews.vn/features/23374/voluntary-drug-detoxification- Subst Use Misuse 2016; in press. a-flop-in-vietnam (accessed Jan 25, 2016). 7 WHO. Assessment of compulsory treatment of people who use 30 Altice FL, Azbel L, Stone J, et al. The perfect storm: incarceration drugs in Cambodia, China, Malaysia and Vietnam: an application of and the high-risk environment perpetuating transmission of HIV, selected human rights principles. Geneva: WHO, 2009. hepatitis C virus, and tuberculosis in Eastern Europe and Central Asia. Lancet 2016; 388: 1228–48. 8 Amon JJ, Pearshouse R, Cohen JE, Schleifer R. Compulsory drug detention in east and southeast Asia: evolving government, UN and 31 Kinlock TW, Gordon MS, Schwartz RP, Fitzgerald TT, O’Grady KE. donor responses. Int J Drug Policy 2014; 25: 13–20. A randomized clinical trial of methadone maintenance for prisoners: results at 12 months postrelease. J Subst Abuse Treat 2009; 37: 277–85. 9 Amon J, Pearshouse R, Cohen J, Schleifer R. Compulsory drug detention centers in China, Cambodia, Vietnam, and Laos: health 32 Rich JD, McKenzie M, Larney S, et al. Methadone continuation and human rights abuses. Health Hum Rights 2013; 15: 124–37. versus forced withdrawal on incarceration in a combined US prison and jail: a randomised, open-label trial. Lancet 2016; 386: 350–59. 10 UN entities. Joint statement: compulsory drug detention and rehabilition centres, 2012. https://www.unodc.org/documents/ 33 Wickersham JA, Zahari MM, Azar MM, Kamarulzaman A, southeastasiaandpacific/2012/03/drug-detention-centre/JC2310_ Altice FL. Methadone dose at the time of release from prison Joint_Statement6March12FINAL_En.pdf (accessed Sept 25, 2016). significantly influences retention in treatment: Implications from a pilot study of HIV-infected prisoners transitioning to the 11 Dolan K, Wirtz AL, Moazen B, et al. Global burden of HIV, viral community in Malaysia. Drug Alcohol Depend 2013; 132: 378–82. hepatitis, and tuberculosis in prisoners and detainees. Lancet 2016; 388: 1089–102. 34 WHO, UNODC, UNAIDS. Technical guide for countries to set targets for universal access to HIV prevention, treatment and care 12 Kamarulzaman A, Reid SE, Schwitters A, et al. Prevention of for injecting drug Users—2012 revision. Geneva: World Health transmission of HIV, hepatitis B virus, hepatitis C virus, and Organization, 2012. tuberculosis in prisoners. Lancet 2016; 388: 1115–26. 35 Mattick RP, Breen C, Kimber J, Davoli M. Buprenorphine 13 Mathers BM, Degenhardt L, Ali H, et al. HIV prevention, treatment, maintenance versus placebo or methadone maintenance for opioid and care services for people who inject drugs: a systematic review of dependence. Cochrane Database Syst Rev 2014; 2: CD002207. global, regional, and national coverage. Lancet 2010; 375: 1014–28. 36 Dennis BB, Naji L, Bawor M, et al. The effectiveness of opioid 14 Tanguay P, Kamarulzaman A, Aramrattana A, et al. Facilitating a substitution treatments for patients with opioid dependence: transition from compulsory detention of people who use drugs a systematic review and multiple treatment comparison protocol. towards voluntary community-based drug dependence treatment Syst Rev 2014; 3: 105. and support services in Asia. Harm Reduct J 2015; 12: 31. www.thelancet.com/lancetgh Vol 5 February 2017 e206 Articles 37 MacArthur GJ, Minozzi S, Martin N, et al. Opiate substitution 42 Osornprasop S, Dahlui M, Kamarulzaman A, et al. Return on treatment and HIV transmission in people who inject drugs: investment and cost-effectiveness of harm reduction program in systematic review and meta-analysis. BMJ 2012; 345: e5945. Malaysia. Directions in development; human development. 38 Adams LM, Kendall S, Smith A, Quigley E, Stuewig JB, Washington, DC: World Bank Group, 2014. http://documents. Tangney JP. HIV risk behaviors of male and female jail inmates worldbank.org/curated/en/4/07/23006930/return-investment-cost- prior to incarceration and one year post-release. AIDS Behav 2013; effectiveness-harm-reduction-program-malaysia (accessed 17: 2685–94. Dec 12, 2015). 39 Merrall ELC, Kariminia A, Binswanger IA, et al. Meta-analysis of 43 UN Office on Drugs and Crime Regional Centre for East Asia and drug-related deaths soon after release from prison. Addiction 2010; the Pacific. Drug-Free ASEAN 2015: status and recommendations, 105: 1545–54. 2015. http://www.unodc.org/documents/southeastasiaandpacific/ 40 Reid G, Kamarulzaman A, Sran SK. Malaysia and harm reduction: Publications/ASEAN_2015.pdf (accessed Sept 25, 2016). the challenges and responses. Int J Drug Policy 2007; 18: 136–40. 44 UNODC. Transition from compulsory centers for drug users to 41 Thomson N. Detention as treatment. Detention of voluntary community-based treatment and services: discussion paper, methamphetamine users in Cambodia, Laos, and Thailand. 2015. https://www.unodc.org/documents/southeastasiaandpacific/ Melbourne: Open Society Institute, 2010. Publications/2015/hiv/Discussion_Paper_on_Transition_from_ CCDUs_Edited_Final4_04Sept15.pdf (accessed Sept 25, 2016). e207 www.thelancet.com/lancetgh Vol 5 February 2017