Setting

There has been substantial heroin misuse in Dublin since the 1970s. In the early 1990s addiction treatment services expanded substantially, moving away from an abstinence model and towards a harm reduction model (Reference Farrell, Howes and VersterFarrell et al, 1999). Many small treatment clinics were opened in communities where opiate misuse was prevalent. General practitioners were recruited and trained to provide treatment for opiate misuse, offering both methadone maintenance and methadone reduction (Reference ButlerButler, 2002). Heroin misuse accounts for the vast majority of presentations to addiction services in Dublin (Reference Smyth, O'Brien and BarrySmyth et al, 2000).

Cuan Dara opened in 1995, operating as a specialist in-patient drug dependency unit focusing primarily on detoxification. Prior to admission, all patients were expected to have commenced therapeutic work with an addiction counsellor in a community-based treatment service. In addition, all patients underwent a psychiatric assessment to determine psychiatric comorbidity and motivation to detoxify. The standard treatment programme lasted 6 weeks. This included a 10-day methadone detoxification and a benzodiazepine detoxification if indicated. Throughout treatment patients were involved in individual therapy and group therapy. This 6-week admission period is longer than in NTORS (Reference Gossop, Marsden and StewartGossop et al, 1998). Patients were encouraged to access one of two forms of after-care following discharge. They could re-attend their local addiction counsellor or they could access an after-care programme in Cuan Dara one evening each week.

Patients

Consecutive admissions to the unit from July 1995 to December 1996 were included if they met the following criteria: primary diagnosis was opiate dependence syndrome, using ICD–10 criteria (World Health Organization, 1992) and they were admitted with the goal of ceasing use of all opiates, both illicit and prescribed. Baseline information was obtained from the semi-structured interview conducted by a psychiatrist on the day of their admission.

Follow-up interview

The core instrument used for data collection during follow-up was the Maudsley Addiction Profile (Reference Marsden, Gossop and StewartMarsden et al, 1998). This yields information on the 30 days prior to interview. Eight experienced addiction outreach workers conducted the interviews. Their expertise ensured that they had the skills and knowledge to locate patients both via treatment services and through drug users’ peer networks. Follow-up interviews took place between July 1998 and March 1999. It was anticipated that the range in time gaps from discharge to follow-up interview would be wide. This was a consequence of the patients being admitted over an 18-month period and followed-up in an opportunistic manner over a 10-month period. Patients who agreed to participate were paid Ir£10 (€12.50). Following interview, those who described ongoing drug use problems were given advice and directed towards appropriate treatment services.

Statistical analysis

The main outcome variable in this study was attainment of abstinence from opiates during the month prior to follow-up. Abstinence implied that patients were neither misusing opiates nor being prescribed methadone. The main predictor variables were those indicating treatment adherence: completion of detoxification; completion of the 6-week in-patient programme; and attendance at after-care for at least 6 months. We also explored the possibility that pre-treatment patient characteristics might predict abstinence at follow-up. Patients followed-up were compared with those lost to follow-up in order to rule out any systematic bias in the follow-up group.

Categorical variables were compared using Pearson's χ² test or Fisher's exact test, as appropriate. Odds ratios (ORs) and their 95% confidence intervals (95% CIs) are reported to indicate the direction and strength of associations. A multivariate analysis was conducted to identify variables that were independently associated with opiate abstinence. All variables were eligible for entry into the final regression equation. The selection method involved using both the forward and backward stepwise selection techniques, using the likelihood ratio test. The P value for entry was set at 0.05 and that for removal at 0.10. Variables entered into the final regression equation were examined for evidence of interaction.

Admission characteristics and treatment adherence

The cohort admitted to Cuan Dara was substantially younger and had a shorter history of opiate use than cohorts from other countries (Reference Gossop, Marsden and StewartGossop et al, 1998; Reference Broers, Giner and DumontBroers et al, 2000; Reference Chutuape, Jasinski and FingerhoodChutuape et al, 2001; Reference Ghodse, Reynolds and BaldacchinoGhodse et al, 2002). More than 90% were unemployed but they had relatively stable accommodation, with over half of the group living with parents. Over one-third had a sibling who used opiates and almost one-fifth reported a history of parental alcohol misuse. The rates of completion of the methadone detoxification and of the full treatment programme are equivalent to those in other in-patient settings (Reference Gossop, Johns and GreenGossop et al, 1986; Reference Ghodse, London and BewleyGhodse et al, 1987; Reference Polkinghorne, Farrell and FryPolkinghorne et al, 1996; Reference Broers, Giner and DumontBroers et al, 2000).

Follow-up

The follow-up rate achieved in this study is equivalent to that in similar studies (Reference Hubbard, Craddock and FlynnHubbard et al, 1997; Reference Gossop, Marsden and StewartGossop et al, 1999). Nevertheless, loss to follow-up is a concern, as those patients who are difficult to locate may be more likely to be using opiates. The absence of any significant difference between the baseline and treatment adherence characteristics of those followed-up compared with those not located suggests selection bias was not prominent.

The period from discharge to follow-up varied substantially in this study owing to methodological issues already discussed. We found no association between duration of follow-up and abstinence. The NTORS demonstrated that the treatment gains obtained at 1 year remained relatively static at years 2 and 5 (Reference Gossop, Marsden and StewartGossop et al, 2003). Although individual patients may alternate between relapse and abstinence during subsequent years, the proportions of patients moving in each direction tend to cancel each other out beyond the first year after treatment.

Mortality

The five deaths that occurred in this young cohort are consistent with international mortality rates of 1–2 per 100 person-years (Reference Oppenheimer, Tobutt and TaylorOppenheimer et al, 1994; Reference Gossop, Stewart and TreacyGossop et al, 2002). It should be noted that one of the risks associated with abstinence-orientated treatments is accidental overdose following relapse due to the reduction in opiate tolerance (Reference Strang, McCambridge and BestStrang et al, 2003).

Methadone maintenance treatment

Over half of the cohort was on methadone maintenance treatment at follow-up. This indicates that many patients relapsed following discharge and subsequently reaccessed treatment. In Switzerland, Broers et al (Reference Broers, Giner and Dumont2000) found that 35% of those admitted for in-patient opiate detoxification were on methadone maintenance when followed-up after 6 months. Other studies have demonstrated that early relapse is a frequent outcome following in-patient treatment (Reference Chutuape, Jasinski and FingerhoodChutuape et al, 2001). The fact that opiate dependence frequently follows a chronic relapsing course highlights the need for an accessible and comprehensive range of therapeutic interventions for this patient group.

Drug misuse outcomes

The NTORS demonstrated a significant decline in heroin misuse among patients offered residential treatment, from 74% at admission to 49% at 1-year follow-up (Reference Gossop, Marsden and StewartGossop et al, 1999). Chutuape et al (Reference Chutuape, Jasinski and Fingerhood2001) and Broers et al (Reference Broers, Giner and Dumont2000) found that about 30% of patients reported abstinence from heroin 6 months after a brief inpatient opiate detoxification. We found that although 89% of the patients were admitted with a primary problem of heroin dependence, only 41% reported recent heroin misuse at follow-up and only 15% report daily heroin use. Although baseline and follow-up data were obtained using different methodologies, our findings support the view that in-patient treatment is effective in reducing heroin misuse. Among those who completed at least the methadone detoxification phase of treatment, 53% denied any opiate misuse at follow-up. Gossop et al (Reference Gossop, Green and Phillips1989) found an almost identical proportion in their 6-month follow-up study.

The reduction in misuse of heroin cannot be entirely attributed to in-patient treatment. Many patients were on methadone maintenance at follow-up and this will also have contributed to the reduced rates of use. While reliance on self-report of substance misuse at follow-up may be considered a weakness of this study design, similar studies have found that self-report correlates highly with results of urine testing (Reference Gossop, Marsden and StewartGossop et al, 1997; Reference DarkeDarke, 1998).

Achievement of abstinence

At follow-up, 23% of participants had achieved their initial treatment goal of abstinence from opiates without the assistance of methadone maintenance. Most studies examining outcome following in-patient treatment report proportions using heroin before and after treatment, without making it clear that those who are abstaining from heroin at follow-up are not receiving methadone maintenance treatment (Gossop et al, Reference Gossop, Green and Phillips1989, Reference Gossop, Marsden and Stewart1999; Reference Broers, Giner and DumontBroers et al, 2000; Reference Chutuape, Jasinski and FingerhoodChutuape et al, 2001). In seeking to clarify this important issue, we found that almost one in four were genuinely abstinent after an average of 2.5 years. This should be a source of optimism to patients, to commissioners of addiction services and to those who deliver similar services. It should be noted, however, that abstinence during the month prior to follow-up interview does not imply abstinence throughout the follow-up period.

In this era of harm reduction, abstinence has become a secondary goal of treatment services. Dublin has embraced the principles of harm reduction, and a well-developed treatment infrastructure existed at the time of this study (Reference Farrell, Howes and VersterFarrell et al, 1999). This included relatively easy access to methadone maintenance treatment. Reservations have long been expressed that improved access to methadone maintenance might reduce the possibility of drug misusers attaining abstinence (Reference Bratter and PennacchiaBratter & Pennacchia, 1978; Reference Gerlach and SchneiderGerlach & Schneider, 1991). This study indicates that abstinence remains an attainable goal and confirms our first hypothesis.

Abstinent patients were more likely to have completed the in-patient treatment programme and more likely to have attended after-care treatment for at least 6 months. The DATOS failed to demonstrate that better outcome was predicted by a longer stay in short-term in-patient treatment programmes such as that delivered in this study (Reference Hubbard, Craddock and FlynnHubbard et al, 1997). This may be due to differences in treatment delivery in the USA and substantial differences in the patient population: the vast majority of patients in DATOS presented with cocaine dependence. There is much other research consistent with our findings that significant improvement in outcome is associated with better treatment adherence and with transfer to long-term out-patient after-care following in-patient addiction treatment (Reference van de Velde, Schaap and Landvan de Velde et al, 1998; Reference Gossop, Marsden and StewartGossop et al, 1999; Reference Chutuape, Jasinski and FingerhoodChutuape et al, 2001; Reference Ghodse, Reynolds and BaldacchinoGhodse et al, 2002). In-patient treatment is an expensive and limited resource. In order to maximise the health gain that such services can deliver we need to identify more effectively those patients who are most likely to persist with treatment. There is also a need to improve our understanding of the factors within different in-patient and after-care programmes that facilitate patient attendance. Finally, there is a need to develop imaginative measures that can actively enhance treatment adherence at all stages of the treatment process (Reference Horwitz and HorwitzHorwitz & Horwitz, 1993; Reference Giuffrida and TorgensonGiuffrida & Torgenson, 1997).

The finding that a family history of substance misuse was associated with a significant reduction in the likelihood of abstinence was unexpected. There are a number of possible explanations for this. Environmental explanations seem most plausible. Two-thirds of those who reported a family history of substance misuse identified a sibling who was misusing opiates. Returning home to an environment with an opiate-using sibling may have made heroin access easier and promoted relapse (Reference Maisto, Pollock and LynchMaisto et al, 2001). There is also a possible contribution of genetic influences. However, it may simply represent a chance finding as a result of a type 2 statistical error in view of the large number of statistical tests conducted in this study.

In common with many other addiction treatment studies, we found that patient pre-admission characteristics account for a very small proportion of the variance in outcome. Consequently, there is minimal evidence to support their use in prioritising access to in-patient treatment.

This study suggests that in-patient treatment can be effective for opiate-dependent patients, particularly when the patient completes treatment and proceeds to access after-care. In addition, evidence from the USA indicates that it can also be a cost-effective option compared with out-patient treatments (Reference French, Salome and KrupskiFrench et al, 2000). In-patient addiction services must strive to develop strategies to improve rates of programme completion. Commissioners of addiction services should ensure after-care is available and drug dependency units should actively facilitate patient transfer to such services following discharge.

CLINICAL IMPLICATIONS

1. ▪ Almost one in four patients report abstinence from opiates, without the assistance of methadone maintenance, 2–3 years after in-patient treatment of opiate dependence.

2. ▪ Abstinence is associated with completion of the in-patient programme and attendance at after-care. Consequently, the health gain from the in-patient treatment may be enhanced by improving rates of programme completion and ensuring easy access to after-care.

3. ▪ Pre-treatment socio-demographic and drug misuse characteristics are poor predictors of outcome and should not be used for selection of patients for abstinence-orientated in-patient treatment.

LIMITATIONS

1. ▪ Different methodologies were used to measure substance misuse characteristics at baseline and follow-up.

2. ▪ The time from discharge to follow-up varied substantially (from 18 months to 42 months).

3. ▪ In common with most studies of outcome following addiction treatment, the failure to include a control group makes it impossible to determine an effect size.