Self-Medication Hypothesis of Substance Use: Testing Khantzian's Updated Theory[Dagger]

Aug 11, 04:00 AM

Current Headlines: By Hall, Danny H Queener, John E

Abstract- Substance use research has been a salient focus for mental health professionals in recent years. Several organizations, including the American Psychological Association, have been pressing for more substance use research, particularly clinically relevant, theory-based investigations. However, there are few theories of substance use, and even fewer with scientific support. One theory is the Self Medication Hypothesis by Khantzian (1977, 1974), a theory with 30 years of research. However, recent modifications in Khantzian's theory (1999) have not been properly tested. Specifically, two areas require further investigation: Khantzian's belief that more negative affect should be related directly to more substance use, and expanding the number of affective states examined, including alexithymia, to better operationalize Khantzian's belief that several painful and ambiguous forms of affect may be implicated in the self-medicating process. The current study assessed anxiety, depression, hostility and alexithymia levels in 70 methadone maintenance treatment patients. Results indicated that affective measures did not have the expected relationship with reported substance use. The authors advocate for the exploration of multiple factors, not merely emotional regulation, in the variability of substance use.

Keywords-addiction, affect, alexithymia, self-medication

Illicit substance and alcohol use has been a major public health concern in many countries (Inaba, Cohen & Holstein 1997). Currently in the United States, the estimated total cost of alcohol and other drug use is more than $240 billion yearly (Martin 2001). Treatment for these concerns has been a primary focus in the social sciences literature for many reasons, but primarily because it is an important health issue. Indeed, the Robert Wood Johnson Foundation, a national nonprofit organization focused on the health care of all Americans, has called substance use "the nation's number one health problem" (Martin 2001; Robert Woods Johnson Foundation 1993).

Despite the national interest in this health problem, there are few theories explaining how addiction originates, is maintained, and how an individual's seemingly illogical and self-destructive behavior can be adaptive and preferable even when faced with overwhelming negative outcomes. One theory that attempted to explain the phenomenon of substance use is the self-medication hypothesis of substance use (Khantzian 1999, 1974), which originally focused on the issue of comorbid substance use and mental health disorders.

SELF-MEDICATION HYPOTHESIS (SMH)

Based in psychodynamic theory, the SMH posits that a person's inability to tolerate strong affect is a primary motivator for substance use. Simply put, substance users use their drug of choice as an external way to modify their emotions. These individuals have an inability to tolerate strong negative affect, or "affect deficit," that may be due to a lack of interna) coping mechanisms (Khantzian 1 977). According to Khantzian, anger is a negative feeling that could be related to the use of illicit substances, especially opioids. Two other negative affective states, depression and anxiety, have been explored in relation to substance use and the SMH (Khantzian 1999). Another affective construct empirically related to substance use and theorized about by Khantzian is alexithymia (Taylor 1997). Alexithymia is defined as an inability to identify or describe to others one's emotional state; thus, feelings are confusing and ambiguous.

In 1999, looking back at nearly 30 years of research on this theory, Khantzian (1999: 246) stated, ". . . the SMH has received wide recognition and acceptance in general-interest publications, and in the professional literature." This recognition was in part due to the empirical research that he and others have done. In one such study, Khantzian and Treece ( 1 985) found in their opioid- dependent sample that 60% of the individuals also met criteria for an Axis II diagnosis. Nearly one third of the participants were diagnosed with antisocial personality disorder, a diagnosis related to feelings of anger, irritability and aggressive behavior. Finally, 93% of the sample had some sort of comorbid disorder, suggesting that, according to the authors, the comorbid disorder may have been causing the substance use disorder. Additional research cited by Khantzian as evidence for his theory includes a study by Rounsaville, Weissman, Kleber and Wilber (1982a). They found similar results indicating nearly 90% of 533 opioid dependent participants had a psychiatric diagnosis in their history. Similarly, Rounsaville and colleagues ( 1 982b) found that a diagnosis of depressive disorder was significantly correlated with continued use of illicit drugs six months into treatment (r = .85, ? < .05). Based on these data and data drawn from other sources (i.e., Khantzian 1977), Rounsaville and colleagues (1982b: 155) endorsed the notion that ". . . depressed addicts use opiates as an attempt at self-treatment for intolerable, dysphoric affects."

Prior incarnations of the SMH theory focused on psychiatric conditions as the causal agent in substance use. These previously mentioned studies demonstrated support for the SMH with significant correlations between substance use variables and mental illness diagnoses. However, recently there has been a change in the theory that has not been tested. Khantzian's (1999: 255) later writings advocate against the use of diagnostic categories as measures of affective problems stating "... I repeatedly emphasize in this chapter, it is not so much a psychiatric condition that one self- medicates, but a wide range of subjective symptoms and states of distress that may or may not be associated with a psychiatric disorder." However, the empirical literature has not tested this modification to Khantzian's theory.

The current study addressed the change in the central tenet of SMH by measuring the relations between substance use and affective states differently than had been done in previous studies- by using continuous instead of categorical measures for negative affect. Using Khantzian's revised theory (1999), a "wide range of subjective symptoms and states of distress" were examined with a substance- using sample. Specifically, two areas were investigated: (1) increasing the number of affective states examined, to better operationalize Khantzian's wide range of affective states fueling an individuals' substance use and (2) focusing on the subjective symptoms of individuals rather than utilizing a mental illness diagnosis as an indicator of negative emotion.

METHOD

Participants

Participants were recruited from a methadone maintenance treatment (MMT) program at a Midwestern substance use treatment facility, and they provided informed consent to the study. AU procedures were reviewed and approved by the University of Akron's institutional review board. Although this was a convenience sample, the sample is similar to the opioid-dependent individuals on which Khantzian based his initial SMH theory and research (1974). Participation in this study was voluntary and participants were assured of their confidentiality. Because of the restrictions associated with the self-report paper-pencil measures, participants were required to have at least a sixth grade education and to be able to comprehend written English.

Participants were asked to rate their affective experiences and also indicate recent drug use in a self-report paper-pencil measure. There were two general hypotheses: ( 1 ) recent affect states, such as anxiety, depression, alexithymia and hostility, would be related positively to drug use severity, and (2) affect states, such as anxiety, depression, alexithymia and hostility, would be related positively to alcohol use severity.

A power analysis was conducted to determine the number of participants necessary for analyses. For this investigation, power was set at .80, alpha was set at .05, and a moderate effect size (Cohen 1977) was assumed. The analysis indicated that 70 participants were needed for this investigation.

The sample was composed of 58.6% women (n = 41), and 4 1 .4% men (n = 29) . The racial composition was 88.6% Caucasian (n = 62) and 1 1 .4% African American (n = 8). The mean age of the participants in this study was 41.21 (SD = 9.65 years). Of the sample, 32.9% (n = 23) reported completing some high school education, 34.3% (n = 24) of the sample indicated graduating high school, 21% (n = 15) cited completing some college education, 7.1% (n = 5) reported being a college graduate, and 4.3% (n = 3) indicated they had entered graduate school.

Measures

Depression. The Beck Depression Inventory- Second Edition (BDI- II; Beck, Steer & Brown 19%) was used to assess participants' perception of depression. The BDI-II is a 21 -item measure with a Likert-type rating from zero (not like me) to three (much like me). Examples of items include "I feel sad much of the time" and "I do not expect things to work out for me." The BDI-II has been well utilized and has demonstrated adequate validity and reliability evidence.

Anxiety. The State-Trait Anxiety Inventory (STAI; Spielberger 1983) was used to assess participants' ratings of anxiety. Trait anxiety (T-anxiety) is the stable, personality based anxiety- proneness of an individual. State anxiety (S-anxiety) is the perception of situations as stressful and dangerous. The STAI is a 40-item instrument that uses a Likert-type format ranging from one (not at all) to four (very much so). Examples of items for the state subscale include "I feel calm" and "I feel secure." Examples of items for the trait subscale include "I feel pleasant" and "I feel satisfied with myself." The STAI also has been well utilized and has demonstrated adequate validity and reliability evidence. Hostility. The Aggression Questionnaire (AQ; Buss & Warren 2000), an updated version of the Buss-Durkee Hostility Index (Buss & Durkee 1957) was used to assess participants' hostility. The AQ is a 34-item, Likert- type format measure that has five subscales: physical aggression, verbal aggression, anger, hostility, and indirect hostility.

The AQ was standardized on a sample of 2,138 individuals aged nine to 88 years from a variety of different settings. The Cronbach's alpha estimates for the five subscales ranged from .71 to .88. The internal consistency for the AQ total score was .94. The AQ total score has been advocated as a broad measure of aggression (Buss & Warren 2000). Although research has been done with this measure in substance-using populations, no psychometric studies have been published regarding the reliability and validity of this measure in this population.

Alexithymia. The 20-item Toronto Alexithymia Scale (TAS-20; Bagby, Parker & Taylor 1994) was used to assess the participants' reported inability to identify and express emotions. The TAS-20 consists of three subscales including: difficulty identifying feelings, difficulty describing feelings, and externally oriented thinking. This measure uses a one through five Likert-type scale format with higher scores indicating more alexithymia. Bagby and colleagues' reliability studies have indicated good internal consistency (a = .8 1 ) as well as good test-retest reliability over a three- week i nterval (r = .77) with the TAS-20 total score. The instrument was correlated negatively with the Openness to Experience dimension of the NEO-PI-R. This is consistent with the alexithymia construct and indicated conceptual overlap with the two constructs ". . . assessing deficiencies in emotional awareness and imaginai activity" (Bagby, Parker & Taylor 1994: 62).

Addiction Severity Index - Drug Composite Score and Alcohol Composite Score. The self-administered Addiction Severity Index (ASI- SA; Rosen et al. 2000) has been modeled on the original, interview- based ASI. Item examples include "How many days did you drink alcohol in the past 30 days?" and "In the past 30 days, how many days have you experienced drug problems?" The items are weighted, as per the manual, and a composite score (CS) is created for each problem area. Each item is weighted based on a similar metric to the parent measure. Rosen and colleagues (2000) utilized a patient population of 341 individuals in substance use treatment to explore the properties of the ASI-SA. The sample was skewed in terms of gender, with 99% men. Over half the sample was Caucasian, 25% African American, 9% Latino, 2% Native American and 7% from other ethnic backgrounds. Of primary importance to the current study are the Drug CS and Alcohol CS. These scales are continuous and range from 0.0 to 1.0. Internal consistency estimates for these scales were .77 for the Alcohol CS and .73 for the Drug CS. Validation of the paper-pencil measure was done by correlating the ASI-SA composite score with the parent measure composite score. These measures were given at least two weeks apart. The correlation between the ASI Alcohol CS and the ASI-SA Alcohol CS was .87 (p < .05); the correlation between the ASI Drug CS and the ASI-SA Drug CS was .73 (p < .01). It is interesting to note that the self-report paper-pencil scores were significantly higher than the ASI interview scores for alcohol (p < .05) and drug use (p < .01), suggesting that patients may feel more comfortable, and thus be more honest, via the use of self-administered measures of substance use than an interview method.

Socially Desirable Responding. The Balanced Inventory of Desirable Responding (BIDR; Paulhus 1991) is a 40-item, Likert type instrument ranging from one (not true) to seven (very true). The instrument has been designed to tap into two constructs: self- deceptive enhancement and impression management. Self-deceptive enhancement (SDE) measures nonsituational personality bias where one is attempting to conceal personal weaknesses. Impression management (IM) is the desire to be perceived in a positive light and connotes more of an active manipulation. Because of the possibility of social desirability when asking people to admit recent substance use, the participants' impression management was assessed and treated as a covariate. The BIDR has been well utilized and has demonstrated appropriate validity and reliability evidence.

Procedure

Fliers were placed at the facility from which MMT participants were recruited. Participants obtained a five-dollar gift certificate to a local grocery store for participation in the study. Previous research suggests that money is a trigger for some addicts (O'Brien, Childress & McLellan 1990), thus a gift certificate was thought to be a more appropriate incentive for participation. Participants voluntarily agreed to participate in the study and were provided with a packet of study materials including an informed consent for participation and a demographic information questionnaire; all study measures were arranged in a counter-balanced order. Participants were informed that they must receive their methadone dose for the day prior to partaking in the study. Having all of the participants complete their dose before taking the study would prevent any possible confounds in the data due to possible medication or placebo effects on mood.

Research Hypotheses

The first research hypothesis was that negative and ambiguous affect (i.e., anxiety, depression, alexithymia and hostility) would be positively related to drug use severity, over and above impression management. The second hypothesis was that negative and ambiguous affect (i.e., anxiety, depression, alexithymia and hostility) would be positively related to alcohol use severity, over and above impression management.

Statistical Treatment

Multiple linear regressions (MLR; McNeil, Newman & Kelly 1996) were used to test hypotheses. For these analyses, alpha was set at p = .05. F tests were then used to determine the percentage of variance accounted for in addiction severity measures (i.e., drug use and alcohol use) by the affective measures over and above impression management. MLR helps the researchers to better understand the relative contribution of each affect state to addiction severity in order to test Khantzian's idea that a wide range of affective states is implicated in substance use severity.

RESULTS

Descriptive Data

The means and Cronbach's alpha estimates for the measures are presented in Table 1. There are no norms for the Drug Composite Score or Alcohol Composite Score scales currently available with this self-administered instrument.

Analyses

Multiple linear regressions (MLR; McNeil et al. 19%) were used to test the hypotheses; none of the hypotheses were statistically significant (Tables 2 and 3). The criterion measures demonstrated lower than expected reliability coefficients. The Cronbach's alpha for the Drug Composite Scale was .702, and it was .653 for the Alcohol Composite Scale (see Table 1). Due to concerns with reliability, post hoc analyses were conducted to remove some reliability errors. Two principal component analyses were conducted on each criterion to remove the error variance of each respective scale (Grimm & Yarnold 1995). Then the regression model was calculated on each subsequent factor in the same manner as previously. A Bonferroni correction was used to control for alpha buildup with these posthoc test (alpha = .0125).

Utilizing an eigenvalue of one and scree test, the analysis resulted in five factors for DCS using a vari max rotation. These five factors accounted for 69.9% of the variance in DCS scores. Regression analyses were calculated in the same fashion as the prior analyses with the rotated factors replacing the enterions. When conducting the regressions with the individual factors, only one factor was significantly predicted by the independent variables (Table 4). The first factor was related to the Aggression Questionnaire measures (p = .001). In total, the five subscales of the AQ accounted for 28.4% of the variance in DCS Factor 1, over and above Impression Management, BDI and STAI. The subscale that was most related to DCS Factor 1 was the Physical Aggression subscale (p = .021). Specifically, this result suggests that Physical Aggression is positively related to DCS Factor I in this sample.

A similar analysis was completed with the ACS. A principal component analysis yielded a one-factor solution that accounted for 59. 14% of the variance in ACS scores. When the same predictors were used to analyze the ACS factor score, there were no significant findings.

A second post hoc analysis investigated the possibility of a curvilinear relationship between depressive affect level and reported drug/alcohol use. In this effort, Beck's cut off scores were utilized. Participants who scores totaled 0-13 were placed in the "minimal" depression range. Those who scored in the 14-19 range were in the "mild" depression range. The moderate total scores ranged from 20-28. The "severe" depression total scores range from 29-63. Regression analyses were conducted for each group, testing to see if scores on the BDI-II were significantly related to DCS above IM at the 0.0125 level. A similar analysis was conducted for the ACS. None of the analyses resulted in significant findings. However, the severe depression group (n = 13) was approaching significance at the .0125 level for predicting DCS (p = .024). DISCUSSION

Putting the current findings in the context of Khantzian's SMH is difficult. Simply put, the hypothesized relationship between substance use and negative and/or ambiguous affect was not found. Also surprising in the context of the SMH, the affective measures in this study did not indicate a high level of emotional problems in this substance-abusing sample. Indeed, except for anxiety, this sample did not differ much from the norms documented in the respective manuals.

These results do not support the theory that addicted individuals use in order to medicate psychological symptoms. Khantzian criticized Freud's contention (1955) that substance use was solely a function of id (i.e., obtaining pleasure). Ironically, Khantzian may have made a similar error in stating that substance use is mostly a function of avoiding unpleasant feelings. It is possible that both could be true for different people (Leshner 1999), and even, at various times, for the same individual. Some individuals may use for the pleasure-enhancing properties of the drug. Some may use for the relief from painful feelings, even though there was no support for this contention in this study. Perhaps some could use for both reasons. However, neither Freud's nor Khantzian's theory include both scenarios.

The SMH appears inadequate in explaining much of the variance in substance use in this exploration. Increased participant size in this investigation could have discovered this relation. However, if that relation exists, it would likely be of a small effect size, given that this investigation used a medium effect size and found no significant findings between criterions and predictors. Additionally, this relation may exist only for a select few substance users-for example, there was some evidence in this investigation that severely depressed patients may show a tendency to use drugs when more depressed. However, neither of these concepts fit with Khantzian's theory. That is, Khantzian's theory details a hypothesized large effect size for negative and ambiguous affective states and substance use. It also suggests that the driving force for alcohol and other drug use is a similar process for all addicted individuals-people use as an attempt to cope with their painful affective states. Therefore, the SMH theory would have to either expand to include these other contingencies, be integrated into another theory, or be discarded.

Further, the SMH utilizes a single (i.e., negative affect) or at best a dual (i.e., painful and ambiguous affect) dimensional perspective of substance use. Thus, the theory would need to find a way to incorporate findings of other researchers about the multidimensional nature of substance use. For example, in a methadone maintenance treatment environment it was found that self- reported substance use (i.e., self-medicating) varied by gender, ethnicity, an individual's use of coping mechanisms, motivation for abstinence and beliefs about methadone (Gollnisch 1997). More of concern to the SMH is the finding by Avants, Margolin, and McKee (2000); the authors cited a significant relationship between negative affect measures and continued cocaine use (r = -.15, p < .05). Essentially, participants who reported more negative affect were less likely to use cocaine while in methadone treatment. The authors indicate the possibility that more negative affect during intake to treatment may help motivate individuals to pursue abstinence from cocaine use. Avants and colleagues (2000) also explored the relative importance of a construct titled "addict- identity," which was significantly related to substance use and not mentioned in the SMH.

Drawing on this empirical data, it is suggested that substance use must be conceptualized as a multifactoral phenomenon with influences from the individual, subcultural, and societal levels. Theories postulating that substance use is merely a function of a person wanting to experience positive emotions (e.g., Freud 1955) or avoid painful or ambiguous emotions (e.g., Khantzian 1999), do not seem to adequately explain the multifaceted addictive cycle.

Implications

Examining the findings of previous SMH research, current findings may suggest that how the researcher asks the question (via diagnostic or continuous affect-level variables) may impact the results of the investigation. Despite this, investigators are urged to utilize continuous data when studying this specific theory because this type of data best operationalizes Khantzian's current tenets. The idea of researchers using the best method to operationalize theoretical underpinnings is an important aspect of the current theory-driven research focus in psychology (Tracey & Glidden-Tracey 1999).

It is also important to note that negative affect could be a beneficial motivational factor. Avants and colleagues (2000) reported a statistically significant negative correlation between negative affect and substance use. It is possible that negative feelings as part of an honest appraisal of the destruction substance use has caused could motivate patients to seek help and comply with suggestions given during treatment. Whether affect is positively or negatively related to substance use for an individual patient, emotional exploration to aid insight should be an important aspect of treatment.

Limitations

The current investigation was limited for a number of reasons, including small sample size and limited geographic area. Additionally, the study utilized individuals who were opioid dependent and participating in a harm reduction treatment modality. Therefore, other individuals not receiving this sort of treatment may be very different than the sample outlined here. Also, the data were gathered only from people who volunteered for the study and may or may not have been representative of the patients in treatment at this facility. Finally, the TAS and the ASI-SA had some concerns associated with their use. First, the TAS subscales were not of adequate reliability. Perhaps this is the reason it is difficult to locate psychometric qualities on the measure in this population. The ASI-SA has a similar problem. Currently, there have been no published articles on the subscales or factor structure. Thus, the psychometrics properties of these measures need to be better understood in this population. Future Research

Recent research and theory has suggested that positive and negative affective states are not mutually exclusive (Folkman & Moskowitz 2000). Perhaps an absence of positive affect, not the presence of negative affect, could be saliently related to substance use and should be empirically ascertained. Another possibility is that measuring negative affect may need to occur more proximally to the actual substance use. Questions like "When I feel angry, I want to use heroin," for example, may be a better method of measuring Khantzian's SMH. Conceivably, individuals are impulsively responding to overwhelming affect as they experience it. Future investigations could also explore the possibility of a curvilinear relationship between reported affect and substance use. As considered in the post hoc analysis, the relationship between depressive symptoms and substance use in people with severe depression maybe clinically and statistically significant. Therefore, it is suggested that research be done on the relationship of substance use to affect in populations of patients reporting high aggression, anxiety or alexithymia. Perhaps the relations posited in the SMH are accurate for individuals with elevated affect states, but not for other individuals who are regulating their emotions more effectively. The utility of both self-administered and clinically-administered methods for assessing affect could be used to further test the assumptions of the SMH. Additionally, the use of projective measures to assess the level of depression, anxiety and hostility could be addressed. It is possible that more subtle methods of assessment could glean a better understanding of the relation of affect to substance use.

The concept of craving has long been in the terminology of recovery. Until recently, however, it has been ignored by much of the scientific literature. Recent developments with Functional Magnetic Resonance Imaging (fMRI) technology have demonstrated that cravings are a physiological as well as psychological phenomenon (Franken 2003). The concept of craving is in its early development and there is much controversy about how to define, and thus measure, the construct. Perhaps as these issues crystallize, future investigations could explore the relationship of negative affect to cravings. Based on anecdotal data, Khantzian's theory would state that there is a strong positive relationship between negative feelings and cravings. Given the abilities of the fMRI to measure changes of blood flow in the reward system of the brain, this question can now be investigated with physiological measures.

In closing, the phenomenon of substance use is complicated and multidimensional. Theories that fail to consider that complexity are theories that, in the least, need renovating, if not discarding. Continued explorations of multiple predictors, including cognitive, affective and behavioral, as well as those that account for other outcomes than substance use, are needed in order to build an integrated theory of substance use.

[dagger] This study was supported by grants P50 DA09253 andT32 DA07250 from the National Institute of Drug Abuse. The authors would like to thank the UCSF Writer 'sTask Force as well as Barbara Havassy, Ph.D. and James L. Sorenson. Ph.D. for their consultation on this manuscript.

REFERENCES Avants, S.K. ; Margolin, A. & McKee, S. 2000. A path analysis of cognitive, affective, and behavioral predictors of treatment response in a methadone maintenance program. Journal of Substance Abuse. 11: 215-30.

Bagby, MR. ; Parker, JDA. & Taylor, G.J. 1994. The twenty-item Toronto Alexithymia Scale: Item selection and cross validation of the factor structure. Journal of Psychosomatic Research 38: 23-32.

Beck, AT., Steer, RA. & Brown, GK. 1996. Beck Depression Inventory-Second Edition. Orlando, FL: Psychological Corporation.

Buss, A.H. & Warren, WL. 2000. Aggression Questionnaire: Manual. Los Angeles, CA.: Western Psychological Services.

Buss, A.H. & Durkee, A. 1957. An inventory for assessing different kinds of hostility. Journal of Consulting Psychology 21 : 343-349.

Cohen, J. 1977. Statistical Power Analysis for the Behavioral Science-Second Edition. New York: Academic Press.

Franken, I.H.A. 2003. Drug craving and addiction: Integrating psychological and neuropsychopharmacological approaches. Progress in NeuroPsychopharmacology & Biological Psychiatry 27: 1-17.

Freud, S. 1955 [1905). Three Essays on the Theory of Sexuality. SE, Vol. 7. London: Hogarth Press.

Folkman S. & Moskowitz JT. 2000. Positive affect and the other side of coping. American Psychologist 55: 647-54.

Gollnisch, G. 1997. Multiple predictors of illicit drug use in methadone maintenance clients. Addictive Behaviors 22: 353-66.

Grimm, GL. & Yarnold, PR. 1995. Reading and Understanding Multivariate Statistics. Washington, D.C.: American Psychological Association.

Inaba, D.S.; Cohen, WE. & Holstein, ME. 1997. Uppers, Downers, All Arounders: Physical and Mental Effects of Psychoactive Drugs. Ashland, Oregon: CNS Publications, Inc.

Khantzian, EJ. 1999. Treating Addiction as a Human Process. London: Aronson

Khantzian, EJ. 1977. The ego, the self, and opiate addiction: Theoretical and treatment considerations. International Review of Psychoanalysis 5: 189-99.

Khantzian, EJ. 1974. Opiate addiction: A critique of theory and some implications for treatment. Journal of American Psychotherapy 28: 59-70.

Khantzian, EJ. & Trecce. C. 1985. DSM-IlI psychiatric diagnosis of narcotic addicts: Recent findings. Archives of General Psychiatry 42: 1067-71.

Leshner, A.l. 1999. Science-based views of drug addiction and its treatment. Journal of the American Medical Association 282 (14): 1314-16.

McNeil, K.; Newman, I. & Kelly, FJ. 1996. Testing Research Hypotheses with the General Linear Model. Carbondale, IL: Southern Illinois University.

Martin, S. 200 1 . Monitor On Psychology: Special Issue on Substance Abuse. Washington, DC: American Psychological Association Press.

O'Brien, CP; Childress, AR. & McLellan, AT. 1990. Integrating systematic cue exposure with standard treatment in recovering drug dependent cases. Addictive Behaviors 1 5: 355-65.

Paulhus, D.L. 1991. Measurement and control of response bias. In: J. P. Robinson; RR. Shaver & L.S. Wrightsman'(Eds.) Measures of Personality and Social Psychological Attitudes: Volume One. San Diego: Academic Press.

Robert Wood Johnson Foundation. 1993. Substance Abuse: The Nation's Number One Health Problem. Princeton, NJ: Institute for Health Policy, Brandeis University.

Rosen, CG.; Henson, BR; Finney, JW. & Moos, RH. 2000. Consistency of self-administration and interview-based Addiction Severity Index composite scores. Addiction 95: 419-25.

Rounsaville, B.J.; Weissman, MM; Kleber, H. & Wilber, C. 1982a. Heterogeneity of psychiatric diagnosis in treated opiate addicts. Archives of General Psychiatry 39: 161-66.

Rounsaville, B.J.; Weissman, M.M.; Crits-Christoph, K., Wilber, C. & Kleber, H. 1982b. Diagnosis and symptoms of depression in opiate addicts. Course and relationship to treatment outcome Archives of General Psychiatry 39: 15156.

Spielberger, CD. 1983. State-Trait Anxiety Inventory (Form Y). Redwood City, CA: Mind Garden.

Taylor, G.J. 1997. Substance use disorders. In: G. Taylor; RM. Bagby & J.D. Parker (Eds.) Disorders of Affect Regulation: Alexithymia in Medical and Psychiatric Illness. Cambridge, UK: Cambridge University Press.

Tracey.TJ.G. AGlidden-Tracey.CE. 1999. Integration of theory, research design, measurement, and statistics: Toward a reasoned argument. The Counseling Psychologist 27: 299-324.

Wasserman DA.; Korcha R.; Havassy BE. & Hall SM. 1999. Detection of illicit opioid and cocaine use in methadone maintenance treatment. American Journal of Drug and Alcohol Abuse 25: 561-71 .

Danny H. Hall, Ph.D.* & John E. Queener, Ph.D.**

* NIDA Postdoctoral Fellow, University of California, San Francisco.

** Associate Professor, Co-training Director, University of Akron, Akron, OH.

Please address correspondence and reprint requests to Danny H. Hall, Ph.D., 4646 John R. Street, Detroit, MI 48021. Tel. 313-576- 1000, email: Danny.Hall2@va.gov

Copyright Haight Ashbury Publications Jun 2007

(c) 2007 Journal of Psychoactive Drugs. Provided by ProQuest Information and Learning. All rights Reserved.