Journal of Behavioral Medicine, Vol. 18, No.

6, 1995

Assessing Chronic Illness Representations: The

Implicit Models of Illness Questionnaire
K a t h l e e n M. S c h i a f f i n o 1,2 and C h r i s t i n e D. Cea I
Accepted for publication: June 7, 1995

Data were collected from three samples using the Implicit Models of Illness
Questionnaire (IMIQ) to assess illness representations as described in the
self-regulation model of common sense illness representations. A factor
structure was identified which displayed some similarities to the common sense
model. This structure was used to examine illness representations of students
and patients concerning three illnesses--rheumatoid arthritis (RA), multiple
sclerosis (MS), and human immunodeficiency virus (HIV). Representations
differed across illnesses and respondent status (patient vs. student). Students
rated individuals as having more personal responsibility for RA or MS than
did patients; moreover, the difference between patient~student ratings was
greater with respect to MS than it was for RA. Patients were more aware of
the variable nature of RA and MS symptoms than were students. This study
demonstrates the value of the IMIQ as a tool for assessing ilhTess cognitions
and suggests that illness representations differ as a function of personal
'ea-perience and personal relevance.
KEY WORDS: chronic illness; cognitions; implicit models; questionnaire.

The study presented here was supported by an Arthritis Foundation Traineeship Grant to
the first author and grants from the National Institute of Health (A/VI36679) and the
Professional Staff Congress--CUNY to Tracey A. Revenson and by a grant to the first author
from Fordham University.
1Psychology Department Fordham University Bronx New York 10458.
-To whom correspondence shou d be addressed.

531
fl16fl-7715/95/12fllld1531507.5[)/l)9 1995 Plenum PublishingC~rporatkm
532 Schiaffino and Cea

INTRODUCTION

A sizable body of research has emerged over the past 10 years which
has, as its primary focus, an interest in the way people think about and
react to the negative life event of illness diagnosis. This research has ex-
amined emotional responses to the diagnosis of illness, attributions
concerning the causes of the illness, and assessment of one's ability to cope
with illness (Parker et al., 1988; Schiaffino and Revenson, 1994; Taylor and
Brown, 1988). In this work, the overarching focus has been on the individ-
ual's attempt to make sense of the illness experience (Affieck et al., 1987;
Wallston and Smith, 1992). This search for meaning is influenced by both
characteristics of the individual and characteristics of the illness. The un-
derstanding a person has concerning the nature of an illness influences later
efforts to cope with it. Therefore, an examination of the cognitive repre-
sentations held by individuals about the nature of illness is essential to
understanding the way that appraisals and coping efforts influence adjust-
ment (Hampson et al., 1994).
Efforts to elaborate illness representations have been pursued with re-
gard to acute, chronic, and hypothetical illnesses (Lau and Hartman, 1983;
Meyer et al., 1985; Nerenz and Leventhal, 1983; Turk et al., 1986). The
most commonly employed model has been the self-regulation or common-
sense model of illness representations proposed by Leventhal and his
associates (Nerenz and Leventhal, 1983; Meyers et al., 1985). According to
this model, a person actively constructs a representation of his or her illness
experience that then regulates illness behavior. Thus, one of the ways in
which individuals come to understand illness is through their own illness
experience. It is the illness representation which emerges from this expe-
rience that forms the basis for the person's emotional response, coping
efforts, and compliance with treatment.
The common-sense model of illness representations (Nerenz and
Leventhal, 1983) identifies four components of how individuals cognitively
represent their illness experience. The first, identity, includes the abstract
label associated with an illness (e.g., "malignant") as well as the concrete
symptoms assumed to accompany the illness (e.g., pain or fever). This iden-
tity or label reflects not only the influence of personal experience, but also
media messages, cultural definitions, and input from informal support sys-
tems (Nerenz and Lcventhal, 1983). For example, for many people in this
country AIDS carries with it a set of abstract labels or images which in-
cludes "rapid d e a t h " and "homosexual disease." While this abstract
representation is no longer accurate, it may nevertheless persist for many
people.
illness Representations 533

Illness representations involve beliefs concerning the second compo-

nent, the time line of the illness, that is, whether the illness is thought of
as acute or chronic, changeable or unchangeable, cyclic or otherwise re-
current. In coming to terms with the illness, the patient then arrives at an
understanding concerning the causes of the illness, and its consequences,
the third and fourth components of the model. Lau and Hartman (1983)
proposed a fifth component, cure, as an additional factor in this model.
Early research on the common-sense model of illness representations
has encompassed both acute and chronic illness. Lau and Hartman (1983)
asked healthy college students about the last time they were sick; the majority
reported experiences with acute everyday illnesses such as a common cold
or the flu. Meyer et al. (1985) examined the way in which individuals think
about hypertension, a chronic illness in which the absence of symptoms does
not mean that the disease is gone; patients who believed they could tell when
their blood pressure was elevated complied less well with treatment.
Methodological approaches to operationalizing illness representations
have included content coding of open-ended responses concerning disease
prototypes (Bishop et al., 1987), unstructured interviews with hypertension
patients (Meyer et al., 1985), and open-ended descriptions of illness epi-
sodes by college students participating in a questionnaire study (Lau and
Hartman, 1983).
In 1986, Turk et al. developed a scale intended to measure illness rep-
resentations. The 45-item Implicit Models of Illness Questionnaire (IMIQ)
was intended to assess components such as those of the common-sense
model of illness representations and to contribute to an understanding about
the existence of a generic cognitive structure which might be consistent
across illnesses and populations. In the initial study the IMIQ was admin-
istered to diabetes educators, diabetic patients, and college students. Factor
analysis across these samples revealed four dimensions of illness repre-
sentation: (1) seriousness, which included nine items assessing knowledge
about the degree to which an illness is serious, contagious, caused by germs
or wealther, chronic, permanent, does not go away on its own, has symptoms
of a common cold, and requires medical attention; (2) personal responsibility,
eight items involving beliefs about who or what might be responsible for
disease onset and/or resolution, such as stress, diet, exercise, not enough
rest, and personal behavior; (3) corztroUability, five items about the extent to
which disease is controllable by the individual or outside forces, is disabling
and painful, and comes back; and (4) changeability, two items reflecting
whether symptoms and other aspects of a disease are changeable and change
over time.
534 Schiaffino and Cea

The Present Study

The research presented here examines the utility of the IMIQ as a

measure with which to assess the common-sense model of illness repre-
sentations. The structure of the IMIQ and its applicability to rheumatoid
arthritis (RA), multiple sclerosis (MS), and human immunodeficiency virus
(HIV) are tested in three samples of individuals. These illnesses share some
similarities with, but also differ from, hypertension and diabetes, chronic
illnesses examined with the IMIQ and the common-sense model.
Like the individual with hypertension, the patient with RA, MS, or
HIV is still considered to have the illness even during periods of symptoms
remission. However, patients with mild symptoms or with symptoms re-
sponding to medication may sometimes think of their chronic illness as an
acute illness, not unlike past episodes of flu or chicken pox. It is only after
repeated symptom flare-ups and remissions, or in the face of severe symp-
toms unresponsive to medication, that these individuals realize the
chronicity and potential uncontrollability of the disease.
RA is an autoimmune illness characterized by joint inflammation,
chronic pain, and stiffness (Genest, 1983). MS is an autoimmune illness char-
acterized by muscle weakness and fatigue (Rao, 1989). HIV seropositivity
indicates that one is infected with the virus that causes AIDS, an autoimmune
deficiency syndrome which leaves individuals vulnerable to a variety of life-
threatening infections. All three of these chronic conditions are characterized
by symptomatology that may vary widely. HIV seropositivity differs from RA
and MS insofar as an individual has the ability to protect him- or herself
from infection, for the most part, by not engaging in risk-related behaviors.
The diagnosis of RA or MS, by comparison, is not preventable or predictable.
The primary purpose of the study was to see if (1) theoretically derived
scales reflecting the components of the common-sense model would dem-
onstrate reliability and validity when considered across populations and
illnesses, and (2) factor analysis of IMIQ items would produce either the
four-factor structure proposed by Turk et al. (1986), a five-factor structure
reflecting the components of the common-sense model of illness repre-
sentations, or a structure different from either of these two.

METHOD

Samples

RA Sample. Participants were recruited from the rheumatic disease

clinic of a major metropolitan hospital and from the private practices of
Illness Representations 535

rhcumatologists associated with that hospital. Patients had been diagnosed

with RA for approximately three years when the IMIQ was administered.
The majority of respondents (n = 63) completing the IMIQ were female
(90%), reflecting the fact that women are affected with RA two to three
times more often than men (Schumacher, 1988). Eighty-one percent of the
respondents were White, 62% were married, and 89% had a high-school
education or more. The median income was between $25,000 and $40,000.
The mean age was 53 (SD = 14).
MS Sample. Adult patients with MS were recruited from the MS clinic
of a major metropolitan hospital. Participants were identified from clinic
records as having been diagnosed with MS within the past 3 years. The
majority of the 101 respondents (90%) were female, reflecting the fact that
women are also affected with MS more often than are men. The sample
was predominantly White (92%); 66% were married, and 96% had a high-
school education or more. The median income ranged between $40,000
and $65,000. The mean age of subjects was 42 (SD = 12). This sample of
MS patients was somewhat younger than the sample of RA patients and
was of a somewhat higher socioeconomic status in terms of income and
education.
Student Sample. Seventy-one undergraduate students enrolled in psy-
chology courses at a private university participated in the remaining portion
of this study in partial fulfillment of class requirements.

Measures

Illness representations were assessed using the IMIQ (Turk et al.,

1986). Subject responded to the 45-item measure on a 5-point scale ranging
from strongly disagree (1) to strongly agree (5).

Procedure

The RA and MS patients each completed the IMIQ along with meas-
ures of functional and psychological status. The students completed the
IMIQ three times, once with regard to RA, once with regard to MS, and
once with regard to HIV (missing data from 1 student resulted in a sample
size of 70 for HIV).
It has been suggested that factor analysis is most effective when used
to identify differences among items, rather than consensus among indi-
viduals (Lau et al., 1989). Therefore, the data employed for factor analysis
here included not only patients representations of their own RA or MS
but also student representations of HIV, RA, and MS. HIV was included
536 Schiaffino and Cea

because of its widespread impact on contemporary society and its particu-

lar relevance to a college population. This disease has been associated
with factors of personal responsibility since it was first labeled in the early
1980s (Moulton et al., 1987). At the present time, HIV appears to be one
of the most threatening of all illnesses but one that can be prevented
through individual behaviors. RA and MS represent unpreventable chronic
illnesses. Both of these may be quite debilitating (e.g., resulting in loss of
motor or cognitive functioning) but typically not fatal. In addition, the
progress of either RA or MS appears to be out of the control of the af-
fected individuals.

RESULTS

Means and standard deviations for individual IMIQ items are pre-
sented for each of the five subsamples (63 RA patients, 101 MS patients,
71 RA representations by students, 71 MS representations by students, and
70 HIV representations by students; N = 376) in the Appendix (Table AI).
Data from all five subgroups were employed in analyses.

Theoretically Derived Illness Representation Scales

One way to determine the utility of the IMIQ with respect to the com-
mon-sense model is to categorize IMIQ items according to the five compo-
nents of the common-sense model of illness representations as described by
Leventhal and his associates: identity, time line, causes, consequences, and
curability. IMIQ items were assigned to one of these components only if
they appeared to reflect clearly the construct in question.
Nine IMIQ items were seen as reflecting the abstract labels and con-
crete symptoms associated with the Identity of illness: symptoms including
fever, stomachache, symptoms on skin, symptoms of a common cold, symp-
toms related to nervous system, painful, contagious, serious, and affecting
many parts of the body. Eight IMIQ items addressed the duration or Time
Line of illness: chronic, constant, permanent, changeable, changes over
time, often comes back, symptoms change seasonally, and goes away by
itself. Eight items specifically concerned the possible Cause of illness:
weather, stress, own behavior, something the person did, germs or virus,
lack of rest, poor diet, and no one being responsible. Six IMIQ items were
concerned with the Consequences of the illness: serious consequences, dis-
abling, affected by stress, requiring medical attention and hospitalization,
and terminal. Finally, six IMIQ items concerned Cures for illness: curable
Illness Representations 537

a n d c u r e d by m e d i c a t i o n , by r e d u c e d stress, by p r o p e r eating habits, by

rest, a n d by exercise. Scales were created by s u m m i n g across all items a n d
dividing by the n u m b e r of valid items, resulting in scales with a possible
r a n g e of 1 to 5.
I n t e r n a l r e l i a b i l i t y for the h y p o t h e s i z e d scales was first assessed
across all five s a m p l e s (N = 376). Reliabilities r a n g e d from a low of .24
for the T i m e Line scale, .55 for Cause, .51 for Identity, a n d .73 for C o n -
s e q u e n c e s , to a high of .82 for C u r e s . I t e m - t o t a l c o r r e l a t i o n s i n d i c a t e d
t h a t s e v e r a l i t e m s w e r e c o r r e l a t i n g n e g a t i v e l y with t h e i r r e s p e c t i v e
scales. T h e r e f o r e , the f o l l o w i n g i t e m s were r e c o d e d : in the I d e n t i t y
s c a l e , " s y m p t o m s in t h e n e r v o u s s y s t e m " a n d " s y m p t o m s a f f e c t i n g
m a n y p a r t s of the b o d y " were reversed; for T i m e Line, "goes away o n
its o w n " and " s y m p t o m s are c o n s t a n t " were reversed; and for Causes,
" c a u s e d by g e r m s " and " n o o n e is r e s p o n s i b l e " were reversed. With these
changes, the i n t e r n a l consistency reliability coefficient for T i m e Line in-
c r e a s e d to .40, that for Causes i n c r e a s e d to .60, a n d that for I d e n t i t y
i n c r e a s e d to .65.
I n t e r n a l consistency reliability coefficients were then calculated for
each of the five scales separately for each of the five subsamples. Descrip-
tive statistics and reliability coefficients for the resulting scales are r e p o r t e d
in Table I. Only the scales related to cures and c o n s e q u e n c e s had relatively
strong i n t e r n a l reliability across all five samples.

Table I. Descriptive Statistics and Internal Reliability for Theoretical Scales

Illness & Mean (SD)
sample Label Time Line Cause Cure Consequences
RA
Students 2.74 (.31) 3.42 (.35) 2.56 (.53) 2.41 (.60) 3.34 (.43)
= .43 ~ = .39 ~ = .77 ~ = .80 ~ = .73
Patients 2.68 (.36) 3.70 (.43) 2.34 (.49) 2.17 (.61) 3.40 (.52)
a = .28 a = ,44 ~ = .70 ~ = .68 ~ = .56
MS
Students 2.82 (.47) 3.44 (.37) 2.29 (.41) 1.99 (.66) 3.79 (.46)
= .63 a = .32 a = .42 a = .86 ~ = .66
Patients 2.07 (.39) 3.82 (.46) 2.27 (.52) 1.80 (.62) 3.39 (.58)
= .41 ~ = .42 ~ = .66 ~ = .78 ~ = .63
HIV
Students 3.11 (.44) 3.47 (.32) 2.59 (.42) 1.97 (.66) 3.61 (.51)
cc = .48 a = .06 c~ = .33 c~ = .87 c~ = .78
538 Schiaffino and Cea

Factor Structure

An initial factor analysis was performed which employed only the 21

items included in Turk's factor structure and which requested a four-fac-
tor solution. The structure which emerged was not consistent with the
structure identified by Turk et al. Therefore, an exploratory factor analysis
was conducted utilizing all 45 items of the IMIQ. A four-factor solution
explaining 46% of the variance was identified using varimax rotation and
was confirmed by factor analysis with an oblique rotation allowing for
the possibility of correlation among the factors (see Table II). One item--
symptoms in the nervous system--was dropped because of double load-
ing.
Two factors included 14 IMIQ items each. The first combined causes
of illness (poor diet, stress/nerves, lack of rest, changes in weather) as welt
as possible cures (proper eating, reduced stress, rest, physical exercise, and
medication). This factor, labeled Curability, reflected the belief that the
illness was curable, not permanent, and not chronic.
The 14 items in the second factor were associated with the individual's
Personal Responsibility for the occurrence of the illness. These items in-
cluded beliefs that the illness was related to something the individual did,
could be avoided, was caused by one's behavior, no one is responsible for
onset, is contagious, and terminal, requires hospitalization, is similar to the
common cold, is characterized by fever and symptoms in the stomach and
on the skin, and is caused by germs or virus. The item "no one is respon-
sible" correlated negatively with the other items, thus indicating individual
responsibility for illness prevention.
Ten items pertaining to the variability of symptoms and the course
of illness comprised the Symptom Variability factor: illness controllability,
illness controllable by individual, changeable symptoms, symptoms related
to stress, weather, and rest, and symptoms affecting many parts of the
body.
Six IMIQ items concerned with the severity of illness were included
in the Serious Consequences factor. Items included serious, serious conse-
quences, painful, disabling, constant, and requiring medical attention.
In the obtained factor structure Leventhal's components c a u s e and
c u r e combine to form the Curability factor, along with three t i m e line items
which also relate to curability--not chronic, not permanent, and goes away
on its own. In fact, Curability appears to be consistent with an acute illness
representation. The Personal Responsibility factor which was obtained here
includes causes, consequences, and identity components of the common-
sense model, which all share the characteristic of being related to the
individual's behavior in terms of responsibility for the cause of the illness;
Illness Representations 539

Table II. Factor Structure IMIQ Responses

Factor 2, Factor 3, Factor 4,
Factor 1, Personal Symptom Serious
Curability Responsibility Variability Consequences
(17.8%) (16.3%) (6.5%) (5.4'70)
IMIQ38 .74 .02 .10 .03
IMIQI9 .74 .03 .04 .01
IMIQ27 .74 ,06 .07 -. 10
IMIQ44 .73 -.05 -. 16 -.03
IMIQ24 .73 .06 .09 -.05
IMIQ32 .72 .08 . 11 -.06
IMIQ20 .69 -.06 .14 .10
IMIQ40 .65 .05 .26 .02
IMIQll .64 -.05 .15 -.03
IMIQ8 .58 .13 -.17 -.05
IMIQ39 .57 .07 -. 10 -. 19
IMIQ 17 -.47 .09 .08 .39
IMIQ3 -.47 -.06 .11 .40
IMIQ21 .44 -.15 .11 -.21
IMIQ28 .01 .82 -.05 .04
IMIQ36 .07 -.81 .18 .01
IMIQ33 -.02 .80 -. 18 .02
IMIQ22 .18 .80 -.05 .06
IMIQ29 -.18 .79 -.20 -.02
IMIQ14 .08 .71 .14 .06
IMIQ15 .20 .67 -.17 .04
IMIQ31 -.22 .65 .04 -.13
IMIQIO -.24 .65 .03 -.02
IMIQ45 -.09 .56 -.39 .31
IMIQ5 .13 .50 -.16 .22
IMIQ12 .28 .49 .05 -.01
IMIQ7 .34 .47 -.18 .09
IMIQ 16 -.06 .40 -.24 .36
IMIQ25 -.03 -. 11 .67 -. 11
IMIQ37 .04 -.24 .60 -. 11
IMIQ18 .37 -.22 .60 -.01
IMIQ4 .11 -.27 .57 .05
IMIQ30 .35 -. 10 .53 .07
IMIQ41 -.19 .06 .52 .05
IMIQ26 .2_3 .26 .44 -.16
IMIQ43 -.33 -. 11 .39 .33
IMIQ35 -.02 -.20 .37 .17
IMIQ1 .17 .20 .34 -.14
IMIQ6 -.02 -.24 -.06 .60
IMIQ9 -.27 .26 -.15 .60
IMIQ13 .26 .02 .10 .56
1MIQ34 -.38 .19 .14 .53
IMIQ2 -.28 .12 -.01 .49
IMIQ23 .14 .13 -.34 .46
540 Schiaffino and Cea

a review of the symptom items which are included in this factor suggest
that this factor most accurately reflects an HIV illness representation.
Turk's factor structure also included a personal responsibility factor, but
the items included more closely match the items in the Curability factor
reported here.
Symptom Variability captures the time line notion of the common-
sense model, but does so with respect to characteristics of the illness
symptoms not the diagnosis or illness label. This factor contains most of
the items in Turk's control factor and changeable factor. It also most
effectively captures the sense of a chronic illness representation. Finally,
Serious Consequences includes items which are consistent with those sug-
gested by the common-sense model; these consequences also more ade-
quately reflect the notion of Illness Label than do the other parts of the
factor solution. Thus, the obtained factor structure is not inconsistent
with the theoretical and empirical evidence concerning Leventhal's com-
mon sense model of illness regulation. Scales were again created for each
of the two patient groups' representations and for the students' illness
representations for each of the three illnesses by summing across factor
items and dividing by the number of valid items. Descriptive statistics
and reliability coefficients for these scales are reported in Table III. In-
ternal consistency reliability coefficients were above .60 for all scales.

T a b l e 111. Descriptive Statistics and I n t e r n a l Reliability for F a c t o r Scales

M e a n (SD)
Personal Symptom Serious
Illness & s a m p l e Curability Responsibility Variability Consequences

RA
Students 2.40 (.53) 2.38 (.46) 3.17 (.41) 3.86 (.47)
= .82 a = .87 ~ = .63 ~ = .62
Patients 2.21 (.50) 2.13 (.48) 3.64 (.54) 3.98 (.61)
= .83 a = .79 ~ = .65 ~ = .64

MS
Students 1.94 (.59) 2.77 (.71) 2.96 (.40) 4.06 (.47)
= .88 a = .78 a = .62 a = .78
Patients 1.93 (.55) 1.91 (.43) 3.49 (.51) 3.55 (.60)
= .85 a = .72 a = .61 a = .61

HIV
Students 1.95 (.62) 3.32 (.80) 2.96 (.54) 3.91 (.51)
cc = .91 c~ = .75 ~ = .65 c~ = .77
Illness Representations 541

Relationship of Illness Representations Across Illnesses and Samples

Since the data were collected with respect to three illnesses from both
patient and nonpatient samples, two important questions can be addressed:
1. Do illness representations vary across illnesses?
2. Do patient and nonpatient representations of the same illness
differ'?
Because they demonstrated stronger reliability, the scales derived from fac-
tor analysis were used to answer these questions.
Two multivariate analyses of variance were performed. In the first
MANOVA, the student sample was used to examine differences in the four
representation scales as a function of illness. It was found that significant
differences existed for the students' representations across the three ill-
nesses. The univariate analyses revealed significant differences among
illness representations with respect to curability [F(2,209) -- 14.41, p <
.001], responsibility [F(2,209) = 35, p < .001], variability [F(2,209) -- 5.19,
p < .01], and consequences [F(2,209) = 3.34, p < .05]. Post hoc examina-
tion of the group differences indicated that the students rated RA as
significantly more curable and more variable than either MS or HIV (see
Fig. 1). The students' assessments of personal responsibility differed across
all three groups; individuals were seen as most responsible for HIV and
least responsible for RA. Finally, the students viewed MS as significantly

Curabiity Responsibility Variability Consequences

l RA i ~ MS ~ HIV

Fig. 1. Illness representations of students for RA, MS, and HIV.

542 S c h i a f f i n o and Cea

more serious than RA; surprisingly, the student evaluations of the serious
consequences of MS were higher than ratings of the consequences of HIV,
although the difference between these two groups was not significant.
The second MANOVA examined the extent to which representations
varied, not only as a function of illness characteristics, but also as a function
of individual experience with the illness. In this MANOVA, the four rep-
resentation scales were considered with respect to illness (RA or MS) and
subject status (patient or student). Overall significance was found for both
interaction effects and main effects. An examination of the univariate F
tests revealed significant interactions between illness and status for both
responsibility [F(1,300) = 24.43, p < .001] and seriousness [F(1,300) =
24.89, p < .001]. Students reported higher levels of responsibility for both
illnesses, but the difference between students and patients was more ex-
treme when the illness being evaluated was MS (see Fig. 2). Students and
patients did not differ dramatically in their ratings of the seriousness of
RA but there was a striking difference for MS, with students rating the
consequences of MS as far more serious than did MS patients themselves.
Only main effects were found for symptom variability: RA was rated
as more variable than MS IF(I,300) = 10.72, p < .001] and patients rated
these illnesses as more variable than did students [F(1,300) = 83.25, p <
.001]. Finally, a main effect of curability was found, with RA viewed as
more curable than MS IF(I,300) = 33.95, p < .001].

5F
4 I
q

Curablity Responsibility Variability Consequences

~1~ RA (lltudents) [~ RA (plltientll)

M8 (etudentll) i M8 (patient-,)

Fig. 2. Comparison of illness representations of students and patients for RA

and MS.
Illness Representations 543

DISCUSSION

People form cognitive models of health and illness which guide the
interpretation of illness and its symptoms (Leventhal et al., 1983). These
cognitive representations of illness are assumed to play an important role
in determining much of health and illness behavior.
The data presented here examined the utility of the Implicit Models
of Illness Questionnaire (IMIQ; Turk et al., 1986) as a measure with which
to assess the constructs identified in the common-sense model of illness
representations (Nerenz and Leventhal, 1983). While factor analysis failed
to support the "generic" four-factor structure proposed by Turk et al., it
did result in the identification of another four factor structure.
Illness causes which provided information regarding possible illness
cures (e.g., my illness was caused by stress and poor diet; therefore, if I
lower my stress and improve my diet, my illness might be cured) tended
to group together on one factor. Moreover, this Curability factor reflected
a real sense of illness as acute and curable rather than chronic. The second
factor, Personal Responsibility, also contained information concerning the
causes of illness. Now, however, the focus was more in terms of placing
blame rather than finding solutions--someone is responsible, I am respon-
sible. This pattern is consistent with at least two studies which have
examined cultural differences in causal explanations for illness and have
discovered that causes are not neatly characterized as either internal or
external; emotional, punitive, natural, and mythical causes have been sug-
gested as quite distinct contributors to personal models of illness (Klonoff
and Landrine, 1994; Landrine and Klonoff, 1994).
The suggestion that the construct of Symptom Variability defined the
third factor underscores the importance of the symptom/diagnosis distinc-
tion in the study of illness representations. Research concerned with illness
attributions and perceived control over illness has noted that perceptions
of causality and control are sometimes quite different when the focus in
on the occurrence or diagnosis of illness compared to the management of
symptoms (Affleck et al., 1987; Nerenz and Leventhal, 1983; Schiaffino and
Revenson, 1995). A crucial step in coping with chronic illness involves the
recognition that the illness is chronic and not acute, that symptom man-
agement does not mean illness cure, but that some part of chronic illness
symptomatology can s o m e t i m e s be s o m e w h a t manageable. Modifiers not-
withstanding, the realization, on the part of chronic illness patients, that
there still is something they can do to help their situation is critical to the
development of illness-specific self-efficacy beliefs, and self-efficacy has
been shown to predict better adjustment (Schiaffino et al., 1992).
544 Schiaffino and Cea

The differences observed here between student representations and pa-

tient representations highlight the role of personal experience in the devel-
o p m e n t of illness representations. Students reported significantly less
symptom variability across all of the illnesses; an understanding of this aspect
of chronic illness most commonly comes from hard personal experience. Future
research might explore the extent to which this variability is assessed differently
by the patient with the illness compared to family or friends; differences in
representations might signal future problems in coping and social support.
The students' perceptions of RA as more curable than the other ill-
nesses may reflect the stereotype or popular label of "minor arthritis pain"
commonly promulgated in the media. Arthritis is seen by many as merely
the minor aches and pain which accompany old age. Unfortunately, the
form of arthritis which was referred to in data collection--rheumatoid ar-
thritis-is most common in younger adults and can be quite severe. In fact,
RA and MS are quite similar illnesses. The differences in student repre-
sentations of these illnesses could be a function of media images: minor
arthritis pain, on the one hand, and public images of Richard Pryor severely
impaired by MS, on the other.
It is important to note that the patterns which emerged for the two
patient samples do not consider differences in disease status which likely
affect representations. The patient samples were relatively close to diag-
nosis and the MS sample was somewhat atypical in its ability to participate
in questionnaire research. More severely affected MS patients often have
cognitive impairments which prohibit such data collection strategies. It is
likely that disease severity would affect illness representations; future re-
search might explore the ways in which both illness characteristics and
patient representations work together to influence coping and adjustment.
Several conclusions can be drawn concerning the usefulness of the
IMIQ in the study of illness representations. A factor structure was iden-
tified here which, while not replicating the original structure, displayed
stability across samples and captured, to some extent, the spirit of the com-
mon-sense model. Nevertheless, the search for an underlying factor
structure common to all illnesses may be somewhat misguided. It may be
that for different illnesses, some items and issues will be more salient than
others; items concerned with duration could be relevant to issues of symp-
tom variability, or illness consequences, or illness identity. The study of
illness cognitions is precisely about understanding the ways in which illness
characteristics come together to create complex illness schemas. The IMIQ
offers a baseline item pool and, at least, a tentative factor structure which
can lend consistency to future research studies and increase the ability to
draw comparisons across studies.
Illness Representations 545

APPENDIX

. . . . . . . ~ ~ . ~ . .

6
oo
-i
. . . . . . . . . . , . . . . . . . . . ~ . . . . . ~

e~

. . . - . . . . ~ ~ ~ . ~

9' - 4 " ~ - , ~ r,*';,'~o-i.-;rq r4r,ic-i~-dri--;oi~";r4 o i e i e i r i ~ i ~ i ~ m ~i~

.~_

.~_
"r"

~ ~ ,- E .~ .o N E ~ ~ .-

u
~ "~~" ~.:r , ~
~,-,~.--.
~ ~ ~ ~ ~.:
~ ~.~ >,
_
,-
~

~ ~ ' ~ ~.'~ ~ - ~ " ~ ~a-o o

546 Schiaffino and Cea

~,~ ~o~, ~.,

ox ~-. o~. -~t-. _~. co . ,_; . ,,o . ,_; . r-- . ._4 o-, o0 ,.o ,_;

l,~: ~O
r4 Ir-i
~ ~r4 IN
r-i ~~ r,,~; O00'~ ~ iz~
r,'; r4 c4 ~ ,~-
r4 r,i H i-q
,-4 ~~ i/h IN
r-i 4 I ~

II
s"

r-.
9
{.j
:-5

[-.

I]

E
O

:> ~,~ ~ .- .~- ~ 0 ~. ~ ~3

o 0 0 ~ t= " ' ~

.o ~ >, >,:> = ~ >, o o ~ ~ c~'-4

o = o = o o o _
Illness Representations 547

ACKNOWLEDGMENT

T h e a u t h o r s w o u l d like to t h a n k Tracey A. R e v e n s o n for h e r r o l e in

the r e s e a r c h r e p o r t e d on t h e R A p a t i e n t s a n d for h e r c o n t i n u e d s u p p o r t .

REFERENCES

Affleck, G., Pfeiffer, C., Tennen, H., and Fifield, J. (1987). Appraisals of control and
predictability in adapting to a chronic disease. J. Person. Soc. PsyehoL 53: 273-279.
Bla[ock, S. J., DeVellis, R. F., Brown, G. K., and Wallston, K. A. (1989). Validity of the
center for epidemiological studies depression scale in arthritis populations. Arth. Rheum.
32: 991-997.
Bishop, G. D., Briede, C., Cavazos, L., Grotzinger, R., and McMahon, S. (1987). Processing
illness information: The role of disease prototypes. Basic Appl. Soc. Psychol. 8: 21-44.
Genest, M. (1983). Coping with rheumatoid arthritis. Can. J. Behav. Sci. 14: 392-408.
Hampson, H. E, Glasgow, R. E., and Zeiss, A. M. (1994). Personal models of osteoarthritis
and their relations to serf-management activities and quality of life. J. Behav. Med. 17:
143-158.
KIonoff, E. A., and Landrine, H. (1994). Culture and gender diversity in commonsense beliefs
about causes of six illnesses. J. Behav. Med. 17: 407-418.
Landrine, H., and Klonoff, E. A. (1994). Cultura[ diversity in causal attributions for illness:
The role of the supernatural. J. Behav. Med. 17: 181-183.
Lau, R. R. (1982). Origins of health locus of control beliefs. J. Person. Soc. Psyehol. 42(2):
322-334.
Lau. R. R., and Hartman, K. A. (1983). Common-sense representations of common illnesses.
Health PsychoL 2: 167-185.
Lau, R. R., Bernard, T. M., and Hartman, K_ A. (1989). Further explorations of common-sense
representations of common illnesses. Health Psychol. 8(2): 195-219.
Meyer, D., Leventha[, H., and Gutmann. (1985). Common-sense models of illness: The
example of hypertension. Health Po,chol. 4(2): 115-135.
Meenan, R. F., Anderson, J. J., Kazis, L. E., Egger, M. J., et al. (1984). Outcome assessment
in clinical trials: Evidence for the sensitivity of a health measure. Arth. Rheum. 27:
1344-1352.
Moulton, J. M.. Sweet, D. M., Temoshok, L., and Mandel, J. S. (1987). Attributions of blame
and responsibility in relation to distress and health behavior change in people with AIDS
and AIDS-related complex. J. Appl. Soc. P~chol. 17(5): 493-506.
Nercnz, D, R., and Leventhal, H. (1983). Self-regulation theory in chronic illness. In Burish.
T. G., and Bradlcy, L. A. (eds.), Coping with Chronic Disease, Academic Press, New York.
Parker, J., Frank, R., Beck, N., et al. (1988). Pain in rheumatoid arthritis: Relationship to
demographic, medical, and psychological factors. J. Rheumatol, 15: 433-437.
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the
general population. AppL Psychol. Measure. 1: 385-401.
Schiaffino, K. M., and Revenson, T. A. (1992), The role of perceived self-efficacy, perceived
control, and causal attributions in adaptation to rheumatoid arthritis: Distinguishing
mediator vs. moderator effects. Person. Soc. Psychol. Bull. 18: 709-718.
Schiaffino, K. M., and Revenson, T. A. (1995). Why me'? The persistence of positive and
negative appraisals over the course of illness. J. AppL Soc. Psychol. 25: 601-618.
Schiaffino, K. M., Revenson, T. A., and Gibofsky, A. (1992). Assessing the impact of
self-efficacy beliefs on adaptation to rheumatoid arthritis. Arth. Care Res. 4: 150-157.
Schumacher, H. R. (ed.) (1988). Primer on tt, e Rheumatic Diseases, 9th ed., Arthritis
Foundation, Atlanta, GA.
548 Schiaffino and Cea

Strickland, B. R. (1989). Internal-external control expectancies: From contingency to creativity.

Am. P.sychol. 4: 1-12.
Smith, C. A., and Wallston, K. A. (1992). Adaptation in patients with chronic rheumatoid
arthritis: Application of a general model. Health Psychol. 11: 151-162.
Taylor, S. E., and Brown, J. D. (1988). Illusion and well-being: A social-psychological
perspective on mental health. P~ychol. Bull. 103: 193-210.
Turk, D. C., Rudy, T. E., and Salovey, P. (1986). Implicit models of illness. J. Behav. Med.
9: 453-474.
W a l l s t o n , K. A., W a l l s t o n , B. S., and D e V e l l i s , R. (1978). D e v e l o p m e n t of the
multidimensional health locus of control (MHLC) scales. Health Educ. Monogr. 6:
160-170.
Wallston, K. A., Smith, R. A., King, J. E., Forsbcr, P. R., Wallston, B. S., and Nagy, V. T.
(1983). Expectancies about control over health: Relationship to desire for control of
health care. Person. Soc. Psychol. Bull. 9: 377-385.