The characteristics of a cohort who tamper with prescribed and diverted opioid medications

Journal of Substance Abuse Treatment 58 (2015) 51–61

Contents lists available at ScienceDirect

Journal of Substance Abuse Treatment

The characteristics of a cohort who tamper with prescribed and diverted opioid medications Briony Larance, Ph.D. a,⁎, Nicholas Lintzeris, Ph.D. b,c, Raimondo Bruno, Ph.D. d, Amy Peacock, Ph.D. d, Elena Cama, B.C.C.J. (Hons.) a, Robert Ali, Ph.D. e, Ivana Kihas, B.Psych. (Hons.) a, Antonia Hordern, M.Health.S.C. (C.D.M.), B.A.App.S.C. (H.I.M.) a, Nancy White, Ph.D. e, Louisa Degenhardt, Ph.D. a,f,g,h a

National Drug and Alcohol Research Centre, UNSW, NSW, Australia Sydney Medical School, Sydney University, NSW, Australia c The Langton Centre, South East Sydney Local Health District (SESLHD) Drug and Alcohol Services, NSW, Australia d School of Medicine, University of Tasmania, Tasmania, Australia e University of Adelaide, South Australia, Australia f School of Population and Global Health, University of Melbourne, Victoria, Australia g Murdoch Children's Research Institute, Australia h Department of Global Health, School of Public Health, University of WA, USA b

a r t i c l e

i n f o

Article history: Received 27 February 2015 Received in revised form 29 May 2015 Accepted 1 June 2015 Keywords: Abuse deterrent Tamper resistant Pharmaceutical opioids Opioids Abuse liability Cohort

a b s t r a c t Aims: To describe the methods and baseline characteristics of a cohort of people who tamper with pharmaceutical opioids, formed to examine changes in opioid use following introduction of Reformulated OxyContin®. Methods: Participants were 606 people from three Australian jurisdictions who reported past month injecting, snorting, chewing or smoking of a pharmaceutical opioid and had engaged in these practices at least monthly in the past 6 months. Baseline interviews were conducted prior to introduction of Reformulated OxyContin® in April 2014. Patterns of opioid use and cohort characteristics were examined according to whether participants were prescribed opioid medications, or exclusively used diverted medication. Results: The cohort reported high levels of moderate/severe depression (61%), moderate/severe anxiety (43%), posttraumatic stress disorder (42%), chronic pain or disability (past 6 months, 54%) and pain (past month, 47%). Lifetime use of oxycodone, morphine, opioid substitution medications and codeine were common. Three-quarters (77%) reported ICD-10 lifetime pharmaceutical opioid dependence and 40% current heroin dependence. Thirteen percent reported past year overdose, and 70% reported at least one past month opioid injection-related injury or disease. The cohort displayed complex clinical profiles, but participants currently receiving opioid substitution therapy who were also prescribed other opioids particularly reported a wide range of risk behaviors, despite their health service engagement. Conclusions: Findings highlight the heterogeneity in the patterns and clinical correlates of opioid use among people who tamper with pharmaceutical opioids. Targeted health interventions are essential to reduce the associated harms. © 2015 Published by Elsevier Inc.

1. Introduction Pharmaceutical opioids have a legitimate role in medicine and make a positive contribution to the health of many people, yet they are not always used in the manner in which they were intended when prescribed (Larance, Degenhardt, Lintzeris, Winstock, & Mattick, 2011). Pharmaceutical opioids are a significant contributor to drug-related morbidity and mortality in many countries (Chen, Hedegaard, & Warner, 2014; Gomes et al., 2014; Larance, Ambekar, et al., 2011; Larance, Degenhardt, et al., 2011; Manchikanti et al., 2012; Roxburgh, Bruno, Larance, & Burns, 2011). These harms are partly attributed to diversion of opioids from those to whom they were prescribed; and tampering for use via routes other than as ⁎ Corresponding author at: National Drug and Alcohol Research Centre, University of New South Wales, Sydney NSW 2052. Tel.: +61 2 9385 0241; fax: +61 2 9385 0222. E-mail address: [email protected] (B. Larance). http://dx.doi.org/10.1016/j.jsat.2015.06.001 0740-5472/© 2015 Published by Elsevier Inc.

intended, particularly via injection (Madadi, Hildebrandt, Lauwers, & Koren, 2013). Pharmaceutical companies are consequently developing formulations intended to be less prone to tampering (e.g., crushing, chewing, snorting, smoking, injecting or dissolving/drinking opioid medications intended for oral administration) (Katz et al., 2011), and misuse. One example of an opioid formulation specifically designed to be tamper-resistant is a reformulated version of controlled-release oxycodone hydrochloride tablets (Reformulated OxyContin®). The tablets are designed to be bioequivalent to the original formulation but employ a controlled-release technology (that makes them difficult to crush) with a hydro-gelling matrix (so the tablet develops into a viscous gel when dissolved in water) (Sellers, Perrino, Colucci, & Harris, 2013), reducing the likelihood of a rapid release of oxycodone following tampering (Cone, Giordano, & Weingarten, 2013). Early US studies indicated lower levels of misuse (Havens, Leukefeld, DeVeaugh-Geiss, Coplan, & Chilcoat, 2014), attractiveness (Sellers et al., 2013) and street price

52

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

(Severtson et al., 2013) among specific populations, and a decrease in OxyContin® poisonings following introduction of the reformulated product (Severtson et al., 2013). In April 2014, Reformulated OxyContin® was introduced onto the Australian market. The National Opioid Medications Abuse Deterrence (NOMAD) study was established to monitor potential changes post reformulation introduction (Degenhardt et al., 2015). There are multiple data collections involved in this study, including formation of a prospective cohort of people who regularly tamper with pharmaceutical opioids, interviewed prior to April 2014. Prospective cohort studies have the advantage of collecting reliable and detailed data close to an event of interest, reducing potential recall bias (Euser et al., 2009). This cohort also addresses a significant gap in the existing literature, in that there has been no comprehensive assessment in Australia of the characteristics, patterns and clinical correlates of opioid use among people who tamper with pharmaceutical opioids. We recruited people who reported obtaining varied pharmaceutical opioids from a range of different sources: both prescribed by a doctor, and diverted from other sources. This is particularly important from a public health perspective, in that those who are prescribed opioids have existing health service engagement, offering greater opportunity for direct harm minimization strategies. As such, the specific aims of this paper are to: 1. describe the protocol for recruitment and strategies for retention of the cohort; 2. describe the baseline demographic and clinical characteristics of the cohort; and 3. examine potential differences among cohort participants according to whether participants were prescribed opioids or used exclusively diverted medication.

2. Materials and methods This study utilizes baseline data collected from the cohort component of the NOMAD study. The broader NOMAD study is detailed elsewhere (Degenhardt et al., 2015). 2.1. Data from people who tamper with pharmaceutical opioids: The NOMAD cohort protocol 2.1.1. Cohort design The NOMAD cohort is a prospective cohort of 606 people who regularly tamper with opioid analgesics. The cohort included people who reported chewing, snorting, smoking and/or injecting opioid medications (see Eligibility criteria below). The cohort was formed and interviewed prior to the introduction of Reformulated OxyContin® (January–March 2014). They will be followed up on two further time points following the change in formulation: 1 to 3 months post-release (May 2014 to July 2014) to capture any changes that occur in the initial period following withdrawal of the old formulation; and 12 to 15 months post-release (May to July 2015), to capture the longer term, more stable patterns of drug use that this group might develop. 2.1.2. Eligibility criteria Eligibility criteria comprised: (a) English language proficiency, (b) aged 18 years or older, (c) reported extra-medical pharmaceutical opioid use (i.e., use outside the bounds of a doctor's prescription (Larance, Ambekar, et al., 2011; Larance, Degenhardt, et al., 2011)) at least monthly in the past 6 months, and (d) reported past month injecting, snorting, chewing or smoking of a pharmaceutical opioid and had engaged in these practices at least monthly in the past 6 months. Participants were excluded if they reported only using their opioid medication in accordance with a doctor's instructions, had only tampered with an opioid substitution therapy (OST) medication (i.e., methadone or buprenorphine/naloxone for the treatment of opioid dependence), had not been a resident of the city/state for 6 consecutive months prior to the interview, or were in prison for the past month.

2.1.3. Recruitment The cohort was recruited from three Australian jurisdictions: Sydney, New South Wales (NSW); Adelaide, South Australia (SA); and Hobart and Launceston, Tasmania (TAS). Participants were recruited through a variety of settings including Needle-Syringe Programs (NSPs) (38%), snowballing and word-of-mouth (36%), OST clinics/prescribers (15%), community pharmacies (6%), advertisements in newspapers and street media (4%), and other health/outreach services (1%). Pharmacy recruitment was conducted utilizing an Australian database of community pharmacies purchased for an existing study (Campbell et al., 2014). The study team telephoned pharmacies in NSW, SA and TAS, and posted packs with flyers and additional information about the study to pharmacists interested in assisting with recruitment. Pharmacists were asked to distribute the flyer to any person filling prescriptions for pharmaceutical opioid medications. All flyers had a unique pharmacy number, and pharmacists were reimbursed $20 for each eligible participant they referred, regardless of whether the person ultimately participated. Interested participants were referred to the research team by making direct contact (by phone) or by opting to leave their details at the recruiting service/pharmacy so the research team could contact them. All interested participants were provided more information about the study and assessed for eligibility by NOMAD staff. Eligible participants went through a voluntary informed consent process before interview times were arranged. 2.1.4. Interview procedure Baseline interviews were conducted from December 2013 to March 2014. Participants completed structured face-to-face interviews, with data entered directly onto laptops utilizing computer-assisted interview software. Interviews took approximately 1.5 hours to complete. Medication prompt cards with photographs of all pharmaceutical opioid medications and a range of benzodiazepines were used to help participants identify the correct medication and tablet size in answering questions about recent use. All interviews were completed by trained interviewers with a degree in behavioral sciences and training in suicide response. Participants were reimbursed AUD$50 on concluding the interview for time and out-of-pocket expenses incurred. 2.1.5. Cohort maintenance strategies A number of methods were used to prevent attrition in the NOMAD cohort. Firstly, the study team attempted to make contact with interested participants within a week of their expression of interest in the study. A follow-up locator form was completed at initial contact, with comprehensive contact details of the participant and at least two secondary contacts (e.g., relatives, friends and medical professionals). Other methods to improve retention included letters thanking the participant for study involvement and regular contact from the study team between time points to check for changes in contact details and provide study updates. 2.2. NOMAD study governance 2.2.1. NOMAD advisory committee An advisory committee was established for the wider NOMAD study. The group included general practitioners, pain specialists, addiction medicine specialists and researchers. The committee met at the beginning of the study to assist in the design of the study and interview schedule, and assist in developing recruitment strategies. 2.2.2. Ethics approval This study has approval from the Ethics Review Committee (Royal Prince Alfred Hospital Zone) of the Sydney Local Health District, Australian Department of Health, AIDS Council of New South Wales, Tasmanian Health and Medical Human Research Ethics Committee, University of Adelaide Human Research Ethics Committee and Southern Adelaide Clinical Human Research Ethics Committee. Access and site approvals were obtained from the following local area health ethics

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

committees governing clinic sites: Sydney Local Health District, South Eastern Sydney Local Health District, South Western Sydney Local Health District and Western Sydney Local Health District. 2.3. Key measures used in this study A detailed description of the full measures collected via the NOMAD cohort interviews is provided elsewhere (Degenhardt et al., 2015). The following lists the measures utilized in this paper. 2.3.1. Cohort characteristics The chronic conditions section from the Composite International Diagnostic Interview (CIDI) was included in the interview to assess whether participants had ever developed a range of physical conditions, such as chronic back or neck problems and diabetes (Kessler & Ustun, 2004). The Brief Pain Inventory (BPI) Severity and Interference subscales were used to assess pain (Tan, Jensen, Thornby, & Shanti, 2004). Depression and generalized anxiety disorder were measured using the PHQ-9 and the Generalized Anxiety Disorder (GAD-7) modules of the Patient Health Questionnaire (Kroenke, Spitzer, Williams, & Lowe, 2010). Validated cut-offs were used for screening tools as follows: symptoms indicating likely presence of moderate to severe depression were defined as a PHQ-9 score of ≥ 10 (Kroenke, Spitzer, & Williams, 2001); and likely symptoms of moderate to severe anxiety were defined as a GAD-7 score of ≥ 10 (Spitzer, Kroenke, Williams, & Lowe, 2006). Post-traumatic stress disorder (PTSD) was measured using the Primary Care PTSD screen (PC-PTSD), with a score N 3 indicating possible presence of PTSD (Prins et al., 2004). 2.3.2. Substance use and use disorders Interviews asked about lifetime and past month prescription and illicit use of pharmaceutical opioids, including types and methods of tampering. Data were also collected on lifetime and past 6 month use of other licit and illicit substances, such as alcohol, heroin and methamphetamine. The Alcohol Use Disorders Identification Test (AUDIT-C) was included, with a score of ≥5 indicative of alcohol-related problems. The Severity of Dependence Scale (SDS) was used to assess severity of dependence on heroin, benzodiazepines and methamphetamines in the past 3 months. Cut off scores of 5 for heroin (Gossop et al., 1995), 7 for benzodiazepines (Gossop et al., 1995) and 4 for methamphetamine (Topp & Mattick, 1997) were used as indicating meeting criteria for current dependence. The drug use disorder section of the CIDI interview was used to determine lifetime and past 12 month ICD-10 dependence diagnoses for pharmaceutical opioids (Kessler & Ustun, 2004). 2.3.3. Drug use behaviors, harms and health service access Interviews collected data regarding past 6 month methods of obtaining diverted (i.e. non-prescribed) pharmaceutical opioids. Participants reported past month injection-related risk behaviors (sharing and re-using needles and other equipment), and experience of any injection-related injury and disease (IRID; classified as non-serious, potentially serious, and serious) after pharmaceutical opioid injection in the past month (see Dwyer et al., 2009). Experience of overdose (lifetime and past 12 months) and access to a range of health services in the preceding 6 months were also assessed.

tapentadol. Four clinically-relevant groups were apparent in the sample, namely (i) ‘non-prescribed’ group (i.e. only used pharmaceutical opioids diverted by someone else, n = 172); (ii) ‘OST only’ group (i.e., were prescribed OST opioid only, n = 186); (iii) ‘OST and other opioids’ group (i.e. were prescribed both OST and other opioid medications, n = 104); and (iv) ‘other opioids only’ group (i.e. not prescribed OST but were prescribed other opioid medications, n = 141, treated as the reference group). Participants in groups (ii), (iii) and (iv) could have additionally used diverted opioid medications, but all participants in these groups were prescribed at least one opioid medication, and all by definition were tampering with opioid medications. Differences between proportions were examined using odds ratios (OR) with 95% confidence intervals (95% CI). For normally-distributed continuous data (e.g. age), means, standard deviation (SD) and parametric tests of significance (independent samples t test) are reported. For skewed data, medians and non-parametric tests of significance (Mann–Whitney U) are reported. 3. Results The research team made contact with 1,176 of 1,321 people who expressed interest in the study. A total of 692 participants were eligible to participate in the cohort. Of those, 606 completed baseline interviews: 303 in NSW, 150 in SA and 153 in TAS. A flowchart of recruitment to the NOMAD cohort is provided in Fig. 1. 3.1. Demographic and clinical characteristics of the NOMAD cohort 3.1.1. Total sample The mean age of the NOMAD cohort was 40 years (SD 9.3), and twothirds were male (69%). The cohort had high levels of socio-economic

Participants referred into study n= 1321

Not assessed (n=145) Could not contact n=123 No longer interested n=22

Participants assessed for eligibility n= 1176 Ineligible (n=484) In prison past month n=11 Had not used/tampered with pharmaceutical opioids n=283 Only used OST medication n=146 Had not used/tampered in past 6 months n=42 Not living in study jurisdictions n=2

Eligible participants n= 692

2.4. Analyses Descriptive analyses were undertaken for the overall sample, and for comparing four independent groups of pharmaceutical opioid users who differed according to medication source (prescribed versus nonprescribed) and type of pharmaceutical opioids used in the past month. Methadone syrup, buprenorphine tablets, and buprenorphine–naloxone tablet or film preparations were classified as OST medications. ‘Other pharmaceutical opioids’ included oxycodone, morphine, methadone tablets (Physeptone®), buprenorphine patches (Norspan®), prescription codeine, fentanyl, tramadol, hydromorphone, dextropropoxyphene and

53

Loss after eligibility (n=78) Refused consent n=2 Could not contact n=68 Unable to consent (e.g. intoxication) n=6 Loss of data n=2

Completed baseline interview n= 606

Fig. 1. Recruitment flowchart for the NOMAD cohort.

54

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

Table 1 Demographic and clinical characteristics of the NOMAD cohort (n = 606), by source and type of opioid prescribed in previous month.

Demographics Mean age (SD) % Male % Unemployed % Pension/Disability benefit % Education b year 10 % Income b $400 per week % Homeless/no fixed abode % Prison history Physical health % Ever diagnosed with… Arthritis or rheumatism Chronic back or neck problems Chest pain Frequent/Severe headaches/migraines Visceral pain Hepatitis C Asthma Stroke Heart attack Sleep apnea Heart disease High blood pressure Chronic lung disease Diabetes Cancer (excl. non-melanoma skin) Pelvic pain % Chronic pain/disability past 6 months % (Non-everyday type) pain past month BPI Severity Score2 (Median, IQR) (N = 279) BPI Interference Score2 (Median, IQR) (N = 276) Current mental health % Moderate/Severe depression % Moderate/Severe anxiety % PTSD

TOTAL sample N = 606

A. ‘Other opioids only’ group1 n = 141

B. ‘OST only’ group n = 186

C. ‘OST and other opioids’ group1 n = 104

D. ‘Not prescribed’ group n = 172

A (ref) vs. B OR (95% CI) or Z-score

A (ref) vs. C OR (95% CI) or Z-score

A (ref) vs. D OR (95% CI) or Z-score

40 (9.3) 69 82 90 29 60 18 62

41 (8.6) 68 83 89 27 54 20 55

41 (8.9) 67 86 90 30 58 16 63

40 (9.3) 63 78 91 27 54 12 72

38 (9.8) 77 81 89 33 69 22 60

t(325) = 0.25 0.96 (0.60–1.53) 1.23 (0.68–2.24) 1.21 (0.60–2.48) 1.15 (0.70–1.86) 1.17 (0.75–1.84) 0.73 (0.41–1.30) 1.44 (0.92–2.25)

t(243) = 0.89 0.80 (0.47–1.36) 0.74 (0.39–1.39) 1.37 (0.58–3.24) 0.99 (0.56–1.76) 0.99 (0.59–1.67) 0.52 (0.25–1.07) 2.14 (1.25–3.69)⁎⁎

t(311) = 3.05⁎⁎ 1.58 (0.96–2.61) 0.88 (0.49–1.57) 1.05 (0.52–2.12) 1.28 (0.79–2.10) 1.86 (1.16–2.96)⁎⁎

30 48 28 39 29 61 30 5 4 14 3 17 8 6 6 11 54 47 5 (2.50)

40 75 37 44 34 64 38 7 5 17 4 17 14 8 7 16 80 73 6 (2.25)

31 39 27 38 26 68 31 3 6 15 2 18 8 5 7 10 46 38 4 (2.25)

35 54 28 52 34 71 35 6 5 18 5 21 6 7 6 17 60 48 6 (2.00)

19 33 23 28 24 46 19 3 1 10 2 15 4 4 2 6 38 35 5 (2.75)

0.66 (0.42–1.05) 0.21 (0.13–0.35)⁎⁎⁎ 0.64 (0.40–1.03) 0.77 (0.50–1.22) 0.70 (0.43–1.14) 1.20 (0.75–1.92) 0.72 (0.45–1.15) 0.43 (0.15–1.22) 1.19 (0.45–3.15) 0.86 (0.47–1.57) 0.49 (0.14–1.76) 1.05 (0.58–1.89) 0.56 (0.27–1.14) 0.59 (0.24–1.47) 0.97 (0.41–2.29) 0.57 (0.29–1.11) 0.22 (0.13–0.36)⁎⁎⁎ 0.23 (0.14–0.37)⁎⁎⁎ Z = −4.88⁎⁎⁎

0.80 (0.47–1.36) 0.39 (0.23–0.68)⁎⁎ 0.68 (0.39–1.19) 1.36 (0.81–2.27) 1.02 (0.59–1.75) 1.36 (0.78–2.35) 0.88 (0.52–1.50) 0.79 (0.28–2.24) 0.95 (0.29–3.08) 1.05 (0.53–2.07) 1.11 (0.33–3.76) 1.35 (0.70–2.58) 0.38 (0.15–1.00) 0.84 (0.31–2.24) 0.79 (0.28–2.24) 1.03 (0.52–2.06) 0.39 (0.22–0.69)⁎⁎ 0.34 (0.20–0.58)⁎⁎⁎

0.34 (0.21–0.57)⁎⁎⁎ 0.16 (0.10–0.27)⁎⁎⁎ 0.51 (0.31–0.85)⁎ 0.50 (0.31–0.81)⁎⁎

Z = −0.56

0.62 (0.38–1.03) 0.47 (0.30–0.75)⁎⁎ 0.39 (0.23–0.65)⁎⁎⁎ 0.38 (0.13–1.15) 0.22 (0.05–1.08) 0.55 (0.28–1.08) 0.52 (0.15–1.89) 0.89 (0.48–1.64) 0.27 (0.11–0.65)⁎⁎ 0.42 (0.15–1.16) 0.30 (0.09–0.99) 0.33 (0.15–0.71)⁎⁎ 0.16 (0.09–0.26)⁎⁎⁎ 0.20 (0.12–0.33)⁎⁎⁎ Z = −3.53⁎⁎⁎

5 (4.29)

6 (3.86)

4 (4.18)

6 (3.43)

4 (3.71)

Z = −4.28⁎⁎⁎

Z = −0.95

Z = −4.60⁎⁎⁎

61 43 42

62 43 45

64 46 46

71 48 45

52 39 36

1.11 (0.70–1.76) 1.12 (0.71–1.76) 1.06 (0.67–1.68)

1.51 (0.87–2.62) 1.20 (0.71–2.02) 1.01 (0.60–1.69)

0.67 (0.42–1.06) 0.84 (0.53–1.34) 0.69 (0.43–1.11)

1.09 (0.63–1.89) 1.24 (0.79–1.95)

Note that those in the ‘not prescribed’ group had ONLY used diverted pharmaceutical opioids, and that participants in the other three groups could have additionally used diverted pharmaceutical opioids (i.e. not prescribed to them by a doctor). PSTD: Post-traumatic stress disorder; OST: opioid substitution therapy; OR = odds ratio; 95% CI = 95% confidence interval; ns = not significant. Bold indicates a significant estimate. 1 ’Other opioids’ include: oxycodone, morphine, methadone tablets, buprenorphine patches, prescription codeine, fentanyl, tramadol, hydromorphone, dextropropoxyphene and tapentadol. 2 BPI score among those who reported pain in past month (N = 282). ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

disadvantage on a range of indicators (Table 1). Over two-thirds of the cohort (71%) had been prescribed an opioid medication in the past month, and just under half (48%) reported receiving past month OST. There was a high prevalence of past month mental health problems: three in five participants reported moderate/severe depression (61%) and two in five reported moderate/severe anxiety (43%) and PTSD (42%) (Table 1). The cohort also reported a high prevalence of physical health conditions (Table 1). Over half the sample reported having chronic pain or disability in the past 6 months (54%) and just under half reported experiencing non-everyday types of pain in the past month (47%). 3.1.2. Comparison of opioid consumer groups Demographic characteristics of the opioid consumer groups were similar, with the exception that the ‘not prescribed’ group were significantly younger and more likely to report low income, and the ‘OST and other opioids’ group were more likely to report a prison history, when compared to the ‘other opioids only’ group. The ‘other opioids only’ group reported higher levels of chronic pain or disability in the past 6 months (80%), past month pain (73%) and chronic back or neck

problems (75%), compared to the other groups. Compared to the ‘other opioids only’ group, the ‘not prescribed’ group were less likely to report a range of chronic health problems (including arthritis or rheumatism, chest pain, frequent and severe headaches, hepatitis C, asthma, chronic lung disease, cancer and pelvic pain). There were no differences in levels of mental health problems for the ‘other opioids only’ group compared to the other three groups. 3.2. Patterns of alcohol and illicit drug use and substance use disorders 3.2.1. Total sample Over three-quarters (72%) reported past 6 month alcohol use on a median of 12 days. The cohort had extensive illicit drug use histories; 99% reported past 6 months injection of illicit drugs, and the majority reported lifetime and past 6 month use of cannabis (99% and 80% respectively), methamphetamine (97% and 73% respectively) and heroin (89% and 64% respectively) (Table 2). Just over three-quarters (77%) met ICD-10 criteria for lifetime pharmaceutical opioid dependence. Two-fifths (40% and 38%) of the cohort met SDS cut-offs for current heroin dependence and AUDIT-C cut off

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

55

Table 2 Drug use disorders among the NOMAD cohort, by source and type of opioid prescribed in previous month.

% Lifetime ICD10 pharmaceutical opioid dependence % past 12 months ICD-10 dependence % Injected (any drug) in past 6 months % Risky drinker (AUDIT-C score ≥3) % Heroin dependence (SDS score ≥5) % Methamphetamine dependence (SDS score ≥4) % Benzodiazepine dependence (SDS score ≥7)

Total N = 606

A. ‘Other opioids only’ group1 n = 141

B. ‘OST only’ group n = 186

C. ‘OST and other opioids’ group1 n = 104

D. ‘Not prescribed’ group n = 172

A (ref) vs. B OR (95% CI) or Z-score

A (ref) vs. C OR (95% CI) or Z-score

A (ref) vs. D OR (95% CI) or Z-score

77

72

78

83

75

1.42 (0.84–2.40)

1.97 (1.03–3.75)⁎

1.19 (0.71–1.99)

51 99 38 40 26

47 99 33 26 23

47 99 35 46 20

58 99 35 51 30

54 98 48 39 32

0.98 (0.62–1.55) 1.34 (0.19–9.65) 1.10 (0.69–1.76) 2.50 (1.54–4.05)⁎⁎⁎

1.53 (0.91–2.57) 1.50 (0.14–16.81) 1.08 (0.63–1.86) 3.00 (1.74–5.17)⁎⁎⁎

1.31 (0.83–2.07) 0.82 (0.14–4.99) 1.88 (1.18–3.00)⁎⁎ 1.83 (1.12–3.00)⁎

0.80 (0.47–1.37)

1.41 (0.79–2.52)

1.54 (0.93–2.57)

19

13

20

37

14

1.63 (0.88–3.01)

3.77 (2.00–7.13)⁎⁎⁎

1.02 (0.53–1.98)

Note that those in the ‘not prescribed’ group had only used diverted pharmaceutical opioids, and that participants in the other three groups could have additionally used diverted pharmaceutical opioids (i.e. not prescribed to them by a doctor). ICD: International Classification of Diseases; AUDIT-C: Alcohol Use Disorders Identification Test short form; OST: opioid substitution therapy; OR = odds ratio; 95% CI = 95% confidence interval; ns = not significant. Bold indicates a significant estimate. 1 ’Other opioids’ include: oxycodone, morphine, methadone tablets, buprenorphine patches, prescription codeine, fentanyl, tramadol, hydromorphone, dextropropoxyphene and tapentadol. ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

scores for risky drinking, one-quarter (26%) for methamphetamine dependence and one-fifth (19%) for benzodiazepine dependence. 3.2.2. Comparison of opioid consumer groups The ‘other opioids only’ group had lower levels of recent heroin use, injection and dependence than the three other groups. The ‘other opioids only’ group was also less likely than the two OST groups (‘OST only’ and ‘OST and other opioids’) to report other forms of illicit drug use and dependence. 3.3. Patterns of pharmaceutical opioid and benzodiazepine use 3.3.1. Total sample The vast majority of the cohort reported lifetime use of oxycodone (99%), and 83% reported past month use on a median of 13 days. Past month oxycodone injection was common (78%), and around one in ten participants reported past month snorting/smoking/chewing of oxycodone (Table 3). Other opioid medications most commonly used by the cohort included morphine, methadone syrup, buprenorphine tablets/film and codeine (Table 3). The least commonly used opioids were hydromorphone, dextropropoxyphene and tapentadol, with reports of past month use rare. Among those only using diverted opioids (‘not prescribed’ group), the most frequently used opioids (among past month users) were oxycodone (median 14 days) and morphine (median 8 days). Lifetime (93%) and past month benzodiazepine use was common (73%), used on a median of 11 days in the past month. Six percent reported past month injection of benzodiazepines, and 12% had snorted/smoked/chewed benzodiazepines.

significantly larger number of different opioid medications via other routes of administration, higher odds of lifetime morphine use, with greater prevalence of other routes of administration for buprenorphine tablets/film and codeine (Table 3). The ‘not prescribed’ group had similar levels of oxycodone use to the ‘other opioids only’ group (although lower levels of past month snorting/ chewing/smoking/dissolving oxycodone). However, the ‘not prescribed’ group reported using fewer different (lifetime and past month) opioids and were less likely to report ever using methadone syrup and tablets or recent use of lower-strength pharmaceutical opioids (i.e., codeine and tramadol). 3.4. Methods of obtaining diverted pharmaceutical opioids 3.4.1. Total sample The most common methods of obtaining diverted pharmaceutical opioids among the NOMAD cohort were buying from a dealer (81%) or a patient who sells them (72%) (Table 4). 3.4.2. Comparison of opioid consumer groups The ‘other opioids only’ and the ‘OST only’ groups did not differ in their methods for obtaining diverted pharamceutical opioids. However, the ‘OST and other opioids’ group reported higher levels of prescription forgery, obtaining prescriptions from multiple doctors and buying from a patient who sells them compared to the ‘other opioids only’ group. In contrast, the ‘not prescribed’ group reported higher odds of buying opioids from a dealer and buying them from a patient who sells them compared to the ‘other opioids only’ group. 3.5. Health service utilization

3.3.2. Comparison of opioid consumer groups Compared to the ‘other opioids only’ group, all three groups reported using significantly larger numbers of different opioid medications in the past month. As expected, those groups receiving substitution treatment for opioid dependence (‘OST only’ and ‘OST and other opioids’) generally reported more recent and frequent use and injection of methadone syrup and buprenorphine tablets/film than the ‘other opioids only’ group. The ‘OST only’ group typically had lower levels of most types of pharmaceutical opioid use in the past month compared to the ‘other opioids only’ group. In general, those using ‘OST and other opioids’ reported similar levels and patterns of oxycodone and other opioid consumption to the ‘other opioids only’ group. However, this group had recently used a

3.5.1. Total sample There were very high levels of health service utilization among the cohort; the majority (84%) of participants reported that they had seen a general practitioner in the past 6 months, and one-third (32%) reported at least one hospital admission in the past 6 months (Table 4). 3.5.2. Comparison of opioid consumer groups The ‘OST only’ group were less likely to see a pain specialist in the past 6 months, and both OST groups (‘OST only’ and ‘OST and other opioids’) reported higher levels of utilizing outpatient drug and alcohol services and addiction specialists in the past 6 months, compared to the ‘other opioids only’ group. Compared to the ‘other opioids only’ group,

56

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

the ‘not prescribed’ group were less likely to report seeing a general practitioner, a pain specialist or being admitted to hospital in the past 6 months. 3.6. Experience of harms 3.6.1. Total sample One in nine participants (13%) had experienced an overdose in the past year. Over two-thirds (70%) had experienced at least one opioid injection-related injury or disease (IRID) in the month prior to the interview. 3.6.2. Comparison of opioid consumer groups The ‘OST and other opioids’ group reported higher levels of past year overdose than the ‘other opioids only’ group. There were no differences in the reporting of injecting risk behaviors or IRIDs by type of opioid prescribed. 4. Discussion The NOMAD cohort is a novel sample of people who tamper with opioid medications. There were high levels of pharmaceutical opioid dependence in this cohort, and considerable illicit drug use and pain histories. Although tampering for injection was reported by the majority of cohort participants, one in ten also reported tampering with opioid medications to administer via other routes (e.g., snorting, smoking, chewing). A key finding was the heterogeneity of cohort, with diverse and complex clinical profiles. In general, the cohort was characterized by marked socio-economic disadvantage, significant physical and mental health comorbidities, pain and high levels of health service utilization. Participants in the ‘OST and other opioids’ group in particular reported a wide range of risk behaviors. However, the cohort as a whole displayed complex patterns of alcohol, pharmaceutical opioid and illicit drug use and dependence. The opioid medications most commonly used by the cohort included oxycodone, morphine, methadone syrup, buprenorphine tablets/film and codeine. The least commonly used opioids were hydromorphone, dextropropoxyphene and tapentadol. Although recent use of oxycodone was high, pharmaceutical opioids (excluding methadone and buprenorphine) were generally not used daily—typically every second day in the past month for oxycodone. Patterns of oxycodone tampering varied within the cohort, with the ‘OST only’ group being less likely to inject oxycodone than the ‘other opioids only’ group, and the ‘not prescribed’ group being less likely to use other routes of administration. In addition, the cohort had typically used a range of different opioids (including heroin). As Reformulated OxyContin® is currently the only tamper-resistant opioid formulation introduced in Australia, it is possible that in a market where a range of different opioids are used and available, displacement of drug use and preferences may occur over time. Consistent with previous studies (Stafford & Burns, 2014), dealers and secondary access from other patients were clearly the primary sources of diverted opioid medications in this cohort. In contrast, prescriptions from multiple doctors or from dishonest prescribers were uncommon. For this reason, prescription monitoring programs will have limited impact in identifying the range of people who use opioids extramedically. However, prescription monitoring programs may have utility in identifying the specific sub-group of OST clients who are also prescribed other opioids (or other risky medication combinations). In the current study, this group reported risky patterns of drug use and high levels of overdose, and targeted interventions are needed to reduce the burden of opioid dependence in this group. This study indicates that people who tamper with pharmaceutical opioids experience considerable harms. There were high levels of poly-sedative use and frequent injection in this group; over two-thirds of the cohort reported experiencing IRID (11% reporting serious IRID),

and more than one in ten reported past year overdose. It is unlikely that adopting abuse-deterrent products as a sole strategy to cease extra-medical pharmaceutical opioid use will be sufficient. Harm minimization strategies are particularly important for this group, and might include OST, safer injecting advice and the provision of naloxone to patients being dispensed strong opioid medications (Dwyer et al., 2009; Lenton et al., 2014). This study is subject to the limitations of self-report, although selfreport has been found to be sufficiently valid and reliable in studies of people who use pharmaceutical opioids extra-medically (Passik, Hays, Eisner, & Kirsh, 2006) and/or use illicit drugs (Darke, 1998). Despite our best efforts, we did not recruit pain patients without extensive illicit drug use histories who were tampering with their opioid medication, so we were unable to quantify the proportion of pain patients who, for example, break/tamper with their controlled-release tablet to make it easier to swallow. However, other Australian data indicate that around 9% of pain patients report altering their dose in some way (e.g. cutting patches or pills in half) and 4% report taking it by an alternative route (e.g. injecting it) (Belcher et al., 2014); suggesting that these practices are rare among pain patients. Data collection was based in three major cities in Australia, where it was known that the availability of heroin differs markedly (Stafford & Burns, 2014). The patterns of drug use and reported rates of extramedical use of pharmaceutical opioid use among people who inject drugs similarly vary across Australian cities (Stafford & Burns, 2014). Additionally, use of pharmaceutical opioids has been found to be higher in rural and regional areas (Day, Conroy, Lowe, Page, & Dolan, 2006). To the extent that OST may be less available in rural or regional areas, it could also be that there is less OST available, and less use of prescribed and diverted OST medications. It is unclear how this may relate to differences among samples of people using pharmaceutical opioids. Furthermore, there does not seem to be any a priori reason to expect that there would be differences in the associations observed with the range of sociodemographic, substance use and mental health variables examined here. 5. Conclusions This is the first large-scale study of trajectories and outcomes among a cohort of people who tamper with pharmaceutical opioids. Ongoing monitoring of harms in this cohort is essential in evaluating the intended and unintended consequences of introducing Reformulated OxyContin®. Heterogeneity in the patterns and clinical correlates of opioid use suggest that the introduction may have differential effectiveness in reducing tampering and associated harms. Many cohort participants were engaged with health services and obtained opioids via prescription, presenting opportunity for interventions to improve health and clinical outcomes. Acknowledgements Many thanks go to the NOMAD cohort participants who generously agreed to share their experiences. Thanks also to staff and services that assisted with recruitment of the NOMAD cohort: the pharmacies who assisted with participant referral; the Sydney Medically Supervised Injecting Centre (MSIC); the Langton Centre, Sydney; the Sydney Wayside Chapel; Kirketon Road Centre, Sydney; Redfern REPIDU; Canterbury REPIDU; Canterbury Opioid Treatment Clinic; St George Drug and Alcohol Service; RPAH Opioid Treatment Clinic; Western Sydney LHD staff; Fleet Street clinic; Jacaranda House; Liverpool Health Connexions; Coopers Cottage Campbelltown; Bankstown Community Health Centre NSP; Scott Street Clinic; Clinic 36; the AIDS Council of NSW; Garden Court Clinic; NSW Users and AIDS Association; Drug And Alcohol Services South Australia; TUHSL; Salvation Army Launceston; Alcohol, Tobacco and other Drugs Council of Tasmania; Anglicare Tasmania; Public Health, Department of Health and Human Services, Tasmania.

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

57

Table 3 Patterns of pharmaceutical opioid use reported by the NOMAD cohort, by source and type of opioid prescribed in previous month.

Total number of pharmaceutical opioids Lifetime Median no. used (Range) Median no. injected (range) Median no. snorted/smoked/ chewed (range) Past month Median no. used (range) Median no. injected (range) Median no. snorted/smoked/chewed/ dissolved (range) Oxycodone % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (range) Among whole sample Among past month users Morphine % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (range) Among whole sample Among past month users Methadone syrup % Used ever % Used past month % Injected past month Median days past month (range) Among whole sample Among past month users Methadone tablets (Physeptone®) % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (Range) Among whole sample Among past month users Buprenorphine tablets/film (OST) % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (Range) Among whole sample Among past month users Buprenorphine patch (Norspan) % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (range) Among whole sample Among past month users

Total N = 606

A. ‘Other opioids only’ group1 n = 141

B. ‘OST only’ group n = 186

C. ‘OST and other opioids’ group1 n = 104

D. ‘Not prescribed’ group n = 172

A (ref) vs. B OR (95% CI) or Z-score

A (ref) vs. C OR (95% CI) or Z-score

A (ref) vs. D OR (95% CI) or Z-score

6 (1–11) 4 (0–10) 0 (0–9)

6 (1–11) 4 (0–8) 0 (0–7)

6 (3–10) 4 (1–9) 0 (0–6)

7 (4–11) 4 (1–10) 1 (0–9)

6 (2–11) 4 (0–9) 0 (0–6)

Z = −1.17 Z = 1.21 Z = 0.88

Z = 1.18 Z = 2.35⁎ Z = 2.35⁎

Z = −3.87⁎⁎⁎ Z = −0.79 Z = 0.64

3 (1–10) 2 (0–10) 0 (0–8)

3 (1–7) 2 (0–5) 0 (0–3)

3 (2–7) 2 (0–5) 0 (0–3)

4 (2–10) 2 (0–10) 0 (0–8)

3 (1–10) 2 (0–9) 0 (0–3)

Z = −2.23⁎ Z = −2.00⁎ Z = −0.85

Z = −4.01⁎⁎⁎ Z = 1.48 Z = 1.91

Z = −3.91⁎⁎⁎ Z = 0.03 Z = −1.06

99 83 78 11

99 86 84 13

98 76 72 9

100 88 80 19

99 84 81 6

0.89 (0.15–5.40) 0.50 (0.28–0.92)⁎ 0.51 (0.30–0.89)⁎ 0.68 (0.34–1.37)

– 1.11 (0.52–2.36) 0.78 (0.41–1.51) 1.60 (0.80–3.21)

2.50 (0.22–27.82) 0.85 (0.45–1.60) 0.84 (0.46–1.50) 0.46 (0.21–1.01)

9 (0–28) 13 (1–28)

14 (0–28) 20 (1–28)

5 (0–28) 9 (1–28)

11 (0–28) 14 (1–28)

10 (0–28) 14 (1–28)

Z = −4.05⁎⁎⁎ Z = −3.43⁎⁎

Z = −1.17 Z = −1.67

Z = −1.36 Z = −1.33

96 66 63 4

94 69 66 5

97 57 55 2

99 68 67 9

94 71 69 3

1.83 (0.62–5.41) 0.59 (0.37–0.94)⁎ 0.64 (0.41–1.01) 0.41 (0.12–1.45)

6.29 (0.77–51.10) 0.96 (0.56–1.67) 1.09 (0.63–1.86) 1.79 (0.64–4.97)

0.99 (0.38–2.58) 1.09 (0.67–1.78) 1.18 (0.74–1.91) 0.57 (0.18–1.82)

2 (0–28) 6 (1–28)

4 (0–28) 11 (1–28)

1 (0–28) 4 (1–28)

3 (0–28) 4 (1–28)

4 (0–28) 8 (1–28)

Z = −4.23⁎⁎⁎ Z = −4.56⁎⁎⁎

Z = −2.14⁎ Z = −3.54⁎⁎⁎

Z = −0.82 Z = −1.88

85 51 28

76 26 18

95 76 33

98 80 42

74 25 20

6.12 (2.82–13.30)⁎⁎⁎ 9.23 (5.55–15.35)⁎⁎⁎ 2.23 (1.31–3.78)⁎⁎⁎

15.88 (3.71–67.89)⁎⁎⁎ 11.31 (6.14–20.83)⁎⁎⁎ 3.34 (1.87–5.98)⁎⁎⁎

0.88 (0.52–1.48) 0.95 (0.57–1.59) 1.17 (0.66–2.06)

1 (0–28) 28 (1–28)

0 (0–28) 3 (1–28)

28 (0–28) 28 (1–28)

28 (0–28) 28 (2–28)

0 (0–28) 3 (1–28)

Z = 10.98⁎⁎⁎ Z = 8.56⁎⁎⁎

Z = 10.19⁎⁎⁎ Z = 7.59⁎⁎⁎

Z = −0.12 Z = 0.35

54 14 12 1

56 16 14 1

59 14 12 0

61 19 13 1

41 9 9 1

1.13 (0.73–1.77) 0.86 (0.47–1.60) 0.84 (0.44–1.60) –

1.20 (0.72–2.01) 1.27 (0.65–2.47) 0.85 (0.40–1.80) 1.34 (0.08–21.67)

0.55 (0.35–0.86)⁎⁎ 0.51 (0.25–1.02) 0.57 (0.28–1.16) 0.81 (0.05–13.02)

0 (0–28) 3 (1–28)

0 (0–28) 3 (1–28)

0 (0–16) 4 (1–16)

0 (0–28) 2 (1–28)

0 (0–22) 3 (1–22)

Z = −0.43 Z = 0.37

Z = 0.55 Z = −1.32

Z = −1.88 Z = 0.56

71 30 16 5

69 25 17 3

72 30 12 4

71 39 14 9

72 28 19 5

1.15 (0.71–1.87) 1.33 (0.81–2.19) 0.71 (0.38–1.33) 1.32 (0.38–4.60)

1.11 (0.63–1.93) 2.01 (1.16–3.49)⁎ 0.85 (0.42–1.72) 3.20 (0.96–10.69)⁎

1.16 (0.71–1.89) 1.20 (0.72–1.99) 1.20 (0.67–2.15) 1.65 (0.49–5.59)

0 (0–28) 14 (1–28)

0 (0–28) 3 (1–28)

0 (0–28) 28 (1–28)

0 (0–28) 17 (1–28)

0 (0–28) 5 (1–28)

Z = 2.09⁎ Z = 2.42⁎

Z = 3.24⁎⁎ Z = 3.69⁎⁎⁎

Z = 0.75 Z = 0.60

15 3 2 b0

31 9 2 0

11 0 0 0

12 3 3 1

9 2 2 0

0.27 (0.15–0.48)⁎⁎⁎ – – –

0.29 (0.14–0.59)⁎⁎⁎ 0.29 (0.08–1.04) 1.35 (0.27–6.81) –

0.21 (0.11–0.41)⁎⁎⁎ 0.23 (0.07–0.73)⁎ 1.08 (0.24–4.91) –

0 (0–28) 28 (1–28)

0 (0–28) 28 (3–28)

– –

0 (0–28) 2 (2–28)

0 (0–28) 23 (1–28)

Z = −4.25⁎⁎⁎ –

Z = −2.05⁎ Z = 0.09

Z = −2.74⁎⁎ Z = 0.23 (continued on next page)

58

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

Table 3 (continued)

Codeine % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (range) Among whole sample Among past month users Fentanyl % Used ever % Used past month % injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (range) Among whole sample Among past month users Tramadol % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (range) Among whole sample Among past month users Hydromorphone % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (Range) Among whole sample Among past month users Dextropropoxyphene % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (range) Among whole sample Among past month users Tapentadol % Used ever % Used past month % Injected past month % Snorted, chewed, smoked, dissolved past month Median days past month (range) Among whole sample Among past month users Benzodiazepines % Used ever % Used past month % Injected past month % Snorted, chewed, smoked past month Median days past month (range) Among whole sample Among past month users

Total N = 606

A. ‘Other opioids only’ group1 n = 141

B. ‘OST only’ group n = 186

C. ‘OST and other opioids’ group1 n = 104

D. ‘Not prescribed’ group n = 172

A (ref) vs. B OR (95% CI) or Z-score

A (ref) vs. C OR (95% CI) or Z-score

A (ref) vs. D OR (95% CI) or Z-score

96 61 2 7

96 76 2 5

96 48 2 5

99 85 5 14

95 49 1 8

0.83 (0.27–2.60) 0.30 (0.19–0.49)⁎⁎⁎ 0.74 (0.15–3.74) 1.07 (0.40–2.89)

3.84 (0.44–33.40) 1.78 (0.92–3.44) 2.29 (0.54–9.81) 2.93 (1.14–7.56)⁎

0.77 (0.25–2.39) 0.31 (0.19–0.50)⁎⁎⁎ 0.53 (0.09–3.24) 1.54 (0.60–3.98)

2 (0–28) 7 (1–28)

5 (0–28) 11 (1–28)

0 (0–28) 4 (1–28)

10 (0–28) 12 (1–28)

0 (0–28) 5 (1–28)

Z = −6.60⁎⁎⁎ Z = −4.69⁎⁎⁎

Z = 1.30 Z = 0.10

Z = −5.85⁎⁎⁎ Z = −3.42⁎⁎

27 9 8 b0

30 10 8 0

22 4 3 0

39 13 12 1

22 9 9 0

0.65 (0.40–1.08) 0.40 (0.16–0.99)⁎ 0.39 (0.14–1.08) –

1.44 (0.85–2.47) 1.28 (0.57–2.84) 1.52 (0.64–3.59) –

0.63 (0.38–1.06) 0.92 (0.43–1.95) 1.19 (0.54–2.66) –

0 (0–28) 2 (1–28)

0 (0–28) 2 (1–28)

0 (0–28) 2 (1–28)

0 (0–28) 2 (1–28)

0 (0–28) 4 (1–28)

Z = −2.07⁎ Z = 0.59

Z = 0.58 Z = 0.84

Z = −0.18 Z = 0.40

54 14 b0 1

68 25 0 1

48 9 0 2

65 21 2 2

42 7 0 0

0.45 (0.28–0.71)⁎⁎ 0.31 (0.17–0.58)⁎⁎⁎ – 1.12 (0.19–6.81)

0.90 (0.53–1.55) 0.83 (0.45–1.52) – 1.34 (0.19–9.70)

0.35 (0.22–0.56)⁎⁎⁎ 0.23 (0.12–0.47)⁎⁎⁎ – 0.44 (0.39–0.50)

0 (0–28) 5 (1–28)

0 (0–28) 10 (1–28)

0 (0–8) 2 (1–8)

0 (0–28) 12 (1–28)

0 (0–10) 3 (1–10)

Z = −4.07⁎⁎⁎ Z = −3.62⁎⁎⁎

Z = −0.56 Z = 0.25

Z = −4.52⁎⁎⁎ Z = −2.74⁎⁎

10 2 1 b0

9 2 2 0

11 0 0 0

15 3 2 1

6 2 1 0

1.17 (0.56–2.44) – – –

1.76 (0.81–3.85) 1.35 (0.27–6.81) 0.89 (0.15–5.42) –

0.60 (0.25–1.41) 0.81 (0.16–4.05) 0.53 (0.09–3.24) –

0 (0–28) 5 (1–28)

0 (0–10) 8 (5–10)

– –

0 (0–28) 2 (1–28)

0 (0–28) 3 (1–28)

Z = −2.01⁎ –

Z = 0.35 Z = 0.51

Z = −0.27 Z = 0.51

12 1 1 b0

12 1 1 0

12 0 0 0

13 2 0 1

11 2 1 0

1.09 (0.55–2.14) – – –

1.10 (0.50–2.40) 2.71 (0.24–30.25) – –

0.90 (0.44–1.84) 2.45 (0.25–23.81) 1.62 (0.15–18.09) –

0 (0–28) 4 (1–28)

0 (0–6) 6 (6–6)

– –

0 (0–10) 6 (1–10)

0 (0–28) 1 (1–28)

Z = −1.16 –

Z = 0.84 Z = 0.00

Z = 0.79 Z = 0.00

b0 0 0 0

0 0 0 0

1 0 0 0

0 0 0 0

0 0 0 0

– – – –

– – – –

– – – –

– –

– –

– –

– –

– –

– –

– –

– –

93 73 6 12

90 75 2 9

95 74 7 18

95 85 8 14

94 64 7 7

2.00 (0.86–4.66) 0.92 (0.55–1.52) 3.10 (0.86–11.20) 2.12 (1.01–4.41)

2.25 (0.76–6.47) 1.82 (0.94–3.51) 3.72 (0.96–14.40) 1.52 (0.65–3.57)

1.65 (0.72–3.75) 0.58 (0.35–0.95)⁎ 3.11 (0.85–11.38) 0.76 (0.32–1.82)

5 (0–28) 11 (1–28)

8 (0–28) 15 (1–28)

5 (0–28) 14 (1–28)

8 (0–28) 10 (1–28)

2 (0–28) 5 (1–28)

Z = −0.73 Z = −0.74

Z = 0.06 Z = −1.72

Z = −4.00⁎⁎⁎ Z = −3.84⁎⁎⁎

Note that those in the ‘not prescribed’ group had only used diverted pharmaceutical opioids, and that participants in the other three groups could have additionally used diverted pharmaceutical opioids (i.e. not prescribed to them by a doctor). OST: opioid substitution therapy; OR = odds ratio; 95% CI = 95% confidence interval; ns = not significant. Bold indicates a significant estimate. 1 ’Other opioids’ include: oxycodone, morphine, methadone tablets, buprenorphine patches, prescription codeine, fentanyl, tramadol, hydromorphone, dextropropoxyphene and tapentadol. ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

59

Table 4 Methods of obtaining pharmaceutical opioids, injection-related risks and harms, and health service utilization among the NOMAD cohort, by source and type of opioid prescribed in previous month.

Methods of obtaining opioid medication (past 6 months) % Forged prescription % Shared with someone % Bought from a dealer % Bought on the internet % Bought from a patient who sells their medications % Stolen from work (hospital, nursing home, clinic etc.) % Stolen from stranger, store or other place % Prescriptions from multiple doctors % Prescriptions from a doctor for medical reason % Prescriptions from a doctor for no medical reason % Prescriptions from a doctor who prescribes dishonestly Drug use risk and harms (past month) % Used a needle after someone else % Used other injecting equipment after someone else % Lent a used needle to someone else % Any opioid injection-related injuries or diseases (IRID)2 % Non-serious IRID3 % Potentially serious IRID4 % Serious IRID5 % Overdose in the past year Health service utilization (past 6 months) % Saw a general practitioner % Saw a pain specialist % Accessed outpatient drug and alcohol services % Saw an addiction specialist % Saw a psychologist % Saw a psychiatrist % Admitted to hospital

TOTAL sample N = 606

A. ‘Other opioids only’ group1 n = 141

B. ‘OST only’ group n = 186

C. ‘OST and other opioids’ group1 n = 104

D. ‘Not prescribed’ group n = 172

A (ref) vs. B OR (95% CI) or Z-score

A (ref) vs. C OR (95% CI) or Z-score

A (ref) vs. D OR (95% CI) or Z-score

3 68 81 3 72

2 64 74 2 64

4 65 77 1 72

9 65 84 3 78

1 74 89 4 75

1.76 (0.45–6.92) 1.07 (0.68–1.70) 1.17 (0.70–1.94) 0.50 (0.08–2.98) 1.45 (0.90–2.34)

4.23 (1.12–16.05)⁎ 1.07 (0.63–1.83) 1.81 (0.95–3.44) 1.34 (0.26–6.76) 1.96 (1.10–3.50)⁎

0.27 (0.03–2.59) 1.65 (1.02–2.69)⁎ 2.84 (1.55–5.23)⁎⁎

2

3

2

0

2

0.55 (0.12–2.51)

–

0.80 (0.20–3.25)

6

4

5

8

7

1.13 (0.39–3.24)

1.83 (0.62–5.46)

1.65 (0.60–4.52)

12 64

12 72

11 86

30 91

2 16

0.86 (0.43–1.70) 2.51 (1.43–4.39)⁎⁎

3.0 (1.55–5.80)⁎⁎ 4.16 (1.91–9.05)⁎⁎⁎

0.13 (0.04–0.44)⁎⁎⁎ 0.08 (0.04–0.13)⁎⁎⁎

14

12

10

36

6

0.81 (0.40–1.62)

3.90 (2.04–7.45)⁎⁎⁎

0.44 (0.20–1.0)

8

4

7

23

2

2.0 (0.69–5.74)

7.92 (2.91–21.59)⁎⁎⁎

0.48 (0.11–2.05)

7 10

6 10

6 9

4 9

9 11

1.04 (0.41–2.66) 0.91 (0.43–1.91)

0.66 (0.19–2.24) 0.84 (0.35–2.03)

1.68 (0.70–4.05) 1.10 (0.53–2.28)

13 70 (422)

13 74 (102)

15 70 (130)

17 73 (76)

9 66 (113)

1.21 (0.63–2.33) 0.83 (0.51–1.37)

1.38 (0.67–2.86) 0.96 (0.54–1.71)

0.72 (0.35–1.49) 0.68 (0.41–1.11)

67 41 11 13

71 40 11 9

67 44 9 12

72 50 16 20

63 35 9 12

0.83 (0.51–1.34) 1.18 (0.75–1.84) 0.78 (0.37–1.63) 1.49 (0.72–3.12)

1.06 (0.60–1.86) 1.51 (0.90–2.52) 1.60 (0.76–3.38) 2.66 (1.24–5.69)⁎

0.69 (0.43–1.11) 0.81 (0.51–1.29) 0.84 (0.40–1.77) 1.47 (0.70–3.11)

84 10 19

87 21 10

89 6 27

87 14 33

72 2 8

1.17 (0.59–2.34) 0.22 (0.10–0.46)⁎⁎⁎ 3.49 (1.81–6.74)⁎⁎⁎

1.00 (0.46–2.16) 0.59 (0.29–1.18) 4.70 (2.33–9.50)⁎⁎⁎

0.37 (0.20–0.67)⁎⁎ 0.09 (0.03–0.26)⁎⁎⁎ 0.79 (0.35–1.75)

33 26 14 32

7 24 15 40

55 27 10 30

75 31 20 38

6 22 12 25

15.30 (7.55–31.01)⁎⁎⁎ 1.15 (0.69–1.92) 0.60 (0.31–1.18) 0.65 (0.41–1.04)

37.61 (17.20–82.24)⁎⁎⁎ 1.42 (0.80–2.52) 1.42 (0.73–2.76) 0.93 (0.55–1.57)

0.87 (0.36–2.12) 0.87 (0.51–1.49) 0.77 (0.40–1.48) 0.49 (0.30–0.81)⁎⁎

1.63 (0.40–6.63) 1.65 (1.01–2.69)⁎

Note that those in the ‘not prescribed’ group had ONLY used diverted pharmaceutical opioids, and that participants in the other three groups could have additionally used diverted pharmaceutical opioids (i.e. not prescribed to them by a doctor). OST: opioid substitution therapy; IRID: injection-related injury or disease; OR = odds ratio; 95% CI = 95% confidence interval. Bold indicates a significant estimate. 1 ’Other opioids’ include: oxycodone, morphine, methadone tablets, buprenorphine patches, prescription codeine, fentanyl, tramadol, hydromorphone, dextropropoxyphene, tapentadol. 2 n = 422 completed the IRID items. 3 Non-serious IRID includes: transient redness, transient swelling, hives, 'dirty hit', hitting an artery, numbness or pins and needles and collapsed/blocked veins. 4 Potentially serious IRID includes: Abscesses, cellulitis, thrombophlebitis, oedema, puffy hands syndrome, and injecting sinus. 5 Serious IRID includes: systemic infections, deep vein thrombosis, gangrene, amputation and venous ulcer. ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

Thanks to the NOMAD study team members who assisted with interviewing of the NOMAD cohort: Simon Wu, Rosemary Cassidy, Amie Francis, Lucy McWilliams, Erica Franklin, Janelle Holden, Hannah de Jonk, Elyse Cromwell, Olivia Schollar-Root, Kate Gray, Ticia Becker, Mawgoshia-Annia Baron and Rebecca Harris. We would also like to thank our co- and associate investigators on the NOMAD study, as well as our members of the NOMAD Advisory Committee: Michael Farrell, Michael Aufgang, Lesley Brydon, Gabrielle Campbell, Apo Demirkol, Malcolm Dobbin, Adrian Dunlop, Angella Duvnjak, Paul Haber, Marianne Jauncey, Robert Kemp, Richard Mattick, Suzanne Nielsen, Nghi Phung, Ann Roche and Hester Wilson. This study received untied educational grant funding from Mundipharma. The funder has no role in the design, conduct or

interpretation of the study's findings. LD and BL are supported by NHMRC research fellowships (#1041472 and #1073858 respectively). The National Drug and Alcohol Research Centre at UNSW Australia is supported by funding from the Australian Government under the Substance Misuse Prevention and Service Improvements Grant Fund. BL, LD, RA and NL have received untied educational grants from Reckitt Benckiser for the post-marketing surveillance of opioid substitution therapy medications in Australia, the development of an opioidrelated behavior scale, and a study of opioid substitution therapy uptake among chronic non-cancer pain patients. RA has received untied educational grants from Reckitt Benckiser for conducting a study into the pharmacogenetic predictors of treatment success. Reckitt Benckiser had no role in the current study.

60

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61

Appendix Table 1. Drug and alcohol use among the NOMAD cohort, by source and type of opioid prescribed in previous month Heroin % Ever used % Used past 6 months % Injected past 6 months Median days used past 6 months (Range) Among whole sample Among those who used in past 6 months (Meth)amphetamine % Ever used % Used past 6 months % Injected past 6 months Median days used past 6 months (Range) Among whole sample Among those who used in past 6 months Cocaine % Ever used % Used past 6 months % Injected past 6 months Median days used past 6 months (Range) Among whole sample Among those who used in past 6 months Alcohol % Ever used % Used past 6 months % Injected past 6 months Median days used past 6 months (Range) Among whole sample Among those who used in past 6 months Cannabis % Ever used % Used past 6 months Median days used past 6 months (Range) Among whole sample Among those who used in past 6 months Tobacco % Ever used % Used past 6 months Median days used past 6 months (Range) Among whole sample Among those who used in past 6 months

89 64 64

86 53 53

93 67 67

98 73 73

81 65 65

2.07 (1.00–4.25) 1.85 (1.18–2.91)⁎⁎ 1.81 (1.15–2.84)⁎

8.57 (1.96–37.56)⁎⁎ 2.45 (1.42–4.24)⁎⁎ 2.45 (1.42–4.24)⁎⁎

0.71 (0.39–1.30) 1.65 (1.04–2.60)⁎ 1.65 (1.04–2.60)⁎

6 (0–180) 48 (1–180)

1 (0–180) 30 (1–180)

7 (0–180) 48 (1–180)

18 (0–180) 49 (1–180)

7 (0–180) 48 (1–180)

Z = 2.42⁎ Z = 0.18

Z = 3.38⁎⁎ Z = 1.11

Z = −2.30⁎ Z = 0.88

97 73 71

97 68 66

97 72 71

99 74 70

97 80 76

0.89 (0.25–3.21) 1.19 (0.74–1.93) 1.29 (0.81–2.07)

3.05 (0.34–27.72) 1.32 (0.75–2.33) 1.24 (0.72–2.15)

0.99 (0.26–3.76) 1.81 (1.08–3.04)⁎ 1.69 (1.03–2.77)⁎

6 (0–180) 12 (1–180)

2 (0–180) 12 (1–180)

5 (0–180) 12 (1–180)

6 (0–180) 10 (1–180)

11 (0–180) 22 (1–180)

Z = 1.00 Z = 0.70

Z = 0.98 Z = 0.34

Z = 3.09⁎⁎ Z = 2.12⁎

74 21 19

66 17 14

81 22 20

81 27 25

67 22 17

2.28 (1.37–3.78)⁎⁎ 1.38 (0.78–2.44) 1.53 (0.84–2.76)

2.22 (1.22–4.04)⁎⁎ 1.86 (1.00–3.46)⁎ 1.98 (1.04–3.80)⁎

1.09 (0.68–1.75) 1.38 (0.77–2.46) 1.26 (0.68–2.33)

0 (0–180) 3 (1–180)

0 (0–150) 3 (1–150)

0 (0–150) 2 (1–150)

0 (0–180) 3 (1–180)

0 (0–180) 3 (1–180)

Z = 1.06 Z = −0.42

Z = 1.93 Z = −0.1.3

Z = 1.08 Z = −0.10

99 72 1

99 68 0

98 73 1

99 67 3

99 77 1

0.33 (0.04–2.98) 1.23 (0.76–1.98) –

0.75 (0.05–12.07) 0.95 (0.55–1.64) –

1.24 (0.08–19.99) 1.53 (0.92–2.53) –

5 (0–180) 12 (1–180)

5 (0–180) 24 (1–180)

3 (0–180) 12 (1–180)

2 (0–180) 9 (1–180)

10 (0–180) 20 (1–180)

Z = 0.01 Z = −1.09

Z = −0.85 Z = −1.26

Z = 1.48 Z = 0.31

99 80

99 86

98 75

99 76

98 83

0.44 (0.05–4.30) 0.50 (0.28–0.89)⁎

0.75 (0.05–12.07) 0.53 (0.28–1.02)

0.41 (0.04–3.97) 0.80 (0.43–1.47)

48 (0–180) 96 (1–180)

48 (0–180) 90 (1–180)

24 (0–180) 90 (1–180)

90 (0–180) 180 (1–180)

71 (0–180) 90 (1–180)

Z = −1.09 Z = 0.88

Z = 0.55 Z = 2.60⁎⁎

Z = 0.30 Z = 1.01

98 95

99 96

98 95

97 94

98 97

0.33 (0.04–2.98) 0.79 (0.28–2.24)

0.24 (0.03–2.38) 0.74 (0.23–2.35)

0.41 (0.04–3.97) 1.25 (0.39–3.96)

180 (0–180) 180 (2–180)

180 (0–180) 180 (8–180)

180 (0–180) 180 (12–180)

180 (0–180) 180 (24–180)

180 (0–180) 180 (2–180)

Z = 0.40 Z = 0.86

Z = 0.39 Z = 0.92

Z = 1.17 Z = 1.18

Note that those in the ‘not prescribed’ group had only used diverted pharmaceutical opioids, and that participants the other three groups could have additionally used diverted pharmaceutical opioids (i.e. not prescribed to them by a doctor). ’Other opioids’ include: oxycodone, morphine, methadone tablets, buprenorphine patches, prescription codeine, fentanyl, tramadol, hydromorphone, dextropropoxyphene and tapentadol. ICD: International Classification of Diseases; AUDIT-C: Alcohol Use Disorders Identification Test short form; OST: opioid substitution therapy; OR = odds ratio; 95% CI = 95% confidence interval; ns = not significant. Bold indicates a significant estimate. ⁎p b 0.05. ⁎⁎p b 0.01. ⁎⁎⁎p b 0.001.

References Belcher, J., Nielsen, S., Campbell, G., Bruno, R., Hoban, B., Larance, B., et al. (2014). Diversion of prescribed opioids by people living with chronic pain: Results from an Australian community sample. Drug and Alcohol Review, 33(1), 27–32. Campbell, G., Mattick, R., Bruno, R., Larance, B., Nielsen, S., Cohen, M., et al. (2014). Cohort protocol: The Pain and Opioids IN Treatment (POINT) study. BMC Pharmacology and Toxicology, 15(17), http://dx.doi.org/10.1186/2050-6511-1115-1117. Chen, L. H., Hedegaard, H., & Warner, M. (2014). Drug-poisoning deaths involving opioid analgesics: United States, 1999-2011. NCHS Data Brief, 166, 1–8.

Cone, E. J., Giordano, J., & Weingarten, B. (2013). An iterative model for in vitro laboratory assessment of tamper deterrent formulations. Drug and Alcohol Dependence, 131(12), 100–105. Darke, S. (1998). Self-report among injecting drug users: A review. Drug and Alcohol Dependence, 51(3), 253–263. Day, C., Conroy, E., Lowe, J., Page, J., & Dolan, K. (2006). Patterns of drug use and associated harms among rural injecting drug users: Comparisons with metropolitan injecting drug users. Australian Journal of Rural Health, 14(3), 120–125. Degenhardt, L., Larance, B., Bruno, R., Lintzeris, N., Ali, R., & Farrell, M. (2015). Evaluating the potential impact of a reformulated version of oxycodone upon tampering, non-

B. Larance et al. / Journal of Substance Abuse Treatment 58 (2015) 51–61 adherence and diversion of opioids: The National Opioid Medications Abuse Deterrence (NOMAD) study protocol. Addiction, 110(2), 226–237. Dwyer, R., Topp, L., Maher, L., Power, R., Hellard, M., Walsh, N., et al. (2009). Prevalences and correlates of non-viral injecting-related injuries and diseases in a convenience sample of Australian injecting drug users. Drug and Alcohol Dependence, 100(1-2), 9–16, http://dx.doi.org/10.1016/j.drugalcdep.2008.08.016 (Epub 2008 Nov 14). Euser, A. M., Zoccali, C., Jager, K. J., Dekker, F. W., Euser, A. M., Zoccali, C., et al. (2009). Cohort studies: Prospective versus retrospective. Nephron, 113(3), c214–c217. Gomes, T., Mamdani, M. M., Dhalla, I. A., Cornish, S., Paterson, J. M., & Juurlink, D. N. (2014). The burden of premature opioid-related mortality. Addiction, 109(9), 1482–1488, http://dx.doi.org/10.1111/add.12598. Gossop, M., Darke, S., Griffiths, P., Hando, J., Powis, B., Hall, W., et al. (1995). The Severity of Dependence Scale (SDS): Psychometric properties of the SDS in English and Australian samples of heroin, cocaine and amphetamine users. Addiction, 90, 607–614. Havens, J. R., Leukefeld, C. G., DeVeaugh-Geiss, A. M., Coplan, P., & Chilcoat, H. D. (2014). The impact of a reformulation of extended-release oxycodone designed to deter abuse in a sample of prescription opioid abusers. Drug and Alcohol Dependence, 139(0), 9–17, http://dx.doi.org/10.1016/j.drugalcdep.2014.02.018. Katz, N., Dart, R. C., Bailey, E. J., Trudeau, J., Osgood, E., & Paillard, F. (2011). Tampering with prescription opioids: Nature and extent of the problem, health consequences, and solutions. The American Journal of Drug and Alcohol Abuse, 37, 205–217. Kessler, R. C., & Ustun, T. B. (2004). The World Mental Health (WMH) Survey Initiative version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). International Journal of Methods in Psychiatric Research, 13(2), 93–121, http://dx.doi.org/10.1002/mpr.168. Kroenke, K., Spitzer, R., & Williams, J. B. W. (2001). The PHQ-9: Validity of a brief depression severity measure. Journal of General Internal Medicine, 16(9), 606–613. Kroenke, K., Spitzer, R. L., Williams, J. B. W., & Lowe, B. (2010). The Patient Health Questionnaire Somatic, Anxiety, and Depressive Symptom Scales: A systematic review. General Hospital Psychiatry, 32(4), 345–359. Larance, B., Ambekar, A., Azim, T., Murthy, P., Panda, S., & Degenhardt, L. (2011a). The availability, diversion and injection of pharmaceutical opioids in South Asia. Drug and Alcohol Review, 30(3), 246–254, http://dx.doi.org/10.1111/j.1465-3362.2011.00304.x. Larance, B., Degenhardt, L., Lintzeris, N., Winstock, A., & Mattick, R. (2011b). Definitions related to the use of pharmaceutical opioids: Extramedical use, diversion, non-

61

adherence and aberrant medication-related behaviours. Drug and Alcohol Review, 30(3), 236–245, http://dx.doi.org/10.1111/j.1465-3362.2010.00283.x. Lenton, S., Dietze, P., Olsen, A., Wiggins, N., McDonald, D., & Fowlie, C. (2014). Working together: Expanding the availability of naloxone for peer administration to prevent opioid overdose deaths in the Australian Capital Territory and beyond. Drug and Alcohol Review, http://dx.doi.org/10.1111/dar.12198 (Epub ahead of print). Madadi, P., Hildebrandt, D., Lauwers, A. E., & Koren, G. (2013). Characteristics of opioid-users whose death was related to opioid-toxicity: A population-based study in Ontario, Canada. PLoS One, 8(4), e60600, http://dx.doi.org/10.60610.61371/journal.pone.0060600. Manchikanti, L., Helm, S., Fellows, B., Janata, J. W., Pampati, V., Grider, J. S., et al. (2012). Opioid epidemic in the United States. Pain Physician, 15(3 Suppl.), ES9–ES38. Passik, S. D., Hays, L., Eisner, N., & Kirsh, K. L. (2006). Characteristics of prescription drug abusers entering drug rehabilitation. Psychiatrie and Pain, 20(2), 5–13. Prins, A., Ouimette, P., Kimerling, R., Camerond, R. P., Hugelshofer, D. S., Shaw-Hegwer, J., et al. (2004). The primary care PTSD screen (PC-PTSD): Development and operating characteristics. Primary Care Psychiatry, 9(1), 9–14, http://dx.doi.org/10.1185/ 135525703125002360. Roxburgh, A., Bruno, R., Larance, B., & Burns, L. (2011). Prescription of opioid analgesics and related harms in Australia. Medical Journal of Australia, 195(5), 280–284. Sellers, E. M., Perrino, P. J., Colucci, S. V., & Harris, S. C. (2013). Attractiveness of reformulated OxyContin® tablets: Assessing comparative preferences and tampering potential. Journal of Psychopharmacology, 27, 808–816. Severtson, S. G., Bartelson, B. B., Davis, J. M., Muñoz, A., Schneider, M. F., Chilcoat, H., et al. (2013). Reduced abuse, therapeutic errors, and diversion following reformulation of extended-release oxycodone in 2010. The Journal of Pain, 14(10), 1122–1130. Spitzer, R. L., Kroenke, K., Williams, J. B. W., & Lowe, B. (2006). A brief measure for assessing generalized anxiety disorder: The GAD-7. Archives of Internal Medicine, 166(10), 1092–1097. Stafford, J., & Burns, L. (2014). Australia Drug Trends 2013: Findings from the Illicit Drug Reporting System (IDRS). Australian Drug Trends Series No. 109. Sydney: National Drug and Alcohol Research Centre, University of New South Wales. Tan, G., Jensen, M. P., Thornby, J. I., & Shanti, B. F. (2004). Validation of the Brief Pain Inventory for chronic nonmalignant pain. The Journal of Pain, 5(2), 133–137. Topp, L., & Mattick, R. P. (1997). Choosing a cut-off on the Severity of Dependence Scale (SDS) for amphetamine users. Addiction, 92(7), 839–845.