A total of 370 male inpatients with a history of alcohol dependence participated in the study. The participants were recruited from nine nationwide hospitals for addiction treatment in Japan (n = 91, National Hospital Organization Kurihama Alcoholism Center, Kanagawa; n = 55, Wakamiya Hospital, Yamagata; n = 50, Komakino Hospital, Tokyo; n = 42, Mie Prefectural Mental Medical Center, Mie; n = 26, Asahiyama Hospital, Hokkaido; n = 17, Ishikawa Prefectural Takamatsu Hospital, Ishikawa; n = 14, National Hospital Organization Hizen Psychiatric Center, Saga; n = 13, Akagi-Kohgen Hospital, Gunma; n = 12, Tohokukai Mental Hospital, Miyagi). Inpatients were provided an average of 80-day treatment programs (group meeting, alcohol education, family treatment programs, psychotherapy, and so on) after detoxification. After recovery from serious physical and mental instability (almost 1 month after hospitalization), informed consent was obtained from the subjects, excluding the patients who had serious cognitive impairment and psychiatric problems, and the ASI-J was administered by psychiatrists who were expert in alcoholism, carefully read the ASI manual (http://www.tresearch.org/resources/manuals/ASIQbyQGuide.pdf), and learned the interview methods by themselves. The average time required for administration was 30 min. Inpatient subjects were requested to answer the questions during the 30 days prior to the start of inpatient treatment. Forty-nine participants were excluded from the final analysis because of lack of reliability (i.e., patient misrepresentation or inability to understand). Methamphetamine abuse data that were compared in the present study were mainly obtained from our previous study that standardized the ASI-J [12]. A total of 116 ASI-J samples with a history of methamphetamine and/or amphetamine abuse were collected from three hospitals and two recovery centers.

We used two types of indexes, including the ASI-J and a series of indexes called the Treatment and Recovery of Alcohol-dependence that determined patient state during and post-treatment. The ASI provides two types of overall scores for each problem area: severity rating (SR) and composite score (CS) [11]. The SR indicates the severity of the problem on the basis of interviewer assessment during in-person interviews using subjective and objective data related to current and lifetime problems. The SR ranges from 0 to 9 points, with a higher score indicating greater problem severity. The CS in each problem area is not a rating but rather a mathematically calculated score based on patient responses to sets of items asking for patient behaviors during the 30 days prior to interview. The CS is calculated using a weighted formula designed to provide an equal contribution from each item and varies from 0 to 1, with a higher score indicating greater problem severity.

The Treatment and Recovery of Alcohol-dependence indexes were administered by the attending physician or deputy at hospital discharge and 3 months and 1 year post-discharge. Each index included 8–18 items that assessed attitudes toward hospital treatment (i.e., Lack of cooperation, Lack of leadership, Rule breaking, Relapse, and Substance abuse), biochemical markers (i.e., glutamic oxaloacetic transaminase [GOT], glutamic pyruvic transaminase [GPT], γ-glutamyl transpeptidase [γ-GTP]), the frequency of relapse after discharge from the hospital, and so on. With regard to scoring the undesirable attitudes toward treatment, the attending physician subjectively rated the five aforementioned items on a 3-point scale (1 = good, 2 = normal, 3 = bad).

First, independence of problem areas was examined by calculating the Pearson’s product-moment correlation coefficient among individual CSs. Second, possibilities of predictive assessments of relapse using the ASI-J were analyzed by comparing CSs between abstinent and relapsed alcohol-dependent individuals via one-way analysis of variance (ANOVA). Associations between CSs and undesirable attitudes toward treatment were examined by the Pearson product-moment correlation coefficient. Third, significant differences in profiles between alcohol-dependent individuals and drug abusers were examined using the χ² test. Comparisons of CS and SR between alcohol-dependent individuals and drug abusers were conducted using one-way ANOVA. With regard to comparisons of CSs in alcohol-dependent individuals among studies, Welch’s t-test was conducted using DA Stat (Shinko Koeki, Tokyo, Japan). Fourth, internal consistency of each ASI area was examined using Cronbach’s alpha value and the Pearson product-moment correlation coefficient between the CS and SR of each area. All analyses, with the exception of comparisons of CSs in alcohol-dependent individuals among studies, were performed with SPSS v. 12.0 for Windows (SPSS Japan, Tokyo, Japan).

The ASI was developed to independently evaluate each of seven problem areas. To assess the independence of the seven areas, the correlation between each CS was analyzed (Table 1). Significant but low correlations were noted in eight relations: between the Medical CS, Alcohol use CS, and Employment CS; between the Employment CS and Legal CS; between the Alcohol use CS, Family/Social CS, and Psychiatric CS; between the Drug use CS, Family/Social CS, and Psychiatric CS; between the Family/Social CS and Psychiatric CS. The Psychiatric CS was significantly related with CSs in three areas. No significant correlations were observed in the other 13 relations. These results indicate substantial independence of the problem areas, with the exception of the Psychiatric CS.

To examine the internal consistency of the CSs in the seven areas, Cronbach’s alpha coefficients were calculated (Table 5). The coefficients ranged from 0.84 in the Psychiatric area to 0.53 in the Family/Social area. Overall, internal consistency of the ASI-J area was acceptable, with the exception of the Family/Social area. Regarding correlation coefficients between the component items of the Family/Social CS, the satisfaction rate for married life and days with trouble within family minimally correlated with other items (−0.03 to 0.22).

Because both the CS and SR assess the level of subject function in a given problem area, examining whether they offer correlated measurements of subject function in each of the problem areas is important [11]. The methods of deriving the CS and SR are based on both similar and different questions within each area. Significant positive correlations between the CS and SR were observed in all problem areas. CS-SR relationships were strong in the Medical area (0.69), Psychiatric area (0.66), and Drug use area (0.62), while weak in the Alcohol use area (0.22) and Employment area (0.29).

The CS and SR in the Alcohol use area were weakly but significantly correlated with some of the biological markers (Table 6). Correlation coefficients between the Alcohol use CS and GOT at hospitalization, between the Alcohol use CS and GPT at hospitalization, and between the Alcohol use CS and γ-GTP at hospitalization were 0.23 (*p < 0.01), 0.17 (*p < 0.01), and 0.25 (*p < 0.01), respectively. Correlation coefficients between the Alcohol use SR and GOT at hospitalization and between the Alcohol use SR and GPT at hospital discharge were 0.13 (*p < 0.05) and 0.17 (*p < 0.01), respectively. These results partially support the concurrent validity of the Alcohol use CS and SR.

The measure of independence of the problem areas illustrates the ability of the ASI-J to classify and quantify wide-ranging problems in addition to alcohol use problems to enable customized treatment. As shown in Table 1, the CSs in ASI areas show relatively few relationships with each other. These data substantiated the relative independence of the problem areas and the ability of the questions to assess these various problems demonstrated in other countries [1,6,7,13–15].

Although each area of the ASI-J is basically independent, the Psychiatric CS in the ASI-J was slightly but significantly related with the Family/Social CS and Drug use CS and Alcohol use CS. This trend was similar to overseas data [1,14,16]. Kosten et al. [14] indicated that most addicts with psychological problems had poor social adjustment and problems in the other ASI areas, suggesting that one or more subgroups of multiproblem addicts who have a variety of psychiatric disorders might be identified by the ASI. The ASI-J would be useful for identifying patients who need early concurrent treatments by determining relationships between their Psychiatric CS and other CSs.

We underscore the function of Family History for customized treatment. Cotton [17] (a review of 32 studies researching frequency of paternal alcoholism) reported a 27% incidence of alcoholism in the fathers of 4329 alcohol-dependent individuals. These data were obviously higher than the rates of alcoholism expected in the general population (2–3%) and in males over 40 years old (6–10%) [18]. Our results showing that problem-drinking was present in 36% of 302 subjects’ fathers and present in 19% of 88 drug abusers’ fathers (Figure 3) were nearly consistent with Cotton’s results. These rates were higher than the general population and higher than the rates reported in Cotton’s study. Furthermore, a study by Iwakura [19] indicated that subjects who grew up in dysfunctional families (e.g., with parents who had alcohol-related problems) struggled with more complicated treatment issues. Therefore, clinicians should thoroughly grasp the patients’ psychosocial histories and provide proper treatment and support. Alcohol dependence in parents has often been assessed by the Children of Alcohol-dependent individuals Screening Test (CAST), which consists of 30 questions. The present results suggest that simple questions in the ASI Family History section, rather than the CAST, are useful for identifying adult children of dysfunctional families with alcoholic problems and assessing the needs for additional treatment.

Drug use, Psychiatric, or Family relationship problems in subjects were related with total undesirable attitude toward treatment (Table 2). Additionally, subjects who had employment problems occasionally lacked cooperation with other inpatients and broke the rules of treatment. The family, criminal, employment, and psychological problems seen among alcohol and drug abuse patients have reportedly been important predictors of response to treatment, with psychiatrically ill and especially antisocial substance abusers particularly likely to show poor treatment response and early relapse, regardless of the treatment modality or setting [20,21]. Our results are consistent with these reports. The moderate relationship between the Drug use CS at the start of hospitalization and relapse during treatment (temporarily leaving the hospital and staying out overnight) suggests difficulty in maintaining abstinence from drug use when the patient is temporarily away from the hospital. For successful hospitalization, improving psychiatric problems as early as possible after hospitalization, controlling family relationships, and providing guidance for staying away from the problem substance during temporary retreats away from the hospital are very important. As McLellan et al. indicated [3], additional services for addiction-related problems at an early stage will improve the outcomes of standard addiction treatments.

Figure 1 shows that the Family/Social CSs at hospitalization were significantly different between subjects who relapsed within 3 months after leaving the hospital and subjects who continued to abstain from drinking. Subjects who had serious family and social relationship problems tended to relapse, suggesting that control of interpersonal relationships during treatment is a key predictor of response to treatment. In contrast, Alcohol use CSs at hospitalization were not effective predictors of relapse and compliance to treatment. These results, together with the results in Table 2, support the view that useful predictors of relapse and compliance to treatment are not the serious drinking itself, but rather the various CSs of the ASI-J at hospitalization, although more data are necessary to confirm these relatively weak relationships.

We tested the usefulness of the ASI-J as a comparison tool to reveal features of the analyzed groups. Argeriou et al. [13] proved sensitivity of the ASI to detect differences in ASI scores across various subgroups (i.e., homeless, near homeless, white, black, Hispanic, men, and women). Our study demonstrated the usefulness of the ASI-J in finding differences in alcohol-dependent individuals in Japan from both drug abusers in Japan and alcohol-dependent individuals in other countries.

Although our data illustrated the ability of the ASI to assess multidimensionally addiction-related problems, the actual numeric value of the CS has no intrinsic meaning [23]. Moreover, CSs are not similarly scaled and therefore cannot be compared between problem areas [13]. In the present study, cutoff points of CSs could not be estimated because the ranges of CSs, especially the Drug use and Legal CSs, were not wide. Additionally, insufficient data exist to definitively know the normal range. The standard of severity in each area is not precisely known, although cutoff points of the Alcohol use CS and Drug use CS (0.17 and 0.16, respectively) have been provided [24]. Therefore, we need to refer to the SR to determine the priority of treatment in seven problem areas, although the SR is a subjective rating and is viable as a clinical summary only for initial treatment planning and referral [5]. Further accumulation of normal data is also needed to define the numerical meaning of the CS.

We compared domestic and overseas CS data for alcohol-dependent individuals. Although these non-comprehensive data showed roughly the features of each country’s data, we could not understand the precise reasons for differences only though CS comparisons. Further study of nationality and the features of treatment facilities would be necessary. International joint investigations of alcohol-dependent individuals in similarly controlled settings using the ASI as a common interview may reveal the features of nationality and facilities in detail.

Regarding the ASI as a prediction tool, although this study suggests its usefulness, further research will be needed, considering the low correlations between CSs and relapse and compliance to treatment in addition to the low follow-up rate (65%) for the data at 3 months post-discharge in the present study.

Finally, the reliability and validity of the ASI-J in alcohol-dependent individuals were not sufficiently examined in the present study. The ASI-J demonstrated good internal consistency, with the exception of the Family/Social area, and criterion validity measured correlations between the Alcohol use CS/SR and biological markers. These results suggest that the Alcohol use CS may be an index for determining severity and reflecting recent alcohol use. Although our team already substantiated inter-rater reliability through training for administration of the ASI and concurrent validity of the ASI-J in drug abusers, we did not examine these variables in alcohol-dependent individuals.