Profiles associated with treatment retention in Japanese patients with methamphetamine use disorder: Preliminary survey



Aims:  To identify profiles associated with treatment retention in Japanese patients with methamphetamine use disorder.

Methods:  The study used a retrospective design based on clinical records. The subjects were 101 patients at the Kanagawa Psychiatric Center, Serigaya Hospital, who were diagnosed as having methamphetamine use disorder. They were divided in two groups, namely those who remained in treatment 3 months after the initial assessment, and those who did not. The primary analysis compared patient profiles between the two groups to detect discriminating variables, which were then submitted for secondary analysis using logistic regression to determine the most relevant predictor of retention.

Results:  Primary analysis indicated that older age, having psychotic symptoms, receiving public assistance, and history of incarceration were associated with treatment retention after 3 months. Secondary analysis showed that positive history of incarceration was the most significant predictor of the outcome.

Conclusions:  History of incarceration had the most significant treatment-retention effect on Japanese patients with methamphetamine use disorder. The development and introduction of integrated programs that link methamphetamine-dependent offenders to drug treatment is recommended in outpatient treatment for Japanese patients with methamphetamine user disorder.

METHAMPHETAMINE ABUSE HAS been the most significant illicit drug problem in Japan since 1945.1,2 First synthesized by a Japanese pharmacologist in 1893, methamphetamine was used mainly by military personnel and factory workers in Japan to ‘enhance energy’ during World War II. After the war, methamphetamine use spread beyond the military, and became distributed throughout society. As a result, prisons and mental hospitals soon became flooded with methamphetamine abusers (the so-called First Epidemic). Since then, Japan has experienced two methamphetamine epidemics, namely the Second Epidemic in the 1970s and 1980s, and the Third Epidemic from the late 1990s to the present. The number of arrests under the Stimulant Control Law (in which the main target drug is methamphetamine) peaked in 1954, amounting to more than 50 000 per year, and still exceeded 13 000 in 2005.3 A nationwide mental hospital survey conducted in Japan in 2002 showed that 55% of all drug-related psychiatric cases were due to methamphetamine abuse.4 In spite of more than 60 years of societal struggle with the drug, psychiatric services for methamphetamine abusers remain almost unchanged: hospitalization for those with psychotic symptoms and treatment with antipsychotic medications.

Compared to the scale of the impact that the methamphetamine problem has had on Japanese society, the energy invested in developing treatment programs has been astonishingly deficient. What has been treated appears to be limited to methamphetamine psychosis. Clinicians and researchers in Japan have long tended to pay little attention to elaborating the treatment strategies targeting the dependence aspect of the problem. For example, there are only a few psychiatric hospitals providing inpatient programs for substance-dependent patients in Japan. Even in such hospitals, there is always an inevitable discontinuity of treatment intensity after discharge, because they have virtually no outpatient programs other than pharmacotherapy and encouraging patients to attend self-help groups. Although privately-funded residential facilities (e.g. the Drug Addiction Rehabilitation Center or DARC) have begun to emerge since the late 1980s, they are still inadequate in number and quality.

Meanwhile, in the USA, the abuse of ‘crystal meth’ and ‘ice’ first became an epidemic in California and Hawaii in the 1980s and 1990s.5 Despite these drugs’ relatively recent emergence, it took less than 20 years to witness a vast amount of clinical research on factors that improve the treatment outcome for methamphetamine abuse and dependence. As the clinical evidence accumulated, it was found that duration of treatment predicted subsequent outcome in methamphetamine6,7 and other substance use disorders.8–10 Those research outcomes were then utilized for the development of various treatment strategies, such as trials and reports on a manualized, comprehensive approach (e.g. the matrix model11) or contingency management12–14 that demonstrated efficacy in treatment retention of stimulant abusers.

Looking back to Japan, clinical evidence regarding treatment retention is, to our knowledge, close to nonexistent. As the framework for drug addiction treatment is still based on acute phase hospitalization, the so-called revolving door phenomenon15 is the usual consequence for Japanese substance abusers nationwide. What is urgently needed is an outpatient program that employs a chronic care perspective,16 so that it can be utilized as a post-discharge follow-up or as a preventative intervention for non-psychotic methamphetamine-dependent patients.

In preparation for development of such a program, basic clinical data on factors relevant to treatment retention are essential. In the current study we conducted a preliminary survey on patient profiles and initial treatment modalities (outpatient vs inpatient) that may be related to treatment retention in Japanese patients with methamphetamine use disorder.



The subjects were 101 Japanese patients diagnosed as having either methamphetamine dependence or methamphetamine-induced psychotic disorder according to DSM-IV-TR.17 They had been admitted for either outpatient or inpatient treatment at Kanagawa Psychiatric Center, Serigaya Hospital, between January and September 2006.

Of the 101 subjects, 23 (22.8%) were women, and 78 (77.2%) were men. The age at initial assessment ranged from 19 to 60 years, with a mean of 36.1 ± 9.3 years.

This survey was conducted as a part of a methamphetamine treatment research study at Serigaya Hospital, which had been approved by the Ethics Committee of the Kanagawa Psychiatric Center.


The present survey used a retrospective design, obtaining the data entirely from clinical charts. The dependent or outcome variable for the survey was treatment retention, determined by the descriptions in the charts. Subjects were classified into the ‘retention group’ if they remained in outpatient or inpatient treatment 3 months after the initial assessment. For termination of the treatment record within 3 months, subjects were classified into the ‘non-retention group’. We applied the threshold value of 3 months for the definition of treatment retention because so far the literature agrees that the length of drug addiction treatment should be at least 12 weeks18 or 90 days19 in order to be effective.

The independent variables selected as potential predictors of treatment retention were sociodemographic factors such as age at initial assessment and gender; and factors related to substance abuse such as age at first use, days between last use and initial assessment, and presence of psychotic symptoms. The number of days between last use and initial assessment was included because generally speaking, the shorter the abstinence period, the more irritability the patients experience, and thus the more inclined they are to relapse.

Living alone or with family members, receiving public assistance, and past history of arrest or incarceration were also included as variables that have potential influence on treatment retention.

We classified subjects into an ‘inpatient group’ if they were admitted within 1 month after the initial assessment. This was to take into account a possible treatment-retention effect of early-phase intensive treatment. Because there is virtually no intensive outpatient program in Japan, we selected early inpatient admission as a substitute. The rest of the subjects were classified as the ‘outpatient group’.

Treatment at Serigaya Hospital

Outpatient treatment for methamphetamine abusers at Serigaya Hospital consists of regular check-ups by the attending psychiatrist, providing pharmacotherapy and brief psychotherapy one–four times per month. There is also an outpatient group therapy session for methamphetamine and other drug abusers held once a week. The atmosphere of each session is similar to self-help groups, except that they do not overtly follow the 12-step system, and they are facilitated by a psychologist or a psychiatric social worker. Medications used in pharmacotherapy are usually antipsychotics for drug-induced hallucinations and delusions, and antidepressants and/or benzodiazepines for depression, anxiety, and other comorbid psychiatric symptoms. Anti-craving medications such as methadone, buprenorphine, naltrexone and acamprosate are not yet approved for use in Japan.

Admission criteria for inpatient treatment were as follows: voluntary request, presence of moderate comorbid psychotic symptoms, or physical illness complications indicating close medical observation. Those patients with severe psychosis who needed treatment in a locked unit, or those with severe physical complications requiring intensive medical treatment were excluded and referred to hospitals elsewhere.

Once admitted to the inpatient program, patients are encouraged to stay for 1 month in the open unit, although abrupt termination before graduation is not rare (unpublished statistical data from the hospital show that of all the drug-dependent patients admitted for inpatient treatment from 2002 to 2004, only 56% completed this successfully). Patients usually receive psychopharmacotherapy together with individual case management and multiple weekly psychoeducational group sessions. They are also recommended to attend self-help groups held in the hospital every week.

Statistical analysis

First, univariate analysis was performed to compare the variables obtained from clinical records between the retention and non-retention groups. Student's t-test was used for continuous variables such as age and days since last use, and Pearson's χ2 test was used for categorical variables such as gender, history of arrest or incarceration, and others. Primary analysis comparing the retention and non-retention groups was then conducted to narrow down the independent variables to those statistically associated with treatment retention.

Further, to identify the predictors directly associated with the retention group by excluding the confounding factors from the results of univariate analysis, backward stepwise logistic regression was used. The inclusion criterion of a liberal P threshold (P < 0.1) was adopted, because the independent variables available for this multivariate analysis were limited. The stepwise approach removed the independent variables for P > 0.1 according to the likelihood ratio. In these analyses, P < 0.05 was accepted as indicating statistical significance. All of the statistical analyses were conducted using SPSS version 15.0J for Windows (SPSS, Chicago, IL, USA).


Patient profiles

Table 1 shows the combined profile data of all subjects, with gender differences. Both women and men tended to start abusing methamphetamine at the mean age of 20–21 years. The number of days between last abuse and hospital admission, however, was significantly higher in male compared to female patients (P < 0.01). Age at first assessment was also significantly older in male patients (P < 0.01).

Table 1. Patient profiles
 Total (n = 101)Female (n = 23)Male (n = 78) F2 P
  • **

    P < 0.01.

Age at initial assessment (years)36.1 ± 9.331.1 ± 7.437.6 ± 9.31.746/0.003**
Age at first abuse (years)21.3 ± 5.820.4 ± 7.221.5 ± 5.30.971/0.409
Days between last abuse and initial assessment632.9 ± 1073.9265.6 ± 493.3741.2 ± 1173.15.591/0.006**
Presence of hallucinations, n (%)50 (49.5)13 (56.5)37 (47.4)/0.5870.444
Living with family, n (%)64 (63.4)17 (73.9)47 (60.3)/1.4720.232
On public assistance, n (%)26 (25.7)5 (21.7)21 (26.9)/0.250.617
History of arrest, n (%)72 (71.3)10 (43.5)62 (79.5)/11.2520.001**
History of incarceration, n (%)43 (42.6)3 (13.0)40 (51.3)/10.6230.001**
Inpatient admission within 1 month, n (%)20 (19.8)6 (26.1)14 (17.9)/0.7410.389

The proportion of patients having psychotic symptoms at initial assessment was approximately 50% for both genders. Approximately two-thirds of all subjects were living with family members, and one-fourth were on public assistance. Approximately one in four (women) or five (men) were admitted to the inpatient program within 1 month of initial assessment.

There were significant gender differences in crime histories. Those having been arrested or incarcerated were significantly more prevalent among male methamphetamine abusers (P < 0.01), reaching approximately 80% and 50%, respectively, compared with 43% and 13% in women.

Although the data are not presented in Table 1, history of incarceration was also significantly associated with receiving public assistance. Of those on assistance (n = 26), 61.5% (n = 16) had a history of incarceration (Pearson's χ2 test; P < 0.05).

Retention group versus non-retention group

A total of 39 patients (38.6%) had remained in treatment 3 months after the initial assessment. The results of the comparison between the retention group and non-retention group are presented in Table 2. In total, three independent variables reached statistical significance, namely age at initial assessment; receiving of public assistance; and history of incarceration.

Table 2. Comparison of retention vs non-retention groups
Total n = 101RetentionNon-retention F2 P
  • *

    P < 0.05,

  • **

    P < 0.01.

Age at initial assessment (years)39.0 ± 8.634.3 ± 9.20.062/0.012*
Age at first abuse (years)20.26 ± 4.221.92 ± 6.64.699/0.124
Days between last abuse and initial assessment805.7 ± 967.5524.2 ± 1129.80.058/0.201
Male, n (%)30 (76.9)48 (77.4)/0.0030.954
Presence of hallucinations, n (%)24 (61.5)26 (41.9)/3.6800.055
Living with family, n (%)23 (59.0)41 (66.1)/0.5280.467
On public assistance, n (%)15 (38.5)11 (17.7)/5.3770.02*
History of arrest, n (%)31 (79.5)41 (66.1)/2.0870.149
History of incarceration, n (%)23 (59.0)20 (32.3)/6.9890.008**
Inpatient admission within 1 month, n (%)8 (20.5)13 (21.0)/0.0030.956

Those in the retention group were significantly older than those in the non-retention group (mean 39 vs 34 years). Also, the retention group had a significantly larger proportion of patients receiving public assistance (38% vs 17%), and significantly more patients with a history of incarceration than in the non-retention group (59% vs 32%).

Each of the factors of age at first abuse, days between last abuse and initial assessment, being male (gender difference), presence of hallucinatory psychosis, living with family, past history of arrest, and inpatient admission within 1 month after initial assessment did not have a statistically significant effect on treatment retention.

Predictors of treatment retention

Stepwise logistic regression analysis was conducted to determine the predictors directly associated with treatment retention. As summarized in Table 3, the analysis of the subjects as a whole included four potential predictors (age at initial assessment; presence of hallucinations; receiving of public assistance; and history of incarceration). Of these four, only history of incarceration was significantly associated with treatment retention (P < 0.05, OR = 2.618). Although weak in significance (P = 0.074, OR = 2.38), those receiving public assistance were twice as likely to remain in treatment as those without.

Table 3. Logistic regression for potential predictors of retention
Total n = 101B P Exp(B)95%CI
  • *

    P < 0.05.

  • Likelihood ratio χ2 = 12.903, d.f. = 4, P = 0.012. CI, confidence interval.

Age at initial assessment0.0200.5161.0200.961–1.082
Presence of hallucinations0.6580.1341.9300.817–4.557
On public assistance0.8670.0742.3800.919–6.162
History of incarceration0.9620.027*2.6181.114–6.152


The present preliminary survey was intended to shed light on the clinical profiles of Japanese patients with methamphetamine use disorders in a hospital setting. Particular emphasis was placed on treatment retention and related gender differences, and this has seldom been studied in the literature on methamphetamine abuse in Japan.

Clinical profiles associated with treatment retention

Primary analysis indicated that three independent variables were significantly associated with treatment retention: age at initial assessment; receiving of public assistance; and history of incarceration.

Those patients who stayed in treatment for 3 months had a mean age of 39 years at initial admission, which was approximately 5 years older than the dropouts. The existing literature also reported a similar age effect on treatment retention among specifically methamphetamine abusers,20 as well as substance abusers in general.21 The difference in age at admission may reflect a relative impulsiveness among younger patients leading to premature cessation of treatment.

There were significantly more recipients of public assistance in the retention group than in the non-retention group, suggesting an impact of welfare support on treatment retention among Japanese methamphetamine abusers. Although the effectiveness of various public welfare interventions on substance-abusing low-income mothers has been reported,22,23 the evidence including male recipients is, to our knowledge, restricted to a study on homeless cases.24

Contrary to our expectations, earlier inpatient admission (within 1 month after the initial assessment) was not associated with longer retention. This result may reflect the patients' lack of motivation to change, which could have been more influential than the protective nature of the inpatient unit.

History of incarceration: Predictor of treatment retention

The results of logistic regression suggested that history of incarceration was the only predictor directly associated with treatment retention for patients with methamphetamine use disorder at Serigaya Hospital. Although receiving public assistance was found to be a statistically significant retaining factor in univariate analysis, it did not reach significance in multivariate analysis, indicating an indirect effect on treatment retention.

In terms of treatment retention, the analysis also indicated a statistically distinct difference between history of incarceration and arrest. One of the reasons why history of incarceration, not arrest, turned out to be the predictor of treatment retention may be related to the expected longer term that one must serve in the case of re-incarceration. The significant association between history of incarceration and receiving of public assistance may have played another role. Those with a history of incarceration were more likely to receive public assistance, which in turn might have led to receiving more supervision for treatment retention.

Last, it is important to emphasize that although history of incarceration had a significant impact on treatment retention in the present study, incarceration per se is not an equivalent of treatment programs, nor can it replace them. As national statistics show, the recidivism rate for Japanese drug offenders was 56.5% in 2006, and the re-incarceration rate within 1 year was also as high as 33.5% in 2005.25 These figures eloquently disprove any simplistic misconception that methamphetamine abusers are better off treated judicially than medically.

Because the target of the present survey was a clinical sample, the outcomes should be interpreted as evidence for the effectiveness of providing addiction treatment and public assistance to methamphetamine-abusing ex-prisoners. In the USA the successful introduction of drug court programs for methamphetamine-dependent subjects has been reported.26 The present results also suggest that there is a need for the development of a similar treatment program that integrates the medical, judicial and public welfare systems in Japan.


The relative paucity of subjects (n = 101) is to be mentioned first as one of the limitations of this survey. The retrospective design of the data collection based on chart records also restricted the amount and possibly the accuracy of the acquired information. The initial diagnoses were not determined by semi-structured interviews. The severity of methamphetamine use disorder and the extent of the motivation for treatment in each patient were not assessed quantitatively. Also, because all of the subjects had been admitted to a single institution (Serigaya Hospital), there is an inevitable sampling bias inherent in the data. Nevertheless, because the clinical data on Japanese methamphetamine abusers are currently far from complete, this survey may hopefully be the first step towards developing future research strategies on drug addiction treatment in Japan.


The present study investigated profiles of Japanese patients with methamphetamine use disorder in order to determine predictors of treatment retention. Univariate analysis found that those who remained in treatment after 3 months were significantly older, more likely to be receiving public assistance, and more likely to have a history of incarceration. Multivariate analysis indicated that history of incarceration was the sole predictor of treatment retention. These results suggest the clinical importance of providing opportunities for addiction treatment and public support for methamphetamine-abusing ex-prisoners in Japan.