- Email: [email protected]
Predictors for completing an inpatient detoxification program among intravenous heroin users, methadone substituted and codeine substituted patients
Drug and Alcohol Dependence 64 (2001) 173– 180 www.elsevier.com/locate/drugalcdep
Predictors for completing an inpatient detoxification program among intravenous heroin users, methadone substituted and codeine substituted patients Markus Backmund a,*, Kirsten Meyer a, Dieter Eichenlaub a, Christian G. Schu¨tz b a
Department of Internal Medicine, General Hospital Schwabing Munich, Kolner Platz 1, 80804 Munchen, Germany b Department of Psychiatry, Uni6ersity of Munich, Nußbaumstraße 7, 80336 Mu¨nchen, Germany Received 9 May 2000; accepted 5 January 2001
Abstract Up to 1999 more opioid dependent patients in Germany were substituted with codeine or dihydrocodeine (summarised as codeine) than with methadone. The current retrospective study compares the differences in detoxification treatment outcome for codeine-substituted patients, methadone-substituted patients and patients injecting illicit heroin. The study is based on the medical records of 1070 patients admitted consecutively for opioid and polytox detoxification between 1991 and 1997. The main hypothesis was that injecting illicit-heroin users would complete detoxification treatment less often than codeine- or methadonesubstituted patients, and that methadone-substituted patients who had received more structured treatment would complete more often than codeine-substituted patients who did not receive any structured treatment beyond the prescription of codeine. We analysed a number of demographic and drug related variables as possible predictors. Our bivariate analyses confirmed our main hypothesis: 50.4% (OR: 1.8) of the methadone-substituted patients, 45.5% (OR: 1.5) of the codeine-substituted patients and 35.9% (OR: 1 comparison group) of the injecting illicit-heroin users completed the detoxification program (P = 0.006). This finding remained significant even after correcting for a number of confounders. Using stepwise multiple logistic regression analyses, we found age, education, history of imprisonment, regular contact with a counsellor, currently being on probation and reported plans for participating in an abstinence treatment program to be significant predictors of completing detoxification treatment. Although the current analysis did not rule out differences in pharmacological effects as a contributing factor, the results are consistent with an interpretation of a dose-response association between psychosocial/psychotherapeutic support and detoxification outcome. More psychosocial/psychotherapeutic support leads to better detoxification treatment response. © 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Opioid addiction; Detoxification; Heroin; Methadone; Codeine
1. Introduction In Germany it was not until the late 1980s that methadone maintenance received increased attention as a treatment option for opioid drug dependence. Relatively high numbers of drug deaths and the spread of infectious diseases — especially HIV and Hepatitis C Readers interested in additional demographic data and information about patient background and current life conditions may refer to the journal internet home page at http://www.elsevier.com/locate/ drugalcdepsuppmatz * Corresponding author. Fax: +49-89-30683934. E-mail address: [email protected] (M. Backmund).
— led to the exploration of new treatment approaches. Apart from an acceleration in the establishment of methadone maintenance programs, this led to the introduction of so-called ‘Qualifizierte Entzugsprogramme (qualified detoxification programs)’ (Go¨rgen et al., 1997; Backmund et al., 1998a). The concept of ‘qualified detoxification’ is characterized by the combination of a medication-supported ‘physical’ detoxification and an integrated psychosocial and psychotherapeutic approach for the treatment of dependence. The goal has been the initiation of abstinence treatment starting from the first day of detoxification and a facilitation of transition to a long-term treatment program.
0376-8716/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 6 - 8 7 1 6 ( 0 1 ) 0 0 1 2 2 - 3
174
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
Overall the numbers of methadone clinics stayed small. Given the ‘legal obstacles’, general practitioners (GPs) felt discouraged from prescribing methadone. By contrast there was no obstacle in prescribing codeine and dihydrocodeine (subsumed under codeine in this report). Until recently this led to more patients being treated with codeine than with methadone. For example, in 1995 an estimated 13 500 addicts were treated with methadone compared to more than 20 000 treated with codeine (Krausz et al., 1998). Others estimated the substitution to be even more skewed towards codeine. Patients in a methadone program were submitted to closer controls and to more ‘structured treatment’ than patients substituted by GP’s without the concomitant stringent regulations. This situation changed in 1998 when the German law on Narcotics and Drugs (Beta¨ ubungsmittelgesetz BtmG) was altered. Before 1998 codeine maintenance was less bureaucratic and on the whole a ‘low threshold therapy’. The 1998 change in the law made the previous generous prescription of codeine illegal. Substitution using codeine was now subject to the same restrictions and regulations as methadone. This included regular checks on substance use and required psychosocial treatment. Codeine was declared a second option substitution drug. In Munich, methadone treatment was introduced at the end of the 1980s. However, the availability of methadone substitution remained restricted. Up until 1998 approximately 3500 patients underwent substitution treatment, with a ratio of 6:1 for codeine versus methadone substitution. Since the change in law, the number of substituted patients has dropped to approximately 2500 and the ratio of codeine to methadone has been reversed, (with a small number of patients now being substituted with buprenorphine). Given this situation, we were in a position to compare patients with no previous addiction treatment (I.V. heroin users), and patients from a methadone-substitution program (the most intense treatment), with an intermediate group of codeine-substituted patients, (regular contact to a physician but no additional supervision, e.g. no psychotherapy and no drug-use control). A few studies have been published on the rate of successful detoxification. Reported success rates range between 46 and 81% (Gossop et al., 1986; Endicott and Watson, 1994; Backmund et al., 1998a; Franken and Hendriks, 1999). Patient factors such as gender, age, duration of drug use, education and employment have not been associated with treatment outcome (San et al., 1989; Armenian et al., 1999). Other studies found that patients who left against medical advice were more likely to be younger, (Jeremiah et al., 1995; Armenian et al., 1999), to be single (De los Cobos et al., 1997; Armenian et al., 1999), and to have severe drug use and severe medical problems as measured by the Addiction Severity Index, (Franken and Hendriks, 1999). Psycho-
pathology and socio-demographic variables did not predict the outcome of detoxification in one particular study (Franken and Hendriks, 1999). Treatment options in opioid-dependent patients are still heavily discussed. In an attempt to optimise the use of resources it is important to know which patients are most likely to be successful given the current detoxification conditions. Does patient preparation within a substitution program, improve treatment outcome? What characterizes patients who leave detoxification programs against medical advice? These factors may suggest treatment alternatives or even changes in the current approach to patients who leave detoxification programs early. It is well known that leaving a program against medical advice is associated with an overall poor prognosis. This may make interventions to encourage completion — or even interventions to allow patients to change to a different setting — even more important.
2. Method
2.1. Sample The sample consisted of all subjects admitted to a detoxification unit between April 1991 and March 1997. All subjects had to be admitted voluntarily and had to meet the current DSM IV diagnosis of opioid dependence: dependence on heroin, methadone or codeine, or a combination of one of these opioids and any other psychotropic drug. Contact with the patients was established by a telephone hotline. Patients could come in of their own volition, and previous contact with a counsellor or physician was not necessary. Intakes for the unit occurred between Monday and Wednesday. Order of intake was determined by perceived severity of drug-associated problems, and by a waiting list. For those patients who were re-admitted, only the first admission within the given time frame was included in the study so as not to violate the assumption of independent observations.
2.2. Treatment setting The detoxification unit is a physically separated ward of a large medical department within a general hospital. It is the former director’s mansion — thus colloquially called the ‘Villa’ — within the hospital compound. A total of 15 beds in seven bedrooms are available for opioid detoxification. The treatment strategy called ‘qualified detoxification’ consists of medication-supported detoxification (almost exclusively methadone), diagnosis and treatment of other diseases (hepatitis B, C, AIDS, pneumonia etc.), and daily group and individual psychotherapy and social therapy with the goal
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
of facilitating the transition to abstinence-treatment following detoxification (Backmund et al., 1998a). The treatment concept is based on a Minnesota Model-like approach. Individual, dynamically oriented brief psychotherapeutic interventions were added. The Villa is conceptualised as a ‘low threshold’ treatment. Patients were admitted as soon as there was an opening. No preparation or pre-conditions were required and thus the group always consisted of patients at different stages of treatment. The treatment regime was independent of the substance used. Opioid detoxification was conducted using methadone as an agonist for tapering dosage so as to reduce withdrawal symptoms (Razani et al., 1975; Bickel et al., 1988; San et al., 1989, 1990). In the first 24 h the methadone dose was fixed according to clinical evaluation with the aim of keeping withdrawal symptoms to a minimum. The initial dose was adjusted by the physician according to the patient’s self-reported opioid dose and the results of the urine screening tests. The current average grade of heroin purity as found by the police was used to estimate the heroin dose (the analyses showed an average purity of 33%). If patients were diagnosed with additional substance dependencies such as benzodiazepines or alcohol, additional treatment with diazepam, carbamazepine or clomethiazol was initiated. The medication-supported detoxification treatment made it possible for most of the patients to participate in the socio-psychotherapeutic part of the program from the day of admission. Therapy involved two group and one individual session each day. Thus the duration of taper was dependent on the initial dose necessary to suppress withdrawal symptoms. The initial dose of methadone differed in the three groups. In the methadone group average the initial dose was the highest (78.6938.9 mg). In the codeine group the average was lower (48.8933.8 mg). Initial dose was lowest in the heroin group (35.69 26.1 mg). Tapering followed a rigid schedule. Methadone was reduced by 5 mg each day except on Sundays. No additional medication was used in cases of pure opioid dependence. Withdrawal symptoms were assessed by the physicians on a regular basis and were based on clinical impressions. The unit was attended by two physicians, one psychologist, six social workers and six nurses. No outside visitors were allowed during hospital stay.
2.3. Definition of completers Patients were considered ‘completers’ if they stayed for a minimum of 14 days, finished the physical detoxification with a negative drug-screening urine analysis, (using an immunoassay (TriageTM®) and KIMS (kinetic interaction of microparticles in a solution), had a complete absence of withdrawal symptoms and had also finished the psychotherapeutic program. The psy-
175
chotherapeutic program starts on the day of admission and ends 8 days after physical detoxification. All patients who did not fulfil any of these criteria were considered ‘non-completers’, including those patients who finished the physical detoxification but not the psychotherapeutic program.
2.4. Opioid groups and data collection Patients were retrospectively allocated to three groups: (i) the methadone group; (ii) the codeine group; and (iii) the heroin group. Allocation was based on self-reported daily consumed opioids, the results of urine screening tests on the day of admission, and the reports of general practitioners in the case of methadone or codeine substitution. Patients fulfilling the conditions of more than one group, e.g. daily use of methadone and daily use of codeine, were included in the group with the more stringent pre-detoxification treatment conditions, (for example in this case, the methadone group). Apart from the main hypothesis that pre-detoxification treatment would predict outcome, other possible predictors of completing the treatment were also analysed. These included: age, sex, education, initial methadone dose, daily consumption of cannabis, cocaine, benzodiazepines, barbiturates and alcohol, duration of opioid use, age of onset of opioid use, history of drug emergency treatment, history of suicide attempts, history of long term abstinence treatment, history of imprisonment, employment status, marital status, drug dependency of significant other, drug dependency in patient’s family, regular contact with a counsellor, self referral or referral from a professional (counselling centre, GP, etc.), waiting list, currently being on probation, court orders and reported plans for abstinence treatment. Assessment was done by interviewers. Patients were interviewed using a standardized protocol. Urine samples were collected on admission day, once every week and then on the final day of treatment. Samples were analysed for methadone, other opiates, cocaine, benzodiazepines, amphetamines and tetrahydrocannabinol.
2.5. Method of analysis After cross-tabulation and bivariate analysis, logistic regression modelling was applied. Of all the patients enrolled in the treatment program between April 1991 and March 1997, analysis was restricted to those coming for the first detoxification. Because preliminary analysis indicated non-linear associations involving age, age at first opioid use, and duration of opioid drug use, these variables were transformed from ordinal to categorical variables. Bivariate analyses were performed for the variable of interest and for all other potentially confounding vari-
176
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
ables. Stepwise multiple logistic regression modelling was used to hold possible confounders constant when estimating the key association. Confounders were added in groups: (1) basic socio-demographics; (2) drug-use history; (3) patient background; (4) current living conditions; (5) circumstances of treatment entry; and (6) future treatment plans. Individual factors were kept in the model if significant. Results are summarized by reporting the models at each step relying on those variables meeting the criterion of inclusion: not entrapping 1 in the 95% confidence interval. Tests for interactions were used to check on the uniformity assumption under which multiple logistic regression estimates are derived.
3. Results From 1991 until 1997 a total of 1685 opioid addicts were admitted to the detoxification unit. For the 1070 patients included in the analyses, the median age was 28 years, 65% were male, and the median duration of opioid use was 10 years. 133 addicts reported daily consumption of methadone (METHADONE GROUP), 661 the daily consumption of codeine (CODEINE GROUP) and 276 the daily injection of heroin (HEROIN GROUP). A total of 43.6% of the patients completed the detoxification treatment program. The bivariate analysis indicated that methadone-dependent and codeine-dependent patients finished the detoxification treatment significantly more often than the heroin-dependent patients (P B 0.01). This is consistent with our main hypothesis. Table 1 presents the bivariate analysis on the key variable and the possible confounding variables predicting the completion of detoxification treatment. As shown, completion of detoxification treatment was predicted by older age, better education, the absence of an addicted partner, the absence of dependency in the family, regular contact with a counsellor, being currently on probation and the reporting of future treatment plans (Table 1). Our key finding persisted even when socio-demographics were included in the model. Sex was not a significant predictor of completing detoxification treatment. Age and education were both significantly associated. Education was divided into three groups: less than, equal to, or more than the mandatory 9 years of schooling. Race was not assessed since almost all the patients were Caucasian (Tables showing these data are available as Supplementary Materials on the journal website, www.elsevier.nl/locate/drugalcdep). Similarly, the drug use variables were included in the second step. None turned out to be significantly involved in predicting outcome. In the third step, patient background was added to the model. This included: history of drug
related emergencies, suicide attempts, imprisonment, and drug treatment. The only significant predictor was history of imprisonment. The next step (‘current life conditions’) consisted of employment, marital status, drug use in significant others, substance dependence in a family member, and counselling. When these variables were added to the model, ‘current counselling’ was the only significant predictor. Circumstances of treatment entry were added next and only one variable, ‘being currently on probation’, significantly added to the model (data in Supplementary Materials). The final model was achieved after the inclusion of ‘reported plans for future abstinence treatment’. In this instance none of the previously entered significant predictors became insignificant and our variable of interest, predetoxification treatment, also remained significant, even if only marginally (Table 2). We also tested for interactions to check on the uniformity assumption of the multiple logistic regression estimates. Not one single significant interaction was found.
4. Discussion The purpose of this paper was to compare detoxification treatment outcomes of methadone- and codeinesubstituted patients with those for patients using intravenous heroin. The hypothesis was that methadone-substituted and codeine-substituted patients would complete the treatment more often than the users of illicit heroin. Furthermore, since codeine-substituted patients are exposed to less treatment they should represent an intermediate group between methadone-substituted patients and heroin-using patients. Since both methadone- and codeine-dependent patients completed the treatment more often than heroin-dependent patients, and since the codeine group was intermediate between the methadone and heroin group, the main hypothesis was confirmed. Similar results were reported in a pilot study with a small sample of 71 patients (Backmund et al., 1998b). A possible explanation is that methadone maintenance (Gro¨ nbladh et al., 1990; Dole et al., 1991Marsch, 1998) and codeine (Krausz et al., 1998) programs stabilise the addicts. Hence, the decision and motivation to begin and finish a detoxification treatment can be sustained by the patient. Codeine also does this but to a lesser degree. Independent of this motivation, the following variables were also associated with increased completion of detoxification treatmen: (1) increase in age predicted an increase in treatment completion; however, this was not a continuous increase, and a ceiling effect occurred at approximately age 28, with no subsequent improvement in completion rate; (2) better education marginally predicted completion; (3) history of imprisonment was a
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
177
Table 1 Cross-tabulation Variable
Completers n
%
O.R.
95% C.I.
P
Main hypothesis Drug groups Methadone Codeine Heroin
133 661 276
50.4 45.5 35.9
1.81 1.49 1
1.19–2.76 1.12–2.00
0.0064 0.0054 0.0065
Basic socio-demographics Sex Female Male
380 690
40.8 45.2
0.83 1
0.65–1.08
0.1624
Age groups B23 years 23–28 years 29–34 years \34 years
216 335 326 190
38.9 38.8 50.3 45.3
0.77 0.77 1.22 1
0.52–1.14 0.53–1.10 0.86–1.75
0.0107 0.1943 0.1489 0.2691
Education Max 9 years, without qualification 9 years, qualified \9 years
93 648 329
33.3 42.3 49.2
0.52 0.76 1
0.32–0.84 0.58–0.99
0.0135 0.0071 0.0389
Initially necessary methadone dose 0–40 mg 41–80 mg 81–200 mg
381 434 255
44.1 40.1 48.0
0.85 0.72 1
0.62–1.17 0.53–0.99
0.1211 0.33 0.04
Daily hash consumption No Yes
878 192
43.2 45.8
0.90 1
0.66–1.23
0.4998
1036 34
44.0 32.4
1.64 1
0.79–3.41
0.1815
Daily benzodiazepine consumption No Yes
596 474
41.8 46.0
0.84 1
0.66–1.08
0.1676
Daily barbiturate consumption No Yes
995 75
43.3 48.0
0.83 1
0.52–1.32
0.4309
Daily alcohol consumption No Yes
772 298
42.5 46.6
0.85 1
0.65–1.11
0.2192
Duration of drug addiction 1 year 2–5 years 6–10 years 11–15 years \15 years
78 394 251 160 187
52.6 40.6 41.8 46.3 46.5
1.27 0.79 0.83 0.99 1
0.75–2.16 0.55–1.12 0.56–1.21 0.65–1.51
0.2509 0.3703 0.1782 0.3279 0.9593
Age onset 1–9.99 years 10–13.99 years 14–17.99 years 18–21.99 years \22 years
18 58 301 390 303
38.9 53.4 41.5 43.1 44.9
0.78 1.41 0.87 0.93 1
0.30–2.07 0.80–2.48 0.63–1.20 0.69–1.26
0.5265 0.6198 0.2321 0.4052 0.6343
Patient background History of emergencies No Yes
697 373
42.0 43.6
0.83 1
0.64–1.07
0.1474
Drug use history
Daily cocaine consumption No Yes
178
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
Table 1 (Continued) Variable
Completers n
%
O.R.
95% C.I.
P
History of suicide attempts No Yes
854 216
42.6 47.7
0.82 1
0.60–1.10
0.1806
History of therapy No therapy Therapy not completed Therapy completed Therapy n.c. and c.
722 123 135 90
43.2 37.4 45.2 53.3
0.67 0.52 0.72 1
0.43–1.03 0.30–0.91 0.42–1.23
0.1377 0.0697 0.0213 0.2315
History of imprisonment Never 1 \1
618 277 175
45.6 43.7 36.6
1.46 1.35 1
1.03–2.06 0.91–1.98
0.1042 0.0334 0.1347
Current life conditions Employment No Yes
772 298
44.3 41.9
1.10 1
0.84–1.44
0.4865
Marital status Single Married, living with a partner Divorced, widowed
701 324 45
43.9 41.4 55.6
0.63 0.56 1
0.34–1.15 0.30–1.06
0.1960 0.1315 0.0742
Drug use of significant others No Yes
356 714
49.4 40.8
1.42 1
1.10–1.84
0.0071
Family member with substance dependency No Yes
400 670
48.8 40.6
1.39 1
1.08–1.79
0.0094
Counselling No Yes
569 501
38.0 50.1
0.61 1
0.48–0.78
0.0001
Circumstances of treatment entry Contact Counsellor, practitioner Directly scene
565 505
42.1 45.3
0.88 1
0.69–1.12
0.2887
Waiting time until admission B3 weeks 3–6 weeks \6 weeks
231 156 683
42.4 39.7 44.9
0.90 0.81 1
0.67–1.22 0.57–1.15
0.4550 0.5044 0.2378
Currently on probation No Yes
742 328
41.4 48.8
0.74 1
0.58–0.96
0.0245
Treatment moti6ation Reported future plan Nothing Work Outpatient abstinence therapy Inpatient abstinence therapy
339 267 149 315
33.9 41.9 46.3 54.3
0.43 0.61 0.73 1
0.32–0.59 0.44–0.85 0.49–1.07
0.0000 0.0000 0.0031 0.1089
very strong predictor; no imprisonment indicated higher rates of completion. This was an interesting finding insofar as the factor was not predictive on its own. To get more information about this phenomenon we used logistic regression and added a number of
possible confounders. By adding the factor, ‘age groups’, the factor ‘history of imprisonment’ became significant. Thus age turned out to be a variable that ‘masked’ the effect of imprisonment as a risk factor; (4) patients currently in some form of counselling had a
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
higher rate of completion. This may indicate that preparation in the form of counselling may induce better outcomes; (5) patients currently on probation also had a higher rate of completion, possibly indicating that avoiding imprisonment might work as a motivator to complete treatment; (6) ‘reported future treatment plans’ consisted of patients reporting plans to start any form of abstinence treatment program after completion of the detoxification program. This included short (3 months) and long term (6–12 months) inpatient or outpatient abstinence treatment. ‘Reported future treatment plans’ can be considered as a measurement of treatment motivation. Thus, we were not surprised to find that this factor had the highest significance level. We were some-
Table 2 Final model Variable
O.R.
95% C.I.
P
Main hypothesis Drug groups Methadone Codeine Heroin
1.42 1.45 1
0.91–2.22 1.07–1.96
0.0495 0.1212 0.0158
Basic socio-demographics Age groups B23 years 23–28 years 29–34 years \34 years
0.55 0.57 1.03 1
0.35–0.85 0.38–0.84 0.71–1.51
0.0002 0.0075 0.0043 0.8609
Education B9 years, without Qualification 9 years, qualification \9 years
0.54 0.81 1
0.33–0.89 0.61–1.07
0.0422 0.0153 0.1343
Patient background History of imprisonment Never 1 \1
2.08 1.68 1
1.41–3.06 1.11–2.53
0.0011 0.0002 0.0132
Current life conditions Counselling No Yes
0.71 1
0.54–0.93
0.0124
Circumstances of treatment entry Currently on probation No Yes
0.69 1
0.51–0.92
0.0107
Treatment moti6ation Reported future plan Nothing Work Outpatient abstinence therapy Inpatient abstinence therapy
0.48 0.65 0.74 1
0.34–0.68 0.45–0.94 0.49–1.12
0.0006 0.0000 0.0031 0.1089
179
what surprised though, that when motivation was entered into the model it hardly changed the significance level of the main predictor for completing treatment (methadone, codeine and heroin group). This indicates that the differences in completion for these three groups is not due to differences in motivation as measured by reported future treatment plans. A number of factors did not predict treatment completion including sex; although females completed detoxification less often than males, the finding was not significant. Initial methadone dose also did not predict completion of detoxification. Initial methadone dose can be regarded as a measurement of the severity of dependence and thus this result can be taken to indicate that severity of dependence is not associated with detoxification outcome. Daily consumption of other drugs, and age of onset, was not predictive. Duration of use was U-shaped with the highest rates of treatment completion amongst those with recent onset (less than 1 year) and those with longer use (more than 10 years). However, this result was not significant. History of emergency treatment and history of suicide attempts were not significant predictors. Although emergency room visits and history of suicide attempts have been reported to increase the likelihood of entering detoxification (Schu¨ tz et al., 1994), they do not seem to be associated with completing a detoxification program. That history of treatment was not a significant predictor was somewhat surprising, since the clinical impression had been that patients with experience of treatment did better. This lack of significance might have been due partially to the fact that the groups with therapy experience were rather small (90, 123, 135 individuals per group). Drug use of significant other and substance abuse in family members were significant as individual factors, but did not reach significance when included in the model. This means that opiate group, age, education, history of imprisonment, and current counselling, accounted for the predictive power of living with a drug-using significant other and substance user in the family. Waiting time was another factor that did not predict completion of treatment. This study is a retrospective investigation using a ‘natural experiment’. The restrictions associated with this approach therefore, apply. For example, patients were not randomly assigned to the three different groups, and factors associated with group assignment might have influenced the outcome. A number of factors not measured in the present study might also be responsible for the differences between these three opiate groups. One such factor could be the differences in withdrawal symptomatology between these groups. Methadone has been reported to be one of the best drugs to reduce opioid withdrawal syndrome, (Razani et al., 1975; Bickel et al., 1988; San et al., 1989, 1990; Wilson and DiGeroge, 1993; Janiri et al., 1994; San et al., 1994; Bearn et al., 1996). Initial methadone dose was lowest among heroin depen-
180
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
dent patients. In this study, heroin dependent patients still completed the treatment less often. Methadone withdrawal symptoms are reported to be more severe than heroin withdrawal symptoms (Gossop and Strang, 1991). We therefore, cannot rule out that differences in pharmacological effects may have played a role. The results support the hypothesis that intensity of previous treatment improves outcome of abstinence induction. However, the current analysis does not rule out that factors such as differences in the pharmacological effects of the different opioids might have contributed to these findings. Drug addicts treated with methadone have a better outcome in detoxification treatment than codeine-substituted patients and patients not in any substitution treatment. The better outcome among methadone treated patient is consistent with the hypothesis that the more psychosocial/psychotherapeutic support these patients receive before starting detoxification, (compared to codeine-substituted patients and heroin-injecting patients), the better the treatment outcome.
Acknowledgements This study was supported by ‘Modellprogramm Kompakttherapie im Verbund der Drogenhilfe 1990– 1995’ from the Department of Health from the Federal Republic of Germany. The authors would like to thank the staff from the detoxification unit ‘Villa’ for their help with data collection and entry.
References Armenian, S.H., Chutuape, M.A., Stitzer, M.L., 1999. Predictors of discharges against medical advice from a short-term hospital detoxification unit. Drug Alcohol Depend. 56, 1 –8. Backmund, M., Eichenlaub, D., Soyka, M., 1998a. Evaluation of a qualified detoxification treatment of intravenous drug addicts in a general hospital — a Federal German model project. Gesundheitswesen 60, 552 – 557. Backmund, M., Meyer, K., Habsch, J., Soyka, M., Eichenlaub, D., 1998b. Qualifizierte Entzugsbehandlung Drogenabha¨ ngiger — Mo¨ gliche Pra¨ diktoren des Therapieverlaufs. Mu¨ nch. med. Wochenschr. 44, 590 –593. Bearn, J., Gossop, M., Strang, J., 1996. Randomised double-blind comparison of lofexidine and methadone in the in-patient treatment of opiate withdrawal. Drug Alcohol Depend. 43, 87 – 91.
Bickel, W.K., Stitzer, M.L., Bigelow, G.E., Liebson, I.A., Jasinski, D.R., Johnson, R.E., 1988. A clinical trial of buprenorphine: comparison with methadone in the detoxification of heroin addicts. Clin. Pharmacol. Ther. 43, 72 – 78. De los Cobos, J.P., Trujols, J., Ribalta, E., Casas, M., 1997. Cocaine use immediately prior to entry in an inpatient heroin detoxification unit as a predictor of discharges against medical advice. Am. J. Drug Alcohol Abuse 23, 267 – 279. Dole, V.P., Newman, R.G., Jospeh, H., 1991. Heroin addiction: a public health problem. Med. N. Amer. 15, 1943 – 1945. Endicott, P., Watson, B., 1994. Interventions to improve the AMA discharge rate for opiate-addicted patients. J. Psychsoc. Nurs. Ment. Health Serv. 32, 36 – 40. Franken, I.H., Hendriks, V.M., 1999. Predicting outcome of inpatient detoxification of substance abusers. Psychiatr. Serv. 50, 813 –817. Go¨ rgen, W., Mo¨ ller, I., Oliva, H., 1997. Einscha¨ tzungen und Ergebnisse einer qualifizierten Entzugsbehandlung fu¨ r Drogenabha¨ ngige aus Patientensicht. Sucht 42, 82 – 91. Gossop, M., Johns, A., Green, L., 1986. Opiate withdrawal: inpatient versus outpatient programs and preferred versus random assignment to treatment. Br. Med. J. 293, 103 – 104. Gossop, M., Strang, I., 1991. A comparison of the withdrawal responses of heroin and methadone addicts during detoxification. Br. J. Psychiatry 158, 697 – 699. Gro¨ nbladh, L., O8 hlund, L.S., Gunne, L.M., 1990. Mortality in heroin addiction impact of methadone treatment. Acta Psych. Scand. 82, 223 – 227. Janiri, L., Mannellli, P., Persico, A.M., Serretti, A., Tempesta, E., 1994. Opiate detoxification of methadone maintenance patients using lefetamine, clonidine and buprenorphine. Drug Alcohol Depend. 36, 139 – 145. Jeremiah, J., O’Sullivan, P., Stein, M.D., 1995. Who leaves against medical advice. J. Gen. Intern. Med. 10, 403 – 405. Krausz, M., Verthein, U., Degkwitz, P., Haasen, C., Raschke, P., 1998. Maintenance treatment of opiate addicts in Germany with medications containing codeine — results of a follow-up study. Addiction 93, 1161 – 1167. Razani, J., Chisholm, D., Glasser, M., Kappeler, T., 1975. Self-regulated methadone detoxification of heroin addicts. An improved technique in an inpatient setting. Arch. Gen. Psychiatry 32, 909 – 911. San, L., Camı´, J., Peri, J.M., Mata, R., Porta, M., 1989. Success and failure at inpatient heroin detoxification. Br. J. Addict. 84, 81 –87. San, L., Camı´, J., Peri, J.M., Mata, R., Porta, M., 1990. Efficacy of clonidine, guanfacine and methadone in the rapid detoxification of heroin addicts: a controlled clinical trail. Br. J. Addict. 85, 141 – 147. San, L., Fernandez, T., Camı´, J., Gossop, M., 1994. Efficacy of methadone versus methadone and guanfacine in the detoxification of heroin-addicted patients. J. Subst. Abuse Treat. 11, 463 –469. Schu¨ tz, C.G., Rapiti, E., Vlahov, D., Anthony, J.C., 1994. Suspected determinants of enrolment into detoxification and methadone maintenance treatment among injecting drug users. Drug Alcohol Depend. 36, 129 – 138. Wilson, R.S., DiGeroge, W.S., 1993. Methadone combined with clonidine versus clonidine alone in opiate detoxification. J. Subst. Abuse Treat. 10, 529 – 535.
Predictors for completing an inpatient detoxification program among intravenous heroin users, methadone substituted and codeine substituted patients Markus Backmund a,*, Kirsten Meyer a, Dieter Eichenlaub a, Christian G. Schu¨tz b a
Department of Internal Medicine, General Hospital Schwabing Munich, Kolner Platz 1, 80804 Munchen, Germany b Department of Psychiatry, Uni6ersity of Munich, Nußbaumstraße 7, 80336 Mu¨nchen, Germany Received 9 May 2000; accepted 5 January 2001
Abstract Up to 1999 more opioid dependent patients in Germany were substituted with codeine or dihydrocodeine (summarised as codeine) than with methadone. The current retrospective study compares the differences in detoxification treatment outcome for codeine-substituted patients, methadone-substituted patients and patients injecting illicit heroin. The study is based on the medical records of 1070 patients admitted consecutively for opioid and polytox detoxification between 1991 and 1997. The main hypothesis was that injecting illicit-heroin users would complete detoxification treatment less often than codeine- or methadonesubstituted patients, and that methadone-substituted patients who had received more structured treatment would complete more often than codeine-substituted patients who did not receive any structured treatment beyond the prescription of codeine. We analysed a number of demographic and drug related variables as possible predictors. Our bivariate analyses confirmed our main hypothesis: 50.4% (OR: 1.8) of the methadone-substituted patients, 45.5% (OR: 1.5) of the codeine-substituted patients and 35.9% (OR: 1 comparison group) of the injecting illicit-heroin users completed the detoxification program (P = 0.006). This finding remained significant even after correcting for a number of confounders. Using stepwise multiple logistic regression analyses, we found age, education, history of imprisonment, regular contact with a counsellor, currently being on probation and reported plans for participating in an abstinence treatment program to be significant predictors of completing detoxification treatment. Although the current analysis did not rule out differences in pharmacological effects as a contributing factor, the results are consistent with an interpretation of a dose-response association between psychosocial/psychotherapeutic support and detoxification outcome. More psychosocial/psychotherapeutic support leads to better detoxification treatment response. © 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Opioid addiction; Detoxification; Heroin; Methadone; Codeine
1. Introduction In Germany it was not until the late 1980s that methadone maintenance received increased attention as a treatment option for opioid drug dependence. Relatively high numbers of drug deaths and the spread of infectious diseases — especially HIV and Hepatitis C Readers interested in additional demographic data and information about patient background and current life conditions may refer to the journal internet home page at http://www.elsevier.com/locate/ drugalcdepsuppmatz * Corresponding author. Fax: +49-89-30683934. E-mail address: [email protected] (M. Backmund).
— led to the exploration of new treatment approaches. Apart from an acceleration in the establishment of methadone maintenance programs, this led to the introduction of so-called ‘Qualifizierte Entzugsprogramme (qualified detoxification programs)’ (Go¨rgen et al., 1997; Backmund et al., 1998a). The concept of ‘qualified detoxification’ is characterized by the combination of a medication-supported ‘physical’ detoxification and an integrated psychosocial and psychotherapeutic approach for the treatment of dependence. The goal has been the initiation of abstinence treatment starting from the first day of detoxification and a facilitation of transition to a long-term treatment program.
0376-8716/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 6 - 8 7 1 6 ( 0 1 ) 0 0 1 2 2 - 3
174
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
Overall the numbers of methadone clinics stayed small. Given the ‘legal obstacles’, general practitioners (GPs) felt discouraged from prescribing methadone. By contrast there was no obstacle in prescribing codeine and dihydrocodeine (subsumed under codeine in this report). Until recently this led to more patients being treated with codeine than with methadone. For example, in 1995 an estimated 13 500 addicts were treated with methadone compared to more than 20 000 treated with codeine (Krausz et al., 1998). Others estimated the substitution to be even more skewed towards codeine. Patients in a methadone program were submitted to closer controls and to more ‘structured treatment’ than patients substituted by GP’s without the concomitant stringent regulations. This situation changed in 1998 when the German law on Narcotics and Drugs (Beta¨ ubungsmittelgesetz BtmG) was altered. Before 1998 codeine maintenance was less bureaucratic and on the whole a ‘low threshold therapy’. The 1998 change in the law made the previous generous prescription of codeine illegal. Substitution using codeine was now subject to the same restrictions and regulations as methadone. This included regular checks on substance use and required psychosocial treatment. Codeine was declared a second option substitution drug. In Munich, methadone treatment was introduced at the end of the 1980s. However, the availability of methadone substitution remained restricted. Up until 1998 approximately 3500 patients underwent substitution treatment, with a ratio of 6:1 for codeine versus methadone substitution. Since the change in law, the number of substituted patients has dropped to approximately 2500 and the ratio of codeine to methadone has been reversed, (with a small number of patients now being substituted with buprenorphine). Given this situation, we were in a position to compare patients with no previous addiction treatment (I.V. heroin users), and patients from a methadone-substitution program (the most intense treatment), with an intermediate group of codeine-substituted patients, (regular contact to a physician but no additional supervision, e.g. no psychotherapy and no drug-use control). A few studies have been published on the rate of successful detoxification. Reported success rates range between 46 and 81% (Gossop et al., 1986; Endicott and Watson, 1994; Backmund et al., 1998a; Franken and Hendriks, 1999). Patient factors such as gender, age, duration of drug use, education and employment have not been associated with treatment outcome (San et al., 1989; Armenian et al., 1999). Other studies found that patients who left against medical advice were more likely to be younger, (Jeremiah et al., 1995; Armenian et al., 1999), to be single (De los Cobos et al., 1997; Armenian et al., 1999), and to have severe drug use and severe medical problems as measured by the Addiction Severity Index, (Franken and Hendriks, 1999). Psycho-
pathology and socio-demographic variables did not predict the outcome of detoxification in one particular study (Franken and Hendriks, 1999). Treatment options in opioid-dependent patients are still heavily discussed. In an attempt to optimise the use of resources it is important to know which patients are most likely to be successful given the current detoxification conditions. Does patient preparation within a substitution program, improve treatment outcome? What characterizes patients who leave detoxification programs against medical advice? These factors may suggest treatment alternatives or even changes in the current approach to patients who leave detoxification programs early. It is well known that leaving a program against medical advice is associated with an overall poor prognosis. This may make interventions to encourage completion — or even interventions to allow patients to change to a different setting — even more important.
2. Method
2.1. Sample The sample consisted of all subjects admitted to a detoxification unit between April 1991 and March 1997. All subjects had to be admitted voluntarily and had to meet the current DSM IV diagnosis of opioid dependence: dependence on heroin, methadone or codeine, or a combination of one of these opioids and any other psychotropic drug. Contact with the patients was established by a telephone hotline. Patients could come in of their own volition, and previous contact with a counsellor or physician was not necessary. Intakes for the unit occurred between Monday and Wednesday. Order of intake was determined by perceived severity of drug-associated problems, and by a waiting list. For those patients who were re-admitted, only the first admission within the given time frame was included in the study so as not to violate the assumption of independent observations.
2.2. Treatment setting The detoxification unit is a physically separated ward of a large medical department within a general hospital. It is the former director’s mansion — thus colloquially called the ‘Villa’ — within the hospital compound. A total of 15 beds in seven bedrooms are available for opioid detoxification. The treatment strategy called ‘qualified detoxification’ consists of medication-supported detoxification (almost exclusively methadone), diagnosis and treatment of other diseases (hepatitis B, C, AIDS, pneumonia etc.), and daily group and individual psychotherapy and social therapy with the goal
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
of facilitating the transition to abstinence-treatment following detoxification (Backmund et al., 1998a). The treatment concept is based on a Minnesota Model-like approach. Individual, dynamically oriented brief psychotherapeutic interventions were added. The Villa is conceptualised as a ‘low threshold’ treatment. Patients were admitted as soon as there was an opening. No preparation or pre-conditions were required and thus the group always consisted of patients at different stages of treatment. The treatment regime was independent of the substance used. Opioid detoxification was conducted using methadone as an agonist for tapering dosage so as to reduce withdrawal symptoms (Razani et al., 1975; Bickel et al., 1988; San et al., 1989, 1990). In the first 24 h the methadone dose was fixed according to clinical evaluation with the aim of keeping withdrawal symptoms to a minimum. The initial dose was adjusted by the physician according to the patient’s self-reported opioid dose and the results of the urine screening tests. The current average grade of heroin purity as found by the police was used to estimate the heroin dose (the analyses showed an average purity of 33%). If patients were diagnosed with additional substance dependencies such as benzodiazepines or alcohol, additional treatment with diazepam, carbamazepine or clomethiazol was initiated. The medication-supported detoxification treatment made it possible for most of the patients to participate in the socio-psychotherapeutic part of the program from the day of admission. Therapy involved two group and one individual session each day. Thus the duration of taper was dependent on the initial dose necessary to suppress withdrawal symptoms. The initial dose of methadone differed in the three groups. In the methadone group average the initial dose was the highest (78.6938.9 mg). In the codeine group the average was lower (48.8933.8 mg). Initial dose was lowest in the heroin group (35.69 26.1 mg). Tapering followed a rigid schedule. Methadone was reduced by 5 mg each day except on Sundays. No additional medication was used in cases of pure opioid dependence. Withdrawal symptoms were assessed by the physicians on a regular basis and were based on clinical impressions. The unit was attended by two physicians, one psychologist, six social workers and six nurses. No outside visitors were allowed during hospital stay.
2.3. Definition of completers Patients were considered ‘completers’ if they stayed for a minimum of 14 days, finished the physical detoxification with a negative drug-screening urine analysis, (using an immunoassay (TriageTM®) and KIMS (kinetic interaction of microparticles in a solution), had a complete absence of withdrawal symptoms and had also finished the psychotherapeutic program. The psy-
175
chotherapeutic program starts on the day of admission and ends 8 days after physical detoxification. All patients who did not fulfil any of these criteria were considered ‘non-completers’, including those patients who finished the physical detoxification but not the psychotherapeutic program.
2.4. Opioid groups and data collection Patients were retrospectively allocated to three groups: (i) the methadone group; (ii) the codeine group; and (iii) the heroin group. Allocation was based on self-reported daily consumed opioids, the results of urine screening tests on the day of admission, and the reports of general practitioners in the case of methadone or codeine substitution. Patients fulfilling the conditions of more than one group, e.g. daily use of methadone and daily use of codeine, were included in the group with the more stringent pre-detoxification treatment conditions, (for example in this case, the methadone group). Apart from the main hypothesis that pre-detoxification treatment would predict outcome, other possible predictors of completing the treatment were also analysed. These included: age, sex, education, initial methadone dose, daily consumption of cannabis, cocaine, benzodiazepines, barbiturates and alcohol, duration of opioid use, age of onset of opioid use, history of drug emergency treatment, history of suicide attempts, history of long term abstinence treatment, history of imprisonment, employment status, marital status, drug dependency of significant other, drug dependency in patient’s family, regular contact with a counsellor, self referral or referral from a professional (counselling centre, GP, etc.), waiting list, currently being on probation, court orders and reported plans for abstinence treatment. Assessment was done by interviewers. Patients were interviewed using a standardized protocol. Urine samples were collected on admission day, once every week and then on the final day of treatment. Samples were analysed for methadone, other opiates, cocaine, benzodiazepines, amphetamines and tetrahydrocannabinol.
2.5. Method of analysis After cross-tabulation and bivariate analysis, logistic regression modelling was applied. Of all the patients enrolled in the treatment program between April 1991 and March 1997, analysis was restricted to those coming for the first detoxification. Because preliminary analysis indicated non-linear associations involving age, age at first opioid use, and duration of opioid drug use, these variables were transformed from ordinal to categorical variables. Bivariate analyses were performed for the variable of interest and for all other potentially confounding vari-
176
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
ables. Stepwise multiple logistic regression modelling was used to hold possible confounders constant when estimating the key association. Confounders were added in groups: (1) basic socio-demographics; (2) drug-use history; (3) patient background; (4) current living conditions; (5) circumstances of treatment entry; and (6) future treatment plans. Individual factors were kept in the model if significant. Results are summarized by reporting the models at each step relying on those variables meeting the criterion of inclusion: not entrapping 1 in the 95% confidence interval. Tests for interactions were used to check on the uniformity assumption under which multiple logistic regression estimates are derived.
3. Results From 1991 until 1997 a total of 1685 opioid addicts were admitted to the detoxification unit. For the 1070 patients included in the analyses, the median age was 28 years, 65% were male, and the median duration of opioid use was 10 years. 133 addicts reported daily consumption of methadone (METHADONE GROUP), 661 the daily consumption of codeine (CODEINE GROUP) and 276 the daily injection of heroin (HEROIN GROUP). A total of 43.6% of the patients completed the detoxification treatment program. The bivariate analysis indicated that methadone-dependent and codeine-dependent patients finished the detoxification treatment significantly more often than the heroin-dependent patients (P B 0.01). This is consistent with our main hypothesis. Table 1 presents the bivariate analysis on the key variable and the possible confounding variables predicting the completion of detoxification treatment. As shown, completion of detoxification treatment was predicted by older age, better education, the absence of an addicted partner, the absence of dependency in the family, regular contact with a counsellor, being currently on probation and the reporting of future treatment plans (Table 1). Our key finding persisted even when socio-demographics were included in the model. Sex was not a significant predictor of completing detoxification treatment. Age and education were both significantly associated. Education was divided into three groups: less than, equal to, or more than the mandatory 9 years of schooling. Race was not assessed since almost all the patients were Caucasian (Tables showing these data are available as Supplementary Materials on the journal website, www.elsevier.nl/locate/drugalcdep). Similarly, the drug use variables were included in the second step. None turned out to be significantly involved in predicting outcome. In the third step, patient background was added to the model. This included: history of drug
related emergencies, suicide attempts, imprisonment, and drug treatment. The only significant predictor was history of imprisonment. The next step (‘current life conditions’) consisted of employment, marital status, drug use in significant others, substance dependence in a family member, and counselling. When these variables were added to the model, ‘current counselling’ was the only significant predictor. Circumstances of treatment entry were added next and only one variable, ‘being currently on probation’, significantly added to the model (data in Supplementary Materials). The final model was achieved after the inclusion of ‘reported plans for future abstinence treatment’. In this instance none of the previously entered significant predictors became insignificant and our variable of interest, predetoxification treatment, also remained significant, even if only marginally (Table 2). We also tested for interactions to check on the uniformity assumption of the multiple logistic regression estimates. Not one single significant interaction was found.
4. Discussion The purpose of this paper was to compare detoxification treatment outcomes of methadone- and codeinesubstituted patients with those for patients using intravenous heroin. The hypothesis was that methadone-substituted and codeine-substituted patients would complete the treatment more often than the users of illicit heroin. Furthermore, since codeine-substituted patients are exposed to less treatment they should represent an intermediate group between methadone-substituted patients and heroin-using patients. Since both methadone- and codeine-dependent patients completed the treatment more often than heroin-dependent patients, and since the codeine group was intermediate between the methadone and heroin group, the main hypothesis was confirmed. Similar results were reported in a pilot study with a small sample of 71 patients (Backmund et al., 1998b). A possible explanation is that methadone maintenance (Gro¨ nbladh et al., 1990; Dole et al., 1991Marsch, 1998) and codeine (Krausz et al., 1998) programs stabilise the addicts. Hence, the decision and motivation to begin and finish a detoxification treatment can be sustained by the patient. Codeine also does this but to a lesser degree. Independent of this motivation, the following variables were also associated with increased completion of detoxification treatmen: (1) increase in age predicted an increase in treatment completion; however, this was not a continuous increase, and a ceiling effect occurred at approximately age 28, with no subsequent improvement in completion rate; (2) better education marginally predicted completion; (3) history of imprisonment was a
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
177
Table 1 Cross-tabulation Variable
Completers n
%
O.R.
95% C.I.
P
Main hypothesis Drug groups Methadone Codeine Heroin
133 661 276
50.4 45.5 35.9
1.81 1.49 1
1.19–2.76 1.12–2.00
0.0064 0.0054 0.0065
Basic socio-demographics Sex Female Male
380 690
40.8 45.2
0.83 1
0.65–1.08
0.1624
Age groups B23 years 23–28 years 29–34 years \34 years
216 335 326 190
38.9 38.8 50.3 45.3
0.77 0.77 1.22 1
0.52–1.14 0.53–1.10 0.86–1.75
0.0107 0.1943 0.1489 0.2691
Education Max 9 years, without qualification 9 years, qualified \9 years
93 648 329
33.3 42.3 49.2
0.52 0.76 1
0.32–0.84 0.58–0.99
0.0135 0.0071 0.0389
Initially necessary methadone dose 0–40 mg 41–80 mg 81–200 mg
381 434 255
44.1 40.1 48.0
0.85 0.72 1
0.62–1.17 0.53–0.99
0.1211 0.33 0.04
Daily hash consumption No Yes
878 192
43.2 45.8
0.90 1
0.66–1.23
0.4998
1036 34
44.0 32.4
1.64 1
0.79–3.41
0.1815
Daily benzodiazepine consumption No Yes
596 474
41.8 46.0
0.84 1
0.66–1.08
0.1676
Daily barbiturate consumption No Yes
995 75
43.3 48.0
0.83 1
0.52–1.32
0.4309
Daily alcohol consumption No Yes
772 298
42.5 46.6
0.85 1
0.65–1.11
0.2192
Duration of drug addiction 1 year 2–5 years 6–10 years 11–15 years \15 years
78 394 251 160 187
52.6 40.6 41.8 46.3 46.5
1.27 0.79 0.83 0.99 1
0.75–2.16 0.55–1.12 0.56–1.21 0.65–1.51
0.2509 0.3703 0.1782 0.3279 0.9593
Age onset 1–9.99 years 10–13.99 years 14–17.99 years 18–21.99 years \22 years
18 58 301 390 303
38.9 53.4 41.5 43.1 44.9
0.78 1.41 0.87 0.93 1
0.30–2.07 0.80–2.48 0.63–1.20 0.69–1.26
0.5265 0.6198 0.2321 0.4052 0.6343
Patient background History of emergencies No Yes
697 373
42.0 43.6
0.83 1
0.64–1.07
0.1474
Drug use history
Daily cocaine consumption No Yes
178
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
Table 1 (Continued) Variable
Completers n
%
O.R.
95% C.I.
P
History of suicide attempts No Yes
854 216
42.6 47.7
0.82 1
0.60–1.10
0.1806
History of therapy No therapy Therapy not completed Therapy completed Therapy n.c. and c.
722 123 135 90
43.2 37.4 45.2 53.3
0.67 0.52 0.72 1
0.43–1.03 0.30–0.91 0.42–1.23
0.1377 0.0697 0.0213 0.2315
History of imprisonment Never 1 \1
618 277 175
45.6 43.7 36.6
1.46 1.35 1
1.03–2.06 0.91–1.98
0.1042 0.0334 0.1347
Current life conditions Employment No Yes
772 298
44.3 41.9
1.10 1
0.84–1.44
0.4865
Marital status Single Married, living with a partner Divorced, widowed
701 324 45
43.9 41.4 55.6
0.63 0.56 1
0.34–1.15 0.30–1.06
0.1960 0.1315 0.0742
Drug use of significant others No Yes
356 714
49.4 40.8
1.42 1
1.10–1.84
0.0071
Family member with substance dependency No Yes
400 670
48.8 40.6
1.39 1
1.08–1.79
0.0094
Counselling No Yes
569 501
38.0 50.1
0.61 1
0.48–0.78
0.0001
Circumstances of treatment entry Contact Counsellor, practitioner Directly scene
565 505
42.1 45.3
0.88 1
0.69–1.12
0.2887
Waiting time until admission B3 weeks 3–6 weeks \6 weeks
231 156 683
42.4 39.7 44.9
0.90 0.81 1
0.67–1.22 0.57–1.15
0.4550 0.5044 0.2378
Currently on probation No Yes
742 328
41.4 48.8
0.74 1
0.58–0.96
0.0245
Treatment moti6ation Reported future plan Nothing Work Outpatient abstinence therapy Inpatient abstinence therapy
339 267 149 315
33.9 41.9 46.3 54.3
0.43 0.61 0.73 1
0.32–0.59 0.44–0.85 0.49–1.07
0.0000 0.0000 0.0031 0.1089
very strong predictor; no imprisonment indicated higher rates of completion. This was an interesting finding insofar as the factor was not predictive on its own. To get more information about this phenomenon we used logistic regression and added a number of
possible confounders. By adding the factor, ‘age groups’, the factor ‘history of imprisonment’ became significant. Thus age turned out to be a variable that ‘masked’ the effect of imprisonment as a risk factor; (4) patients currently in some form of counselling had a
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
higher rate of completion. This may indicate that preparation in the form of counselling may induce better outcomes; (5) patients currently on probation also had a higher rate of completion, possibly indicating that avoiding imprisonment might work as a motivator to complete treatment; (6) ‘reported future treatment plans’ consisted of patients reporting plans to start any form of abstinence treatment program after completion of the detoxification program. This included short (3 months) and long term (6–12 months) inpatient or outpatient abstinence treatment. ‘Reported future treatment plans’ can be considered as a measurement of treatment motivation. Thus, we were not surprised to find that this factor had the highest significance level. We were some-
Table 2 Final model Variable
O.R.
95% C.I.
P
Main hypothesis Drug groups Methadone Codeine Heroin
1.42 1.45 1
0.91–2.22 1.07–1.96
0.0495 0.1212 0.0158
Basic socio-demographics Age groups B23 years 23–28 years 29–34 years \34 years
0.55 0.57 1.03 1
0.35–0.85 0.38–0.84 0.71–1.51
0.0002 0.0075 0.0043 0.8609
Education B9 years, without Qualification 9 years, qualification \9 years
0.54 0.81 1
0.33–0.89 0.61–1.07
0.0422 0.0153 0.1343
Patient background History of imprisonment Never 1 \1
2.08 1.68 1
1.41–3.06 1.11–2.53
0.0011 0.0002 0.0132
Current life conditions Counselling No Yes
0.71 1
0.54–0.93
0.0124
Circumstances of treatment entry Currently on probation No Yes
0.69 1
0.51–0.92
0.0107
Treatment moti6ation Reported future plan Nothing Work Outpatient abstinence therapy Inpatient abstinence therapy
0.48 0.65 0.74 1
0.34–0.68 0.45–0.94 0.49–1.12
0.0006 0.0000 0.0031 0.1089
179
what surprised though, that when motivation was entered into the model it hardly changed the significance level of the main predictor for completing treatment (methadone, codeine and heroin group). This indicates that the differences in completion for these three groups is not due to differences in motivation as measured by reported future treatment plans. A number of factors did not predict treatment completion including sex; although females completed detoxification less often than males, the finding was not significant. Initial methadone dose also did not predict completion of detoxification. Initial methadone dose can be regarded as a measurement of the severity of dependence and thus this result can be taken to indicate that severity of dependence is not associated with detoxification outcome. Daily consumption of other drugs, and age of onset, was not predictive. Duration of use was U-shaped with the highest rates of treatment completion amongst those with recent onset (less than 1 year) and those with longer use (more than 10 years). However, this result was not significant. History of emergency treatment and history of suicide attempts were not significant predictors. Although emergency room visits and history of suicide attempts have been reported to increase the likelihood of entering detoxification (Schu¨ tz et al., 1994), they do not seem to be associated with completing a detoxification program. That history of treatment was not a significant predictor was somewhat surprising, since the clinical impression had been that patients with experience of treatment did better. This lack of significance might have been due partially to the fact that the groups with therapy experience were rather small (90, 123, 135 individuals per group). Drug use of significant other and substance abuse in family members were significant as individual factors, but did not reach significance when included in the model. This means that opiate group, age, education, history of imprisonment, and current counselling, accounted for the predictive power of living with a drug-using significant other and substance user in the family. Waiting time was another factor that did not predict completion of treatment. This study is a retrospective investigation using a ‘natural experiment’. The restrictions associated with this approach therefore, apply. For example, patients were not randomly assigned to the three different groups, and factors associated with group assignment might have influenced the outcome. A number of factors not measured in the present study might also be responsible for the differences between these three opiate groups. One such factor could be the differences in withdrawal symptomatology between these groups. Methadone has been reported to be one of the best drugs to reduce opioid withdrawal syndrome, (Razani et al., 1975; Bickel et al., 1988; San et al., 1989, 1990; Wilson and DiGeroge, 1993; Janiri et al., 1994; San et al., 1994; Bearn et al., 1996). Initial methadone dose was lowest among heroin depen-
180
M. Backmund et al. / Drug and Alcohol Dependence 64 (2001) 173–180
dent patients. In this study, heroin dependent patients still completed the treatment less often. Methadone withdrawal symptoms are reported to be more severe than heroin withdrawal symptoms (Gossop and Strang, 1991). We therefore, cannot rule out that differences in pharmacological effects may have played a role. The results support the hypothesis that intensity of previous treatment improves outcome of abstinence induction. However, the current analysis does not rule out that factors such as differences in the pharmacological effects of the different opioids might have contributed to these findings. Drug addicts treated with methadone have a better outcome in detoxification treatment than codeine-substituted patients and patients not in any substitution treatment. The better outcome among methadone treated patient is consistent with the hypothesis that the more psychosocial/psychotherapeutic support these patients receive before starting detoxification, (compared to codeine-substituted patients and heroin-injecting patients), the better the treatment outcome.
Acknowledgements This study was supported by ‘Modellprogramm Kompakttherapie im Verbund der Drogenhilfe 1990– 1995’ from the Department of Health from the Federal Republic of Germany. The authors would like to thank the staff from the detoxification unit ‘Villa’ for their help with data collection and entry.
References Armenian, S.H., Chutuape, M.A., Stitzer, M.L., 1999. Predictors of discharges against medical advice from a short-term hospital detoxification unit. Drug Alcohol Depend. 56, 1 –8. Backmund, M., Eichenlaub, D., Soyka, M., 1998a. Evaluation of a qualified detoxification treatment of intravenous drug addicts in a general hospital — a Federal German model project. Gesundheitswesen 60, 552 – 557. Backmund, M., Meyer, K., Habsch, J., Soyka, M., Eichenlaub, D., 1998b. Qualifizierte Entzugsbehandlung Drogenabha¨ ngiger — Mo¨ gliche Pra¨ diktoren des Therapieverlaufs. Mu¨ nch. med. Wochenschr. 44, 590 –593. Bearn, J., Gossop, M., Strang, J., 1996. Randomised double-blind comparison of lofexidine and methadone in the in-patient treatment of opiate withdrawal. Drug Alcohol Depend. 43, 87 – 91.
Bickel, W.K., Stitzer, M.L., Bigelow, G.E., Liebson, I.A., Jasinski, D.R., Johnson, R.E., 1988. A clinical trial of buprenorphine: comparison with methadone in the detoxification of heroin addicts. Clin. Pharmacol. Ther. 43, 72 – 78. De los Cobos, J.P., Trujols, J., Ribalta, E., Casas, M., 1997. Cocaine use immediately prior to entry in an inpatient heroin detoxification unit as a predictor of discharges against medical advice. Am. J. Drug Alcohol Abuse 23, 267 – 279. Dole, V.P., Newman, R.G., Jospeh, H., 1991. Heroin addiction: a public health problem. Med. N. Amer. 15, 1943 – 1945. Endicott, P., Watson, B., 1994. Interventions to improve the AMA discharge rate for opiate-addicted patients. J. Psychsoc. Nurs. Ment. Health Serv. 32, 36 – 40. Franken, I.H., Hendriks, V.M., 1999. Predicting outcome of inpatient detoxification of substance abusers. Psychiatr. Serv. 50, 813 –817. Go¨ rgen, W., Mo¨ ller, I., Oliva, H., 1997. Einscha¨ tzungen und Ergebnisse einer qualifizierten Entzugsbehandlung fu¨ r Drogenabha¨ ngige aus Patientensicht. Sucht 42, 82 – 91. Gossop, M., Johns, A., Green, L., 1986. Opiate withdrawal: inpatient versus outpatient programs and preferred versus random assignment to treatment. Br. Med. J. 293, 103 – 104. Gossop, M., Strang, I., 1991. A comparison of the withdrawal responses of heroin and methadone addicts during detoxification. Br. J. Psychiatry 158, 697 – 699. Gro¨ nbladh, L., O8 hlund, L.S., Gunne, L.M., 1990. Mortality in heroin addiction impact of methadone treatment. Acta Psych. Scand. 82, 223 – 227. Janiri, L., Mannellli, P., Persico, A.M., Serretti, A., Tempesta, E., 1994. Opiate detoxification of methadone maintenance patients using lefetamine, clonidine and buprenorphine. Drug Alcohol Depend. 36, 139 – 145. Jeremiah, J., O’Sullivan, P., Stein, M.D., 1995. Who leaves against medical advice. J. Gen. Intern. Med. 10, 403 – 405. Krausz, M., Verthein, U., Degkwitz, P., Haasen, C., Raschke, P., 1998. Maintenance treatment of opiate addicts in Germany with medications containing codeine — results of a follow-up study. Addiction 93, 1161 – 1167. Razani, J., Chisholm, D., Glasser, M., Kappeler, T., 1975. Self-regulated methadone detoxification of heroin addicts. An improved technique in an inpatient setting. Arch. Gen. Psychiatry 32, 909 – 911. San, L., Camı´, J., Peri, J.M., Mata, R., Porta, M., 1989. Success and failure at inpatient heroin detoxification. Br. J. Addict. 84, 81 –87. San, L., Camı´, J., Peri, J.M., Mata, R., Porta, M., 1990. Efficacy of clonidine, guanfacine and methadone in the rapid detoxification of heroin addicts: a controlled clinical trail. Br. J. Addict. 85, 141 – 147. San, L., Fernandez, T., Camı´, J., Gossop, M., 1994. Efficacy of methadone versus methadone and guanfacine in the detoxification of heroin-addicted patients. J. Subst. Abuse Treat. 11, 463 –469. Schu¨ tz, C.G., Rapiti, E., Vlahov, D., Anthony, J.C., 1994. Suspected determinants of enrolment into detoxification and methadone maintenance treatment among injecting drug users. Drug Alcohol Depend. 36, 129 – 138. Wilson, R.S., DiGeroge, W.S., 1993. Methadone combined with clonidine versus clonidine alone in opiate detoxification. J. Subst. Abuse Treat. 10, 529 – 535.