Table 1 Characteristics of included studies of validation testing
First author and publication year | Participant’s characteristics | Study characteristics | Survey instrument/questionnaire characteristics | Validation methods | Key findings | ROB assessment |
|---|---|---|---|---|---|---|
Alexander and Leung [18] | • Aged ranged from 18 to 54 with mean age of 33.2 • Most of participants were male (n = 64, 58.7%) • 63% White, 20% Black, and 17% Hispanic | Non-population-based study evaluated the concurrent, convergent, and discriminant validity between the MSI-X and five other instruments in the alcohol and drug program | MSI-X • Specific substance use: Marijuana use • Sample size: n = 107 • Response rate = 13.4% • Survey duration: September 2001–June 2002 | Construct (convergent) validity • Comparison measure: Compared with SASSI-3 overall “decision rule” for those most likely to be diagnosed with substance dependence or substance abuse • Statistical analysis: Two separate independent sample t-tests Criterion (concurrent) validity • Comparison measure: Comparing with DSM-G-MI, DAST-20, SASSI-3, and ASI interviewer severity rating scales and selected variables • Statistical analysis: Correlation coefficients | Those who had a high probability of substance dependence had a significantly (t = 3.256, df = 43.388, p = .002) higher number of MSI-X problems (mean = 6.2) than those who had a low probability for substance dependency (mean = 2.4). The concurrent validity analysis reveals excellent, strong significant positive association between MSI-X scores and the DSM-G-MI (r = 0.811; p < .001), good to moderately strong significant positive associations with the DAST-20 scores (r = 0.531; p < .001), and two specific SASSI-3 subscale scores, the Face Valid Other Drugs (FVOD) subscale scores (r = 0.626; p < .001) and correctional (COR) subscale (r = 0.522; p < .001) | 5 |
Appleby, Dyson [19] | • Mean age of 32 years | Non-population-based study conducted in the inpatient clinical setting and used CUAD by comparing CUAD severity ratings with severity ratings from the other instruments used in the study and by determining its sensitivity and specificity in relation to SCID diagnoses of substance use disorders | CUAD • Specific substance use: Not reported • Sample size: 100 • Response rate: Not reported • Survey duration: Not reported | Construct validity • Comparison measure: Comparing CUAD severity ratings with severity ratings from other instruments used in the study: SMAST, DAST, ASI • Statistical analysis: Chi-square statistical analysis Criterion validity (concurrent validity) • Comparison measure: Comparing CUAD severity ratings with severity ratings from ASI alcohol and drug scales • Statistical analysis: Sensitivity | This study showed a highly significant positive correlation between the CUAD alcohol subscale and the SMAST and an even stronger association between the CUAD drug subscale and the DAST. Additionally, high positive correlations between CUAD and ASI alcohol and drug scales and both current substance use measures provided additional evidence of concurrent validity. Seventy-three patients had a SCID lifetime diagnosis of an alcohol use disorder, and the CUAD correctly identified 51 of them (70% sensitivity), which also showed evidence criterion validity | 4 |
Boothroyd, Peters [20] | • Mean age of 40.1 (SD = 12.2) • 25.4% were males, and 74.6% were females | Population-based study examined the SSI’s psychometric properties within Medicaid recipients in Florida | SSI-SA • Specific substance use: Not reported • Sample size: n = 6664 • Response rate: 35–45% • Survey duration: 1998–2008 | Construct validity (convergent validity) • Comparison measures: Examine the association of respondents’ SSI-SA classification at cutoff of 4 vs. at or above the cutoff 4 with their scores on a measure of general functioning, an indicator of current substance use, and quality of life • Statistical analysis: Independent t-tests sensitivity, specificity, PPV, NPV, ROC curve analysis | The SSI-SA had excellent internal consistency (0.85). Evidence of the SSI’s validity was strong. Using the recommended SSI-SA cutoff score of 4 or higher to indicate the presence of a substance abuse problem, the SSI-SA had respectable sensitivity (0.82) and specificity (0.90) | 4 |
Broderick, Richmond [21] | • Mean age (SD) = 43 (15) years • Most of participants were males (57% vs. 43%) • Hispanic (37%), White (41%), and Black (18%) | Non-population-based study conducted in the inpatient clinical setting to evaluate two brief screen questions to assess the degree to which these single-item screening questions detected risky substance use compared to a longer, validated screening tool | Two-item brief screen (survey name not reported) • Specific substance use: Illicit drugs and marijuana • Sample size: n = 1218 • Response rate: n = 72% • Survey duration: August 25–October 31, 2010 | Construct validity • Comparison measure: ASSIST • Statistical analysis: Sensitivity and specificity | Sensitivity values for the marijuana and street drug questions were 72% and 40%, respectively. Specificity values for the marijuana and street drug questions were 96% and 99%, respectively | 5 |
Carter, Yu [6] | • Mean age (SD) = 49 (14.9) • Most participants were female (61.9%) • White (92.5%), Black/African American (4.5%), and Hispanic (0.9%) | Non-population-based study conducted in the community inpatient clinical setting to validate the performance of TAPS tool compared to a reference-standard substance use assessment | TAPS • Specific substance use: Not reported • Sample size: n = 1523 • Response rate: Not reported • Survey duration: November 2019 to October 2020 | Construct validity • Comparison measure: ASSIST • Statistical analysis: AUC of ROC curve | The TAPS tool showed fair or better discrimination between moderate risk use and high-risk use for tobacco, alcohol, and prescription opioids (AUCs: 0.75–0.97) and fair or better discrimination between low-risk and moderate-risk use in five of eight subscales, including tobacco, alcohol, marijuana, stimulants, and heroin (AUCs: 0.70–0.92) | 5 |
Chasnoff, Wells [22] | All pregnant women 18 years of age or older | Non-population-based study conducted in the inpatient clinical setting to validate the 4P’s Plus to identify women whose substance-use levels fall below the DSM-IV criteria but still at risk from any level of use of alcohol or illicit drugs | 4P’s Plus • Specific substance use: Not reported • Sample size: n = 228 • Response rate: 59% • Survey duration: Not reported | Criterion validity (predictive validity) • Comparison measure: Comparing the 4P’s Plus positive and negative screens with positive and negative clinical assessment • Statistical analysis: Sensitivity, specificity, PPV, NPV | The overall reliability for the 5-item measure was 0.62. Sensitivity and specificity were very good at 87 and 76%, respectively. Positive predictive validity was low (36%), but negative predictive validity was quite high (97%). Of the 31 women who had a positive clinical assessment, 45% were using less than 1 day per week | 5 |
Dennis and Davis [23] | • Mean age (SD) = 31 (11) years • 59% of the study population were males, and 41% were females • White: 64%, African American/Black: 10%, and Other race: 10% | Non-population-based study conducted in the outpatient clinical setting to examine the psychometric properties of core GAIN-Q3 assessment | GAIN-03 • Specific substance use: Not reported • Sample size: n = 10,167 • Response rate: Not reported • Survey duration: 2002–2010 | Construct validity (convergent and discriminant validity) • Comparison measure: Compared the correlation between the screener and the full GAIN-I scale • Statistical analysis: Pearson correlation matrix between the shortened screener and the full-length scores, ROC analysis | Despite the condensed lengths of the screening measures compared with their longer versions, the reliability estimates are within the good to excellent range (0.7 to 0.9) in terms of internal consistency for 7 of the 10 screeners for adults. Moreover, there is strong evidence for the measures’ convergent and discriminant validity and efficiency (i.e., maximum information gathered in as few items possible) relative to the full-length scales as well as relative to other scales in the full GAIN-I | 5 |
Dezman, Gorelick [24] | • All participants aged over 18 years old • Most participants were male (n = 803, 72.0%) • Caucasian: n = 627 (56.2%), African American: n = 481 (43.1%), Hispanic: n = 7 (0.6%) | Non-population-based study conducted in the trauma inpatient clinical setting to test characteristics of a 4-item drug CAGE questionnaire to detect DUDs | CAGE • Specific substance use: Not reported • Sample size: n = 1115 • Response rate: 58% • Survey duration: September 1994–November 1996 | Criterion validity • Comparison measure: Comparison with SCID-generated DUD diagnoses as the standard • Statistical analysis: Sensitivity, specificity, PPV, NPV, and AUC | The drug CAGE screen had an AUC = 0.9. Each individual question had a high AUC (0.78–0.87). The drug CAGE questionnaire had high AUCs across all sociodemographic and injury mechanism subgroups, both for each individual question and overall | 5 |
Duncan, Sacks [25] | The study showed self-reported demographic characteristics of the study sample composed of African American, Latinos, and Whites. The majority of the sample in each racial and ethnic group was male. The average age was 36.03 years for African Americans, 32.74 years for Latinos, and 36.02 years for Whites. Males: 56.03% for African Americans, 60.8% for Latinos, and 36.02% for Whites | Non-population-based study conducted in the prison substance abuse treatment programs to test the stability of the performance of the CODSI-MD and SMD across three racial/ethnic groups of offenders entering prison substance abuse treatment programs | CODSI-MD • Specific substance use: Not reported • Sample size: n = 353 • Response rate: Not reported • Survey duration: Not reported | Criterion validity • Comparison measure: Compare three racial or ethnic groups of offenders entering prison substance abuse treatment programs • Statistical analysis: Sensitivity, specificity, PPV, NPV, positive and negative likelihood ratio | No statistical differences in sensitivity or specificity for either CODSI-MD or SMD across the African American, Latino, and White prisoner groups | 4 |
CODSI-SMD • Specific substance use: Not reported • Sample size: n = 353 • Response rate: Not reported • Survey duration: Not reported | Criterion validity • Comparison measure: Compare three racial or ethnic groups of offenders entering prison substance abuse treatment programs • Statistical analysis: Sensitivity, specificity, PPV, NPV, positive and negative likelihood ratio | |||||
Han, Sherman [26] | • Mean age (SD) = 47.1 (13.4) years • Most participants were males (72.0%) • White: 56.2%, African American/Black: 40.8%, American Indian/Alaska Native: 1.8%, Asian: 2.5%, Other race: 22.1% | Non-population-based study conducted in the inpatient clinical setting to assess the diagnostic accuracy of the SUBS in comparison to ASSIST and AUDIT-C to identify unhealthy and high-risk alcohol and drug use among hospitalized current smokers | SUBS • Specific substance use: Not reported • Sample size: n = 439 • Response rate: Not reported • Survey duration: Not reported | Criterion validity (concurrent) • Comparison measure: AUDIT-C for alcohol use and ASSIST for drug use • Statistical analysis: ROC analysis, sensitivity, specificity | The SUBS had a sensitivity of 98% (95% CI 95–100%) and specificity of 61% (95% CI 55–67%) for unhealthy alcohol use, a sensitivity of 85% (95% CI 80–90%) and specificity of 75% (95% CI 78–87%) for illicit drug use, and a sensitivity of 73% (95% CI 61–83%) and specificity of 83% (95% CI 78–87%) for prescription drug nonmedical use. For identifying high-risk use, a higher cutoff (response of “3 or more days” of use indicates a positive screen), the SUBS retained high sensitivity (77–90%), and specificity was 62–88% | 4 |
Harris, Ellerbe [27] | Not reported | Non-population-based study conducted in the inpatient and outpatient clinical setting to assess the specification validity of the 2009 HEDIS substance use disorder initiation and engagement | HEDIS • Specific substance use: Not reported • Sample size: n = 2726 • Response rate: Not reported • Survey duration: Not reported | Criterion validity (specification validity) • Comparison measure: Not reported • Statistical analysis: PPV | The PPV were excellent (> 9 0%) for residential and outpatient records selected from addiction treatment programs but more modest for records generated in non-addiction settings and were highly variable across facilities | 1 |
Hasin, Greenstein [28] | • More participants were females than males (54.5% vs. 45.5%) • White: 67.8%, African American/Black: 22.5%, American Indian/Alaska Native: 2.1%, Asian: 0.7%, and Other race: 6.9% | Population-based study used a test–retest design to compare concordance of the NESARC-III survey questions with a semi-structured interview, the Psychiatric Research Interview for Substance and Mental Disorders, DSM-5 version (PRISM-5) administered by a clinician | AUDADIS • Specific substance use: Not reported • Sample size: n = 712 • Response rate: 92.5% • Survey duration: June 2012–July 2013 | Criterion validity (concurrent validity) • Comparison measure: Physician diagnosis • Statistical analysis: Intraclass correlation coefficients | Concordance of the AUDADIS-5 and the PRISM-5 for DSM-5 diagnoses of substance use disorders ranged from fair to good (κ = 0.40–0.72). Concordance on dimensional scales was excellent (ICC ≥ 0.75) for the majority of DSM-5 SUD diagnoses and fair to good (ICC = 0.43–0.72) for most of the rest | 5 |
Hasin, Keyes [29] | For frequent marijuana users: 66.9% males and 33.1 females, 75.4% White, 11.4% African American/Black, 3.7% American Indian/Alaska Native, 2.2% Native Hawaiian or Other Pacific Islander For marijuana use only users: 58.2% males and 41.8% females, 71.1% White, 15.3% African American/Black, 2.7% American Indian/Alaska Native, 2.7% Native Hawaiian or Other Pacific Islander | Sub-population-based study and participants were selected from the NESARC-III sample to investigate the factor structure, clinical validity, and psychiatric correlates | NESARC • Specific substance use: Marijuana • Sample size: n = 3732 • Response rate: n = 81% • Survey duration: 2001–2002 | Construct validity (convergent validity) • Comparison measure: AUDASIS-IV • Statistical analysis: Factor analysis Criterion validity (predictive validity) • Comparison measure: DSM-IV marijuana dependence criteria • Statistical analysis: Binomial regression | Both marijuana withdrawal symptoms were associated with significant distress/impairment, substance use to relieve/avoid marijuana withdrawal symptoms, and quantity of marijuana use. Panic and personality disorders were associated with anxiety symptoms in both frequent marijuana users and marijuana-only users | 4 |
Hser, Shen [30] | • The overall mean age was 32.6 years (SD = 7.6) • Two-thirds (66.0%) of the DATOS sample were male • Almost half of the patients (46.6%) were African American, 38.3% were white, 12.5% were Hispanic, and the remainder (2.7%) were either Asian or others | Non-population-based study conducted in the inpatient and outpatient clinical setting to develop a lifetime severity index for cocaine use disorder and examine its predictive validity of posttreatment outcome using data from the nation Drug Abuse Treatment Outcome Study | LSI-Cocaine • Specific substance use: Cocaine use disorder • Sample size: n = 2107 • Response rate: Not reported • Survey duration: 1993–1999 | Criterion validity (concurrent validity) • Comparison measure: Not reported • Statistical analysis: Concordance agreement | A higher value of the index, indicating greater severity, predicted a greater likelihood of relapse (the odds ratios were 5.7 for high severity and 4.4 for medium severity, relative to low severity) and shorter time to relapse. Similarly, the polytomous logistic analysis indicated that the index predicted levels of posttreatment cocaine use (odds ratios of daily use were 47.8 for the high severity and 18.8 for medium severity; the corresponding odds ratios of weekly use were 6.75 and 5.10 and for less than weekly use were 3.35 and 3.57) | 3 |
Jackson, Covell [31] | • Mean age (SD) = 37 (7.8) • 71% males and 29% females • 41% were White, and 59% were African American/Black | Non-population-based study conducted in the outpatient clinical setting to clarify the concordance rates of self-report of drug use compared to results of urine screens for drugs of abuse among individuals diagnosed with both a mental health and a substance use disorder | Self-reported drug use • Specific substance use: Not reported • Sample size: n = 196 • Response rate: Not reported • Survey duration: 1992–1998 | Criterion validity (concurrent validity) • Comparison measure: Urine drug screen • Statistical analysis: Concordance agreement (concordance between self-reported and urine drug screen) | The concordance between self-report and results from urine screens was high. Estimates for the likelihood of use of marijuana and cocaine within the past 30 days were 15% and 32%, respectively, based on urine screens, 25% and 35% based on self-report, and 28% and 43% based on information from both sources combined | 5 |
Joyner, Wright [32] | • Mean age = 30 years • 70% males and 30% females • 22% were White; 78% were African American/Black | Non-population-based study conducted in inpatient clinical setting to examine the ability of the ASI to provide valid and reliable data within a homeless population of drug misusers | ASI • Specific substance use: Not reported • Sample size: n = 23 • Response rate: 100% • Survey duration: May–June 1991 | Criterion validity (concurrent validity) • Comparison measure: Correlation between ASI section and corresponding composite measure • Statistical analysis: Pearson product-moment correlation coefficients | Only three of the coefficients did not meet this criterion. Years of education and employment income were more highly correlated with composites outside the area of employment. The legal composite rating was more highly correlated with days of cocaine/crack use rather than length of last incarceration. The mean composite score in the study sample was 0.102 for those with no alcohol problem, 0.403 for those with alcohol as a secondary problem, and 0.694 for those identifying alcohol as their primary problem | 4 |
Kellogg, Ho [33] | All drug positive participants: Mean age (SD) = 40.0 (9.8), most participants were male (64%), Caucasian (38%), African American (29%), Hispanic (26%), Other/mixed race (7%), mean education year (SD) = 12.5 (2.5) Normal volunteers: Mean age (SD) = 35.7 (14.1), male participants 49%, Caucasian (56%), African American (16%), Hispanic (14%), Other/mixed race (1%), Asian (13%), mean education year (SD) = 15.7 (3.2) | Non-population-based study conducted in the inpatient and outpatient clinical setting to assess the validity of the DRG scale through multiple analyses | PAI DRG • Specific substance use: Not reported • Sample size: n = 370 • Response rate: Not reported • Survey duration: Not reported | Construct validity • Comparison measure: Compared with ASI drug composite scores and ASI drug severity ratings • Statistical analysis: Kruskal–Wallis test Content validity • Comparison measure: Not reported • Statistical analysis: Sensitivity and specificity | The Kruskal–Wallis tests revealed that there were significant correlations among the PAI DRG scale and the ASI scales related to the frequency of use, negative consequences of use, and need and desire for treatment | 5 |
Kupetz, Klagsbrun [34] | Not reported | Non-population-based study conducted in the inpatient and outpatient clinical setting to develop a self-assessment survey instrument for the purpose of detecting abuse of both alcohol and other drugs and to assess validity | SUAS • Specific substance use: Not reported • Sample size: n = 5745 • Response rate: 88% • Survey duration: Not reported | Construct validity (convergent validity) • Comparison measure: Medical chart • Statistical analysis: Prevalence, prevalence percent agreement, detection | The SUAS was far superior to patient charts in identifying heavy drug and/or alcohol use, especially when both quantity-frequency and social consequences criteria were considered. Also, the SUAS did at least in simply identifying current heavy use of substances, although the numbers are too small to permit definite conclusions. However, an internal validity, which is the correlation of various parts of the SUAS, was not included in the present study | 3 |
Leonhard, Mulvey [35] | • Mean age (SD) = 34.46 (9.82) • More participants were males (72.2% vs. female 27.1%) • White (47.4%), African American/Black (38.1%), American Indian/Alaska Native (0.7%), Asian (0.5%), Other race (11.3%) | Non-population-based study conducted in a busy online inpatient alcohol and drug abuse clinical setting without tight procedural controls to investigate the internal consistency and validity | ASI • Specific substance use: Not reported • Sample size: n = 8984 • Response rate: 90% • Survey duration: Not reported | Construct validity (convergent and discriminant validity) • Comparison measure: Interviewer severity ratings and composite scores • Statistical analysis: Pearson correlation coefficients | Validity analyses showed good promise. Correlation matrices for both composite scores and severity ratings demonstrated good evidence for discriminant validity in measuring the ASI’s seven dimensions, medical severity, employment problems, alcohol and drug use, legal problems, social difficulties, and psychiatric problems | 4 |
McGovern and Morrison [36]* | • Mean age = 36 years, male: 67% and female: 33% • Most participants were White (50%), followed by Black (42%) and other (1%) | Non-population-based study conducted in the inpatient clinical setting to examine criterion-related validity by comparing the CUAD’s concordance rates with psychiatrist-derived diagnoses | CUAD • Specific substance use: Not reported • Sample size: n = 129 • Response rate: Not reported • Survey duration: Not reported | Criterion validity (concurrent validity) • Comparison measure: Comparing between CUAD with the MAST and DAST instruments • Statistical analysis: Comparison values in parentheses (%) | The CUAD total severity score is positively associated with all three criterion measures at highly significant levels. The CUAD total severity score is able to significantly discriminate the level of care assignments (p < .001). The validity of the CUAD is reported and appear satisfactory | 4 |
• Mean age = 35.8 years • Most participants are male (n = 259, 74%) • Most participants were White (70%), followed by Black (25%) and Hispanic (5%) | Non-population-based study conducted in the inpatient clinical setting to further investigate the validity of the CUAD by comparing two frequently used measures, MAST and DAST | CUAD • Specific substance use: Not reported • Sample size: n = 348 • Response rate: Not reported • Survey duration: Not reported | Construct validity • Comparison measure: The comparison was made by contrasting the CUAD derived DSM-III-R substance use disorder diagnoses with the chart diagnosis determined by the unit psychiatrists • Statistical analysis: Pearson correlation coefficients Criterion validity (concurrent validity and predictive validity) • Comparison measure: Explored by testing the capacity of the CUAD total severity score to distinguish among patients assigned to three levels of substance abuse treatment: inpatient, partial hospitalization, and outpatient • Statistical analysis: Relating the instrument concurrently with subjects scored on the MAST and DAST and predictively with the assignment of patients to varying levels of care | 4 | ||
McNeely, Cleland [37] | • Age ranges from 21 to 65, mean age (SD) = 46 (12), median age = 48 • More participants were female (n = 236, 51.4%), followed by male (n = 211, 48.1%) • More participants were Black/African American (n = 238, 51.0%), followed by Hispanic (n = 93, 20.2%), White (n = 88, 19.1%), Other (n = 38, 8.2%) • Most of participants had education on some college or trade school (n = 116, 25.3%) | Non-population-based study conducted in the inpatient clinical setting to test the sensitivity and specificity of SISQs for alcohol and other drug use, as well as their feasibility, for self-administration | SISQs • Specific substance use: Not reported • Sample size: n = 459 • Response rate: Not reported • Survey duration: June–July 2012 (site A), November 2013–June 2013 (site B) | Construct validity (discriminant validity) • Comparison measure: Not reported • Statistical analysis: Sensitivity, specificity, positive and negative diagnostic likelihood ratios, ROC curves | The SISQ drug had sensitivity of 71.3% (95% CI 62.4–79.1) and specificity of 94.3% (95% CI 91.3–96.6), AUC = 0.83 (95% CI 0.79–0.87), for detecting unhealthy drug use, and sensitivity of 85.1 (95% CI 75.0–92.3) and specificity of 88.6% (95% CI 85.0–91.6), AUC = 0.87 (95% CI 0.83–0.91), for drug use disorder | 4 |
McNeely, Strauss [38] | • Age ranges from 21 to 65, mean age (SD) = 46 (11.8), median age = 49 • Males and females were equally distributed (n = 292, 49.8%) • More participants were Black/African American (n = 293, 50.2%), followed by Hispanic (n = 127, 21.7%), White (n = 109, 18.7%), Other (n = 51, 8.7%) • Most of participants had HS grad or GED degree (n = 199, 34.0%) | Non-population-based study conducted in the adult inpatient clinical setting to evaluate the validity and test–retest reliability of the SUBS | SUBS • Specific substance use: Not reported • Sample size: n = 586 • Response rate: Not reported • Survey duration: April 2011–April 2012 (site A); June–July 2012 (site B) | Criterion validity (concurrent validity) • Comparison measure: Compared with participants with saliva drug testing, the reference standard measures were used in combination to identify unhealthy use and substance use disorder, for each of the four substance classes in the SUBS • Statistical analysis: Pearson correlation coefficients | For unhealthy use of illicit or prescription drugs, sensitivity was 82.5% (95% CI 75.7 to 88.0) and specificity 91.1% (95% CI 87.9 to 93.6). Analyses of area under the receiver operating curve (AUC) indicated good discrimination (AUC 0.74–0.97) for all substance classes | 4 |
Miele, Carpenter [39] | • Mean age (SD) = 35.6 (8.7) • Most participants were males (62%), non-Hispanic origin White (61%), African American (26%), and 13% Hispanic • 23% had not completed high school | Non-population-based study conducted in the inpatient clinical setting to investigate the test–retest reliability, internal consistency, diagnostic concordance, and concurrent validity of the SDSS’s ICD-10 dependence and harmful use scales | SDSS • Specific substance use: Marijuana use • Sample size: n = 180 • Response rate: Not reported • Survey duration: Not reported | Criterion validity (concurrent validity) • Comparison measure: Compared with ICD-10 dependence diagnosis • Statistical analysis: Bivariate correlations between CUDIT-R composite scores and measures of the frequency of marijuana use per week and marijuana-related consequences | The ICD-10 dependence and harmful use severity scales were significantly associated with GAS scores for heroin. For cannabis, the 3 ICD-10 dependence scales and the frequency of harmful use symptoms scale were significantly associated with the number of days of cannabis use | 4 |
O'Hare, Cutler [40] | • Mean age was 44.5 years old • More participants were females (54.2%) than males (45.8%) • The majority of participants were White (90.7%) | Non-population-based study conducted in the outpatient clinical setting to examine preliminary validity and reliability of the SSPI and test the concurrent validity of the SRSA, QFI, and one-item frequency of marijuana use index by examining their intercorrelations | SSPI • Specific substance use: Not reported • Specific mental health: Multidimensional psychosocial distress scale • Sample size: n = 227 • Response rate = 76.4% • Survey duration: Spring–summer of 1999 | Content validity (factorial validity) • Comparison measure: Not reported • Statistical analysis: Principal components analysis with varimax rotation Construct validity (specific type not reported) • Comparison measure: Self-related SRSA, a QFI, and a one-item index measuring the frequency of marijuana use • Statistical analysis: Coefficient correlations (Spearman’s rho)-correlations among substance abuse indices | Correlations among the one-item self-reported substance abuse index (SRSA), QFI (average drinks per week), and frequency of marijuana use were moderate and significant. Correlations among SSPI subscales and these three substance abuse indices tended to be insignificant overall, or if significant, very low | 4 |
SRSA, QFI, and one-item index measuring the frequency of marijuana use • Specific substance use: Substance abuse • Sample size: n = 227 • Response rate = 76.4% • Survey duration: Spring–summer, 1999 | Construct validity (specific type not reported) • Comparison measure: Not reported • Statistical analysis: Coefficient correlations (Spearman’s rho): correlations with brief SSPI subscales | |||||
Peters, Greenbaum [41] | • Mean age (SD) = 32.6 (10.2) years • All the participants were males • White (32.7%), African American (44.9%), and Other race (22.5%) | Non-population-based study conducted among a sample of 400 male inmates in the Holiday Transfer Facility to evaluate the effectiveness of eight different substance abuse screening instruments included overall accuracy, PPV, and sensitivity | ASI, DAST, SASSI-2, SSI, and TCUDS • Specific substance use: Not reported • Sample size: n = 400 • Response rate = 75% • Survey duration: Feb–April 1996 | Criterion validity (concurrent validity) • Comparison measure: SCID • Statistical analysis: Sensitivity, specificity, PPV, NPV | The TCUDS, SASSI-2, and SSI were examined with respect to their utility to identify either alcohol or drug disorders, while other screening instruments (including independent alcohol and drug screens from the ASI) were examined with respect to alcohol or drug disorders. The positive predictive value in detecting either alcohol or drug dependence was highest for the TCUDS, followed by the ADS/ASI drug, the SSI, and the SASSI-2. Sensitivity of the multipurpose instruments in detecting either alcohol or drug dependence was highest for the SSI, followed by the ADS/ASI drug, the SASSI-2, and the TCUDS | 5 |
Ramsay, Abedi [42] | • Mean age (SD) = 23.5 (4.6) years • More participants were males (70%) • The majority of participants were African American (91.7%). Other race (5%) | Non-population-based study conducted in the inpatient clinical setting to describe the development and initial validation of two new, multidimensional measures of substance use, LSUR and LSUR-12, and to provide structured visual tools to aid recall and enhance the validity of the data | LSUR • Specific substance use: Not reported • Sample size: n = 60 • Response rate: Not reported • Survey duration: Not reported | Construct validity (hypothesis testing validity) • Comparison measure: Comparing LSUR and LSUR-12 scores for each substance use and a set of key scores from the LSUR and LSUR-13 • Statistical analysis: Spearman correlations, independent samples Student’s t-tests, and Mann–Whitney U-tests | Lifetime and past 12-week doses measured by the LSUR and LSUR-12 were highly correlated with the HONC and FTND scores (ranging from ρ = 0.796 to ρ = 0.852). The number of years of education completed was negatively correlated with the lifetime nicotine dose, as predicted (ρ = − 0.351). The LSUR can be used to obtain total alcohol consumption for the 5 years preceding assessment, which had an even stronger correlation with the RETROSUB (ρ = 0.768, p < .001, n = 43). The LSUR can also estimate the number of days of cannabis use in the 5 years preceding the assessment, and for this, a stronger correlation was noted with the number of days of drugs use measured by the RETROSUB (ρ = 0.726, p < .001, n = 41) | 4 |
LSUR-12 • Specific substance use: Not reported • Sample size: n = 60 • Response rate: Not reported • Survey duration: Not reported | Construct validity (hypothesis testing validity) • Comparison measure: Comparing LSUR and LSUR-12 scores for each substance use and a set of key scores from the LSUR and LSUR-13 • Statistical analysis: Spearman correlations, independent samples Student’s t-tests, and Mann–Whitney U-tests | |||||
Rosenberg, Drake [43] | • Mean age (SD) = 38.03 (8.82) years • More participants were females (52.2% vs. 48.8% for males) • The majority of participants were White (98.4%). Other race (1.6%) | Non-population-based study conducted in the outpatient clinical setting to test the validity of DALI compared with clinician diagnosis and other screens for substance use disorders | DALI • Specific substance use: Not reported • Sample size: n = 247 • Response rate: Not reported • Survey duration: 1994–1996 | Criterion validity (concurrent validity) • Comparison measure: Clinician diagnosis • Statistical analysis: ROC analysis | ROC curves showed that the DALI functioned significantly better than traditional instruments for both alcohol and drug use disorders | 5 |
Salyers, Bosworth [44] | • Mean age (SD) = 42.3 (10.1) • More participants were females (64.8% vs. 35.2% for males) • White (47.1%), African American (44.6%), and Other race (5.5%) | Non-population-based study conducted in the inpatient and outpatient setting with several mental illness to examine the internal factor structure of the SF-12, test–retest reliability, and convergent and divergent validity by comparing SF-12 scores to other indexes of physical and mental health | SF-12 • Specific substance use: Not reported • Sample size: n = 801 • Response rate: Not reported • Survey duration: June 1997–December 1998 | Construct validity (specific type of validity not reported) • Comparison measure: Not reported in detail, only specify making comparison between SF-12 to other indexes of physical and mental health • Statistical analysis: Coefficient correlations | Each of the physical health indexes were significantly related to PCS as well as to MCS. However, psychiatric hospitalization and substance use disorder were associated with MCS but not with PCS. Chronic health problems and doctor visits for physical health were more strongly related to PCS than to MCS. However, the correlations for physical health hospitalizations did not differ significantly between PCS and MCS. Mental health indexes of self-reported overall mental health, psychiatric hospitalization, and substance use disorder were more strongly related to MCS than to PCS | 1 |
Schultz, Bassett [45] | • Mean age (SD) = 20.03 (1.51) • Most of participants were females (70.1%) and White (87.8%) | Sub-population-based study conducted in a large, public southeastern university to evaluate the internal consistency, concurrent and discriminant validity, and item performance of CUDIT-R among college students who reported recent marijuana use | CUDIT-R • Specific substance use: Marijuana use • Sample size: n = 229 • Response rate: Not reported • Survey duration: Not reported | Construct validity (discriminant validity) • Comparison measure: DSM-5 diagnostic severity levels • Statistical analysis: Sensitivity, specificity, ROC curve Criterion validity (concurrent validity) • Comparison measure: Compared to scores and measures of the frequency of marijuana use and marijuana-related consequences • Statistical analysis: Bivariate correlations between CUDIT-R composite scores and measures of the frequency of marijuana use per week and marijuana-related consequences | The CUDIT-R showed good internal consistency and concurrent validity with cannabis-related outcome measures including frequency of use, cannabis-related consequences, and total DSM-5 criteria endorsed. The CUDIT-R also showed evidence of discriminant validity across DSM-5 severity classifications, achieved high levels of sensitivity (0.929) and specificity (0.704), and excellent area under the receiver operating characteristics curve when using a cutoff score of six. All items displayed high levels of discrimination and varied in terms of difficulty and information provided | 5 |
Schwartz, McNeely [46] | • Mean age (SD) = 46.0 (14.7) • More participants were females (n = 1124, 56.2%) compared to males (n = 874, 43.7%) • More participants were African-American (n = 1112, 55.6%), followed by White (n = 667, 33.4%), Other race (n = 113, 5.7%), multiracial (n = 66, 3.3%), Asian (n = 35, 1.8%) | Non-population-based study conducted in the inpatient clinical setting to examine the performance of a TAPS tool compared to the WHO ASSIST by conducting concurrent validity | TAPS • Specific substance use: Not reported • Sample size: n = 2000 • Response rate: Not reported • Survey duration: August 2014–April 2015 | Criterion validity (concurrent validity) • Comparison measure: Comparison to the full WHO ASSIST as part of a large, multi-site study in eastern US primary care patients • Statistical analysis: Sensitivity, specificity, and PPV and NPV were calculated. ROC curves were computed, and the AUC was examined | For illicit drugs, sensitivities were ≥ 0.82 and specificities ≥ 0.92. The TAPS (at a cutoff of 1) had good sensitivity and specificity for moderate-risk tobacco use (0.83 and 0.97) and alcohol (0.83 and 0.74). Among illicit drugs, sensitivity was acceptable for moderate risk of marijuana (0.71), while it was low for all other illicit drugs and nonmedical use of prescription medications Specificities were 0.97 or higher for all illicit drugs and prescription medications | 5 |
Smith, Bennett [47] | • Age ranges from 18–25, mean age (SD) = 21.0 (2.39), 35% (N = 3396) were females • More participants were Whites (n = 5196, 53%), followed by African American (n = 1563, 16%), Latino (n = 1907, 19%), and Other (n = 1147, 12%) | Non-population-based study conducted in the outpatient clinical setting to examine the sensitivity and specificity of the SDScrY for three past year criterion variables, AOD, AUD, or DUD, in predicting emerging adults (18–25) substance use disorders | GAIN Short SDScrY • Specific substance use: Not reported • Sample size: n = 804 • Response rate: Not reported • Survey duration: Not reported | Criterion validity (predictive validity) • Comparison measure: Compared to the past year AOD, AUD, and DUD • Statistical analysis: Sensitivity, specificity, and the ROC curve | Analyses revealed a high correlation between the SDScrY screener and its longer parent scale (r = 0.95, p < 0.001). Sensitivity (83%) and specificity (95%) were highest at a cutoff score of two (AUC = 94%) on the SDScrY for any past year substance use disorder. Sensitivity (85%) was also high at a cutoff score of two on the SDScrY for any past year alcohol disorder | 4 |
Smith, Cheng [48] | • Age ranges from 21–86, median age = 49, 54% were females • The majority of participants (63%) identified themselves as Black or African American | Non-population-based study conducted in the inpatient clinical setting to test the validity of a single question for unhealthy alcohol use to detect other drug use | SIP-DU • Specific substance use: Not reported • Sample size: n = 286 • Response rate: 73% • Survey duration: October 2006–June 2007 | Content validity • Comparison measure: Not reported • Statistical analysis: Sensitivity, specificity, likelihood ratios, and AUC curve | The single-question screen at a cutoff of one or more times (the value considered a positive test for alcohol screening) was 67.6% sensitive (95% confidence interval (CI), 50.2–82.0%) and 64.7% specific (95% CI, 58.4–70.6%) for the detection of a current drug use disorder. It appeared slightly less sensitive (62.6%; 95% CI, 52.3–72.2%) and more specific (72.7%; 95% CI, 65.8–79.0%) for the detection of current drug use (although CIs overlapped). If oral fluid test results were considered, the sensitivity for detecting current drug use was lower (58.8%; 95% CI, 47.6–69.4%) and the specificity higher (80.3%; 95% CI, 72.5–86.7%) | 2 |
Smith, Schmidt [49] | Total study population • Age: Age range from 21 to 86, mean (SD) = 49 (12.3) • Sex: n = 115 (54.2%) were women • Race: American Indian/Alaskan native: n = 8 (2.8%), Asian: n = 7 (2.4%), Black: n = 179 (62.6%), Native Hawaiian: n = 3 (1.1%), White: n = 49 (17.1%), unknown: n = 40 (14.0%), Hispanic: n = 46 (16.1%) • Education: Some high school: n = 81 (28.3%), high school graduate: n = 107 (37.4%), some college: n = 59 (20.6%), college graduate: n = 28 (9.8%), postgraduate education: n = 11 (3.9%) Consent to oral fluid testing • Age: Age ranges from 21 to 86, mean (SD) = 49.3 (12.8) • Sex: n = 135 (56.2%) were women • Race: American Indian/Alaskan native: n = 5 (2.1%), Asian: n = 5 (2.1%); Black: n = 153 (63.8%), Native Hawaiian: n = 2 (0.8%), White: n = 42 (17.4%), unknown: n = 33 (13.8%), Hispanic: n = 38 (15.8%) • Education: Some high school: n = 68 (28.4%), high school graduate: n = 86 (35.8%), some college: n = 50 (20.8%), college graduate: n = 26 (10.8%), postgraduate education: n = 10 (4.2%) | Non-population-based study conducted in the inpatient clinical setting to validate a single-question screening test for drug use and drug use disorders | SIP-DU • Specific substance use: Not reported • Sample size: n = 286 • Response rate: 73% • Survey duration: October 2006–June 2007 | Content validity • Comparison measure: DAST-10 • Statistical analysis: Sensitivity, specificity, likelihood ratios, and AUC curve | The single screening question was 100% sensitive (95% confidence interval [CI], 90.6–100%) and 73.5% specific (95% CI, 67.7–78.6%) for the detection of a drug use disorder. It was less sensitive for the detection of self-reported current drug use (92.9%; 95% CI, 86.1–96.5%) and drug use detected by oral fluid testing or self-report (81.8%; 95% CI, 72.5–88.5%). Test characteristics were similar to those of the DAST-10 and were affected very little by participant demographic characteristics | 5 |
Tarter and Kirisci [50] | Group I: Alcohol and drug abusers • Age: Mean (SD) = 40.9 (5.7) • Sex: Male: n = 56 (47.1%), female: n = 63 (52.9%) • Race: Euro-American: n = 104 (87.4%), African American: n = 14 (11.8%), Other: n = 1 (0.8%) Group II: Normal controls • Age: Mean (SD) = 41.18 (4.79) • Sex: Male: n = 59 (49.6%), female: n = 60 (50.4%) • Race: Euro-American: n = 108 (90.8%), African American: n = 11 (9.2%), Other: not reported | Non-population-based study conducted in the inpatient clinical setting to evaluate the structure and psychometric properties of the adult version of the DUSI, to determine the capacity of the DUSI to discriminate substance abusers from non-substance abusing individuals, and to test the sensitivity of the DUSI for detecting individual who qualify for a DSM-III-R diagnosis of abuse or dependence | DUSI • Specific substance use: Not reported • Sample size: n = 238 (Group I); n = 299 (Group II) • Response rate: 73% • Survey duration: Not reported | Construct validity • Comparison measure: Compared to other measures, such as Multidimensional Personality Questionnaire, Family Assessment Measure, and General Health Questionnaire • Statistical analysis: The scales of the DUSI were correlated with other measures to examine the construct validity of the scales | Each of the 10 DUSI domains is unidimensional. Inter-item, split half, and internal reliability ranged from good to excellent. A score of 4 or higher on the substance use domain correctly classified 80% of the substance abusers, whereas a score of 3 or less accurately detected 100% of the normal control subjects. These results demonstrate that the DUSI is a practical and psychometrically sound screening instrument | 5 |
Tiet, Leyva [51] | • Mean age (SD) = 62.2 (12.6) years • The majority of participants were males (95.25%) • Whites (53.7%) and other race (45.05%) | Non-population-based study conducted in the outpatient clinical setting to create and validate a two-item screen for drug use from the ASSIST (excluding tobacco and alcohol) and to improve the efficiency of screening of drug misuse in primary care | ASSIT drug • Specific substance use: Not reported • Sample size: n = 1283 • Response rate: Not reported • Survey duration: Feb 2012–April 2014 | Criterion validity (concurrent validity) • Comparison measure: MINI and the inventory of drug use consequences • Statistical analysis: Sensitivity, specificity | Based on the development sample, the ASSIST-Drug was 94.1% sensitivity and 98.6% specific for drug use disorders. Based on the validation sample, it was 95.4% sensitive and 87.8% specific | 4 |
Tiet, Leyva [52] | • Mean age (SD) = 62.2 (12.6) years • The majority of participants were males (95.25%) • Whites (53.7%) and Other race (45.05%) | Non-population-based study conducted in the outpatient clinical setting to examine the concurrent diagnostic accuracy of the SoDU in helping to detect marijuana use disorder | SoDU • Specific substance use: Marijuana use • Sample size: n = 1283 • Response rate: Not reported • Survey duration: Feb 2012–April 2014 | Criterion validity (concurrent validity) • Comparison measure: MINI • Statistical analysis: ROC analysis, sensitivity, specificity | The SoDU was 100% sensitive and 87.5% specific. When tested in subgroups of patients varying in age, gender, race/ethnicity, marital status, education level, and PTSD status, the SoDU maintained 100% sensitivity in all subgroups; specificity ranged from 76.26 to 94.34% | 4 |
Tiet, Leyva [53] | Total • Mean age (SD) = 62.2 (12.6) years • The majority of participants were males (95.25%) • Whites (53.7%) and Other race (45.05%) • N = 969 (75.5%) completed higher than high school education Validation • Mean age (SD) = 62.63 (13.01) • The majority of participants were males (95.3%) • Whites (57.5%) • N = 485 (75.8%) completed higher than high school education | Non-population-based study conducted in the outpatient clinical setting to develop and validate the SoDU for diagnostic accuracy by conducting item performance analyses and to examine the performance | SoDU • Specific substance use: Not reported • Sample size: Total: n = 1283, validation: n = 640 • Response rate: Not reported • Survey duration: Feb 2012–April 2014 | Criterion validity (concurrent validity) • Comparison measure: The Mini-International Diagnostic Interview was used as the criterion for DUDs, and the Inventory of Drug Use Consequences was used as the criterion for NCDU • Statistical analysis: ROC analysis, sensitivity, specificity | The screening instrument was 100% sensitive and 93.73% specific for DUDs; when replicated in the second half of the sample, it was 92.31% sensitive and 92.87% specific. The screening instrument was 93.18% sensitive and 96.03% specific for NCDU; when replicated in the second half of the sample, it was 83.17% sensitive and 96.85% specific | 4 |
Tiet, Leyva [5] | • Mean age (SD) = 62.2 (12.6) years • The majority of participants were males (95.25%) • Whites (53.7%) and Other race (45.05%) | Non-population-based study conducted in the outpatient clinical setting to validate brief DAST and to test if the briefer version of the DAST is practical for routine use in primary settings | DAST • Specific substance use: Not reported • Sample size: n = 1283 • Response rate: Not reported • Survey duration: Feb 2012–April 2014 | Criterion validity (concurrent validity) • Comparison measure: Two criterion measures were used: MINI for drug use disorders and the InDUC for negative consequences of drug use, which includes individuals who may or may not meet criteria for a drug use disorder • Statistical analysis: ROC analysis, sensitivity, specificity | The DAST-2 was 97% sensitive and 91% specific for DUDs in the development sample and 95% sensitive and 89% specific in the validation sample. It was highly sensitive and specific for DUD and negative consequences for drug use in subgroups of patients | 4 |
Tiet, Schutte [54] | • Mean age (SD) = 50.59 (9) years • The majority of participants were males (86.4%) • White (50%), African American/Black (31.8%), and Other race (9.9%) | Non-population-based study conducted in the outpatient clinical setting to conduct AUC, sensitivity, specificity, efficiency, PPV, and NPV of the PCL, PC-PTSD, and five abbreviated versions of the PCL in detecting PTSD for patients seeking treatment in substance use disorder specialty treatment | PTSD PCL-C, PTSD PCL-Bliese-4, PTSD-LS-2, PTSD PCL-LS-3, PTSD-PCL-LS-4, PTSD-PCL-LS-6, PC-PTSD • Specific substance use: Not reported • Specific mental health: PTSD • Sample size: n = 242 • Response rate: Not reported • Survey duration: August 2003–December 2004 | Criterion validity (concurrent validity) • Comparison measure: C-DIS-IV • Statistical analysis: AUC, cut-point sensitivity, specificity, PPV, NPV, efficient, test + % | Based on the C-DIS-IV, prevalence of PTSD was found to be 36.7 and 52.9% in the SUD and MH samples, respectively. The PCL, PC-PTSD, and five abbreviated versions of the PCL were found to have adequate psychometric properties for screening patients in SUD (AUC ranged from 0.80 to 0.86) and MH (AUC ranged from 0.77 to 0.80) outpatient treatment settings | 4 |
Westermeyer, Crosby [55] | • Mean age (SD) = 30.3 (10.7) • Males (95.25%) and female (43.1%) | Non-population-based study conducted in the outpatient clinical setting to conduct concurrent validity of the M-SAPS compared with three factors: psychiatric-behavioral problems, social-interpersonal problems, and addiction-dependence symptoms | M-SAPS • Specific substance use: Not reported • Sample size: n = 642 • Response rate: Not reported • Survey duration: Not reported | Construct validity (convergent validity) • Comparison measure: Not reported • Statistical analysis: Factor analysis Criterion validity (concurrent validity) • Comparison measure: Psychiatric rating, social problem rating scales, assistive use scales • Statistical analysis: Pearson correlation coefficients | All six psychiatric rating scales were most strongly correlated with psychiatric-behavior problems. Additionally, all six comparisons with two social problem measures vs MMADST and Axis V coping scales across the three M-SAPS factors were highly significant. More severe psychosocial stressors in the last year and more substance dependence were associated with higher addictive use symptoms scores | 5 |
Wickersham, Azar [56] | • Mean age (SD) = 47.2 (8.3) • The majority of participants were males (70.1%) • White (19.6%), African American/Black (54.6%), and Other race (25.8%) | Non-population-based study and participants were recruited from a novel jail-release program to conduct the initial validation of the RODS among a sample of 97 newly incarcerated, HIV-positive individuals by comparing the MINI as the primary measure of opioid dependence | RODS • Specific substance use: opioid use • Sample size: n = 97 • Response rate: Not reported • Survey duration: 2009–2011 | Criterion validity (concurrent validity) • Comparison measure: MINI • Statistical analysis: Concordance analysis, sensitivity, specificity, PPV, NPV | The RODS showed good-to-strong sensitivity (0.97), specificity (0.76), positive predictive value (0.69), and negative predictive value (0.98), while concordance analysis revealed moderate diagnostic agreement (κ = 0.67) | 4 |
Woicik, Stewart [57]* | Mean age (SD) = 20 (3.1) years | Sub-population-based study and participants were undergraduate but did not specify the setting to conduct analysis of the internal structure of two versions of the SURPS and to conduct concurrent, discriminant, and incremental validity compared to other theoretically relevant personality and drug use criterion measures | SURPS • Specific substance use: Opioid use • Sample size: n = 195 • Response rate: Not reported • Survey duration: Not reported | Construct validity (convergent and discriminant validity) • Comparison measure: Personality scales of NEO-FFI, MAST, PH scores • Statistical analysis: Pearson correlation coefficients Criterion validity (predictive validity) • Comparison measure: Personality scales • Statistical analysis: Hierarchical regression analyses | The first set included the five subscales of the NEO-FFI and accounted for 5% (R = .22, pb.08, ΔR2 = .05) of the variance in MAST scores and 8% of the variance in Ph scores (R = 0.28, pb.01, ΔR2 = .08) | 4 |
• Mean age (SD) = 19.3 (3.1) years • More females (55.13%) than males (44.87%) | Non-population-based study and participants were recruited from Stony Brook University to conduct test–retest reliability and validity with respect to measuring personality vulnerability to reinforcement-specific substance use patterns | SURPS • Specific substance use: Opioid use • Sample size: n = 390 • Response rate: Not reported • Survey duration: Not reported | Construct validity (convergent and discriminant validity) • Comparison measure: DMQ • Statistical analysis: Pearson correlation coefficients | To test the equivalence of item measurements across samples, we applied a more stringent procedure in which all paths (i.e., factor loadings, variances, and covariances) in the model were constrained to be equal for both samples | 4 | |
Zanis, McLellan [58] | • Mean age = 39 • Sex: All participants were males • 91% were African American, 8% Caucasian, and 1% Latino; • Education: Average education was 12.4 years | Non-population-based study conducted in the outpatient clinical setting with a sample of 98 homeless substance users awaiting temporary housing placement and shelter to examine the reliability and validity of the ASI | ASI • Specific substance use: Not reported • Sample size: n = 98 • Response rate: Not reported • Survey duration: Not reported | Construct validity (discriminant validity) • Comparison measure: Compared the ASI composite score and severity ratings to other tests, including MAST, RAB alcohol, RAB drug, BDI, and SCL-90 • Statistical analysis: Correlations between ASI subscale and appropriate comparable test Criterion validity (concurrent validity) • Comparison measure: (1) Assessed by comparing a test measure to a known conceptually similar standard measure at the same point in time; in (2) a separate evaluation of the concurrent validity of the ASI drug scale, authors examined data from 25 subjects that had drug metabolites detected in a urine sample obtained during the first interview and compared this result with their self-reported use of drugs during the 30-day assessment period covered by the ASI interview • Statistical analysis: Correlations | Both composite score and severity rating measures were found to be quite independent with low intercorrelations. Three of the seven ASI composite scores were tested for and found to have moderate concurrent validity: alcohol (r = 0.31 to 0.36), drug (r = 0.46), and psychiatric (r = S3 to 0.66). Composite score interitem correlations were 0.70 or greater in each of the domains except for employment (0.50) and family (0.52) | 5 |
Zanis, McLellan [59] | • Mean age (SD) = 62 (6.95) years • The majority of participants were males (74.2%) • White (43.5%), African American/Black (56.5%) | Non-population-based study conducted in the inpatient clinical setting to examine aspects of reliability, validity, and utility of ASI among individuals with severe and persistent mental illness and concurrent substance abuse disorders | ASI • Specific substance use: Not reported • Sample size: n = 62 • Response rate: Not reported • Survey duration: Not reported | Criterion validity (concurrent validity) • Comparison measure: Compared with the urine sample at the conclusion of the first ASI interview • Statistical analysis: Not reported in detail | The past 30-day self-reported drug use questions of the ASI have poor concurrent validity when compared with urine screens, which showed that the ASI has limited validity | 5 |