Antisocial personality disorder

INTRODUCTION

There is a growing realization among researchers that antisocial personality disorder (ASPD) is highly correlated with substance abuse (1-4). Approximately 90% of persons diagnosed with ASPD are substance abusers (5, 6). Studies have also shown the incidence rate of ASPD to be about 40% to 50% in samples of male substance abusers (7). De Leon and Rosenthal (8) reported that about 30% to 40% of all those admitted to a therapeutic community (TC) obtain a diagnosis of ASPD.

Despite substantial literature on the topic, one continued difficulty lies in the diversified definitions of ASPD. There is no accepted universal operational definition for ASPD, although there is general agreement for much of the behavioral criteria used to detect the disorder. Available diagnostic measures reflect different assumptions regarding a diagnosis of ASPD and often depend on the preference of the researchers or the sample population. Yet, the degree to which specific diagnostic instruments differ in their assessment of ASPD and the variation in assessment among different study populations is not known.

Two of the most popular diagnostic scales for assessing personality disorders are the Structured Clinical Interview for the DSM-III-R, Axis II (SCID-II), and the Millon Clinical Multiaxial Inventory, second edition (MCMI-II). This study assessed the degree to which the MCMI-II (a self-report inventory) and the SCID-II (a clinical assessment) differ in their assessment of ASPD among substance abusers.

The next section provides a description of the disorder of ASPD and the criteria used to diagnose ASPD and overviews the SCID-II and the MCMI-II. A description of the study methodology and results and a discussion of the findings are also included.

DIAGNOSING ANTISOCIAL PERSONALITY DISORDER

The term antisocial personality is often used interchangeably with psychopath or sociopath and is connotative of many forms of deviant behavior. Davison and Neale describe those with an antisocial personality as "disruptive individuals whose deep-seated ethical and moral maladjustments frequently bring them into serious conflict with their associates and with society" (9, p. 260). People with ASPD have no inhibitions about committing antisocial acts; they have no fear of impending punishments; they often act without regret or shame; they are unable to learn from experience; and they are impulsive and aggressive (10). This disorder is much more prevalent among males than females, but prevalence is generally the same for whites and blacks (11, 12). Persons with ASPD also have elevated rates of all other diagnoses and psychosocial problems (12, 13).

Antisocial personality disorder is outlined in the Diagnostic and Statistical Manual of Mental Disorders (fourth edition), commonly known as the DSM-IV (14). The DSM-IV is multiaxial, with personality disorders included under Axis II classifications. The essential feature of ASPD is "a pervasive pattern of disregard for, and violation of, the rights of others that begins in childhood or early adolescence and continues into adulthood" (14, p. 645). Disregard for others is evidenced by such specific behaviors as deceitfulness, impulsiveness, and recklessness. Operational criteria, such as criminal activity, violence, job troubles, and marital difficulties were added to improve the reliability of psychiatric diagnosis.

Although social scientists most often use DSM criteria to assess ASPD, many have raised concerns about possible limitations of the DSM. For example, Gerstley and associates (7) contend that a limitation of the DSM diagnosis of ASPD is the overemphasis on observable behavioral criteria instead of underlying personality traits. The authors suggest that basing the diagnosis on behavioral criteria gives too large a role to criminality. Criminal behavior and ASPD have a significant and positive relationship; however, criminal activity is not a necessary condition of the disorder, and many persons diagnosed with ASPD have no record of arrest (12). In response to this type of criticism, the number of adult symptoms in the DSM was increased with the addition of "lack of remorse" (12).

Hare (15) has also argued that the DSM fails to represent the traditional concept of psychopathy and should be modified to better reflect the "Cleckley psychopath syndrome." Cleckley (10) provides one of the most extensive clinical descriptions of the antisocial personality and describes those with antisocial personalities as having an inner emptiness. Cleckley's checklist places much more emphasis on personality traits, such as total self-centeredness and lack of remorse, guilt, or empathy. Hare's research attempted to produce a more objective scale. Hare developed a checklist from a factor analysis of ratings of Cleckley's criteria known as the Psychopathy Checklist (PCL). This scale includes items pertaining to history of social deviance and is rated from case history data and a structured interview.

Structured interviews, semistructured interviews, and self-report inventories were created in the late 1980s and used widely to maximize the reliability and validity of DSM diagnoses (16). The format of these types of interviews and inventories are based on questions that are keyed into each of the diagnostic criteria of the DSM. A true structured interview asks questions in a specific way and leaves little or no room for interviewers to use their own clinical judgment (17). Semistructured interviews also have specified questions, but allow interviewers to use their own skills in eliciting more information, and include more open-ended questions (17). Structured and semistructured interviews decrease the information variance because they require clinicians (or trained psychometricians) to ask the same questions. The development of self-report inventories, on the other hand, was prompted by a need for diagnostic procedures that were less time consuming and less costly and that could accurately diagnose psychiatric disorders (18).

Two of the most popular diagnostic scales for assessing personality disorders are the SCID-II and the MCMI-II. Both of these diagnostic scales are often used with substance-abusing populations and were administered to the current study sample.

Structured Clinical Interview for the DSM-III-R

The SCID was developed in 1985 and is a semistructured interview for making Axis I and Axis II diagnoses using DSM-III-R classifications. The first scale (SCID-I) consists of a present mental state interview that provides differential diagnosis for DSM-III-R Axis I disorders (primarily mood and substance abuse disorders). The second interview (SCID-II) assesses all 12 adult personality disorders outlined in the DSM-III-R. The SCID-II ASPD section consists of a series of scaled questions and is often preceded by a 113-item paper-and-pencil questionnaire. Consistent with the DSM-III-R, the scaled questions and the questionnaire focus predominantly on antisocial behavior in childhood that has continued into adulthood.

There are various clinical and investigative uses for the SCID. It is often used to diagnose past and current mental disorders in a study population. It is administered by a clinician or a trained interviewer, and the populations studied include psychiatric patients, medical patients, subjects in the community, and substance abusers (19). The SCID includes an overview section that allows the subject to describe the development of his or her current illness. It also includes a design that allows researchers to eliminate major diagnostic classes that are irrelevant to their study.

The SCID was developed simultaneously with the revised version of the DSM-III and has been extensively field tested (19). The SCID has undergone several revisions and has gained widespread use, including utilization as a criterion measure in several studies to cross validate a number of other diagnostic instruments and clinical diagnoses (20-23).

Social scientists usually use the kappa coefficient to determine an accurate measure of reliability and validity that assesses the degree of total agreement beyond that expected by chance (24). Kappa ranges between 0 and 1, with 0 indicating agreement no better than chance, and 1 indicating perfect agreement. It has been suggested that coefficients above 0.75 indicate excellent agreement, those between 0.60 and 0.74 indicate good agreement, those between 0.40 and 0.59 indicate fair agreement, and those below 0.40 indicate poor agreement (25). Kappa coefficients are usually only calculated for diagnoses occurring 5% or more of the time within appropriate sample sizes because kappa has high variability with low base rates (26).

The reliability of the SCID has been tested in several clinical and nonclinical populations. Because the SCID is a semistructured interview and requires the judgment of the clinician or trained interviewer, the reliability can vary according to the circumstances in which it is being used (19). However, studies have reported good-to-excellent test-retest reliability of the diagnosis of ASPD generated by SCID-II in psychiatric and substance-abusing populations (18, 27, 28).

Although data on the validity of the SCID vary according to the scale (SCID-I or SCID-II), the specific disorder, and the criterion measure, a SCID-II diagnosis of ASPD in substance-abusing populations often yields the highest agreement with the criterion diagnosis, with kappas ranging from 0.75 to 0.95 (16, 29-31). Data on the validity of the SCID are limited, partly because of the relative newness of the interview.

The Millon Clinical Multiaxial Inventory

The publication of the DSM-III-R criteria for Axis I and Axis II disorders led to the development of a number of self-report instruments designed to assess personality disorders and other clinical syndromes (18). One of the first diagnostic scales measuring ASPD to have a drug abuse scale embodied in it was the Millon Clinical Multiaxial Inventory (MCMI). This self-report inventory consists of 175 true/false items designed to assess basic personality styles, severe personality disorders, and clinical syndromes. The MCMI also measures personality characteristics associated with substance abuse (32). The MCMI was developed by Millon to make operational the pathologies of personality that he presented in Modern Psychopathology (33). The initial scale was similar to the DSM-III distinction between Axis I and Axis II disorders. Millon published a revised version (MCMI-II) in 1987 to maximize the similarity between the MCMI scales and DSMI-II-R criteria (18, 34). The latest version of the MCMI (MCMI-III) was developed in 1994 in coordination with the DSM-IV (35). Although the MCMI assesses antisocial behaviors in childhood and adulthood, the majority of the questions are designed to assess pathological personality traits.

The first 10 scales of the MCMI are designed to detect Basic Personality Patterns (schizoid, avoidant, dependent, histrionic, narcissistic, antisocial, compulsive, passive-aggressive, aggressive/sadistic, and self-defeating). There are also three Pathological Personality Patterns (schizotypal, borderline, and paranoid). There are nine Clinical Syndrome Scales (anxiety, somatoform, hypomania, dysthymia, alcohol abuse, drug abuse, psychotic thinking, psychotic depression, and psychotic delusions). The MCMI-II and MCMI-III include an additional scale that measures subjects' response tendencies (the internal validity scale), and MCMI profiles are generally computer scored by National Computer Systems (36). Raw scores are convened into base rate (BR) scores "which are transformed scores based on known prevalence rates of the disorders and syndromes" (36, p. 231). MCMI-II diagnostic assignments usually begin with a BR score exceeding 70. Scores of 75-84 suggest a chronic and moderately severe level of personality functioning, while BR scores of 85 and above are often indicative of a more decompensated personality pattern.

While the MCMI was intended to be used with psychiatric patient populations, it is now being utilized in a variety of research populations. The MCMI-II has recently been used with substance-abusing populations to identify a substance abuse problem, to assess the personality of the known abuser, or to predict treatment response (37). The reliability and validity of this test have been the subject of several reviews (32, 37-41).

Craig (38) presents a selected review of the empirical literature pertaining to test-retest reliability of the MCMI-I, MCMI-II, and MCMI-III. The author reports median stability estimates for MCMI scales across the three versions of the test. Reliability scores for a diagnosis of ASPD were as follows: 0.60 for the MCMI-I, 0.91 for the MCMI-II, and 0.93 for the MCMI-III. Craig concluded from his review that there is gradual improvement across the three generations of the MCMI in test-retest reliability, and that correlations for ASPD are generally reliable.

Craig and Weinberg (37) conducted an extensive review of studies that used the MCMI with substance abusers to assess the validity of the self-report interviews. The authors concluded that, among substance-abusing individuals, personality disorder prevalence rates are similar to those derived by structured psychiatric interviews. Two other studies reported that use of the MCMI among substance-abusing populations appears to have acceptable research validity with regard to the diagnosis of ASPD (36, 41). According to Craig and Weinberg, this test "has become a popular instrument for the assessment of personality, and for an objective assessment of DSM-III-R personality disorders and clinical syndromes" (p. 249).

The current study is the first to compare the self-report inventory (the MCMI-II) to a semistructured clinical interview (the SCID-II) for a diagnosis of ASPD among substance abusers in a TC population. However, a previous study compared the diagnoses by the two scales of personality disorders in a sample of 55 psychiatric outpatients (18). The authors found low agreement for a diagnosis of ASPD between the two scales after determining predictive power ratings, collectively termed the "operating characteristics" of a test. A positive predictive power rating (i.e., the ratio of true-positive results to all positive results) of 0.29 was reported. The SCID-II diagnosed 4% of the sample with ASPD and the MCMI-II diagnosed 13% of the sample with ASPD. In fact, 8 of the 12 MCMI-II scales yielded prevalence rates in excess of their SCID-II counterparts. The authors reported that two other objective inventories (the Personality Disorder Questionnaire and the MMPI Personality Disorder Scale), both designed specifically for the differential diagnosis of DSM personality disorders, also yielded prevalence rates of ASPD in excess of the SCID-II.

In contrast, there was perfect agreement (1.0) between the MCMI-II and the SCID-II for no ASPD (i.e., negative predictive power). Thus, a negative test result was usually an accurate indication that the client did not have the disorder. The authors also reported an overall diagnostic power rating of 0.91 (i.e., composite of both true ASPD positive test results and true ASPD negative test results). Yet, overall agreement can be misleading as an indication of validity because the high rate of true negatives will inflate the overall value. Conclusions are tentative due to the very low prevalence rates of ASPD within this sample, which decrease positive predictive power and increase overall diagnostic power. These findings suggest that determining the operating characteristics of a test may not be the most useful measures of an instrument's diagnostic efficiency. It was also suggested that the limitations of the various DSM diagnostic scales reflect the actual limitations of the DSM-III-R, such as the overemphasis on behavioral criteria on the Axis II scale (42).

Millon (43) has argued that the MCMI is an operational measure of his biopsychosocial theory of personality pathology, and that one should not expect perfect correlations between MCMI Personality Disorder Scales and similar measures of DSM personality disorders. Although Millon has revised his test to be more compatible with DSM-IV criteria, there will still be differences due to the additional assessment by the MCMI of psychopathological personality traits, as well as antisocial behaviors (44).

The higher prevalence rates of ASPD generated by the MCMI-II compared to the SCID-II in psychiatric patient populations from the prior literature suggests that these two measures are far from identical. The different types of information collected by the two scales (pathological personality traits versus antisocial behaviors) may generate low agreement on a diagnosis of ASPD. Thus, we hypothesized that there will be minimal agreement between the MCMI-II and the SCID-II diagnosis of ASPD in this substance-abusing population due to the primary focus of the SCID on behavioral criteria ([kappa] < 0.40). We also tried to determine if there are any client characteristics or behaviors related to the agreement between the MCMI-II and SCID-II diagnosis of ASPD.

METHOD

Data came from the District of Columbia Treatment Initiative (DCI) study (described below). The next section describes the subsample of DCI clients for the present study analyses, followed by the study procedures.

District of Columbia Treatment Initiative

The DCI experiment was designed to conduct a drug abuse treatment outcome study that assessed the value of providing enhanced treatment services (45). The treatment programs included two inpatient TCs managed by Second Genesis, Incorporated, a private nonprofit group that has administered TC treatment for the past 25 years. The TCs accepted clients from February 1992 through January 31, 1994. The two treatment facilities were called "standard inpatient" or "abbreviated inpatient" programs. The standard inpatient program was designed to reflect TC treatment that was customarily available in the United States and consisted of approximately 10 months of inpatient care followed by 2 months of outpatient care. The abbreviated inpatient program provided 6 months of inpatient care, 6 months of outpatient care, and a wide range of extra services (46).

Subjects

A total of 412 clients was randomly assigned to the standard or abbreviated inpatient programs. This study focuses on a subsample of 314 DCI clients who were administered the SCID-II and the MCMI-II at treatment admission. Of these 314 clients, 39 cases were scored as invalid via the computer-generated scoring system for the MCMI-II and were subsequently excluded, leaving a final subsample of 275 clients.

Clients ranged in age from 20 to 54 years, with a mean age of 32.8 years. Approximately 72% of the sample were male, and the majority were black (98%). Clients completed an average of 11 years of education, and 67% had never been married. Of the sample, 92% had a history of prior arrest, with an average of 8.6 adult arrests prior to treatment admission. This was primarily a cocaine-abusing sample; 47% of the clients were diagnosed with the use of cocaine/crack as their primary drug disorder, and 46% had problems with both cocaine and heroin.

Procedure

People who sought treatment at the Central Intake Division (CID) run by the D.C. Alcohol and Drug Abuse Services Administration (ADASA) or who were ordered by the court to obtain treatment were eligible to volunteer to participate in the DCI. Clients were also recruited by the Second Genesis Treatment staff from Building 12 of the D.C. General Hospital; from the D.C. ADASA 24-hour walk-in substance abuse clinic; from the ADASA Criminal Justice Division, in which those involved with the criminal justice system could enter treatment; from the 7-day drug detoxification ward at the D.C. General Hospital; from a community resource center in the Marshall Heights Section of the District of Columbia; from waiting lists for Second Genesis treatment programs (those voluntarily waiting and those mandated by a court to receive inpatient treatment); and from correctional facilities in the District of Columbia.

Extensive personal interview and psychometric data were collected from each respondent at intake. The Diagnostic Unit staff pretested early versions of the intake interview (the Individual Assessment Profile, IAP) and conducted the reliability/validity lAP study prior to collection of this sample. The IAP data provide detailed information concerning all facets of a client's life, including medical condition, employment, drug use, alcohol use, illegal activity, family relations, psychiatric condition and severity, sexual practices, and human immunodeficiency virus (HIV) transmission risk behaviors (47).

Immediately after the IAP interview, clients received the Reading Comprehension Subtest of the Peabody Individual Achievement Test-Revised, which measured the client's reading grade level. Clients who were found to be only marginally literate were not asked to proceed with any written psychological tests (45), but were eligible for treatment (20% read below the sixth grade level) (48). Those who had an appropriate reading grade level were administered a battery of psychological tests, which included the Beck Depression Inventory, the Brief Symptom Inventory, the State-Trait Anger Expression Inventory, the Trail Making Test, the MCMI-II, and the SCID-I and SCID-II.

Initial testing for the MCMI-II did not begin until 2 weeks postadmission to allow for detoxification and stabilization of mental status. For this study, ASPD was diagnosed with a BR score of 75 or greater. Base rate scores of 75-84 suggest a chronic and moderately severe level of personality functioning. While BR scores of 85 and above are often indicative of a more decompensated personality pattern. Cases were excluded from the comparisons (n = 39) if the Validity scale from the MCMI-II rated a score as "invalid." A score of 1 on the Validity scale is usually scored "valid with caution"; this means that the subject claimed one absurd item to be true about himself/herself (49). Scores above 1 are an indication that all other results are uninterpretable or random choices, and the user is encouraged to disregard all of the findings for that subject (49).

Clients were administered the SCID 3 weeks after admission. The Diagnostic Unit staff trained treatment facility psychometricians to administer the psychological test. The SCID was administered in a single setting; it takes 75 minutes to complete. The SCID-II was used in conjunction with the SCID-I and was preceded by a 113-item paper-and-pencil questionnaire.

RESULTS

SCID-II and MCMI-II Prevalence of Antisocial Personality Disorder

Both the SCID-II and the MCMI-II were administered to 275 DCI clients at treatment admission. The SCID-II diagnosed 47% of the target sample with ASPD and 53% with no ASPD, while the MCMI-II diagnosed 68% with ASPD and 32% with no ASPD. Clients with "no disorders" and those with "other disorders excluding ASPD" are combined into the "no ASPD" groups. The SCID-II diagnosed 61 clients with disorders other than ASPD and 84 clients with no disorder. The MCMI-II diagnosed 77 clients with disorders other than ASPD and 11 clients with no disorders. The degree of agreement between the MCMI-II and the SCID-II diagnosis of ASPD was assessed via a chi-square analysis and the kappa coefficient and is presented in the next section.

Agreement Between MCMI-II and SCID-II Diagnosis of Antisocial Personality Disorder

Our hypothesis was that there will be minimal agreement between the MCMI-II and the SCID-II diagnosis of ASPD due to the primary focus of the SCID on behavioral criteria ([kappa] < 0.40). This hypothesis was supported. Table 1 presents the degree of agreement between an MCMI-II diagnosis of ASPD and an SCID-II diagnosis of ASPD. The chi-square test of independence shows that the two scales are significantly related (p < .01); however, the kappa statistic shows that the amount of agreement is small (0.27). Cells on the diagonal contain the concordant responses. Off-diagonal cells contain the nonconcordant responses. The two scales only agreed on a diagnosis of ASPD on 39% of the cases. Moreover, the agreement on no occurrence of ASPD was even less evident, 24% of the cases. Of the overall scores, 63% from the two scales match. The SCID-II diagnosed only 57% of the cases diagnosed as ASPD by the MCMI-II, while the MCMI-II diagnosed 82% of the cases diagnosed as ASPD by the SCID-II. Findings indicate that the two scales are far from identical, with the MCMI-II diagnosing ASPD more often than the SCID-II counterpart.

The research question was whether any client characteristics or behaviors were related to the agreement between the MCMI-II and the SCID-II diagnosis of ASPD. Descriptive analyses were conducted (chi-square correlations and t tests) to explore client characteristics or behaviors that might be related to the agreement between the two scales on a diagnosis of ASPD (see Table 2). Self-report and official record information that were examined in these bivariate analyses included age at treatment admission, gender, education, age of first arrest, age first used alcohol or marijuana, criminal justice status (e.g., parole/probation, in jail, or on bail), total prior adult arrests, multiple drug dependencies, and primary drug disorder. Race/ethnicity was not examined because 98% of the sample was black.

Clients the MCMI-II and the SCID-II diagnosed as having ASPD (n = 107) were compared with those for whom the two scales disagreed on the diagnosis of ASPD (SCID-II only ASPD n = 23 and MCMI-II only ASPD n = 80) and with those for whom the scales agreed there was no ASPD (n = 65).

Clients who both scales agreed had ASPD were youngest at admission to treatment (mean age 31.2 years, p < .01), had less education (mean 10.6 years, p < .01), and were youngest at first arrest (mean age 18.0 years, p < .01) compared with the other groups. Age of first arrest ever appeared to be high for all groups; however, there were large standard deviations. The mean age of first arrest for the entire sample was 20.4 years with a standard deviation of 7.1 years (age of self-reported first arrest ranged from 8 to 45 years). Clients who both scales agreed had ASPD also were youngest the first time they used alcohol (mean age 13.5 years, p < .05) and marijuana (mean age 14.6 years, p < .05) compared with the other groups. The "agree-had-ASPD" group also had the highest average BR score on the MCMI-II (mean BR 98.7), followed by the "MCMI-II-only ASPD" group (mean BR 91.9). The "SCID-II-only ASPD" group, however, generated an average BR score similar to the "agree-no-ASPD" group (66.0 vs. 64.4). This finding suggests that the MCMI-II does not distinguish between the SCID-II ASPD group and the no ASPD group.

Significant differences were also found for criminal justice supervision status of the clients, number of prior adult arrests, and multiple drug dependencies. However, the SCID-II-only ASPD group was most likely to be associated with these behaviors, followed by the agree-had-ASPD group. Of the ASCID-II-only ASPD group, 63%were under some form of supervision at admission compared with 79% of the agree-had-ASPD group, 52% of the MCMI-II-only ASPD group, and 69% of the agree-no-ASPD group (p < .01). The SCID-II-only ASPD group also had the highest average number of prior adult arrests (mean = 11.5) compared with 9.6 for the agree-had-ASPD group, 6.7 for the MCMI-II-only ASPD group, and 8.4 for the agree-no-ASPD group (p < .05). And, the SCID-II-only ASPD group had the highest average number of drug dependencies (mean 3.7, p < .01) compared with the other groups. No differences among the four groups were found for gender, past use of needles, or primary drug diagnosis at admission. However, the MCMI-II-only ASPD group was the only group that was most likely to have cocaine dependency (60%) rather than use of cocaine with heroin.

These comparisons indicated that client characteristics of those for whom the two scales agree had ASPD are consistent with the more deviant and earlier behavior that is associated with a diagnosis of ASPD. The SCID-II-only ASPD group also consistently manifested the behaviors associated with this disorder. The MCMI-II-only ASPD group appears to be most similar to the agree-no-ASPD group with regard to these behaviors. It is apparent from these findings that the entire sample had extensive criminal and substance-abusing histories throughout the late teen and early adult years. However, it is the existence of childhood antisocial behaviors that should truly differentiate the clients with ASPD from the clients without ASPD. Unfortunately, the data do not allow for these comparisons. Furthermore, the existence of dysfunctional personality traits (e.g., lack of empathy, guilt, and extreme self-centeredness) are also not revealed by the data. The large BR scores for the MCMI-II-only ASPD clients compared with the ASCID-II-only ASPD clients may be an indication of the different types of information collected by the two scales. The MCMI-II includes questions about antisocial behaviors and pathological personality traits, while the SCID-II focuses mainly on behavioral characteristics.

DISCUSSION

It was hypothesized that there would be minimal agreement (beyond that expected by chance) between the MCMI-II and the SCID-II on a diagnosis of ASPD in this substance-abusing population. This hypothesis was supported; a small degree of agreement ([kappa] = 0.27) was found between the two scales, with the MCMI-II diagnosing ASPD in 31% more cases than the SCID-II counterpart. The strength of the kappa statistic is often dependent on the prevalence rates of the disorders examined; however, given the overall high prevalence of ASPD in this sample, it is clear that these two scales are far from identical. This finding is consistent with those of Guthrie and Mobley (18) and others, who previously found that the MCMI-II yielded higher prevalence rates of ASPD in psychiatric populations than the SCID-II.

The low diagnostic agreement between the MCMI-II and the SCID-II may be due to the different types of information collected by the two scales. It has been suggested that the SCID-II overemphasizes observable behavioral characteristics, giving too large a role to criminality in the diagnosis of ASPD (7), whereas the MCMI-II focuses on capturing pathological personality traits, as well as antisocial behaviors (34). In fact, the MCMI-II BR score for clients diagnosed as having ASPD by the SCID-II only was 66.0, which was well under the 75 or greater cutoff used in this study to qualify for a MCMI-II diagnosis of ASPD and was similar to the mean BR score of clients diagnosed as not having ASPD by both scales (mean BR = 64.4). This finding indicates that, on average, the MCMI-II scale did not distinguish between the clients diagnosed as having ASPD by the SCID-II only and clients diagnosed as not having ASPD by both scales.

One might have predicted that the SCID-II would overdiagnose ASPD in this highly criminal population because of the focus of the SCID-II on deviant behaviors. Yet, the SCID-II diagnosed less ASPD in this population than the MCMI-II. It is possible that clients with extensive criminal backgrounds underreport criminal activity in response to the direct nature of the SCID-II questions (50). The MCMI-II may avoid this problem to some degree because the questions assessing pathological personality traits are more indirect in nature and not as obvious to the respondent. The fact that the SCID-II diagnosis of ASPD requires the existence of childhood antisocial behaviors that continue into adulthood also serves to distinguish between those with and without the disorder in highly criminal populations.

Client behaviors of those the two scales agreed had ASPD exhibited the early deviant behaviors that are hypothesized to be indicative of this disorder (e.g., average age first used alcohol 13.5 years and first used marijuana 14.6 years). Clients diagnosed with ASPD by the SCID-II only, however, most often manifested the adult criminal behaviors (e.g., criminal justice supervision, number of prior adult arrests, and multiple drug dependencies) associated with the disorder. In contrast, clients diagnosed with ASPD by the MCMI-II only were very similar to those without a diagnosis of ASPD with regard to criminal and substance-abusing histories. In fact, those diagnosed by both the SCID-II and the MCMI-II as not having ASPD often had criminal behavior in excess of the clients diagnosed with ASPD by the MCMI-II only. This finding could indicate that behavioral criteria alone may not be an accurate reflection of ASPD in substance-abusing populations in which the majority of clients have extensive criminal histories. Clients diagnosed with ASPD by the MCMI-II are probably more likely than the other groups to manifest pathological personality traits such as self-centeredness and lack of remorse, empathy, or guilt. Unfortunately, this possibility could not be examined with the available data.

Absent an accepted criterion for ASPD, it is difficult to conclude that the discrepancies between the two scales result from flaws in one measure or the other. Yet, the possibility that the SCID-II ASPD client sample fails to represent the traditional concept of psychopathy cannot be ruled out. Many researchers have relied on the SCID diagnostic scale as a criterion measure to cross validate other diagnostic scales; however, this may not be the most accurate diagnostic instrument for all populations. It is possible that the focus of the MCMI-II on pathological personality traits more accurately diagnoses ASPD in substance-abusing populations with extensive criminal histories. In fact, these are the types of persons coming into TC populations every day, especially with the renewed interest in legal coercion to treatment for substance-abusing offenders (51).

Some limitations of this study should be kept in mind when interpreting the results. First, because 98% of the sample are black clients, examining the racial or ethnic correlates of the SCID-II and MCMI-II diagnosis was not possible. Thus, the generalizability of the findings regarding diagnosis and prevalence rates of ASPD to more ethnically diverse substance-abusing populations is limited. Second, this study population was a highly homogeneous sample. Both clients with and without ASPD had substantial histories of drug abuse and crime, making distinctions regarding diagnosis between the two groups very difficult. The majority of this sample had histories of alcohol and drug abuse since adolescence, multiple drug dependencies at treatment admission, and eight or more prior arrests on average. Third, clients without ASPD may have been diagnosed with various other Axis I and Axis II disorders. It is difficult to know the degree to which the existence of other disorders confounded the distinction between those with and without ASPD.

In conclusion, researchers asking similar questions about the association of ASPD to treatment outcomes, but using different diagnostic scales, may obtain conflicting answers. It is difficult to draw definite conclusions regarding the proper measurement of ASPD in substance-abusing samples without additional empirical evidence. In light of the high prevalence rates of ASPD in substance-abusing populations, future research should continue to explore the many issues surrounding the diagnosis of ASPD, as well as its relationship to treatment outcomes.

ACKNOWLEDGMENT

The DCI was a cooperative agreement among the D.C. Alcohol and Drug Abuse Services Administration (ADASA); the National Institute on Drug Abuse (NIDA); Caliber Associates by contract; the Center for Substance Abuse Treatment (CSAT); Koba Associates, Incorporated, in collaboration with the Research Triangle Institute (RTI); the Institute for Behavior Resources (IBR); and Second Genesis, Incorporated.

We are grateful to the staff and clients at Second Genesis for their participation. We also appreciate the assistance of Jerome Jaffee, Barry Brown, Herman Diesenhaus, Gary Palsgrove, John Carver, Samuel Karson, Robert Gesumaria, and the Addiction Prevention and Recovery Administration (APRA). Finally, special thanks are due to the interviewers who conducted the follow-up interviews, who worked day and night to locate the clients. Without the cooperation of all of these parties, this study would not have been possible.

REFERENCES

(1.) Abram, K.M. The Effect of Co-occurring Disorders on Criminal Career: Interaction of Antisocial Personality, Alcoholism, and Drug Disorders. Int. J. Law Psychiatry 1989, 12, 133-148.

(2.) Harry, B.; Steadman, H.J. Arrest Rates of Patients Treated at a Community Health Center. Hospital Community Psychiatry 1988, 39, 862-866.

(3.) Hodgins, S.; Cote, G. The Criminality of Mentally Disordered Offenders. Crim. Justice Behav. 1993, 20 (2), 115-129.

(4.) Teplin, L.A. Psychiatric and Substance Abuse Disorders Among Male Urban Jail Detainees. Am. J. Public Health 1994, 84 (2), 290-293.

(5.) Forrest, G.G. Chemical Dependency and Antisocial Personality Disorder: Psychotherapy and Assessment Strategies; Hawthorne Press: New York, 1991.

(6.) Tims, F.M.; De Leon, G.; Jainchill, N., Eds. Therapeutic Community: Advances in Research and Application. NIDA Res. Monogr. 1986, 144.

(7.) Gerstley, L.J.; Alterman, A.I.; McLellan, A.T.; Woody, G.E. Antisocial Personality Disorder in Patients with Substance Abuse Disorders: A Problematic Diagnosis? Am. J. Psychiatry 1990, 147 (2), 173-178.

(8.) De Leon, G.; Rosenthal, M.S. Treatment in Residential Therapeutic Communities. In Treatments of Psychiatric Disorders; Karasu, T., Vol. Ed.; American Psychiatric Press: Washington, DC, 1989; Vol. 2, 1380-1398.

(9.) Davison, G.C.; Neale, J.M. Abnormal Psychology, 5th Ed.; Wiley and Sons: New York, 1990.

(10.) Cleckley, H. The Mask of Sanity, 5th Ed.; Mosby: St. Louis, MO, 1976.

(11.) Nigg, J.T.; Goldsmith, H.H. Genetics of Personality Disorders: Perspectives from Personality and Psychopathology Research. Psychol. Bull. 1994, 115 (3), 346-380.

(12.) Robins, L.N.; Tipp, J.; Przybeck, T. Antisocial Personality. In Psychiatric Disorders in America: Antisocial Personality; Robins, L., Regier, D., Eds.; Free Press: New York, 1991; 258-290.

(13.) Brooner, R.; Kidorf, M.; King, V.; Bigelow, G. Preliminary Results from a Controlled Clinical Trial of Behavior Therapy for Antisocial Opioid Abusers. NIDA Res. Monogr. 1996, 162, 323.

(14.) American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th Ed.; American Psychiatric Association: Washington, DC, 1994.

(15.) Hare, R.D. Psychopathy and Violence. In Violence and the Violent Individual; Hays, J.R., Roberts, T.K., Solway, K.S., Eds.; Jamaica, NY, 1981.

(16.) Kranzler, H.; Kadden, R.; Babor, T.; Tennen, H.; Rounsaville, B. Validity of the SCID in Substance Abuse Patients. Addiction 1996, 91 (6), 859-868.

(17.) Lesser, I. Cultural Considerations Using the Structured Clinical Interview for DSM-III for Mood and Anxiety Disorders Assessment. J. Psychopathol. Behav. Assess. 1997, 19 (2), 149-160.

(18.) Guthrie, P.C.; Mobley, B.D. A Comparison of the Differential Diagnostic Efficiency of Three Personality Disorder Inventories. J. Clin. Psychol. 1994, 50 (4), 656-665.

(19.) Spitzer, R.L.; Williams, J.B.; Gibbon, M.; First, M.B. User's Guide for the Structured Clinical Interview for DSM-III-R; American Psychiatric Press, Inc.: Washington, DC, 1990.

(20.) Albanese, M.J.; Bartel, R.L.; Bruno, R.F.; Morgenbesser, M.W.; Schatzberg, A.F. Comparison of Measures Used to Determine Substance Abuse in a Inpatient Psychiatric Population. Am. J. Psychiatry 1994, 151 (7), 1077-1078.

(21.) Hyler, S.; Skodol, A.; Kellman, H.; Oldham, J.; Rosnick, L. Validity of the Personality Diagnostic Questionnaire--Revised: Comparison with Two Structured Interviews. Am. J. Psychiatry 1990, 147, 1043-1048.

(22.) Malow, R., West, J.; Williams, J., Sutker, P. Personality Disorders Classification and Symptoms in Cocaine and Opioid Addicts. J. Consult. Clin. Psychol. 1989, 57, 765-767.

(23.) Ross, H., Swinson, R.; Larkin, E.; Doumani, S. Diagnosing Comorbidity in Substance Abusers Computer Assessment and Clinical Validation. J. Nerv. Ment. Dis. 1994, 182 (10), 556-563.

(24.) Cohen, J. A Coefficient of Agreement for Nominal Scales. Educ. Psychol. Meas. 1960, 20, 37-46.

(25.) Cicchetti, D.; Sparrow, S. Developing Criteria for Establishing Interrater Reliability of Specific Items: Applications to the Assessment of Adaptive Behavior. Am. J. Ment. Defic. 1981, 86, 127-137.

(26.) Perry, J. Problems and Considerations in the Valid Assessment of Personality Disorders. Am. J. Psychiatry 1992, 149, 1645-1653.

(27.) Ross, H.; Swinson, R.; Doumani, S.; Larkin, E. Diagnosing Comorbidity in Substance Abusers: A Comparison of the Test-Retest Reliability of Two Interviews. Am. J. Drug Alcohol Abuse 1995, 21 (2), 167-185.

(28.) Spitzer, R.; Williams, J.; Gibbon, M.; First, M. Instruction Manual for the Structured Clinical Interview for DSM-III-R (SCID, 5/1/89 Revision); Biometric Research Department New York State Psychiatric Institute: New York, 1989.

(29.) Kranzler, H.; Rounsaville, B.; Tennen, H. Validity of the SCID in Substance Abuse Patients. NIDA Res. Monogr. 1995, 2, 35.

(30.) Kranzler, H.; Tennen, H.; Babor, T.; Kadden, R.; Rounsaville, B. Validity of the Longitudinal, Expert, All Data Procedure for Psychiatric Diagnosis in Patients with Psychoactive Substance Use Disorders. Drug Alcohol Depend. 1997, 45, 93-104.

(31.) Skodol, A.E.; Rosnick, L.; Kellman, D.; Oldham, J.M.; Hyler, S.E. Validating Structured DSM-III-R Personality Disorder Assessments with Longitudinal Data. Am. J. Psychiatry 1988, 145, 1297-1299.

(32.) Calsyn, D.A.; Saxon, A.J.; Daisy, F. Validity of the MCMI Drug Abuse Scale with Drug Abusing and Psychiatric Samples. J. Clin. Psychol. 1990, 46 (2), 244-246.

(33.) Millon, T. Modern Psychopathology: A Biosocial Approach to Maladaptive Learning and Functioning; Saunders: Philadelphia, 1969.

(34.) Millon, T. Manual for the Millon Clinical Multiaxial Inventory-II, 2nd Ed.; National Computer Systems, Inc.: Minneapolis, MN, 1987.

(35.) Millon, T., Ed. The Millon Inventories: Clinical and Personality Assessment; Guilford Press: New York, 1997.

(36.) Craig, R.J.; Olson, R.E. MCMI Comparisons of Cocaine Abusers and Heroin Addicts. J. Clin. Psychol. 1990, 46 (2), 230-237.

(37.) Craig, R.J.; Weinberg, D. Assessing Drug Abusers with the Millon Clinical Multiaxial Inventory: A Review. J. Subst. Abuse Treat. 1992, 9, 249-255.

(38.) Craig, R.J. A Selected Review of the MCMI Empirical Literature. In The Millon Inventories: Clinical and Personality Assessment; Millon, T., Ed.; Guilford Press: New York, 1997; 303-326.

(39.) Craig, R.J. Sensitivity of MCMI-III Scales T (Drugs) and B (Alcohol) in Detecting Substance Abuse. Subst. Use Misuse 1997, 32 (10), 1385-1393.

(40.) Lesswing, N.J.; Dougherty, R.J. Psychopathology in Alcohol-Dependent and Cocaine-Dependent Patients: A Comparison of Findings from Psychological Testing. J. Subst. Abuse Treat. 1993, 10 (1), 53-57.

(41.) McMahon, R.C.; Davidson, R.S.; Gersh, D.; Flynn, P. A Comparison of Continuous and Episodic Drinkers Using the MCMI, MMPI, and A1CEVAL-R. J. Clin. Psychol. 1991, 47 (1), 148-159.

(42.) Mellsop, G.; Varghese, F.; Joshua, S.; Hicks, A. The Reliability of Axis II of DSM-III-R. Am. J. Psychiatry 1982, 139, 1360-1361.

(43.) Millon, T. The MCMI Provides a Good Assessment of DSM-III Disorders: The MCMI-II Will Prove Even Better. J. Pers. Assess. 1985, 49, 379-391.

(44.) Millon, T.; Davis, R. Putting Humpty Dumpty Together Again: Using the MCMI in Psychological Assessment. In Integrative Assessment of Adult Personality; Beautler, L., Berren, M., Eds.; Guilford Press: New York, 1995; 240-279.

(45.) Hoffman, J.; Schneider, S.; Koman, J.; Flynn, P.; Luckey, J.; Cooley, P.; Wish, E.; Diesenhaus, H. The Centralized Intake Model for Drug Abuse Treatment: The Role of Computerized Data Management. Computers Human Behav. 1995, 11 (2), 215-222.

(46.) Nemes, S.; Wish, E.; Messina, N. Comparing the Impact of Standard and Abbreviated Treatment in a Therapeutic Community: Findings from the District of Columbia Treatment Initiative Experiment. J. Subst. Abuse Treat. 1999, 17 (4), 339-347.

(47.) Flynn, P.; Hubbard, R.; Luckey, J.; Phillips, C.; Fountain, D.; Hoffman, J.; Koman, J. The Individual Assessment Profile (IAP): Standardizing the Assessment of Substance Abusers. J. Subst. Abuse Treat. 1995, 12, 213-221.

(48.) Karson, S.; Gesumaria, R. Program Description and Outcome of a Enhanced, Six Months Residential Therapeutic Community. In Community as Method: Therapeutic Communities for Special Populations and Special Settings; De Leon, G., Vol. Ed.; Greenwood Publishing Group, Inc.: Westport, CT, 1997.

(49.) Thompson, C. Personality Disorders. In The Instruments of Psychiatric Research; Thompson, C., Ed.; John Wiley and Sons: New York, 1989.

(50.) Renneberg, B.; Chambless, D.; Dowdall, D.; Fauerback, J.; Gracely, E. The Structured Clinical Interview for DSM-III-R, Axis II and the Millon Clinical Multiaxial Inventory: A Concurrent Validity Study of Personality Disorders Among Anxious Outpatients. J. Pers. Disorders 1992, 6 (2), 117-124.

(51.) De Leon, G. Legal Pressure in Therapeutic Communities. J. Drug Issues 1988, 18 (4), 625-640.

Nena Messina, (1) Eric Wish, (2), * Jeffrey Hoffman, (3) and Susanna Nemes (3)

(1) Bureau of Governmental Research, Center for Applied Policy Studies, and (2) Center for Substance Abuse Research, University of Maryland, College Park, Maryland (3) Danya International, Incorporated, Silver Spring, Maryland

* Corresponding author. Center for Substance Abuse Research (CESAR), 4321 Hartwick Road, Suite 501, College Park, MD 20740. Fax: (301) 403-8342; E-mail: Ewish@CESAR. umd.edu

COPYRIGHT 2001 Marcel Dekker, Inc.
COPYRIGHT 2001 Gale Group