Upload
lynhan
View
228
Download
0
Embed Size (px)
Citation preview
Bond UniversityePublications@bond
Humanities & Social Sciences papers Faculty of Humanities and Social Sciences
5-11-2000
Discriminant validity of the Illness BehaviorQuestionnaire and Million Clinical MultiaxialInventory-III in a heteregenous sample ofpsychiatric outpatientsGregory J. BoyleBond University, [email protected]
Loick Le DéanBond University
Follow this and additional works at: http://epublications.bond.edu.au/hss_pubs
Part of the Psychology Commons
This Journal Article is brought to you by the Faculty of Humanities and Social Sciences at ePublications@bond. It has been accepted for inclusion inHumanities & Social Sciences papers by an authorized administrator of ePublications@bond. For more information, please contact Bond University'sRepository Coordinator.
Recommended CitationGregory J. Boyle and Loick Le Déan. (2000) "Discriminant validity of the Illness BehaviorQuestionnaire and Million Clinical Multiaxial Inventory-III in a heteregenous sample of psychiatricoutpatients" Journal of Clinical Psychology, 56 (6), 779-791: ISSN 0021-9762.
http://epublications.bond.edu.au/hss_pubs/588
Discriminant Validity of the Illness Behavior Questionnaire and Millon Clinical Multiaxial
Inventory-Ill in a Heterogeneous Sample of Psychiatric Outpatients
Gregory J. Boyle
Department of Psychology, Bond University
and
Department of Psychiatry, University of Queensland
Australia
and
Loick Le Dean
University of Aston, United Kingdom
___________________________________________________________________________
_ Correspondence concerning this article should be addressed to Gregory J. Boyle, Ph.D.,
Department of Psychology, Bond University, Gold Coast, Queensland, 4229, Australia.
Abstract
The discriminant validity of measures of abnormal illness behaviours and
psychopathology was examined in three samples differing in illness prone- ness: a sample
of young healthy university students. (N = 38), a general community sample (n = 36),
and a sample of clinical psychiatric outpatients (N = 36). Adjustment to illness was
measured using the Illness Behaviour Questionnaire (lBO: Pilowsky & Spence. 1994),
while the Millon Clinical Multiaxial Inventory-Ill (MCMI-III, Millon. 1994) was used to
measure clinical syndromes and personality. MANCOVAs were performed across the
three groups on the lBO and the MCMI-III categories separately. As expected clinical
outpatients obtained significantly higher scores than did nonclinical groups on most of the
lBQ scales, suggesting discernible discriminant validity. However, the lack of discrimination
between groups on several of the MCMI-III scales raises questions about the test validity
of this multidimensional instrument.
Introduction
The test validity of measures of clinical symptoms and personality remains
problematic for both clinicians and personality researchers alike (see Carson, 1991;
Clark, Watson, & Reynolds, 1995; Nelson Gray, 1991; Sher & Trull, 1996).
Perception of changes in moods associated with an illness may play an important
role in implementing adaptive behaviours (Turk & Salovey, 1995). Also,
individuals judge the severity of symptoms against prior experiences and stored
schemata of an illness (Leventhal, 1983; Turk, Holz- man, & Kerns, 1986). Illness
behaviours refer to the ways in which individuals perceive and report symptoms
associated with their conditions. Such a concept has often been used in clinical
practice to distinguish between individuals suffering from a physical ailment and
those whose somatic complaints are mainly attributable to psychological factors
(Pilowsky & Spence, 1994; Turk & Salovey, 1995).
Illness Behaviour Questionnaire
Pilowsky and Spence (1994) designed the illness Behaviour Questionnaire
(IBQ), a self-report inventory that assesses an individual's ideas, affects, and
attributions of clinical symptoms. The IBQ is a 62-item questionnaire with a
dichotomous yes/no response scale that yields scores on seven subscales
measuring an individual's attitudes toward illness. The IBQ is primarily used to
detect abnormal illness behaviours and to identify physical complaints that are
manifestations of psychosomatic disorders. Examples of IBQ items are "Do you
ever think that you have an illness which is punishment for something you have
done wrong in the past?" and "Do you get the feeling that people are not taking
your illness seriously enough when you are sick?" Clinically, the IBQ to some
extent distinguishes between patients with a physical syndrome and those who
have a largely psychosomatic component to their illness behaviours. Several of the
items pertain to psychological attributes, rather than to physical illness behaviours
per se (e.g., "Do you worry or fuss over small details that seem unimportant to
others?" and "Do you often find that you get depressed?"
The IBQ was derived from factor-analytic studies of chronic pain patients (Pilowsky
& Spence, 1994). Recent factor analyses have provided some support for the construct
validity of the seven subscales in the instrument (Pilowsky & Spence, 1994; Zonderman,
Heft, & Costa, 1985). Although some studies have failed to provide unequivocal support
(Main & Waddell, 1987; Stretton, Salovey, & Mayer, 1992), other studies have established
the concurrent validity of the IBQ with various measures of anxiety and depression (Grassi,
Rosti, Albierti, & Marangolo, 1989; Harkins, Price, & Braith, 1989; Pilowsky & Spence,
1994). Pilowsky and Spence reported moderate to high stability coefficients over a 1 to 12
week retest period ranging from .87 to .67, with a mean coefficient of .80. The test manual
also provides correlations between patients' scores and others' perceptions of patients'
responses ranging from .50 and .78 (mean=.63).
Patients suffering from physical complaints without demonstrable organic
pathology (viz., psychiatric outpatients) usually exhibit significantly higher scores on the
IBQ scales labelled Disease Conviction, Affective Disturbance (involving both anxiety
and depression), and Irritability than do patients with somatic complaints where organic
pathology is established (Pilowsky, Smith, & Katsikitis, 1987). More recent studies
(Pilowsky & Katsikitis, 1994; Pilowsky & Spence, 1994) have shown that psychiatric
outpatients also tend to exhibit higher levels on General Hypochondriasis and Affective
Inhibition, than do patients with somatic disorders.
Millon Clinical Multiaxial Inventory-III
The Millon Clinical Multiaxial Inventory-III (MCMI-III; Millon, 1994) is a
175-true/false-item inventory that yields scores on 11 personality patterns, three
severe personality patterns, seven clinical syndromes, and three severe clinical
syndromes. Clinical symptoms are assumed to be transient maladaptive states of the
individual's basic personality pattern that emerge under perceived stressful situations.
Several factor analyses have provided limited support for the purported MCMI
structure (Choca, Shanley, & van Denburg, 1993; Craig & Weinberg, 1993; Lorr,
Retzlaff, & Tarr, 1989; Lorr, 1993; Lorr, Strack, Campbell, & Lamnin, 1990; Retzlaff,
Lorr, Hyer, & Ofman, 1991). However, the claimed factor structure of the instrument
may in part be an artifact produced by item overlap. While Millon asserted that item
overlap is consistent with the clinical model of psychopathology, overlapping items
cause scales to be linearly dependent (Gibertini & Retzlaff, 1988; Lumsden, 1988;
Retzlaff & Gibertini, 1987). Indeed, in clinical settings, Cantrell and Dana (1987), and
Piersma (1986) found only limited support for the use of the instrument as a tool in
the screening of psychopathology
The MCMI-III manual reports overall test-retest reliability coefficients ranging
from .95 to .82, with an average of .90 over a 5 to 14 day retest interval. However, no
long term stability coefficients are provided. The manual reports Cronbach alphas ranging
from .95 (Debasement scale) to .66 (Compulsive scale), with a mean of .83: No fewer
than 20 of the scales have alphas exceeding .80, suggesting the possibility of some
item redundancy in the MCMI-ill subsca1es (see Boyle, 1991; Clark & Watson, 1995;
Cortina, 1993).
Studies have generally supported the concurrent validity of the MCMI scales
(e.g., Craig & Olson, 1992; Dyer, 1994; Gabrys et al., 1988; McCann, 1991; Montag
& Comrey, 1987; Morey & Levine, 1988; Schuler, Snibbe, & Buckwalter, 1994;
Terpylak & Schuerger, 1994). Also, there is some evidence of discriminant validity
(e.g., McMahon, Flynn, & Davidson, 1985; Morey, Blashfield, Webb, & Jewell,
1988).
Using the MCMI, Vollrath, Almes, and Torgersen (1994a) found that
individuals with personality disorders tend to rely excessively on dysfunctional coping
strategies, to discard social support, to use disengagement, and to vent negative
emotions. Similarly, individuals with elevated clinical syndrome scores tend to discard
adaptive emotion- focused coping (e.g., seeking social support), and to use
maladaptive emotion-focused coping strategies such as disengagement (Vollrath et al.,
1994b).
Leaf, Alington, Ellis, DiGiuseppe, and Mass (1992) found that individuals
diagnosed with Schizoid, Avoidant, Dependent, Passive-Aggressive, Schizotypal, and
Borderline personality disorders on the MCMI were characterized by acute clinical
syndromes. Such individuals reported more distress, more negative life events, and
more social problems than individuals without disorders. Leaf et al. corroborated
previous studies in which normal subjects have consistently scored more highly on
the MCMI Narcissistic scale than did psychiatric patients (Dubro, Wetzler, & Kahn,
1988; Reich & Troughton, 1988b; Wetzler, Kahn, Cahn, van Praag, & Asnis, 1990).
As the psychiatric condition improved, the score on the scale increased (Joffe &
Regan, 1988; McMahon, Flynn, & Davidson, 1985; Reich & Troughton, 1988a).
Whyne-Berman and McCann (1995) investigated personality disorders and
defence mechanisms measured by the MCMI. Antisocial personality scores were
significantly related to acting out, Compulsive traits were associated with reaction
formation, and Passive-Aggressive personality scores were linked with displacement.
Paranoid personality tendencies were significantly correlated with projection, while
Dependent personality characteristics suggested reliance on introjection. The MCMI
Dependent scale appears to be important in determining psychological adjustment, as
individuals with elevated Dependent personality were found to have been repeatedly
hospitalized (Overholser, Kabakoff, & Norman, 1989).
Millon claimed that the MCMI-III is not intended for use with nonclinical
populations, and that "base rate scores" are prefaced on the assumption that the
respondent is a member of the "clinical population," broadly defined. However, in
order to assess the utility of the MCMI-III instrument for use with nonclinical
populations, the present study examines the personality and symptom characteristics
of clinical psychiatric outpatients as compared with a nonclinical group of healthy,
young students, and also with a non-clinical sample from the general adult
community.
Hypotheses
The first hypothesis (H1) proposed that the clinical outpatients would
demonstrate significantly more elevated scores than would either of the nonclinical
groups on the IBQ and MCMI-III scales (even though non-patients may tend to score
more highly than patients on some of the MCMI-III personality scales). It was
assumed that a continuum exists ranging from healthy adjustment to psychological
maladjustment. H2 proposed that on the IBQ and MCMI-III scales, a general adult
community group would perform mid- way between samples of clinical outpatients and
healthy university undergraduates, and that scores would differ significantly between
student and community samples. Since the student sample was young and healthy, it
was expected that they would exhibit lower levels of illness, chronic pain, illness
proneness behaviours, and psychological maladjustment than would older individuals,
especially those from lower socioeconomic and educational backgrounds selected
randomly from the adult community at large.
Method
Participants
All data was collected in accord with the Code of Ethics of the World
Medical Association after approval had been obtained from the Bond University
Ethics Review Committee. Informed consent was obtained from the research
participants after the nature of the testing procedures had been explained to them.
The total sample consisted of 110 participants from the Gold Coast area of
Queensland who had responded to the items in the MCMI-III and IBQ
questionnaires. The overall age of respondents (50 men; 60 women) ranged from 19
to 72 years (M = 33.12 years, SD = 12.95 years).
In order to test the utility of the IBQ and MCMI-III instruments, three
approximately equal-sized groups differing on a continuum of illness-proneness
behaviours and psychological maladjustment were administered both instruments.
Individuals comprising the two nonclinical groups and the clinical outpatients'
group all freely chose to fill out the MCMI-III and IBQ instruments.
A college sample consisted of 38 predominantly young, healthy individuals
obtained by randomly surveying undergraduate students enrolled at Bond
University. In general, this student group which comprised 18 males and 20 females
aged ranged from 19 to 45 years (M = 24.53 years, SD = 6.97 years) was expected
to be quite low on physical illness, chronic pain, illness-proneness behaviours and
psychological maladjustment. Although the MCMI-III instrument was designed
primarily to discriminate between different psychiatric syndromes, the present
comparison between outpatients and nonclinical groups was considered a useful test
of the discriminant validity of the MCMI-III and IBQ instruments.
A general community sample consisted of 36 adults who were recruited by
surveying individuals from the general adult community at large, located in the
Gold Coast region of Queensland. Their participation was mainly solicited through
personal contacts of the investigators, and therefore did not represent a truly
random sample. They were not paid for their participation, nor were they screened
explicitly for psychiatric disorders. This sample comprised a much greater diversity
of individuals from differing socioeconomic and educational backgrounds than was
the case with the undergraduate student sample, whose members generally came
from more educated backgrounds. The general community sample, which consisted
of 13 males and 23 females aged between 19 and 72 years (M = 38.86 years, SD =
15.56 years), because of its overall greater age composition was expected to be
somewhat more prone to physical illness, chronic pain, illness behaviours, and
psychological maladjustment than was the university student sample.
The clinical outpatients group (n = 36) comprised individuals referred to
two Gold Coast psychologists for evaluation primarily because of psychosocial
dysfunction that warranted examination by an appropriately trained mental health
practitioner-e.g., employment, relational problems, and/or DSM-IV diagnosable
psychopathology (predominantly Axis I and II disorders). Most of them suffered
from depressive mood, and/or severe personality disorders. Few of the clinical
outpatients exhibited discernible Axis III medical conditions. The outpatients who
chose to participate in the study did so freely after the nature of the testing
requirements had been explained to them. They were informed prior to accepting
therapy that they would be asked to complete a battery of psychological tests, and
that their responses to the IBQ and MCMI-III instruments would be used for
research purposes as well as for psychotherapy. The clinical outpatients sample
consisted of 19 males and 17 females whose ages ranged from 22 to 60 years (M =
36.44 years, SD = 10.16 years).
Results
The overall alpha coefficient for the IBQ was .62, and for the MCMI-III
subscales alphas were as follows: Personality Patterns (.66), Severe Personality
Patterns (.80), Clinical Syndromes (.84), and Severe Clinical Syndromes (.67).
Evidently, the high alpha coefficients for the Severe Personality Patterns and
Clinical Syndromes may suggest the possibility of some item redundancy in these
scales (cf. Boyle, 1991). Means and standard deviations for the undergraduate
students, general community, and clinical outpatient groups on the IBQ and
MCMI-III scales are presented in Tables 1 and 2 (also Figs. 1 and 2), respectively.
A one-way between-groups ANOVA revealed that age differed significantly
across groups, F(2,107) = 16.92, p < .001. Two-tailed t-test comparisons
(Bonferroni corrected) indicated that the student sample was significantly younger
than both the general adult community sample, t(106) = 5.41, p < .05, and the
clinical sample, t(106) = 4.50, p < .05. In view of the significant main effect, age
was treated as a covariate in the separate between-subjects MANCOVAs (cf.
Huberty & Morris, 1989) performed on the IBQ scales and Modifying Indices,
Personality Patterns, Severe Personality Pathologies, Clinical Syndromes, and
Severe Clinical Syndromes of the MCMI-III. No significant gender differences
were apparent between the three samples.
After adjustment for age differences, the combined scores on both the IBQ and
MCMI-III instruments differed significantly between groups. The multivariate
effects (Bonferroni corrected) were: IBQ, F(14,200) = 6.48, p < .05; MCMI-III:
Modifying Indices, F(6,208) = 8.98, p < .05; Personality Patterns, F(22,192) =
3.38, p < .05; Severe Personality Patterns, F(6,208) = 4.82, p = .05; Clinical
Syndromes, F(l4,200) = 6.83, p < .05; and Severe Clinical Syndromes, F(6,208)
= 12.31, p < .05, respectively. For the IBQ instrument, ANOVAs (Bonferroni
corrected) suggested that Disease Conviction, F(2,106) = 32.45, p < .05; Affective
Disturbance, F(1,106) = 23.72, p < .05; and Irritability, F(2,106) = 23.83, p < .05,
contributed to the overall IBQ main effect. Clinical outpatients scored more highly
than both college students and individuals from the community at large on Disease
Conviction (M = 51.81, M = 15.55, and M = 19.22,
Table 1
IBQ Group Means and Standard Deviations
Students
Community
Outpatients
Scale
Mean
(SD)
Mean
(SD)
Mean
(SD)
General Hypochondriasis
15.34
(14.59)
15.94
(19.37)
28.83
(25.74)
Disease Conviction 15.55 (12.45) 19.22 (18.48) 51.81 (30.25)
Psychological vs. Somatic Perception 46.32 (19.79) 44.44 (16.64) 47.78 (20.59)
Affective Inhibition 40.00 (32.22) 49.44 (34.64) 62.78 (34.19)
Affective Disturbance (Dysphoria) 28.42 (32.76) 37.78 (32.70) 76.11 (26.97)
Denial 45.26 (26.58) 40.47 (32.65) 47.78 (30.34)
Irritability 33.79 (20.59) 32.50 (26.08) 61.97 (31.71)
Table 2
MCMI-Ill Group Means and Standard Deviations
Students (N = 38) Community (N = 36) Outpatients (N = 36)
Scale Mean (SD) Mean (SD) Mean (SD)
Modifying Indices
Disclosure 41.42 (17.51) 53.97 (19.50) 61.11 (19.20)
Desirability 74.50 (13.09) 65.72 (17.65) 58.89 (16.15)
Debasement 31.53 (24.04) 48.50 (23.63) 67.75 (17.78)
Personality Patterns
Schizoid 33.55 (23.46) 48.50 (23.41) 60.31 (22.85)
Avoidant 29.89 (26.04) 42.86 (30.24) 57.78 (24.85)
Depressive 21.58 (25.32) 37.31 (31.12) 57.92 (25.70)
Dependent 27.29 (23.06) 46.00 (29.28) 63.31 (26.70)
Histrionic 74.79 (18.78) 6o.64 (24.27) 47.75 (22.57)
Narcissistic 81.29 (20.29) 65.69 (20.60) 54.28 (19.49)
Antisocial 41.37 (21.23) 52.11 (22.53) 47.64 (23.77)
Aggressive (Sadistic) 44.16 (22.49) 51.53 (24.56) 51.00 (23.95)
Compulsive 60.18 (15.67) 55.31 (19.29) 52.56 (18.95)
Passive-Aggressive (Negativistic) 32.61 (22.02) 52.47 (27.08) 56.75 (28.11)
Self-Defeating 27.16 (29.92) 35.58 (32.55) 47.97 (32.12)
Severe Personality Patterns
Schizotypal 29.61 (28.12) 41.94 (27.33) 52.53 (24.74)
Borderline 27.08 (24.96) 41.17 (28.33) 54.81 (25.76)
Paranoid 35.18 (26.77) 49.44 (28.00) 51.67 (25.25)
Clinical Syndromes
Anxiety 34.68 (28.96) 48.42 (30.62) 79.03 (23.76)
Somatoform 16.89 (21.41) 32.75 (29.47) 61.69 (26.98)
Bipolar: Manic 50.61 (23.33) 53.72 (20.29) 55.03 (20.92)
Dysthymia 13.61 (20.79) 32.47 (33.51) 65.53 (24.62)
Alcohol Dependence 36.47 (29.40) 47.64 (27.82) 44.42 (30.27)
Drug Dependence 45.87 (21.50) 50.03 (25.29) 43.83 (24.48)
Post-Traumatic Stress Disorder 21.79 (22.74) 29.81 (26.34) 62.67 (15.92)
Severe Clinical Syndromes
Thought Disorder 28.97 (24.19) 37.97 (28.07) 59.81 (22.07)
Major Depression 12.71 (18.60) 30.06 (31.91) 64.14 (27.92)
Delusional Disorder 27.66 (29.25) 38.00 (29.92) 31.92 (29.42)
50
0
100
90
80
70
"' 60
--Students
_ Community Sample
40
30
20
10
0
GH DC PS AI AD D
IBQ Scales
--- Outpatients
Fig. I. Mean scale profiles on the IBQ. OH: General Hypochondriasis; DC:
Disease Conviction; PS: Psychological versus Somatic Perception; AI: Affective
Inhibition; AD: Affective Disturbance (Dysphoria); D: Denial; I: Irritability.
120
110
100
90
" 80
70 )
---< -- Students
- Community Sample
--Outpatients
60
50
13
MCMI-Scales
Fig. 2. Mean scale profiles on the MCMI-III. X: Disclosure; Y: Desirability; Z: Debasement; 1:
Schizoid; 2A: Avoidant; 2B: Depressive; 3: Dependent; 4: Histrionic; 5: Narcissistic; 6A:
Antisocial; 6B: Aggressive (Sadistic); 7: Compulsive; 8A: Passive-Aggressive (Negativistic); 8B:
Self-Defeating; S: Schizotypal; C: Borderline; P: Paranoid; A: Anxiety; H: Somatoform; N: Bipolar
(Manic); D: Dysthymia; B: Alcohol Dependence; T: Drug Dependence; R: Post-Traumatic Stress
Disorder; SS: Thought Disorder; CC: Major Depression; PP: Delusional Disorder.
respectively); Affective Disturbance (M = 76.11, M = 28.42, and M =
37.78,(respectively); and Irritability (M = 61.97, M = 33.79, and M = 26.08,
respectively).
For the MCMI-Ill instrument, the variable that contributed most to the
Modifying Indices main effect was Debasement, F(2,106) = 27.81, p < .05. Clinical
patients (M =67.75) scored more highly on Debasement than both students (M =
31.53) and community individuals (M = 48.50). The univariate main effect for
Disclosure was F(2,106) = 16.46, p < .05, and for Desirability F(2, 106) = 8.25, p <
.05. Clinical outpatients scored more highly than did students on Disclosure (M =
61.11 and M = 41.42), and lower on Desirability (M = 58.89 and M = 74.50). In
contrast, community individuals scored more highly than students on Debasement and
on Disclosure (M = 53.97), and lower on Desirability (M = 65.72).
For the MCMI-III Personality Patterns significant (Bonferroni corrected) main
effects occurred for the Schizoid, F(2,106) = 12.64, p < .05; Avoidant, F(2,106) =
10.79, p <.05; Depressive, F(2,106) = 16.31, p < .05; Dependent, F(2,106) = 16.05,
p < .05; Histrionic, F(2,106) = 14.14, p < .05; Narcissistic, F(2,106) = 14.59, p <
.05; and Passive-Aggressive, F(2, 106) = 13.71, p < .05, personality scales. Simple
contrasts showed that clinical outpatients scored significantly lower than did the
student sample on Histrionic (M = 47.75 and M = 74.79), Narcissistic (M = 54.28
and M = 81.29), and Compulsive (M = 52.56 and M = 60.18) personality scales,
and significantly higher on most of the remaining personality scales. Community
individuals scored lower than did college students on Histrionic (M = 60.64) and
Narcissistic (M = 65.69) personality scales, and more highly on the remaining scales.
14
Clinical outpatients scored higher than did individuals from the community at large on
Depressive (M = 57.92 and M = 37.31), Dependent (M = 63.31 and M = 46.00),
Avoidant (M = 57.78 and M = 42.86), and Schizoid (M = 60.31 and M = 48.50)
personality scales, and lower on Histrionic (M = 47.75 and M = 60.64) and
Narcissistic (M = 54.28 and M = 65.69) personality scales, respectively.
Two of the Severe Personality scales of the MCMI-Ill contributed
significantly to group differences (Bonferroni corrected). The Borderline personality
scale appeared particularly important, F(2, 106) = 12.63, p < .05, as did the
Schizotypal scale, F(2,106) = 8.82, p < .05. Simple contrasts showed that college
students scored significantly lower than both clinical outpatients and community
individuals on all three scales. Community individuals scored lower than clinical
outpatients on the Borderline personality scale only (M = 41.17 and M = 54.81).
Four out of the seven Clinical Syndromes scales contributed significantly to the
MCMI-III main effect. The univariate effects (Bonferroni corrected) were: Anxiety,
F(2, 106) = 26.43, p < .05; Somatoform Complaints, F(2,106) = 28.67, p < .05;
Dysthymia, F(2,106) = 37.56, p < .05; and Post-Traumatic Stress Disorder, F(2,
106) =
37.56, p < .05, respectively. Clinical outpatients obtained significantly higher scores
than did community individuals and college students on Dysthymia (M = 65.53, M =
37.47, and M = 13.61); Post-Traumatic Stress Disorder (M = 62.67, M = 29.81, and
M = 21.79); Somatoform Complaints (M = 61.69, M = 32.75, and M == 16.89); and
Anxiety (M = 79.03, M = 48.42, and M = 34.68). Individuals from the community at
large scored significantly higher than did students on Somatoform, Bipolar, and
Dysthymia scales.
The significant MCMI-III Severe Clinical Syndrome main effect was mainly
due to differences in Major Depression, F(2,106) = 36.97, p < .05, and Thought
Disorder, F(2,106) = 17.19, p < .05. Clinical outpatients scored more highly than
15
both community members and college students on Major Depression (M = 64.14, M =
30.06, and M = 12.71) and Thought Disorder (M = 59.81, M = 37.97, and M =
28.97).
Discussion
The three groups obtained scores that differed markedly on several of the IBQ
and MCMI- III scales, lending partial support to Hl. Clinical outpatients reported
significantly more elevated scores on MCMI-III scales labelled Debasement,
Schizoid, Avoidant, Depressive, Dependent, Self-Defeating, Schizotypal,
Borderline, Anxiety, Somatoform, Dysthymia, Post-Traumatic Stress Disorder,
Thought Disorder, and Major Depression, than did both nonclinical samples,
suggesting greater emotional and personal difficulties, as would be expected.
Similarly, clinical patients exhibited elevated scores on Disease Conviction,
Affective Disturbance (Dysphoria), Irritability, General Hypochondriasis, and
Affective Inhibition on the IBQ. Apparently, such individuals are convinced that
they suffer from an illness; they exhibit high levels of anxiety, depression, tension,
and sadness; and they report resentment and interpersonal difficulties (cf.
Pilowsky et al., 1987; Pilowsky & Spence, 1994). They also tend to display
exaggerated health concerns, being overly preoccupied with physical symptoms, as
well as difficulties in expressing emotions (Pilowsky & Katsikitis, 1994). Likewise,
Vollrath et al. (1994b) reported that clinical outpatients have less emotional support,
and tend to disengage more from goals, and to focus on and vent more emotions than
nonclinical samples.
The Depressive personality scale of the MCMI-III was sensitive in
differentiating between clinical and nonclinical individuals. Similarly, the
Dependent personality scale was a critical indicator of psychological
maladjustment. Earlier studies have found the Dependent personality scale to be
16
sensitive and specific in the assessment of dependency (Miller, Streiner, &
Parkinson, 1992), and to display good convergent validity (Morey & Levine, 1988;
Torgersen & Almes, 1990), as well as satisfactory concurrent and predictive validity
(Overholser & Freiheit, 1994). The finding that the nonclinical groups obtained
higher scores on narcissistic personality suggests that self-esteem tends to be lower
in a group of mixed psychiatric outpatients than in members of the general adult
population at large.
It is not surprising that these measures would discriminate between outpatient
and nonclinical samples. What is surprising and interesting (and from a practical
point of view raises questions regarding test validity) is the lack of discrimination
between some of the MCMI-III personality scales, clinical syndrome scales, and
the Delusional Disorder scale. In terms of clinical utility, this apparent lack of
discriminant validity may suggest that only some items in these scales can
discriminate adequately, and that some items may be poorly written and may not be
measuring the syndromes they are purported to measure. These questions must
await further research on larger samples than used in the present study, given that item
analyses involving small sample sizes are notoriously unreliable (Boyle, Stankov, &
Cattell, 1995).
Little support for H2 (that the community group would obtain scores midway
between those obtained by the clinical and student sample) was found in relation to
the IBQ subscales. In support of H2 though, virtually all of the MCMI-III
subscales differed significantly between the two nonclinical groups, suggesting
(contrary to Millon's assertion) that this multidimensional instrument may be useful for
assessing nonclinical samples.
However, severe personality pathologies measured by the MCMI-III failed to
discriminate clinical outpatients from the sample of community adults. While in
accord with H2, students differed from the other two groups on each of the scales, only
17
Borderline personality discriminated between the clinical and community adults.
Surprisingly, previous studies have found these scales to have good convergent
validity, and to be quite sensitive to psychological maladjustment (Craig &
Weinberg, 1993; Millon, 1994; Piersma, 1993; Torgersen & Alnres, 1990; Widiger &
Corbitt, 1993). One possibility is that the present study used clinical outpatients rather
than inpatients who might have displayed more severe maladaptive features. However,
the proposition by Millon and Davis (1996) and others (e.g., Clarke, Minas, & Stuart,
1991; .Mayou & Hawton, 1986) that psychiatric morbidity in the community may be
more widespread than expected cannot be discounted.
Only half of the personality patterns and none of the severe personality patterns
differed significantly between the community and clinical groups, while half of the
clinical syndromes and two of the three severe clinical syndromes exhibited
corresponding significant differences. Consequently, the differential sensitivity of the
personality and symptoms subscales of the MCMI-III instrument may be deficient.
Summary and Conclusions
The profile of clinical outpatients on the IBQ was consistent with previous
studies which found that psychiatric patients tend to report more abnormal illness
behaviour than non-clinical groups. The IBQ discriminated efficiently between the
clinical sample and non-clinical samples. The scales labelled Disease Conviction,
Affective Disturbance, and Irritability appeared very sensitive to the report of
abnormal illness behaviour by clinical outpatients. General Hypochondriasis and
Affective Inhibition also appeared to play a role, although to a lesser extent, in
discriminating clinical from nonclinical individuals.
The profile of clinical outpatients on the MCMI-Ill also seemed fairly
consistent with earlier findings. The Debasement modifying index, the Anxiety,
Somatoform, Dysthymia, and Post-Traumatic Stress Disorder clinical-syndrome scales, as
18
well as the Thought Disorder and Major Depression severe-clinical-syndrome scales
appeared to be particularly sensitive to psychological maladjustment.
Although the present study found limited support for discriminant validity, the
utility of the MCMI-III personality scales remains to be further explored. For the
severe personality scales, only the Borderline scale differentiated clinical outpatients
from either of the nonclinical groups. Similarly, only half of the personality patterns
(and none of the severe personality patterns) discriminated between community adults
and clinical outpatients. Since the present results provide only partial support for the
MCMI-III clinical syndrome scales in the screening of psychopathology, the
discriminative validity of the instrument necessarily remains somewhat uncertain.
Consequently, the MCMI-III in its present form should be regarded more as a
research instrument which has only limited utility for clinical practice and diagnostic
considerations.
References
Boyle, G.J. (1991). Does item homogeneity indicate internal consistency or item
redundancy in psychometric scales? Personality and Individual Differences, 11,
291-294.
Boyle, G.J., Stankov, L., & Cattell, R.B. (1995). Measurement and statistical models in
the study of personality and intelligence. In D.H. Saklofske & M: Zeidner (Eds.),
International hand- book of personality and intelligence. New York: Plenum.
Cantrell, J.D., & Dana, R.H. (1987). Use of the Millon Clinical Multiaxial Inventory
(MCMI) as a screening instrument at a community mental health center. Journal
of Clinical Psychology, 43, 366_:375.
Carson, R.C. (1991). Dilemmas in the pathway of the DSM-IV. Journal of Abnormal
Psychology, 100, 302-307.
19
Chick, D., Martin, S.K., Nevels, R., & Cotton, C.R. (1994). Relationship between
personality disorders and clinical symptoms in psychiatric inpatients as measured
by the Millon Clinical Multiaxial Inventory. Psychological Reports, 74, 331-336.
Choca, J.P., Shanley, L.A., & VanDenburg, E. (1993). Interpretative guide to the
Millon Clinical Multiaxial Inventory. Washington, DC: APA.
Clark, L.A., & Watson, D. (1995). Constructing validity: Basic issues in objective scale
development. Psychological Assessment, 7, 309-319.
Clark, L.A., Watson, D., & Reynolds, S. (1995). Diagnosis and classification of
psychopathology: Challenges to the current system and future directions. Annual
Review of Psychology, 46, 121-153.
Clarke, D.M., Minas, I.H., & Stuart, G.W. (1991). The prevalence of psychiatric
morbidity in general hospital inpatients. Australian and New Zealand Journal of
Psychiatry, 25, 322-329.
Cortina, J.M. (1993). What is coefficient alpha? An examination of theory and
applications. Journal of Applied Psychology, 78, 98-104.
Craig, R.J., & Olson, R.E. (1992). Relationship between MCMI-ll scales and normal
personality traits. Psychological Reports, 71, 699-705.
Craig, R.J., & Weinberg, D. (1993). MCMI: Review of the literature. In R. J. Craig
(Ed.), The Millon ·clinical Multiaxial Inventory: A clinical research information
synthesis (pp. 23-70). Hillsdale, NJ: Erlbaum.
Dubro, A., Wetzler, S., & Kahn, M.A. (1988). A comparison of three tests for Axis II
diagnosis. Journal of Personality Disorders, 2, 256-266.
Dyer, F.J. (1994). Factorial trait variance and response bias in MCMI-ll personality
disorder scale scores. Journal of Personality Disorders, 8, 121-130.
Gabrys, J.B., Utendale, K.A., Schumph, D., Phillips, N., Peters, K., Robertson, G.,
Sherwood, G., O'Haire, T., Allard, I., Clark, M., & Laye, R.C. (1988). Two
20
inventories for the measurement of psychopathology: Dimensions and common
factorial space on Millon's clinical and Eysenck's general personality scales.
Psychological Reports, 62, 591-:601.
Gibertini, M., & Retzlaff, P.D. (1988). Factor invariance of the Millon Clinical
Multiaxia1 Inventory. Journal of Psychopathology and Behavioral Assessment, 10,
65-74.
Grassi, L., Rosti, G., Albierti, G., & Marangolo, M. (1989). Depression and abnormal
illness behaviour in cancer patients. General Hospital Psychiatry, 11, 404-411.
Harkins, S.W., Price, D.D., Braith, J. (1989). Effects of extraversion and neuroticism
on experimental pain, clinical pain, and illness behavior. Pain, 36, 209-218.
Huberty, C.J., & Morris, J.D. (1989). Multivariate analysis versus multiple univariate
analysis. Psychological Bulletin, 105, 302-308. ·
Joffe, R., & Regan, J. (1988). Personality and depression. Journal of Psychiatric
Research, 22, 279-286.
Leaf, R.C., Alington, D.E., Ellis, A., DiGiuseppe, R., & Mass, R. (1992). Personality
disorders, underlying problems, and clinical syndromes. Journal of Personality
Disorders, 6, 134-152. Leventhal, H. (1983). Behavioral medicine: Psychology in
health care. In D. Mechanic (Ed.),
Handbook of health, health care, and the health professions (pp. 709-743). New York:
Free Press.
Lorr, M. (1993). Dimensional structure of the Millon Clinical Multiaxial Inventory. In
R.J. Craig (Ed.), The Millon Clinical Multiaxial Inventory: A clinical research
information synthesis (pp. 81-89). Hillsdale, NJ: Erlbaum.
Lorr, M., Retzlaff, P.D., & Tarr, H.C. (1989). An analysis of the MCMI-I at the item
level. Journal of Clinical Psychology, 45, 884-890.
21
Lorr, M., Strack, S., Campbell, L., & Larnnin, A. (1990). Personality and symptom
dimensions of the MCMI-II: An item factor analysis. Journal of Clinical
Psychology, 46, 749-754.
Lumsden, E.A. (1988). The impact of shared items on the internal-structural validity of
the MCMI. Educational and Psychological Measurement, 48, 669-678.
Main, C.J., & Waddell, G. (1987). Psychometric construction and validity of the
Pilowsky Illness Behaviour Questionnaire in British patients with chronic low
back pain. Pain, 28, 13-25.
Mayou, R., & Hawton, K. (1986). Psychiatric disorder in the general hospital. British
Journal of Psychiatry, 149, 172-190.
McCann, J.T. (1991). Convergent and discriminant validity of the MCMI-11 and MMPI
personality disorder scales. Personality Assessment, 3, 9-18.
McMahon, R.C., Flynn, P.M., & Davidson, R.S. (1985). Stability of the personality
and symptoms scales of the Millon Clinical Multiaxial Inventory. Journal of
Personality Assessment, 49, 231-234.
Miller, H.R., Streiner, D.L., & Parkinson, A. (1992). Maximum likelihood estimates of
the ability of the MMPI and MCMI personality disorder scales and the SIDP to
identify personality disorders. Journal of Personality Assessment, 59, 1-13.
Millon, T. (1994). Millon Clinical Multiaxial Inventory (3rd ed.). Minneapolis, MN:
National Computer Systems.
Millon, T., & Davis, R.D. (1996). An evolutionary theory of personality disorders. In
J.F. Clarkin & M.F. Lenzenweger (Eds.), Major theories of personality disorder
(pp. 221-346). New York: Guilford.
Montag, I., & Cornrey, A.L. (1987). Millon MCMI scales factor analyzed and
correlated with MMPI and CPS scales. Multivariate Behavioral Research, 22, 401-
413.
22
Morey, L.C., Blashfield, R.K., Webb, W.W., & Jewell, J. (1988). MMPI scales for
DSM-ill personality disorders: A preliminary validation study. Journal of Clinical
Psychology, 44, 47-50. Morey, L.C., & Levine, D.J. (1988). A multitrait-
multimethod examination of Minnesota Multiphasic Personality Inventory
(MMPI) and Millon Clinical Multiaxial Inventory (MCMI). Journal of
Psychopathology and Behavioral Assessment, 10, 333-344.
Nelson-Gray, R.O. (1991). DSM-IV: Empirical guidelines from psychometrics. Journal
of Abnormal Psychology, 100, 308-315.
Overholser, J.C., & Freiheit, S.R. (1994). Assessment of interpersonal dependency
using the Millon Clinical Multiaxial Inventory-IT (MCMI-11) and the Depressive
Experience Questionnaire. Personality and Individual Differences, 17, 71-78.
Overholser, J.C., Kabakoff, R., & Norman, W.H. (1989). The assessment of
personality characteristics in depressed and dependent psychiatric inpatients.
Journal of Personality Assessment, 53, 40-50.
Piersma, H.L. (1986). The Millon Clinical Multiaxial Inventory (MCMI) as a
treatment outcome measure for psychiatric inpatients. Journal of Clinical
Psychology, 42, 493-499.
Pilowsky, I., & Katsikitis, M. (1994). A classification of illness behavior in pain clinic
patients. Pain, 57, 91-94.
Pilowsky, I., Smith, Q., & Katsikitis, M. (1987). Illness behavior and general practice
utilization: A prospective study. Journal of Psychosomatic Research, 37, 177-183.
Pilowsky, I., & Spence, N. (1994). Manual for the Illness Behaviour Questionnaire (3rd
ed.). Department of Psychiatry, University of Adelaide.
Reich, J., & Troughton, E. (1988a). Comparison of DSM-ill personality disorders in
recovered depressed and panic disorder patients. Journal of Nervous and Mental
Disease, 176, 300-304.
23
Reich, J., & Troughton, E. (1988b). Frequency of DSM-ill personality disorders in
patients with panic disorder: Comparison with psychiatric and normal control
subjects. Psychiatry Research, 26, 89-100.
Retzlaff, P.D., & Gibertini, M. (1987). Factor structure of the MCMI basic personality
scales and common-item artifact. Journal of Personality Assessment, 51, 588-594.
Retzlaff, P.D., Lorr, M., Hyer, L., & Ofman, P. (1991). An MCMI-II item-level
component analysis: Personality and clinical factors. Journal of Personality
Assessment, 57, 323-334. Schuler, C.E., Snibbe, J.R., & Buckwalter, J.G.
(1994). Validity of the MMPI personality disorder scale (MMPI-PD). Journal
of Clinical Psychology, 50, 220-227.
Sher, K.J., & Trull, T.J. (1996). Methodological issues in psychopathology
research. Annual Review of Psychology, 47, 371-400.
Stretton, M.S:, Salovey, P., & Mayer, J.D. (1992). Assessing·health concerns.
Imagination, Cognition and Personality, 12, 115-137.
Terpylak, 0., & Schuerger, J.M. (1994). Broad factor scales of the 16PF Fifth Edition
and Millon personality disorder scales: A replication. Psychological Reports,
74, 124-126.
Torgersen, S., & Almes, R. (1990). The relationship between the MCMI
personality scales and DSM-III, Axis II. Journal of Personality Assessment,
55, 698-707.
Turk, D.C., Holzman, A.D., & Kerns, R.D. (1986). Chronic pain. In K.A. Holroyd
& T.L. Creer (Eds.), Self-management approaches to the prevention and
treatment of chronic disease (pp. 441- 472). San Diego, CA: Academic Press.
Turk, D.C., & Salovey, P. (1995). Cognitive-behavioral treatment of illness
behavior. In P.M. Nicassio & T.W. Smith (Eds.), Managing chronic illness: A
biopsychosocial perspective (pp. 245- 284). Washington, DC: APA.
24
Vollrath, M., Alnres, R., & Torgersen, S. (1994a). Coping and MCMI-II personality
disorders. Journal of Personality Disorders, 8, 53-63.
Vollrath, M., Alnres, R., & Torgersen, S. (1994b). Coping and MCMI-II symptom
scales. Journal of Clinical Psychology, 50, 727-736. .
Wetzler, S., Kahn, R.S., Cahn, W., van Pniag, H.M., & Asnis, G.M. (1990).
Psychological test characteristics of depressed and panic patients. Psychiatry
Research, 31, 179-192.
Whyne Berman, S.M., & McCann, J.T. (1995). Defence mechanisms and
personality disorders: An empirical test of Millon's theory. Journal of
Personality Assessment, 64, 132-144.
Widiger, T.A., & Corbitt, E.M. (1993). The MCMI-II personality disorder scales
and their relationship to DSM-III-R diagnosis. In R.J. Craig (Ed.), The Millon
Clinical Multiaxial Inventory: A clinical research information synthesis (pp.
181-201). Hillsdale, NJ: Erlbaum.
Zonderman, A.B., Heft, M.W., & Costa, P.T., Jr. (1985). Does the Illness Behaviour
Questionnaire measure abnormal illness behaviour? Health Psychology, 4,
425-436.