A Critical Review of the Substance Abuse Subtle Screening Inventory

alyssa fredricks image3Introduction

The Substance Abuse Subtle Screening Inventory (SASSI) was written by Glen Miller and originally released in 1985 by the SASSI Institute.  It is currently on its third edition, and was most recently revised in 1997 (Fernandez, 2004).  The SASSI-3 is a psychometric instrument that aids in the identification and treatment of individuals who may struggle with substance-use disorders.  According to the SASSI Institute, SASSI-3 starter kits range in price from $125-$165.  A typical starter kit includes the test manual, scoring key, pencil and paper versions of the test, and a user’s guide.  In addition, audio versions of the assessment are also provided for clients’ who may not be able to complete the written version of the test due to reading difficulties.  The publishers require first time users to purchase a starter kit; however, additional copies of the test may be purchased separately.  Currently, the test is available in a variety of different formats including pencil and paper, scantron, software and online versions.  The SASSI Institute also provides a Spanish translation of SASSI-3, as well as a separate assessment (the SASSI-A2) which is specific to identifying substance related disorders in individuals ranging in age from 12-18 years

Discussion

The Substance Abuse Subtle Screening Inventory was developed in response to the need for a brief, easy to administer, objectively scored, accurate, and cost effective screening tool for substance use disorders.  The instrument is designed to identify individuals who have a high probability of having a substance dependence disorder (Lazowski & Miller, 1999).  The test is based on “subtle” or “indirect” approaches to screening for substance dependence problems; therefore, the test assumes that an individual with a substance use disorder will not or cannot acknowledge their problems if asked directly about their substance abuse related behaviors.  As a result, the SASSI-3 is was constructed using an empirical approach to formulate questions that commonly distinguish a non-substance abuse population from a substance abuse population without inquiring specifically about substance abuse (Clements, 2002).  However, the format of the SASSI has been updated to include face valid measures in addition to subtle screening.

The format of the SASSI-3 is divided into 10 subscales.  The SASSI-3 is printed on two sides of a single sheet of paper.  One side consists of 67 true/false questions that should be administered first.  These 67questions are considered to be subtle items, and are subdivided into eight subscales: symptoms of substance abuse (SYM), obvious attributes (OAT), subtle attributes (SAT), defensiveness (DEF), supplemental addictions measure (SAM), family vs. control subjects (FAM), correctional (COR), and random answering pattern (RAP).  The backside of the SASSI-3 is composed of the face valid alcohol (FVA) and face valid other drug (FVOD) scales.  These two scales measure face validity (Clements, 2002).  On average, the test takes approximately 15 minutes to administer and score.  The test was originally designed to be hand-scored but is also available in electronic formats.

The SASSI-3 is an appropriate assessment tool for test users 18 years of age and older.  The test taker should also be performing on a minimum 3.2 grade reading level in order to ensure accurate results and comprehension of the test material.  The test taker should also have a clear understanding of the English language, though an equivalent form of the test is available to Spanish speaking populations.  Additional versions of the SASSI-3 are available on audio formats for test takers who may have a visual impairment or reading difficulties.  The test may be administered individually or in a group.  Finally, the SASSI-3 is an appropriate clinical tool across a variety of different demographics including gender, marital status, and age (Lazowski et al, 1999).

Due to the complexities of properly administering, assessing, and interpreting the SASSI-3, The SASSI Institute has set up specific requirements in accordance with the American Psychology Association for the types of clinicians who are able to use and purchase the test.  According to the SASSI Institute, “the use of the SASSI-3 for clinical purposes is limited to individuals with training and experience in the area of assessment.  However, individuals who do not have professional training can administer and score the instrument if there is proper supervision.”  Therefore, clinicians who are most likely using this type of test include licensed psychologists, psychiatrists or other types of medical doctors, psychiatric nurses, addictions counselors, mental health counselors, rehabilitation counselors, social workers and individuals who are currently undergoing SASSI training under a licensed professional.  In addition, common settings for the administration of the SASSI-3 include substance abuse treatment facilities, hospitals, behavioral health centers, prisons/jails, probation settings, drug court, military, welfare/child protective services, assessment, sex offender treatment facilities and private practices.

Clinical questions that are commonly answered through the administration of the SASSI-3 include, “Does the individual suffer from a substance use disorder?”  Feldstein and Miller (2006) explain, “The usual purpose of the screening is to identify people who warrant more careful evaluation to confirm or disconfirm substance-related disorders.  There has been particular interest in this instrument to circumvent denial and accurately detect substance-use disorders regardless of the respondent’s honesty or awareness.”

The SASSI-3 is an inappropriate tool for clients who are under 18 years of age or read below a 3.2 grade level.  Additionally, individuals who are currently experiencing psychotic symptoms or suffer from severe mental retardation may be inappropriate clients for the SASSI-3.  In contrast, an appropriate client for the SASSI-3 would be over the age of 18 with basic verbal comprehension skills.  The statistical analysis shows consistency across a wide range of demographics information.  Therefore, an appropriate test-taker for the SASSI-3 would have clinically significant accuracy as a function of demographical information across gender, education, treatment setting, ethnicity, and age (Lazowski et al, 1999).

The clinical utility of the SASSI-3 is somewhat questionable since the test manual fails to indicate the clinical utility of the instrument compared to other measures of substance abuse.  However, other studies have reported a moderate to high coefficient alpha when the SASSI-3 was compared to alternative screening instruments (Pittenger, 2004).  Similarly, the SASSI’s interchangeable use of substance dependency, substance abuse, and substance misuse also negatively impacts the clinical utility by not clearly defining the construct being measured.

The clinical sample set for the SASSI-3 was based on a data set of 2015 respondents which were provided by clinicians working in various service settings throughout the United States including addiction treatment centers, general psychiatric hospitals, dual diagnosis hospitals, vocational rehabilitation programs, and sex offender treatment programs.  In addition, prisoners in correctional facilities and research subjects who had a known family history of substance abuse also participated in the study (Lazowski et al, 1999).   Then, a subset group of 839 individuals who had both SASSI scores and DSM diagnoses was created, and further divided in half.  Since the subgroup division was randomized, it was approximated that each sub-sample contained approximately equal of individuals diagnosed with having or not having a substance use disorder (Fernandez, 2004).  Lastly, data from the subgroups was interpreted to make revisions and updates to the scoring procedures, as well as cross-validating the newly derived rules.  However, in assessing the standardization of the SASSI-3, it is important to take into demographic information cross-validation.    Likewise, the demographic characteristics of development and cross-validation samples are strikingly equal.  For example, the normative samples of the development and cross-validation group show similar percentages among vital characteristics such as gender, data source, clinical diagnosis, marital status, employment status, education, ethnic group, and age (Lazowski et al, 1999).

According to the Substance Abuse Subtle Screening Inventory Manual, the SASSI has high test-retest reliability based on a two-week period.  The stability coefficient ranges from .92-1.00 based on ratings for the 10 SASSI subscales.  Therefore, the total coefficient alpha for the entire SASSI-3 was .93 depiction strong internal consistency (Lazowski et al, 1999).  Thus, it can be conclude that the SASSI-3 is evaluating the singular construct of substance abuse effectively.  However, the manual additionally breaks down the coefficient alphas by subsections, which reports much lower coefficient alphas on certain scales.  However, since the coefficient alpha is based on the correlation among items, it can be argued that using a coefficient alpha statistic is not significant, since the SASSI-3 was designed to measure dimensional constructs across a multitude of scales, as opposed to just measuring a singular construct.

For the validity analysis of the SASSI-3, the comparison classifications correspond to the presence or absence of a substance abuse disorder.  Based on the developmental sample, Lazowaki et al (1999) reports, “The SASSI-3 correctly identifies 94.6% of the people who were diagnosed as having a substance use disorder and correctly identifies 93.2% of those who were diagnosed as not having a substance use disorder.”  However, the high percentages of validity within the development group may be contributed to the formulation of grading criteria and decision rules were formulated around this particular subgroup.  Nevertheless, the cross-validation sample group also demonstrated high validity statistics identifying and differentiating clinical substance use diagnoses.  Within the cross-validation sample, The SASSI-3 correctly identifies 93.2% of the people who were diagnosed as having a substance use disorder and correctly identifies 93.2% of those who were diagnosed as not having a substance use disorder.  Therefore, the overall accuracy of the validity statistic when the two subgroups were combined was 93.9% (Lazowaki et al, 1999).

However, though the test manual reports high accuracy, the results should be interpreted with some caution.  First of all, the manual only reports test-retest reliability over a two-week period.  Minimal information is given about the stability of the instrument.  Similarly, the criterion validity may be distorted due to use a sample population with a high baseline of substance abuse.  Would the criterion validity be the same if it was measured in a population with lower baseline for substance use disorders? (Pittenger, 2004).

Overall, the Substance Abuse Subtle Screening Inventory is an effective and useful tool in clinical practice for identifying and treating potential substance-use disorders.  The instrument purports high validity and reliability statistical values.  Additionally, the subtleness of the SASSI is helpful when dealing with clients who may be resistant or deny having a substance use problem due scales, such as the defensiveness scale, designed specifically to identify “fake good” responses and other forms of resistance.  The SASSI-3 is also convenient to use and quick to score.    However, weaknesses of the test include its failure to differentiate between substance dependency, substance abuse and misuse, and limited interpretations of results.   Fernandez (2004) explains, “An even more serious handicap of the SASSI-3 is that it permits only a dichotomous interpretation of high versus low probability of substance dependence.  Have the authors perhaps undersold the instrument by ignoring the possibility that the number of decision rules satisfied might be related to the confidence level in diagnosing a substance dependence disorder?  At least, if statements about medium probability of the disorder were possible, the instrument would be more clinically useful.”  Likewise, nearly 80% of the sample group reported having a diagnosis of substance dependence, therefore, skewing the accuracy of the SASSI-3 (Fernandez, 2004).

In conclusion, though the SASSI-3 is a quick and convenient way to screen for substance use disorders, the results should be individualized and tailored to the client through a full clinical assessment that includes a multitude of different clinical tools.  Though the SASSI manual reports high accuracy, there is still a measurement of error.  The manual is also limited in its statistical analysis and assessed research methods.  Therefore, it is important to consult the DSM-IV-TR criteria for substance use disorders when making a diagnosis in conjunction with the SASSI-3.

References

Clements, R. (2002).  Psychometric properties of the Substance Abuse Subtle Screening    Inventory-3.  Journal of Substance Abuse Treatment, 23, 419-423.

Feldstein, S. W., & Miller, W. R. (2006).  Does subtle screening for substance abuse          work?  A review of the Substance Abuse Subtle Screening Inventory (SASSI).        Addiction, 201 (1), 41-50.

Fernandez, E. (2004).  Review of the Substance Abuse Subtle Screening Inventory-3.       Mental Measurements Yearbook, 15.

Hood & Johnson (2007). Assessment in Counseling (4th Edition). Alexandria,          VA: American Counseling Association

Lazowski, L.E.; & Miller, F.G. (1999).  The adult substance abuse subtle screening            inventory (3rd Ed.) manual.  Springville, IL: The SASSI Institute.

Miller, S. C., Woodson, J., Howell, R. T., & Shields, A. L., (2009).  Assessing the reliability of scores produced by the Substance Abuse Subtle Screening Inventory.     Substance Use and Misuse, 44, 1090-1100.

Pittenger, D.J. (2004).  Review of the Substance Abuse Subtle Screening Inventory-3.      Mental Measurements Yearbook, 15.

This entry was posted in Papers and tagged , , . Bookmark the permalink.

Comments are closed.