One year outcomes for heroin dependence: Findings from the Australian Treatment Outcome Study (ATOS)

Drug and Alcohol Dependence 83 (2006) 174–180

One year outcomes for heroin dependence: Findings from the Australian Treatment Outcome Study (ATOS) Maree Teesson a,∗ , Joanne Ross a , Shane Darke a , Michael Lynskey b , Robert Ali c , Alison Ritter d , Richard Cooke c a b

National Drug and Alcohol Research Centre, University of New South Wales, NSW, Australia Department of Psychiatry, Washington University School of Medicine, St. Louis, United States c Drug and Alcohol Services Council, SA, Australia d Turning Point Alcohol and Drug Centre, Vic., Australia

Received 15 April 2005; received in revised form 9 November 2005; accepted 13 November 2005

Abstract Aim: To determine 1 year outcomes for drug use, criminality, psychopathology and injection-related health problems in those entering treatment for heroin dependence in Australia. Design: Longitudinal prospective cohort study. Participants: Seven hundred and forty five individuals entering treatment (methadone/buprenorphine maintenance therapy; detoxification; residential rehabilitation) and 80 heroin users not seeking treatment. Setting: Sydney, Melbourne and Adelaide, Australia. Findings: A total of 657 individuals were re-interviewed at 1 year, 80% of the original sample. There were substantial reductions in heroin and other drug use across all three treatment modalities. The majority of those who had entered treatment were heroin abstinent at 1 year (maintenance therapy 65%, detoxification 52%, residential rehabilitation 63%) compared to 25% of the non-treatment sample. The reduction in heroin use among the treatment samples was paralleled by reductions in poly drug use. There were also substantial reductions in risk-taking, crime and injection-related health problems across all treatment groups, and less marked reductions among the non-treatment group. Psychopathology was dramatically reduced among the treatment modalities, while remaining stable among the non-treatment group. Positive outcomes at 1 year were associated with a greater number of cumulative treatment days experienced over the 1 year follow-up period (‘treatment dose’) and fewer treatment episodes undertaken in that time (‘treatment stability’). Conclusions: At 1 year, there were impressive reductions in drug use, criminality, psychopathology and injection-related health problems following treatment exposure. The positive findings were associated with a greater “dose” of treatment, and with more treatment stability over the follow-up period. © 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Heroin dependence; Prospective cohort study; Treatment outcome

1. Introduction Heroin dependence is a persistent, and in many cases, lifelong condition (Goldstein and Herrera, 1995; Hser et al., 2001; Oppenheimer et al., 1994). The persistence of heroin dependence is a major public health concern, as heroin use is associated with a range of serious harms including overdose, blood borne virus transmission, psychopathology, criminality, trauma and suicide

∗

Corresponding author. Tel.: +61 2 9385 0333; fax: +61 2 9385 0222. E-mail address: [email protected] (M. Teesson).

0376-8716/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.drugalcdep.2005.11.009

(Darke and Ross, 2002; Hser et al., 2001; Mills et al., 2005; Oppenheimer et al., 1994). Mortality among heroin users is high, with annual rates in the order of 1–3% (Davoli et al., 1997; Hulse et al., 1999). Heroin use is also strongly associated with crime, largely in support of the drug use (Godfrey et al., 2004; Hser et al., 2001). Despite the significant burden of heroin dependence, large scale, longitudinal, multi-site treatment outcome studies of heroin dependence are rare. The few studies conducted in the United States and Europe have demonstrated that individuals in the major treatment modalities can substantially reduce their use of illicit drugs and criminal activity (Anglin et al., 1997; Godfrey

M. Teesson et al. / Drug and Alcohol Dependence 83 (2006) 174–180

et al., 2004; Hubbard et al., 1989, 1997; Simpson and Sells, 1982). The UK based National Treatment Outcome Research Study (NTORS; Gossop et al., 1998) found that at 1 year posttreatment entrance current heroin abstinence had increased from 21% to 48% (Gossop et al., 2002). Consistently, the US Drug Abuse Treatment Outcome Study (DATOS) found weekly or more frequent heroin use had reduced from 89% to 28% by 1 year post-treatment entrance (Hubbard et al., 1997). NTORS also demonstrated significant improvements in physical and mental health across treatment modalities (Gossop et al., 2002). The factors associated with treatment outcome have been examined in a number of studies. Consistent in the literature is the finding that longer retention times are associated with better outcome (Flynn et al., 2003; Gossop et al., 1999; Hubbard et al., 1997, 1989; Sanchez-Carbonell et al., 1988; Simpson et al., 1997), reflecting the overall ‘treatment dose’ individuals receive (Joe et al., 1999). A history of previous treatment has also been associated with poorer outcome among DATOS participants and may reflect the potentially more severe drug use characteristics of repeat treatment seekers (Anglin et al., 1997; Hser et al., 1999). Similarly, in NTORS, it has recently been reported that multiple enrolments over a 12-month period were associated with poorer outcome (Gossop et al., 1999). Heroin use is remarkably persistent, and cyclical in nature. Sheehan et al. (1993) for instance, reported that three quarters of opiate users who entered treatment were heroin-abstinent at 12month follow-up. However, when use patterns over the entire follow-up period were examined, it was revealed that only a quarter had been abstinent for more than 80% of the time. More people are treated in Australia for opioid dependence than for any other illicit drug class (Shand and Mattick, 2002). Nearly 7 in every 1000 Australian adults are dependent on heroin, a rate similar to that found in the UK (European Monitoring Centre for Drugs and Drug Addiction, 1999; Hall et al., 2000). The Australian Treatment Outcome Study (ATOS) commenced in 2001 to examine treatment outcome in a large cohort of heroin dependent individuals entering treatment in Australia. The characteristics of the ATOS sample have been reported in detail elsewhere (Darke et al., 2003; Ross et al., 2005). An important design feature of ATOS is the collection of data from a group of heroin dependent individuals not seeking treatment. This paper details the findings at 1 year follow-up for patients recruited into maintenance therapy (MT), detoxification (DTX) and residential rehabilitation (RR), as well as for a group who had not entered treatment at baseline (NT). Outcomes are reported for heroin and other drug use, risk-taking, physical health, psychopathology and criminal behaviour. In addition, the study aimed to ascertain whether there were baseline demographic and drug use characteristics that predicted reduced drug use, and examined the role of treatment in achieving reduced drug use. Specifically, the aims of the current study were: 1. To determine the drug use, criminality, physical and mental health of ATOS participants at 12-month follow-up. 2. To determine the role of treatment in the achievement of improved outcomes.

175

2. Method 2.1. Procedure ATOS is a longitudinal study of entrants to treatment for heroin dependence, recruited from randomly selected treatment agencies, and a comparison group of heroin users not in treatment (non-treatment group). Baseline data were collected between February 2001 and August 2002. Subjects were recruited from 38 agencies treating heroin dependence in Sydney, Adelaide and Melbourne. These agencies were randomly selected from within treatment modality. A list of treatment agencies was constructed from all available sources and they were randomly selected by drawing names from a hat representing treatment modality. They comprised 13 methadone/buprenorphine maintenance (MT) agencies, 9 detoxification (DTX) facilities, 8 residential rehabilitation units (RR) and 8 agencies which provided both maintenance and detoxification services. In addition, a comparison group of heroin users not currently in treatment, or seeking treatment, were recruited from needle and syringe programs in the Sydney regional health areas from which treatment entrants were recruited. Eligibility criteria were: (i) no treatment for heroin dependence in the preceding month, (ii) no imprisonment in the preceding month, (iii) agreed to give contact details for followup interviews, (iv) 18 years or over and (v) sufficient English to complete the interview. The final sample consisted of 825 current heroin users: 277 entering maintenance therapy, 288 entering detoxification, 180 entering residential rehabilitation units and 80 non-treatment subjects. All subjects were paid up to A$ 30 for completing the baseline interview, which took approximately 90 min to complete, with follow-up interviews at 3 months and 1 year taking 30 min each. 2.2. Measures A structured interview addressed demographics, treatment history, drug use history and heroin overdose history. Drug use, needle risk-taking, injection-related health problems and criminal behaviours over the month preceding interview were measured using the Opiate Treatment Index (OTI) (Darke et al., 1992). Drug use estimates in the OTI are ratios based upon recent use episodes, and are expressed as Q scores. A Q score of 1.0 indicates an average of one use episode per day, 2.0 would indicate use, on average, twice a day, etc. DSM-IV diagnoses of past month Major Depressive Disorder, Post-Traumatic Stress Disorder (PTSD) and an ICD 10 screen of Borderline Personality Disorder (BPD) were obtained using the Composite International Diagnostic Instrument (World Health Organization, 1993). Diagnoses of antisocial personality disorder (ASPD) were obtained from the Diagnostic Interview Schedule (Robins et al., 1981), modified to obtain DSM-IV diagnoses. General physical health and mental health was measured using the Short Form-12 (SF-12: Ware et al., 1996). The follow-up interview matched the baseline interview with the addition of treatment information and exclusion of drug use history, PTSD, BPD and ASPD. Participants were asked how many times they had com-

176

M. Teesson et al. / Drug and Alcohol Dependence 83 (2006) 174–180

menced treatment, in any modality for heroin dependence since baseline interview, and the time spent in each treatment episode. The number of treatment episodes and the cumulative number of treatment days (in any modality) was calculated. 2.3. Statistical analyses Where data were highly skewed medians were reported. Highly skewed continuous data were analysed using the Mann–Whitney U statistic. Means were reported for non-skewed data. In order to examine change over time in the major domains of drug use, crime, physical and mental health, a series of linear and logistic regressions were conducted using measures obtained at 1 year as the dependent variables. Age, gender, a baseline measure of the dependent variable, the cumulative number of days in any treatment for heroin dependence over the follow-up period and the total number of treatment episodes over the follow-up period were entered into all regression models. Logistic regressions were conducted to determine factors independently associated with dichotomous variables. Independent predictors of continuous variables were determined using linear regression. All analyses were conducted using SPSS for Windows, Version 12.0 (SPSS, 2003). 3. Results 3.1. Sample characteristics A total of 657 (80%) of the cohort were successfully reinterviewed at 1 year follow-up. Follow-up rates for the four index groups were: 82% (MT), 82% (DTX), 78% (RR) and 66% (NT). At baseline, the mean age of the 657 participants was 29.5 years (S.D. 7.9, range 18–56), and 64% were male. The sample had completed a mean of 10.1 years of secondary education (S.D. 1.6, range 2–13), 33% had completed a trade/technical course, and 7% a university degree. Thirty seven percent had a prison history. The main sources of income were: social security allowances (50%), criminal activity (20%) and wage/salary (18%). Eighty nine percent had been enrolled in treatment for opiate dependence prior to ATOS. The median number of previous treatment episodes was 4 (range 0–218), and 76% of participants reported more than one prior episode.

The mean age of first intoxication was 13.5 years (S.D. 3.2, range 2–34) and 19.6 years (S.D. 5.2, range 9–43) for first heroin use. The mean length of heroin use career at baseline was 9.9 years (S.D. 7.6, range <1–35 years) and the median OTI heroin use score was 2.0 (range 0.0–33.3). The sample had used a mean of 9.2 (S.D. 1.1, range 2–11) drug classes in their lives, and 4.9 (S.D. 1.7, range 1–10) in the month preceding interview. Twenty eight percent met criteria for current Major Depressive Disorder, and 40% for a lifetime diagnosis of PTSD. A diagnosis of BPD was received by 45% of the sample, and ASPD by 71%. Comparisons of those re-interviewed with those lost to follow-up indicated there were no significant differences in age, heroin use, previous treatment enrolment, criminal involvement or global mental health. The only notable difference between the current sample and those lost to follow-up was that the NT group were less likely than the treatment groups to have been followed-up (81% versus 66%, OR 0.45, 95% CI: 0.27–0.75). 3.2. Treatment exposure at 1 year Sixty percent were enrolled in a treatment program at 1 year (Table 1). Post-index treatment episodes were reported by 47% of the MT index group, 88% of the DTX index group and 68% of the RR index group. Seventy-four percent of the NT group had been exposed to treatment during the 1 year follow-up period, most commonly MT. The median number of treatment days over the follow-up period for the whole sample was 156: MT 334 days, DTX 78 days, RR 138 days and NT 30 days (Table 1). The NT group spent significantly less days in treatment over the follow-up period than those entering the study through treatment (30 days versus 170 days, Mann–Whitney U = 7747.0, p < 0.001). 3.3. Changes in heroin use There were substantial declines in heroin use between baseline and 1 year follow-up (Table 2). The median OTI score declined from 2.0 to 0.0 (Z = 17.3, p < 0.001). Past month heroin abstinence rates climbed dramatically in all three treatment modalities: from 3% to 65% in the MT group, from 0% to 52% in DTX and from 2% to 63% in RR. There were substantial reductions in the number of heroin use days in the last

Table 1 Treatment status at 1 year Index group MT (N = 227) Currently in treatment (%)

All (N = 657) DTX (N = 236)

RR (N = 141)

NT (N = 53)

79

49

49

47

59

One year treatment exposure (%) MT DTX RR

100 7 3

53 100 25

18 27 100

53 21 13

62 46 32

Any treatment (%) Post-index treatment (%) Median treatment episodes since baseline interview Median cumulative treatment days since baseline interview

100 47 1 334

100 88 3 78

100 68 2 138

74 74 1 30

97 69 2 156

M. Teesson et al. / Drug and Alcohol Dependence 83 (2006) 174–180

177

Table 2 Outcomes at 1 year

Heroin use Abstinent from heroin (% last month) Heroin use days (mean last month) OTI Heroin score (median last month) Poly drug use Drug classes used (mean last month)

MT (N = 227)

DTX (N = 236)

RR (N = 141)

NT (N = 53)

Total (N = 657)

BL

1 year

BL

1 year

BL

1 year

BL

1 year

BL

1 year

3 19.1 1.5

65 2.9 0

0 22.8 2.0

52 6.0 0

2 17.3 1.5

63 4.1 0

0 21.8 2.0

25 10.3 0.3

2 20.3 2.0

57 4.9 0

4.7

3.2

4.9

3.4

5.1

2.8

5.5

4.4

4.9

3.3

Overdose Past year (%)

22

Physical health SF-12 (mean) Injection-related health problems (%)

43.6 71

Risk-taking Daily injecting (%)

78

Mental health SF-12 (mean) Current Major Depressive Disorder (%) Attempted suicide (% last month)

32.3 26 3

Criminal activity Any crime in preceding month (%)

45

4

24

16

38

19

21

15

26

13

42.8 74

48.3 28

45.0 84

51.0 28

43.8 76

46.9 59

43.6 75

48.7 28

84

23

75

15

83

36

80

17

43.4 11 1

31.2 32 8

39.8 18 3

28.4 31 6

40.2 13 1

37.7 13 2

40.3 15 4

31.5 28 5

41.2 14 2

19

59

28

61

27

60

40

55

26

47.9 21 7

month in all three treatment groups. Heroin had been used on a mean of 2.9 days by the MT group, 6.0 days by the DTX group and 4.1 days by the RR group. There was a smaller decline amongst the NT group, who had used on a mean of 10.3 days. None of these subjects reported abstinence in the past month at baseline (a function of the NT inclusion/exclusion criteria in the study), while 25% reported past month abstinence at 1 year follow-up. Changes in the heroin abstinence rates were paralleled by a reduction in median heroin OTI scores from baseline to 1 year follow-up. Substantial reductions were noted among the MT (1.5 versus 0.0), DTX (2.0 versus 0.0) and RR (1.5 versus 0.0) groups. A less marked reduction in heroin OTI score was evident among the NT group (2.0 versus 0.3). Participants in NSW and SA were asked about their use of heroin during the 1 year follow-up period (N = 570). While the prevalence of heroin use in the preceding month was low among treatment groups, the majority had used heroin within the year: MT (82%), DTX (92%), RR (72%) and NT (100%). Overall, 14% of participants in NSW and SA were abstinent for the entire year, all of whom had entered treatment at baseline. Time to follow-up was unrelated to continuous abstinence status (abstinent: 368.6 days versus others: 372.8 days, p > 0.5). Factors associated with the number of heroin use days in the month preceding the follow-up interview were examined using multiple linear regression. The model was significant F5,639 = 12.9, p < 0.001) and accounted for 9% of the variance. A greater number of baseline heroin use days (β = 0.1, t = 3.5, p < 0.001), fewer treatment days (β = −0.2, t = 5.7, p < 0.001) and more treatment episodes over the follow-up period (β = 0.1, t = 2.7, p < 0.01) were independently associated with more frequent heroin use at 1 year.

3.4. Changes in other drug use There were notable reductions in the use of other drugs among all treatment modalities (Table 2). The sample had used a mean of 3.2 (S.D. 1.7, range 0–8) drug classes in the month preceding the 1 year interview, a significant decline from baseline (4.9 versus 3.2, t569 = 19.9, p < 0.001). A higher level of baseline poly drug use (β = 0.3, t = 8.0, p < 0.001), fewer treatment days (β = −0.1, t = 2.1, p < 0.05) and more treatment episodes (β = 0.1, t = 2.3, p < 0.05) were independently associated with using a greater number of drug classes at 1 year. There was no effect of age or gender. The model was significant (F5,640 = 16.9, p < 0.001) and accounted for 11% of the variance. 3.5. Heroin overdose There were notable reductions in rates of non-fatal heroin overdose among treatment modalities from baseline to 1 year follow-up (Table 2). A logistic regression model examining 1 year overdose experience at 1 year was significant (χ2 = 125.7, d.f. = 5, p < 0.001), and had a good fit (Hosmer–Lemeshow χ2 = 3.5, d.f. = 8, p = 0.90). The factors associated with overdose at 1 year were having overdosed in the year prior to baseline (OR = 6.52, 95% CI: 3.76–11.28), and having more treatment episodes over the follow-up period (OR = 1.52, 95% CI: 1.34–1.71). 3.6. Physical health SF-12 physical health scores improved slightly across all groups at 1 year and were closer to the general population norm than they were at baseline (Table 2). A linear regres-

178

M. Teesson et al. / Drug and Alcohol Dependence 83 (2006) 174–180

sion examining physical health scores at 1 year follow-up was conducted controlling for age, gender, baseline SF-12 physical health score, cumulative number of treatment days and number of treatment episodes over the 1 year follow-up period. The model was significant (F5,636 = 17.6, p < 0.001) and accounted for 12% of the variance. Higher baseline SF-12 scores (β = 0.3, t = 8.1, p < 0.001) and younger age (β = −0.1, t = 2.6, p < 0.01) were independently associated with higher physical SF-12 score (better health) at 1 year follow-up. Obvious improvements in injection-related health were seen across all treatment modalities, and to a lesser extent among the NT group. Factors associated with having injection-related health problems in the month preceding 1 year interview were examined using a logistic regression. The model was significant (χ2 = 49.3, d.f. = 5, p < 0.001), and had a good fit (Hosmer–Lemeshow χ2 = 3.0, d.f. = 8, p = 0.93). Injectionrelated health problems at baseline (OR 2.75, 95% CI: 1.68–4.49), and a greater number of treatment episodes (OR 1.18, 95% CI: 1.09–1.29) were independently associated with injection-related health problems at 1 year. 3.7. Needle risk-taking All treatment groups showed marked declines in the proportions injecting daily (Table 2). A smaller decline was noted in the NT group, where 36% were still injecting daily at follow-up. A logistic regression model examining factors associated with daily injecting at 1 year was significant (χ2 = 55.0, d.f. = 5, p < 0.001) and had a good fit (Hosmer–Lemeshow χ2 = 3.0, d.f. = 8, p > 0.5). Daily or more frequent injecting at baseline (OR 4.31, 95% CI: 1.9–9.6), and more treatment episodes over the follow-up period (OR 1.11, 95% CI: 1.0–1.2) were independently associated with daily or more frequent injecting at 1 year.

attempted suicide with rates appearing to fall in each of the three treatment groups, but remaining relatively stable in the non-treatment group. A logistic regression model examining current Major Depressive Disorder status at 1 year was significant (χ2 = 61.0, d.f. = 5, p < 0.001) and had a good fit (Hosmer–Lemeshow χ2 = 3.22, d.f. = 8, p = 0.92). Being female (OR 2.37, 95% CI: 1.44–3.90), Major Depressive Disorder at baseline (OR 3.77, 95% CI: 2.35–6.05), and a greater number of treatment episodes over the follow-up period (OR 1.15, 95% CI: 1.05–1.27) were all independently associated with Major Depressive Disorder at 1 year. 3.9. Criminal activity In parallel with the reductions in drug use and improvements in mental health, there were marked reductions in the percentage of respondents who had committed crime in the month preceding interview (Table 2). In comparison, there was a less marked drop in crime among the non-treatment group. Property crime (16%) and dealing (11%) remained the most common forms of crime reported by the sample at 1 year. A logistic regression model examining factors associated with criminal activity at 1 year was significant (χ2 = 85.8, d.f. = 5, p < 0.001) and had a good fit (Hosmer–Lemeshow χ2 = 3.057, d.f. = 8, p = 0.93). Having committed crime in the month prior to baseline (OR 3.67, 95% CI: 2.38–5.65) and having a greater number of treatment episodes over the follow-up period (OR 1.18, 95% CI: 1.08–1.29) were independently associated with crime at 1 year. Younger age (OR 0.97, 95% CI: 0.95–1.00) was of marginal significance in predicting crime at 1 year. There was no effect of gender or cumulative treatment days. 4. Discussion

3.8. Mental health All treatment groups showed substantial improvement in mental health between baseline and 1 year follow-up (Table 2). A linear regression examining SF-12 mental health scores at 1 year follow-up was conducted, controlling for age, gender, baseline SF-12 mental health score, the cumulative number of treatment days and the number of treatment episodes over the followup period. The model was significant (F5,636 = 21.8, p < 0.001), and accounted for 14% of the variance. Higher baseline SF12 scores (β = 0.3, t = 7.0, p < 0.001), younger age (β = −0.1, t = 3.0, p < 0.005), a greater cumulative number of treatment days (β = 0.1, t = 2.6, p < 0.01) and fewer treatment episodes (β = −0.2, t = 5.1, p < 0.001) were independently associated with better mental health at follow-up. Current Major Depressive Disorder dropped from 26% to 11% among those entering MT, from 32% to 18% among those entering DTX and from 31% to 13% among those entering RR (Table 2). Current Major Depressive Disorder remained stable in the non-treatment group, but it should be acknowledged that they had a lower baseline prevalence (13% at baseline and 15% at 1 year). A similar pattern of results emerged for past month

The first major finding of this study concerned the feasibility of conducting longitudinal research within this population. At 1 year follow-up, 80% of ATOS participants were successfully re-interviewed. This response rate is equivalent to that of other longitudinal studies of treatment seeking drug use populations (Gerstein et al., 1994; Gossop et al., 1997; Hubbard et al., 1989, 1997; Simpson and Sells, 1982), and indicates that, despite the often chaotic lifestyle of people with opioid dependence, it remains feasible to conduct longitudinal studies of treatment outcomes in this group. The major finding, in terms of outcome, was that there were considerable reductions in heroin use across all three treatment modalities, and a less notable reduction among the NT group. There were also substantial reductions in injection frequency, needle risk-taking behaviours, criminal involvement, injectionrelated health problems and heroin overdose. There were large declines in levels of Major Depressive Disorder in all treatment groups, and improvements in general mental health. In contrast, the psychological health of the NT group varied little. Positive outcomes at 1 year tended to be associated with a greater cumulative number of treatment days experienced over

M. Teesson et al. / Drug and Alcohol Dependence 83 (2006) 174–180

the follow-up period (treatment dose), fewer treatment episodes undertaken in that time (treatment stability) and better baseline functioning. The finding that longer exposure to treatment significantly improved outcome is consistent with the previous findings (Flynn et al., 2003; Gossop et al., 1999; Hubbard et al., 1989, 1997; Sanchez-Carbonell et al., 1988; Simpson et al., 1997). The findings regarding treatment stability are also similar to those reported in the UK-based NTORS study, in which postindex treatment enrolment was associated with poorer outcome (Gossop et al., 1999). The comprehensive data on treatment exposure collected illustrates the complex treatment careers of heroin users. Importantly, 74% of those in the non-treatment group also accessed some form of treatment during the follow-up period. However, the NT group had spent significantly fewer days in treatment for their heroin dependence over the follow-up period than the treatment groups. Treatment exposure in this group may potentially explain some of the observed improvements in drug use and other outcomes. There were dramatic reductions in heroin use over the followup period. Both the extent of treatment exposure over the follow-up period and stability of treatment (number of treatment episodes) were important predictors of the level of heroin use at 1 year. An association between multiple treatment enrolments over a 1 year period and poorer outcome has also been reported in the United Kingdom. Past month abstinence rates among the index treatment groups ranged from 52% to 65% at 1 year follow-up, with only 25% of the NT group being abstinent. Similar levels of abstinence at 1 year post-treatment entry have been reported internationally (Gossop et al., 2000; Hubbard et al., 1997). While general physical health appeared to improve across all treatment modalities, treatment related factors such as extent of exposure and stability over the follow-up period were not significantly associated with better health at 1 year. Conversely, a large decline in injection-related health problems was noted across all treatment groups, and better injection-related health at 1 year was associated with stability of treatment over the study period. Mental health improved markedly across the treatment groups, but remained relatively unchanged in the NT group. A similar trend towards improvement was seen in relation to current Major Depressive Disorder. As was the case with heroin use, stability of treatment (i.e. fewer treatment episodes) during the follow-up period was associated with better general mental health, and a reduced likelihood of depression at 1 year. Greater treatment exposure over the study period (i.e. more treatment days) was also an important predictor of better general mental health. There were substantial reductions in criminal activity among the treatment groups between the baseline and 1 year interviews, and a less marked reduction among the NT group. These findings parallel the existing literature showing that successful engagement with treatment services is accompanied by substantial reductions in criminal activity (Hubbard et al., 1997). Such reductions can be considered a major benefit of treatment and are typically attributed to a reduction in drug use. In the

179

current study, stability of treatment was found to be the treatment factor associated with less criminal involvement at 1 year. Clearly, treatment for heroin dependence has a positive impact that extends beyond reducing heroin use. ATOS is the first Australian study to have examined medium term treatment outcomes for heroin users. The study found impressive reductions in drug use, criminality, psychopathology and injection-related health problems across all index treatment groups. The study attests to the robustness of similar findings across different systems and cultures. Overall, the general finding is that all groups had substantially improved since baseline. These findings suggest that longer-term retention in fewer treatment episodes substantially improves the health and welfare of heroin users. Acknowledgements This research was funded by the National Health and Medical Research Council, the Australian Government Department of Health and Ageing, and the Victorian Department of Human Services. The authors wish to thank Katherine Mills, Anna Williamson, Kate Hetherington, Evelyn Wilhelm, Alys Havard, Sandra Fairbairn, Nicola Henderson, Josephine Weekley, Amy Swan, Tracey Holt, Sanja Pahoki and all ATOS participating agencies. References Anglin, M.D., Hser, Y., Grella, C.E., 1997. Drug addiction and treatment careers among clients in the Drug Abuse Treatment Outcome Study (DATOS). Psychol. Addict. Behav. 4, 308–323. Darke, S., Ross, J., 2002. Suicide among heroin users: rates, risk factors and methods. Addiction 97, 1383–1394. Darke, S., Hall, W., Wodak, A., Heather, N., Ward, J., 1992. Development and validation of a multi-dimensional instrument for assessing outcome of treatment among opiate users: The Opiate Treatment Index. Br. J. Addict. 87, 733–742. Darke, S., Ross, J., Teesson, M., Lynskey, M., 2003. Health service utilisation and benzodiazepine use among heroin users: findings from the Australian Treatment Outcome Study (ATOS). Addiction 98, 1129–1135. Davoli, M., Perucci, C., Rapiti, E., Bargagli, A., D’Ippoliti, D., Forastiere, F., Abeni, D., 1992. A persistent rise in mortality among injection drug users in Rome, 1980 through 1992. Am. J. Public Health 87, 851– 853. European Monitoring Centre for Drugs and Drug Addiction, 1999. Extended annual report on the state of the drugs problem in the European Union 1999. Office for Official Publications of the European Communities, Luxembourg. Flynn, P.M., Joe, G.W., Broome, K.M., Simpson, D.D., Brown, B.S., 2003. Recovery from opioid addiction in DATOS. J. Subst. Abuse Treat. 25, 177–186. Gerstein, D., Harwood, H., Suter, N., 1994. Evaluating recovery services: The California Drug and Alcohol Treatment Assessment (CALDATA) General Report. Department of Alcohol and Drug Programs, California. Godfrey, C., Stewart, D., Gossop, M., 2004. Economic analysis of costs and consequences of the treatment of drug misuse: 2-year outcome data from the National Treatment Outcome Research Study (NTORS). Addiction 99, 697–707. Goldstein, A., Herrera, J., 1995. Heroin addicts and methadone treatment in Albuquerque: a 22 year follow-up. Drug Alcohol Depend. 40, 139–150. Gossop, M., Marsden, J., Stewart, D., Lehmann, P., Wilson, A., Segar, G., 1998. Substance use, health and social problems of clients at 54 drug

180

M. Teesson et al. / Drug and Alcohol Dependence 83 (2006) 174–180

agencies: intake data from the National Treatment Outcome Research Study (NTORS). Br. J. Psychiatry 173, 166–171. Gossop, M., Marsden, J., Stewart, D., Treacy, S., 2002. Change and stability of change after treatment of drug misuse: 2 year outcomes from the National Treatment Outcome Research Study (UK). Addict. Behav. 27, 155–166. Gossop, M., Marsden, J., Stewart, D., Edwards, C., Lehmann, P., Wilson, A., Segar, G., 1997. The National Treatment Outcome Research Study in the United Kingdom: six-month follow-up outcomes. Psychol. Addict. Behav. 11, 324–337. Gossop, M., Marsden, J., Stewart, D., Rolfe, A., 1999. Treatment retention and 1 year outcomes for residential rehabilitation programmes in England. Drug Alcohol Depend. 57, 89–98. Gossop, M., Marsden, J., Stewart, D., Rolfe, A., 2000. Reductions in acquisitive crime and drug use after treatment of addiction problems: 1 year follow-up outcomes. Drug Alcohol Depend., 165–172. Hall, W., Ross, J., Lynskey, M., Law, M., Degenhardt, L., 2000. How many dependent heroin users are there in Australia? Med. J. Aust. 173, 528–531. Hser, Y., Grella, C.E., Hsieh, S., Anglin, M.S., Brown, B., 1999. Prior treatment experience related to process and outcomes in DATOS. Drug Alcohol Depend. 57, 137–150. Hser, Y., Hoffman, V., Grella, C., Anglin, M.D., 2001. A 33-year follow-up of narcotic addicts. Arch. Gen. Psychiatry 58, 503–508. Hubbard, R., Marsden, M., Rachal, J., Harwood, H., Cavanaugh, E., Ginzburg, H., 1989. Drug Abuse Treatment: A National Study of Effectiveness. University of North Carolina Press, Chapel Hill. Hubbard, R., Craddock, S., Flynn, P., Anderson, J., Etheridge, R., 1997. Overview of 1-year follow-up outcomes in the Drug Abuse Treatment Outcome Study (DATOS). Psychol. Addict. Behav. 11, 261–278. Hulse, G., English, D., Milne, E., Holman, C., 1999. The quantification of mortality resulting from the regular use of illicit opiates. Addiction 94, 221–229. Joe, G.W., Broome, K.M., Simpson, D.D., 1999. Retention and patient engagement in models for different treatment modalities in DATOS. Drug Alcohol Depend. 57, 113–125.

Mills, K., Lynskey, M., Ross, J., Teesson, M., Darke, S., 2005. Post traumatic stress disorder among people with heroin dependence in the Australian Treatment Outcome Study (ATOS): prevalence and correlates. Drug Alcohol Depend. 77, 243–249. Oppenheimer, E., Tobutt, C., Taylor, C., Andrew, T., 1994. Death and survival in a cohort of heroin addicts from London clinics: a 22 year follow-up. Addiction 89, 1299–1308. Robins, L., Helzer, J., Croughan, J., Ratcliff, K., 1981. National Institute of Mental Health Diagnostic Interview Schedule: its history, characteristics, and validity. Arch. Gen. Psychiatry 38, 381–389. Ross, J., Teesson, M., Darke, D., Lynskey, M., Ali, R., Ritter, A., Cooke, R., 2005. The characteristics of heroin users entering treatment: Findings from the Australian Treatment Outcome Study (ATOS). Drug Alcohol Rev. 24, 411–418. Sanchez-Carbonell, J., Cami, J., Brigos, B., 1988. Follow-up of heroin addicts in Spain (EMETYST project): results 1 year after treatment admission. Br. J. Addict. 83, 1439–1448. Shand, F., Mattick, R.P.M., 2002. Results from the 4th National Clients of Treatment Service Agencies census: changes in clients’ substance use and other characteristics. Aust. N.Z. J. Public Health. 26, 352–357. Sheehan, M., Oppenheimer, E., Taylor, C., 1993. Opiate users and the first years after treatment: outcome analysis of the proportion of follow-up time spent in abstinence. Addiction 88, 1679–1689. Simpson, D., Joe, G.W., Brown, B.S., 1997. Treatment retention and followup outcomes in The Drug Abuse Treatment Outcome Study (DATOS). Psychol. Addict. Behav. 4, 294–307. Simpson, D., Sells, S., 1982. Effectiveness of treatment for drug abuse: an overview of the DARP research program. Adv. Alcohol Subst. Abuse 2, 7–29. SPSS Inc., 2003. SPSS for Windows (Version 12.0). Chicago. Ware, J., Kosinski, M., Keller, S., 1996. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med. Care. 34, 220–233. World Health Organization, 1993. Composite International Diagnostic Interview (Version 1.1). World Health Organization, Geneva.