Diagnostic reliability of the Semi-structured Assessment for Drug Dependence and Alcoholism (SSADDA)

Abstract

The Semi-structured Assessment for Drug Dependence and Alcoholism (SSADDA) is a diagnostic instrument developed for studies of the genetics of substance use and associated disorders. The SSADDA provides more detailed coverage of specific drug use disorders, particularly cocaine and opioid dependence, than existing psychiatric diagnostic instruments. A computerized version of the SSADDA was developed to permit direct entry of subject responses by the interviewer. This study examines the diagnostic reliability of the SSADDA for substance use disorders and for other DSM-IV disorders that are commonly associated with substance use disorders.

Methods

Two hundred and ninety-three subjects (mean age = 39 yr, 52.2% women) were interviewed twice over a 2-week period in two sub-studies examining the inter-rater (n = 173) or test–retest reliability (n = 120) of the SSADDA. The κ statistic and Yule's Y were used to measure reliability.

Results

The reliability of most substance dependence diagnoses was good to excellent, although the reliability of substance abuse diagnoses was substantially lower. The reliability of the associated psychiatric diagnoses varied from fair to excellent.

Conclusions

The SSADDA yields reliable diagnoses for a variety of psychiatric disorders, including alcohol and drug dependence. Although developed for use in genetic studies, its broad and detailed coverage of disorders and computer-assisted format will allow it to be used in a variety of applications requiring careful diagnostic assessment.

Introduction

The Semi-structured Assessment for Drug Dependence and Alcoholism (SSADDA) was derived from the Semi-structured Assessment for the Genetics of Alcoholism (SSAGA). The SSAGA is a comprehensive psychiatric interview schedule developed for use in the collaborative study on the genetics of alcoholism (COGA). It assesses the physical, psychological, social, and psychiatric manifestations of alcohol abuse and dependence and a variety of psychiatric disorders in adults (Bucholz et al., 1994).

The SSADDA was developed for use in linkage studies of cocaine and opioid dependence, in order to provide detailed coverage of these and other DSM-IV diagnoses that commonly co-occur with these disorders. Cocaine and opioid dependence, which in the SSAGA were evaluated as part of a general section on drug use disorders, are covered in separate sections in the SSADDA. The SSADDA includes detailed questions on the onset and recency of symptoms for these two major drugs of abuse, and adds some additional items specific to these drugs (e.g., assessment of symptoms of cocaine-induced paranoia). In addition, the SSADDA contains sections covering attention deficit hyperactivity disorder (ADHD) and pathological gambling, both of which are theoretically and clinically relevant to drug dependence (Petry, 2002, Wilens, 2004). Finally, a section on environmental covariates considered likely to have an impact on drug and alcohol dependence risk was added.

The SSADDA was first developed as a paper interview and then as a computer-assisted version, allowing direct entry of subject responses by the interviewer. It may be obtained for free at http://genetics.bumc.bu.edu/ssadda. The computerized format includes core features, such as automatic ‘skip-outs,’ a cross-checking function to identify inconsistent responses, a running tabulation of diagnostic criteria, and a check for out-of-range values, features that serve to streamline the interview process and aid in the collection of accurate information. The computerized version of the SSADDA also allows for direct upload of data to a database, thereby eliminating the time-consuming steps of data entry and verification (with their attendant potential for errors) and permitting the ready generation of DSM-IV diagnoses (American Psychiatric Association, 1994) using scoring algorithms.

The SSADDA allows a trained (non-clinician) interviewer to identify a variety of substance use and psychiatric disorders by collecting information on the onset and recency of symptoms and on their severity and duration. A useful feature of the SSAGA, which was retained in the SSADDA, is its assessment of the relationship between alcohol and drug problem clusters and the occurrence of other psychiatric disorders (Bucholz et al., 1994). Information including the dates of occurrence of alcohol and drug use; clustering of problems; periods of abstinence from alcohol and drug use; and dates of depressive episodes, dysthymic periods, manic episodes, and other psychiatric disorders are correlated. This makes it possible to categorize the respondent's history of psychiatric disorders as being either completely independent of substance problems, or including at least some symptoms occurring in temporal association with substance use.

Although the reliability of diagnostic interviews has been shown to vary considerably (Bryant et al., 1992, Burke, 1986, Griffin et al., 1987, Williams et al., 1992, Easton et al., 1997), the SSAGA was found to have good reliability (Bucholz et al., 1994, Bucholz et al., 1995), with values of κ ranging from 0.70 to 0.90 for lifetime DSM-III-R substance dependence diagnoses. The SSAGA also showed good reliability for lifetime depression (κ = 0.65) and antisocial personality disorder (κ = 0.70).

Because substantial changes (i.e., abridgement of the alcohol section, expansion of some drug sections, addition or modification of some sections on psychiatric disorders, and change in format from paper-and-pencil to computer-adapted) were made in the course of adapting the SSADDA from the SSAGA, we evaluated the inter-rater and test–retest reliability of the SSADDA. This was accomplished by conducting within- and cross-center interviews with a diverse sample of subjects, including subjects recruited from substance abuse and psychiatric treatment settings, and subjects recruited from the community.