The Addiction Severity Index (ASI) is a measurement tool to assess the problems experienced by people who use substance through six domains, namely health/medical status, employment, drug and alcohol use, legal, family/social relationships, and psychiatric status. This study was conducted to obtain data validation, pertaining validity and reliability of the ASI-5-BM.The study population was clients from the AADK drug treatment and rehabilitation centers. A total of 209 participants were selected from five addiction treatment centers (C & C) and a district rehabilitation center (CCSC).Three instruments were used for validation purposes, namely ASI-5-BM, 12-Item Short Form Health Survey (SF-12) and Beck Depression Inventory (BDI). The study found high and significant coefficient scores for ‘convergent’ validity. Clients who use drugs are closely correlated with health/medical problems (r=0.145; p<0.05); family problems (r =0.282, p <0.001); and psychiatric morbidities (r=0.207, p<0.01). The use of alcohol is also significantly associated with health/medical problems (r=0.179, p<0.01) and legal problems (r=0.215, p<0.01). The “concurrent” validity was measured with the SF-12, which examines the physical and mental health status of an individual, and the BDI, to confirm the psychiatric domain in the ASI-5-BM. SF-12 has a significant relationship with the medical (r=-0.261, p<0.01) and psychiatric (r=-0.318, p<0.001) domains. BDI was also found to be closely correlated with the psychiatric domain (r=0.219; p<0.01). In terms of reliability, the internal consistency of data through Cronbach’s alpha was satisfactory (between 0.581-0.873), and external reliability through inter-rater (0.283-0.7) and test-retest (0.251-0.946) was also good. In conclusion ASI-5-BM had good psychometric values and can be used in Malaysia.

ASI, addiction severity index; PWUD, people who use drugs; BDI, beck depression inventory; ICCE, international center for credentialing and education for addiction professionals; UTC, universal treatment curriculum.

Addiction Severity Index (ASI) is a WHO and UNODC recognized clinical and research based measurement tool, which has been widely used for the purposes of assessing and evaluating problems experienced by people who use drugs (PWUD). The assessment and evaluation conducted by the ASI identifies the severity of substance use and its harm to the individual.¹ The ASI-5 that was developed in 1992 is a modified version of the instrument.² Currently the ASI has developed a sixth version that is utilized for clinical and research purposes.³ ASI-5 assesses seven domains or aspects of life that can be affected as a result of substance use, namely medical status, employment status, alcohol use, drug use, legal status, family/social relationships, and psychiatric status,^1,4 but only 6 domains when the use of drugs and alcohol are combined. The assessment to evaluate the severity and seriousness of substance use is done over 2 separate time periods, in the previous 30 days and lifelong. For research purposes, the composite score for each domain is also utilized in the summarized scores of individual ASI domains. Since its development, the ASI¹ has been widely utilized to assess the disruption and deterioration caused by substance abuse in individuals.¹ The ASI has been translated and validated in China,⁵ Japan,⁶ Turkey,⁷ France,⁸ Netherlands⁹ and other countries, justifying ASI as an efficient assessment tool. The translated instrument is used to formulate treatment and rehabilitation plans¹⁰ at a variety of settings, such as addiction clinics, rehabilitation centers, mental health clinics and prisons, in a variety of population (people who use substances and alcohol) who volunteer for treatment and rehabilitation and substance abusers with dual diagnosis.

Reliability

Much of the ASI psychometric studies done in the U.S. show strong reliability data with 'inter-rater reliability’ of 0.89 and above.^1,11 In addition, the ASI is also proven to be sensitive in detecting changes in clients during treatment¹² while predicting treatment and recovery outcomes.¹³ For that purpose, local adaptations of the ASI enable execution of psychometric tests to evaluate the suitability and strength of the instrument in assessing addiction severity, to evaluate change during treatment and to monitor outcomes of the recovery program. In a survey of internal consistency among 115 alcohol abusers using Cronbach's alpha coefficient, Demirbas et al.⁷ found satisfactory result, namely 0.74 for medical status; 0.64 for employment status; 0.70 for legal status; 0.69 for alcohol use; 0.64 to family/social relationships; and 0.77 for psychiatric status. Studies on French speaking Swiss addicts also showed good internal consistency, between 0.58s-0.81.

A study by Leonhard et al.¹⁴ on 8,984 participants in the US revealed good Cronbach's alpha coefficient for internal consistency, namely 0.89 for medical status; 0.65 for employment status; 0.84 for alcohol use; 0.69 for drugs use; 0.65 for legal status; 0.74 for family/social relationships and 0.84 for psychiatric status. A Chinese study by Luo et al.⁵ also found good internal consistency of 0.86 for medical status; 0.55 for employment status; 0.79 for alcohol use; 0.49 for drug use; 0.76 for legal status; 0.66 for family/social relationships; and 0.71 for psychiatric status. In another Dutch study among 142 addicts, the internal consistency was also good, namely 0.81 for medical status; 0:58 for employment status; 0.92 for alcohol use; 0.70 for drug use; 0.71 for legal status; 0.53 for family/social relationships; and 0.84 for psychiatric status.⁹ In Japan, the internal consistency of the ASI data was also found to be good; 0.79 for medical status; 0.67 for employment status; 0.67 for alcohol use; 0.49 for drug use; 0.76 for legal status; 0.66 for family/social relationships; and 0.71 for psychiatric status.⁶

Demirbas et al.,⁷ in testing external reliability through test-retest correlation among 31 subjects for 10 days, found correlation values of 0.85 for medical status; 0.75 for employment status; 0.84 for legal status, 0.85 for alcohol use; 0.83 for family / social relationships and 0.91 for psychiatric status. Good correlation values (0.70 - 0.92) were also obtained by Luo et al.⁵ of between 0.70-0.92. In testing inter-rater reliability among 33 subjects, Demirbas et al.⁷ found a high range of correlations, between 0.74-0.99, assigned by three raters. In China, Luo et al.⁵ found an inter-rater correlation of between 0.74-0.98among addicts.

Validity

The validity of ASI obtained in the US and other countries showed good results.ASI has good 'concurrent' and 'discriminant' validity (>0.92).⁴ Leonhard et al.¹⁴ also noted significant correlation between the ASI domains except between the medical with drug use domain and the legal with the psychiatric domain. Demirbas et al.⁷ reported satisfying coefficient correlation scores in the rating of severity and ASI composite scores, namely, 0.89 for the medical; 0.54 for employment; 0.59 for alcohol use; 0.87 for legal; 0.61 or family/social; and 0.81 for psychiatric domain. Discriminant validity between Beck Depression Inventory (BDI) and the use of alcohol was also found to be good (r=0.45, p<0.01).

Inversely in China, Luo et al.¹⁵ found multiple domains in significantly related to each other, namely medical with drug use and legal; employment with alcohol use, drug use and family/social relationships; alcohol use with drug use, legal issues and family/social relationships; drug use with legal domain: and legal with family/social relationships and psychiatric domain. This finding is surprising based on the significant correlation between domains observed in many other countries. A suggested factor affecting the results of this research could have been the selection of samples among alcohol users whom had no serious dependence. Similar results in a Swiss study by Krenz et al.⁸ reported significant correlation only between certain domains, namely, legal with occupational domain (r=-0.280, p<0.05); drug use with family/social relationship domain (r=0.271, p<0.05) and family/social relationship with psychiatric status domain (r=0.385, p<0.01). Based on previous studies, the AS-5 demonstrated good reliability and validity in both versions, original and translated.

The objective of this study was to obtain data validation on the validity and reliability of the translated ASI-5 Bahasa Malaysia Version (ASI-5-BM).

This study utilized a variety of approaches in the data collection phase, beginning with the training of officers and counsellors from the National Anti-Drug Agency (AADK) through interviews, questionnaires and data analysis. The study population consisted of clients from the public drug rehabilitation centers. Samples were selected from five public addiction treatment centers and a district rehabilitation center, as shown in Table 1.

Study administration & instruments

Three instruments were used in this validation study. The first is ASI-5-BM. The ASI version used for the validation study is ASI-5 UNODC Treatment version utilized by the ‘International Center for Credentialing and Education for Addiction Professionals’ (ICCE) Colombo Plan in their Universal Treatment Curriculum (UTC). It was translated into Bahasa Malaysia by "back translation" method.^16,17 Second is the SF-12 (12-Item Short Form Health Survey), developed by Ware et al.¹⁸ from the longer 36-Item Short Form Health Survey (SF-36),as a summarized version of the physical and mental health component for the U.S. population. The test-retest reliability (2 weeks) was good (0.89 to 0.76 for physical and mental) and the validity of the 'convergent' obtained was also good, ranging from 0.62 - 0.88. The SF-12 was also utilized by Denis et al.¹⁹ to obtain data validity of ASI. The third is the BDI (Beck Depression Inventory). The BDI that was founded by Aaron T Beck²⁰ contains 21 items. It has been widely used to examine the severity of depression. The internal consistency of the BDI is 0.9 and test re-test reliability ranged from 0.73-0.96.²¹ BDI was also used by Demirbas et al.⁷ to validate the ASI among alcohol users and by Hendriks et al.⁹ among drug addicts.

The ASI-5-BM was administered by officers and counsellors specially trained to administer the instrument. The training was provided by three experts (a psychologist, a psychiatrist and a physician) with more than 10 years experience of administering the instrument for the assessment of substance addiction. Training of the trainers was conducted over two days in the form of lectures, practical demonstrations, exercises, reflective and practical sessions till satisfactory standards were achieved. The trained ASI trainers were expected to conduct at least three additional ASI interviews before training others. Following this, the officers and counsellors at the individual centers would continue conducting training session on administering the ASI-BM-5 for purposes of assessing severity of substance addiction.

Participants

Participants of this study consist of clients undergoing treatment and rehabilitation at six public treatment and rehabilitation centers under the National Anti-Drugs Agency (AADK) as shown in Table 1. Participants were selected among clients newly admitted to treatment and recovery centers. Clients undergo detoxification for 14 days. After this period, the counsellor would determine the client’s stability to undergo the ASI-5-BM interview. All 209 participants had given consent to be interviewed. The trained counsellors interviewed the participants by administrating the ASI-5-BM. The range of the interview time was between 50 - 80 minutes. After the ASI-5-BM interview, clients proceeded to complete the 12 Item Short Form Health Survey (SF-12) and Beck Depression Inventory (BDI).

Analysis

Analysis of validity was performed using the internal and external methods (convergent and concurrent). As for reliability, analysis of internal consistency using Cronbach's alpha, as well as external reliability factors, such as test-retest correlation and inter-rater analysis was performed.

A total of 209 participants were interviewed by officers and counsellors who have been specially trained to administer the ASI-5-BM. Information pertaining to the participants are shown in Table 2.

Validity

'Convergent' validity refers to the degree to which two measures of construct that theoretically should be related are in fact related. To establish the 'convergent' validity a correlation analysis was performed between domains within the ASI-5-BM. Correlation analysis showed high and significant coefficient scores. Analysis Table 3 found that participants who use drugs were closely correlated with health/medical problems (r=0.145; p<0.05); family/relationship problems (r=0.282, p<0.001); and psychiatric problems (r=0.207, p<0.01). The use of alcohol was also significantly correlated with health/medical problems (r=0.179, p<0.01) and legal problems (r=0.215, p<0.01). In addition, participants who were facing fewer legal problems also had fewer psychiatric problems (r=0.234, p<0.01) and those with family/social problems also had psychiatric problems (r=0.410, p<0.001).Theoretically, all these domains were closely correlated with each other.

The second step was to measure the "concurrent" validity which measures the extent to which results of a test/measure correlates with a measure that has previously been validated, on the same construct. To examine the validity of the ASI-5-BM, two instruments with genuine capabilities to measure domains within the ASI were identified for purposes of comparison, namely the SF-12 and the BDI. The SF-12 that examines the physical and mental health status of an individual is then compared against the health/medical domain and the psychiatric domain of the ASI-5-BM, while the BDI was used to confirm the psychiatric domain of the ASI-5-BM. Table 4 shows that participants who scored high in the health/medical domain (multiple health/medical problems) scored low in the SF-12 (r=-0.261, p<0.01).This meant that the 'concurrent' validity was established. Concurrently, those who scored high on the SF-12 (healthy) were less likely to use drugs (r=-0.126, p<0.05); less likely to use alcohol (r=-0.182, p<0.01), had less legal problems (r=-0.193, p<0.05), less family/social relationship issues (r=-0.170, p<0.05) and less psychiatric problems (r=-0.318, p<0.001). These results supported the 'concurrent' validity. Participants who scored high on the BDI also scored high in the psychiatry domain of the ASI-5-BM (r=0.291, p<0.01). This also suggested that the 'concurrent' validity was established. There were no other significant correlations observed.

Reliability

Reliability is related to the quality of a measurement. Three forms of analysis were conducted, an internal consistency and two analysis of external reliability. The internal consistency examined the extent to which a set of items measured the construct consistently within the group. This was measured with Cronbach's alpha. The analysis results can be seen in Table 5. A reliable measurement is also a consistent measurement, which meant that similar results could be obtained consistently or almost similar. The two methods used are firstly, the inter-rater reliability to measure the degrees of agreement among raters. The second method utilized is the test-retest reliability that examines consistency of a measurement over a period of time. The inter-rater correlation between raters across all domains of the ASI-5-BM. Table 6 was found to be positive and significant (r=0.231 to 0.700, p<0.05). Despite the different raters, the scores were consistent.

For the test-retest correlation, a period of 14 days retest was set in accordance with a previous study.¹⁵ The interviewer randomly interviewed five participants from each of the centers and then re-interviewed them 14 days later. Test-retest correlation of ASI-5-BM composite scores in Table 7 were in the range of 0.251-0.946. Similar analysis was also performed for SF-12 and BDI. Both scales measured good Cronbach's alpha values (Table 8) and test-retest correlation (Table 9). It can be concluded from the reliability and validity tests data that the ASI-5-BM version of the instrument has good psychometric values.

The aim of this study was to obtain validation data, i.e. the validity and reliability of the ASI-5 instrument that has been translated to Bahasa Malaysia Version (ASI-5-BM). The validity and reliability analysis of ASI-5-BM has been conducted and found to have good psychometric characteristics. Much data about the validity of the ASI has been specified in the US and other countries. The ASI has good 'concurrent' and 'discriminant' validity (>0.92).¹ Convergent analysis found clients who use drugs and/or alcohol to have health/medical problems, family/social problems, and/or psychiatric problems. All these domains are theoretical correlated to one another.^11,19 In Turkey, Demirbas et al.⁷ recorded a satisfactory correlation for the severity rating and composite ASI scores (0.54 - 0.87) for all ASI-5 domains.

The discriminant validity of the BDI for the use of alcohol was also good (r=0.45, p<0.01).⁷ In the US, Leonhard et al. (2002), found significant correlation between domains in the ASI-5 excepting in the medical domain with drug use and the legal domain with psychiatry problems. Although the study by Luo et al.¹⁵ found many domains unrelated to each other, but his study was limited to alcohol users who may not have had significant family, legal or psychiatric problems. The Krenz et al.⁸ study with Swiss samples reported that significant correlation was found only between legal and occupational domains (r=-0.280, p<0.05); drug use and family/social relationship domains (r=0.271, p<0.05) and family/social relationship and psychiatric domains (r=0.385, p<0.01). Our analysis also showed that the ASI-5-BM medical domain has significant correlation with SF-12, BDI and the psychiatric domain. This means that 'concurrent' validity for ASI-5-BM has been acquired.

The reliability of our study was found to be consistent with a good Cronbach's alpha estimates for the ASI-5-BM, in keeping with data from the US,¹⁴ Netherlands,⁹ China,^5,15 Switzerland⁸ and Japan.⁶ Alpha values for the medical and occupation domain are lesser than 0.7. For the medical domain (0.581), this is probably due to the reporting of how many times there were hospitalized as opposed to how many days that they were in the hospital. For the occupation domain (0.634), there may be some variation on the full-time employment as opposed to non-full time. The inter-rater analysis showed good consensus between two raters to an independent evaluation of the severity of each domain in the ASI-5-BM. Demirbas et al.⁷ and Luo et al.⁵ also found similar good inter-rater correlations in their studies, namely between 0.71-0.99 and between 0.74-0.98 respectively. This data gives us confidence that the evaluation of a same situation is consistent among multiple raters. The test-retest correlation was also found to be good in this study, which is in keeping with similar results from the Demirbas et al.⁷ and Luo et al.⁵ studies. The test-retest reliability examines the variation in measurement over time, where the instrument gave similar values when re-utilized by the same interviewer on the same client, even between assessments of 14 days apart. These findings substantiates that the ASI-5-BM has good and stable psychometric properties as stated by Stoffelmayer et al..^22,23 Based on our findings, we conclude that the ASI-5-BM can be used in Malaysia.

National Anti Drug Agency, Ministry of Home Affairs, Malaysia.

The author declares no conflict of interest.

1. McLellan AT, Luborsky L, Woody GE, et al. An improved diagnostic evaluation instrument for substance abuse patients, The Addiction Severity Index. J Nerv Ment Dis. 1980;168(1):26–33.
2. McLellan AT, Kushner H, Metzger D, et al. The fifth edition of the Addiction Severity Index. J Subst Abuse Treat. 1992;9(3):199–213.
3. Kessler F, Cacciola J, Alterman A, et al. Psychometric properties of the sixth version of the Addiction Severity Index (ASI–6) in Brazil. Revisiting Brasileira de Psiquiatria. 2012;34(1):24–33.
4. McLellan AT, O'Brien CP, Kron R, et al. Matching patients to appropriate substance abuse treatment: a conceptual and methodological approach. Drug Alcohol Depend. 1980;5(3):189–195.
5. Luo W, Wu Z, Wei X. Reliability and validity of the Chinese version of the addiction severity index. J Acquir Immune Defic Syndr. 2010;55(Suppl 1):S121–S125.
6. Haraguchi A, Ogai Y, Senoo E, et al. Verification of the Addiction Severity Index Japanese version as a treatment–customization, prediction and comparison tool for alcohol dependent patient. Int J Environ Res Public Health. 2009;6(8):2205–2225.
7. Demirbaş H, Özgür Ilhan İ, Doğan YB, et al. Reliability and validity of the Turkish version of the Addiction Severity Index in Male Alcohol Dependents. Noro Psikiyatr Ars. 2014;51(3):216–221.
8. Krenz S, Dieckmann S, Favrat B, et al. French version of the addiction severity index. 5th Edition: validity and reliability among Swiss opiate–dependent patients. French validation of the Addiction Severity Index. Eur Addict Res. 2004;10(4):173–179.
9. Hendriks VM, Kaplan CD, van Limbeek J, et al. The Addiction Severity Index: Reliability and validity in a Dutch addict population. J Subst Abuse Treat. 1989;6(2):133–141.
10. Denis CM, Cacciola JS, Alterman AI. Addiction Severity Index (ASI) summary scores: Comparison of the Recent Status Scores of the ASI–6 and the Composite Scores of the ASI–5. J Subst Abuse Treat. 2013;45(5):444–450.
11. McLellan AT, Luborsky L, Cacciola J, et al. New Data from Addiction Severity Index: Reliability and validity in 3 centers. J Nerv Ment Dis. 1985;173(7):412–423.
12. McLellan AT, Woody GE, Luborsky L, et al. Increase effectiveness of substance abuse treatment. A prospective study for patient–treatment matching. J Nerv Ment Dis. 1983;171(10):597–605.
13. McLellan AT, Luborsky L, Woody CP, et al. Is effective treatment for substance abuse? Journal of the American Medical Association. 1983;247:1422–1428.
14. Leonhard C, Mulvey K, Gastfriend DR, et al. The Addiction Severity Index: A field study of internal consistency and validity. J Subst Abuse Treat. 2000;18(2):129–135.
15. Luo W, Guo CX, Han DL, et al. Reliability and validity of Chinese version of the Addiction Severity Index among drug users in the community. Biomed Environ Sci. 2012;25(6):684–689.
16. Brislin RW. Back–translation for Cross–Cultural Research. Journal of Cross–Cultural Psychology. 1970;1(3):185–216.
17. Mahmood NM. Cross–cultural limitation of back translated test used in management and social science research. Journal of Social Development. 2004;8:45–62.
18. Ware J Jr, Kosinski M, Keller SD. A 12–item Short–Form Health Survey: Construct of scales and preliminary tests of reliability and validity. Med Care. 1996;34(3):220–233.
19. Denis CM, Cacciola JS, Alterman AI. Addiction Severity Index scores summary: comparison of the recent status of the ASI scores–6 and the composite scores of the ASI–5. J Subst Abuse Treat. 2013;45(5):444–450.
20. Beck AT, Ward CH, Mendelson M, et al. An inventory for measuring depression. Arch Gen Psychiatry. 1961;4:561–571.
21. Beck AT, Steer RA, Garbin MG. Psychometric properties of the Beck Depression Inventory: Twenty–five years of evaluation. Clinical Psychology Review. 1988;8(1):77–100.
22. Stöffelmayr BE, Mavis BE, Kasim RM. The longitudinal stability of the Addiction Severity Index. J Subst Abuse Treat. 1994;11(4):373–378.
23. Cacciola JS, Alterman AI, McLellan AT, et al. Initial evidence for the reliability and validity of a "Lite" version of the Addiction Severity Index. Drug Alcohol Depend. 2007;87(2–3):297–302.