Generalized Cancer Anxiety Index
An Index of Generalized Cancer Anxiety
for Use in Health and Risk Communication Surveys.
A paper submitted to the Science Communication Interest Group of the Association
for Education in Journalism and Mass Communication, for presentation at the
annual convention, New Orleans, LA, August 1999.
Craig W. Trumbo, Ph.D.
Prathannna Kannaovakun, Doctoral Candidate in Mass Communication
Department of Agricultural Journalism
The University of Wisconsin-Madison
440 Henry Mall
Madison WI 53706
(608) 262-5902
[log in to unmask]
An Index of Generalized Cancer Anxiety
for Use in Health and Risk Communication Surveys.
A paper submitted to the Science Communication Interest Group of the Association
for Education in Journalism and Mass Communication, for presentation at the
annual convention, New Orleans, LA, August 1999.
ABSTRACT (75 words)
This article describes the development and validation of an index measuring the
construct of Generalized Cancer Anxiety for application in health and risk
communication research. Various literatures are searched for the concept of
cancer anxiety. The resulting set of 18 questions are evaluated using a survey
of university employees (pre-test) and four surveys of communities involved in
cancer rate investigations. A subset of 4-6 question items are identified with
good reliability, test-retest stability, and validity.
This research was supported by a grant from the National Science Foundation,
SBR-97-14960.
This article will describe the development and validation of an index to measure
the construct of Generalized Cancer Anxiety (GCA) as observed in general
population samples. Such an index could prove to be very useful in survey
research conducted on health and risk communication involving cancer. Any effort
to model mechanisms of information processing, information credibility, judgment
or decision-making in a health context involving cancer should include some
measurement of GCA in order to better isolate the predictive mechanisms of
specific interest. It is also likely that GCA could figure directly into a
variety of communication processes involving cancer information. For example,
studies have shown that anxiety over cancer appears to suppress systematic
handling of cancer information (Jepson and Chaiken, 1990) and that risk judgment
surrounding public health controversies involving cancer tends to be driven by
heuristic processes, especially in the sub-populations least concerned with the
risk (Trumbo, 1999).
But a thorough search of the literature fails to turn up a carefully evaluated
index of this construct. Cancer anxiety, or "cancerphobia," has appeared as a
concern in a good number of publications. Each of these studies offers a unique
measurement of the construct, and in each case finds the measurement adequate
(cancer anxiety questions are typically modeled after some other standardized
instrument). Yet no effort appears to have been made to establish a uniform tool
for capturing this construct in general population survey samples. This research
will undertake that task, and will evaluate the resulting index as it functions
in concert with a set of information processing, communication, and credibility
models.
Cancer Anxiety in the Literature
Many terms have been used to describe anxiety about cancer and have often been
used interchangeably. The examples include fear of cancer, concern about cancer
(Berman & Wandersman, 1990), cancerphobia (Berman & Wandersman, 1990; Brown &
Lees-Haley, 1992), and worry about cancer (Easerling & Leventhal, 1989). These
terms are often used to represent a dimension subsumed under more global
constructs such as attitudes to cancer, knowledge of cancer (Dent & Goulston,
1982), health anxiety (Speckens et al., 1996), or disease phobia (Kellner et
al., 1987; Pilowsky, 1967). Indeed, these several studies seem to suggest the
concepts share certain common characteristics involving responses to this
dreaded disease, that is, anxiety seemed to be a pervasive response to cancer.
To avoid confusion over the term used to describe psychology of cancer,
generalized cancer anxiety (GCA) is introduced here as a broader concept. GCA is
designed to tap individual's attitudes to, and beliefs about cancer associated
with his or her emotional responses to cancer cues. GCA represents a continuum
where on one end is an absence of concern about cancer, the other end is
hypochondriacal concerns about the disease.
GCA can be seen as a rational response, which is associated with fear of death,
disfigurement, disability, dependence, and disruption of key relationships
(Holland & Cullen, 1986). The authors suggest that these fears explain why
cancer seems so frightening and may cause preoccupations or phobia. In addition,
GCA may result from mass hysteria, social proof, reporting bias, and litigation
rather than rational responses to threat (Brown & Lees-Haley, 1992). GCA
involves the perception of risk or vulnerability, perception of severity, and
environment cues. Prior experience with cancer, knowledge about cancer, and
attitudes toward health may contribute to individuals' generalized cancer
anxiety. (Gutteling, Seydel, & Weigman, 1986).
Perceived Vulnerability. Several factors influence public perception of
vulnerability to cancer. Berman & Wandersman (1990) pointed out that the
pervasiveness of cancer incidences as well as multiple causes (hereditary,
environment, lifestyle) lead to the belief that "everything causes cancer."
Environmental threats such as drinking water contamination or hazardous waste
sites can have a great influence on public perceptions of vulnerability to
cancer. Love Canal and Three Mile Island are two famous instances among hundreds
of hazardous waste sites or accidents that have caused strong community
concerns. Empirical evidence has shown that cancer anxiety or fear of cancer is
more prevalent in communities located near hazardous waste sites, chemical
waste, or toxic exposure sites even though there was no greater incidence of
serious illness, cancer or death in the communities near the sites (Dunne, et
al., 1990).
Media reports of cancer death can also provide fuel for worry about cancer
(Easerling & Leventhal, 1989). Schwartz et al., (1985) argue that "if the press
overemphasizes the risks, it can create unnecessary concern." A literature
review by Brown and Lees-Haley (1992) suggests that reporting bias may play a
role in higher symptom rates. Perceived vulnerability to cancer can also be
induced by noncancer symptoms, such as tiredness or headaches, interpreted as
symptoms of cancer. Environmental cues such as smokestack emissions can also
elicit worry about cancer.
Individuals' prior experience with cancer can also influence estimation of their
chance of getting cancer. Gutteling, Seydel, & Weigman (1986) found that
subjects with prior experiences with cancer estimated their susceptibility to
cancer as being higher than the respondent with fewer prior experiences.
Perceived Severity. Cancer is a fear-provoking medical condition. In a study by
Gutteling, Seydel, & Wiegman (1987), 90 percent of the respondents perceived
cancer to be more serious and more threatening than other diseases. Berman &
Wandersman (1990) suggested that cancer is the most feared disease because it
results in prolonged suffering, pain, disfigurement, and social isolation.
Fatalism about control of cancer was identified as one of four dimensions of
subjects' attitudes to cancer (Dent & Goulston, 1982).
Katz, et al., (1987) asked subjects to rate cancer, AIDS, diabetes, and heart
disease in terms of their competence, moral worth, dependence, depression, and
morbidity. Cancer was always described as the most painful condition and, next
to AIDS, the least understood medically and most deadly. Cancer patients were
always seen as more depressed than AIDS patients. Cancer anxiety was moderately
related to perceptions of cancer patients. Specifically, the higher the
subject's anxiety score, the more unfavorable were their evaluations of cancer
patients.
Effects of Cancer Anxiety. Evidence has shown that persons with health anxiety
demonstrate more negative cognition. In other words, they interpret diagnostic
information more negatively than non-health anxious people. They also perceive
themselves having a higher risk for disease (Hadjistavropoulos, et al., 1998).
More pertinent to this study, Jepson & Chaiken (1990) have shown that anxiety
over cancer appears to suppress systematic processing of cancer information.
Trumbo (1999) found that risk judgment surrounding public health controversies
involving cancer tends to be driven by non-systematic processes, especially in
the sub-populations least concerned with the risk.
Katz et al., 1987 discuss how cancer victims are avoided because their presence
arouses anxiety and feelings of vulnerability. In addition, fear of cancer and
worry about cancer have been shown to discourage preventive behaviors such as
participation in cancer screening programs (Donato et al., 1991), or delay
treatment seeking (Love, 1991). Dent and Goulston (1982) found that one
dimension of attitudes to cancer is subjects' defensive avoidance. The denial
was reflected in ways in which the subjects did not wish to know if they had
cancer, did not want their friends to know, and did not want to think about
cancer.
Gutteling, Seydel, and Weigman (1986) found that when compared to women with
higher levels of anxiety about cancer, those with a low level of fear knew more
about the disease and demonstrated more intention to behave preventively. They
had lower estimations of their vulnerability to cancer, and perceived cancer as
less threatening than did those with higher levels of anxiety. Furthermore, in a
study of women with a family history of breast cancer, Kash, et al., (1992)
found that increased cancer anxiety decreased regular clinical examinations.
Generalized Cancer Anxiety shares common characteristics with hypochondriasis
or a persistent preoccupation with the fear of having, or the belief that one
has, a serious disease despite medical reassurance. Hypochondriacs, like other
members of the population, perceive cancer as a frightful disease (Brown &
Lees-Haley, 1992).
The Whitely Index (WI) and Illness Attitude Scales (IAS) are two questionnaires
exploring attitudes, fears, and beliefs of hypochondriasis. The WI was the first
questionnaire constructed by Pilowsky (1976) to assess hypochondriasis. The
content of the questionnaire assesses worry or fear of disease, the experience
of somatic symptoms, attention paid to symptoms, hypochondriacal beliefs and
attitudes other than fears and beliefs about disease.
Factor analysis of the 14 items yielded three separate factors: Bodily
Preoccupation, Disease Phobia, and Conviction of the Presence of Disease with
nonresponse to reassurance. However, the principal component analysis by
Speckens et al. (1996) yielded only one-factor solution in three populations
studied (general medical outpatients, general practice patient, and subjects
from the general population). Similarly, Barsky et al. (1986) showed that the
subscales of the WI were highly intercorrelated.
Speckens et al (1996) also found that the WI can discriminate between
hypochondriacal and non-hypochondriacal subjects and predicts a lower rate of
recovery from the symptoms. The WI was not associated with medical care
utilization. Similar findings were observed in the studies conducted by Beaber
and Rodney 1984, and Brasky et al. 1986.
The Illness Attitude Scale (IAS) is another questionnaire designed to measure
psychopathology associated with hypochondriasis. The questionnaire was developed
by Keller et al. (1987) and consists of nine subscales: worry about illness,
concern about pain, health habits, hypochondriacal beliefs, thanatophobia,
disease phobia, bodily preoccupation, treatment experience, and effects of
symptoms. However, factor analysis from the study by Specken et al., (1996)
revealed that the items yielded only two components: Health Anxiety and Illness
Behavior. The Health Anxiety subscale of the IAS discriminates well between
hypochondriacal and non-hypochondriacal patients. Scores on the WI and Health
Anxiety subscale of the IAS declined significantly from general medical
outpatients, through general practice patients to subjects from the general
population.
The Trait Scale of the State-Trait Anxiety Inventory (Spielbeger, 1970) has
been widely used in assessing clinical anxiety in patients. In general,
psychoneurotic and depressed patients have higher scores on this scale. Trait
anxiety is defined as "relatively stable individual differences in anxiety
proneness." The scale is also widely used outside of medical settings, such as
on college students and military recruits.
A number of studies of cancer anxiety have employed the State-Trait Anxiety
Inventory to measure general forms of anxiety or to validate the fear of cancer
measure (e.g. Gutteling, Seydel, & Wiegman, 1987; Steptoe, Horti, & Stanton,
1986). In the latter study, trait anxiety did not have an influence on the
likelihood of having concern about cancer. However, among women who were
worried, those with higher trait anxiety also reported more concern.
Methods
First, the collection of articles in which cancer anxiety has appeared was
examined for survey questions. Some 28 questions were assembled and boiled down
to 18 questions by eliminating redundancies. We identified four sub-constructs
that the questions fell into: preoccupation or worry about cancer, fear of
social isolation that might be caused by the disease, the perceived seriousness
of cancer as a disease, and vulnerability to cancer. A question set was
developed that mixed these four concepts with a balance of positive and negative
statements. To help assess criterion validity, personal experience with the
disease is measured, and a set of two questions are included to evaluate how the
respondents perceive their cancer risk compared to the general population. All
questions appear in Table 1.
Second, the health psychology literature was examined to identify closely
related constructs for which there might be well-established measurement
instruments. Two important tools were located: The State-Trait Anxiety Inventory
(STAI) (Spielberger, 1983) and the Whitley Index for Hypochondriasis (Pilowski,
1967; Speckens et al., 1996). A valid measure of GCA should be somewhat related
to - yet distinct from - both of these constructs. The STAI is being used only
to measure trait anxiety, and the Whitley Index for general health anxiety.
The first analysis is designed to perform initial testing of the question set
and to determine an appropriate subset of questions to use in a larger survey.
For this effort, 128 non-academic staff persons at a large Midwestern university
were randomly selected from its phone directory (genders equal). This group was
surveyed by mail in November, 1998. A total of 77 completed questionnaires were
returned (60 percent). The response is gender balanced (53 percent female), has
a range of age from 24 years to 65 (f = 45, SD = 9). Three months later, the set
of GCA questions were sent again to the respondents to evaluate short-term
stability. Of the 73 mailings (there was one retirement reported and three came
back without identification numbers), 61 individuals responded to the follow-up
(84 percent). The re-test response has the same age characteristics, but is
biased toward females (59 percent).
Subsequent to an analysis of the campus data, a shortened version of the GCA was
included in a set of four surveys of communities involved in cancer
investigations. Epidemiological investigations of local cancer rates were
recently completed in these four sites: Seattle-Tacoma, WA, where aircraft
exhaust is of concern near the airport; Marion, OH, where concern exists over
leukemia rates and the presence of Army ammunition manufacturing; Niagara Falls,
NY, where a neighborhood is concerned about cancer rates and near-by industrial
dumping; and Myrtle Beach, SC, which has become known locally as a "leukemia hot
spot" and some are concerned about water quality. Samples of 400 households were
selected at each site, and questionnaires were sent in early March, 1999. In
addition to the GCA, the questionnaires include items on risk perception,
information processing, and credibility that may be used to evaluate the GCA's
construct validity. These survey returns are now arriving, and the results will
be reported in a revision to this paper.
Results and Discussion
The State-Trait Anxiety Inventory preformed reliably (alpha = .91), as did the
Whitley Index (alpha = .82). Means and standard deviations for the GCA questions
are reported in Table 1. Note that the mean values are from recoded variables,
so that in all cases high values represent higher anxiety (questions are in
their original forms). Responses are consistently biased toward the less anxious
end of the scale (grand mean = 2.7), probably representing the commonly observed
"optimistic bias." Analysis of the GCA question set is performed in four steps:
evaluation of the results of an exploratory factor analysis, evaluation of the
full index, evaluation of the four subconcepts defined a priori, and evaluation
of a reduced index. Additionally, the reduced index will be evaluated as it
performs in the field application described previously (that analysis to be
added to this paper's revision).
Exploratory Factor Analysis. The scree plot from the factor analysis strongly
suggests that only two factors should be considered. Loading patterns support
this conclusion, with factors three and four cross loading excessively (Table
2). While factor five does not cross load, it accounts for a small percentage of
variance. The first two factors together account for 38 percent of total
variance.
An exploratory factor analysis is most useful when the resulting factors present
interpretable patterns. In this case, interpretability is somewhat equivocal.
The first factor is a mixture of variables representing elements of worry and
variables representing elements of social isolation. The second factor is
two-thirds social isolation and one item representing seriousness of the
disease. Experimentation with alternate methods of extraction and rotation
returned very similar results.
To evaluate the factors, additive indices were created. Table 3 reports their
evaluation. Factor one has good reliability in the first test, but poor
reliability in the second. Factor two is somewhat less reliable in both tests.
Test-retest correlation for both factors is sufficient but not excellent. Factor
one does not correlate with either the Whitley Index or the STAI, but has a
negative correlation with the variable "odds" (perceived relative odds for
cancer, negative values indicate that individuals believe they are more prone to
cancer than are others). Factor two correlates with the STAI.
The Full Index. The full set of 18 questions was also summed to a single index.
Reliabilities at both the first and second test are good, as is test-retest
correlation. The full GCA correlates moderately with the STAI, but not the
Whitley Index. This suggests that generalized cancer anxiety (by this index) is
not unrelated to trait anxiety, and is distinct from health anxiety. The strong
correlation with the variable odds suggests some external validation for the
full GCA.
A Priori Subconcepts. The development of the GCA questionnaire was centered
around the identification of salient concepts in the literature. Researchers
felt comfortable with the face validity of the four subconcepts identified. To
strengthen these subconcepts, each subset of variables was examined by factor
analysis (alpha extraction) to maximize potential reliability. This resulted in
four variables being excluded: media was excluded from the worry subgroup, other
diseases was excluded from the seriousness subgroup, and both uneasy and others
were excluded from the social isolation subgroup. Table 3 indicates the set of
variables included in each of the subgroups.
Reliability is strongest for the subconcepts capturing disease seriousness and
worry. Reliability for the social isolation subgroup is not especially strong,
and is unacceptable for the vulnerability subgroup. Test-retest correlations are
fairly good for all but vulnerability. Correlations among these subconcepts show
that disease seriousness and social isolation are associated quite strongly, and
that vulnerability is not associated with any of the other subconcepts. Of the
four subconcepts, disease seriousness presents the most compelling set of
relationships with the other indices. This subgroup is strongly associated with
trait anxiety and moderately associated with both health anxiety and perceived
relative odds for cancer. The social isolation subgroup is moderately correlated
with trait anxiety, but distinct from health anxiety or perception of relative
odds for cancer. Worry is not correlated with any other index; vulnerability
only presents a perplexing negative correlation with trait anxiety.
A Reduced Index. In many research situations, it is desirable to have an index
of no more than a half-dozen items. An important goal of this analysis is
therefore to derive such a reduced index. Several approaches might be taken. The
first factor might be used. But while it presets fair reliability and stability,
it does not have the desired relationship with the other indices. Generalized
cancer anxiety should be related to, yet distinct from, both trait anxiety and
general health anxiety. The best candidate for a reduced index with these
characteristics is clearly the disease seriousness subconcept. And with only
four questions, it also leaves room to consider another subconcept.
Vulnerability can be ruled out for several obvious reasons. Social factors wins
out handily over worry thanks to it strong correlation with disease seriousness
and association with the STAI.
Confirmatory factor analysis was used to evaluate this choice (using AMOS 3.2).
Figure 1 present the final results of this effort. Several other models were
evaluated en route to this final structure. First, a simple three factor model
including the worry subset was evaluated. It fit well (Chi-square = 26, df = 24,
p = .37, AGFI = .88). The model was then trimmed to include only the variables
from the social isolation and disease seriousness subgroups (a subsequent
iteration also eliminated one of the social isolation variables, which was
redundant). This simpler model also fit well (Chi-square = 8.7, df = 8, p = .37,
AGFI = .91). Finally, the structure depicted in Figure 1 was evaluated. This
includes the Whitley Index and the STAI. This model fits quite well. The social
isolation component is unique from health anxiety and trait anxiety, but
strongly associated with the disease seriousness component (which does load on
both trait and health anxiety). This is held to be a satisfying reduced index of
six items.
Exploratory Evaluation. An additional set of five questions measured the
respondents' experience with cancer (Table 1), ranging from having had the
disease, to having lost a friend, or having a family member with the disease.
These questions were included for exploratory purposes. However, one might
expect a generalized concept of cancer anxiety to be somewhat resilient to
actual experience. This is in fact what was observed. Each of the 10 indices
evaluated in Table 3, plus the variables "odds," were separately regressed on
the set of five cancer experience variables. Only two of the resulting 50 betas
were significant at the .05 level; those without friends having been diagnosed
with cancer scored lower in anxiety on the social isolation (beta = -.27) and
vulnerability (beta = -.24) subcomponents. Notably, the Whitley Index also had
no association with cancer experience.
Age and gender were also included for exploratory purposes. Each of the indices
was regressed on the respondent's age: none of the resulting associations were
significant. Each was also regressed on gender (male = 1). The full index,
Factor 1 and the social isolation subcomponent all presented correlations with
gender significant at the .05 level in which males are higher in anxiety (with
about five percent variance explained in each case).
Field Application. To evaluate the performance of the reduced index, the set of
six questions was included in the previously described surveys sent to four
communities involved in cancer investigations. A 30 percent response rate was
achieved prior to the second mailing. Expectations are for a 50-60 percent
return by the end of April (for a data set of 600-700 cases). Analysis of these
data will be included in the final revision of this paper. This analysis will
include an evaluation of the reduced index's reliability in four separate
samples, and its ability to account for variance in risk judgment - held to be
important for evaluating the index's construct validity. Additionally, the GCA
will be evaluated in concert with a variety of variables addressing information
processing and information credibility.
Conclusions
Several approaches have been taken to optimize this proposed index of
generalized cancer anxiety. Other approaches might be taken. While the full set
of 18 questions performs acceptably, a subset of four questions presents the
optimal balance of reliability, validity, and parsimony: Once cancer invades the
body, it is almost impossible to get rid of it; Cancer is always an extremely
painful disease; If I were to be diagnosed with cancer, it is very likely that I
would survive; By the time you know that you have cancer, it's usually too late
to cure it. The reliability of the resulting GCA index that focuses on the
seriousness of the disease, and its short-term test-retest stability, falls
short of the ideal of a well-developed index (note that the STAI weighs in at
.91). However, the alpha coefficients are strong enough to support this
introduction of the GCA index into the literature. The strength of fit in the
confirmatory factor analysis with structural equation modeling offers additional
support. Improvements in question wording (perhaps simplification) could be one
route to boosting reliability.
The resulting index offers promise in terms of its validity as well. Our
approach to establishing a base for an initial question set rested on a careful
search of the extant literature. The resulting four subcomponents, and
especially the final two used, offer a degree of content validity in that they
capture the primary ideas associated with cancer anxiety as expressed in a range
of published research.
An evaluation of criterion validity rests on the argument that the concept of
generalized cancer anxiety should be associated with both trait anxiety and
general health anxiety. It should also be associated with other variables
capturing how individuals perceive their likelihood of having cancer as compared
to others. The final subgroup of questions performed well in this respect. It
appears that GCA is more strongly related to trait anxiety (which is held to be
an enduring aspect of personality), and is about as strongly associated with
trait anxiety as is the Whitley Index. However, the association between the GCA
Index and the Whitley Index is weaker. A Venn diagram is easily imagined in
which both the GCA Index and the Whitley Index share 19 and 24 percent variance
(respectively) with trait anxiety, and only six percent common variance between
themselves. Confirmation with the variable "odds," in which higher anxiety is
associated with a pessimism about cancer, adds support for criterion validity.
An evaluation of construct validity will have to wait for the incoming survey
data to be analyzed.
Table 1. GCA questions grouped by concept and other external questions.
Codewords key to other tables. All GCA questions measured on 1-7 agree/disagree
scale and recoded so that high values equal higher anxiety.
preoccupation and worry
f sd
worry 2.4 1.0 I seldom worry about getting cancer.
thoughts 3.2 1.3 I have never thought that I might have some form of cancer.
concern 2.9 1.2 Cancer does not concern me any more than other serious disease.
media 2.1 0.9 I am upset when I see a TV program or newspaper article about
cancer.
word 2.6 1.1 The word cancer itself does not scare me.
perceived seriousness
f sd
invade 2.4 1.0 Once cancer invades the body, it is almost impossible to get
rid of it.
painful 2.7 0.9 Cancer is always an extremely painful disease.
survival 2.8 0.9 If I were to be diagnosed with cancer, it is very likely that
I would survive.
too late 2.3 1.0 By the time you know that you have cancer, it's usually too
late to cure it.
other disease 3.1 1.0 Many other diseases are more serious than cancer.
social isolation
f sd
talk is hard. 2.4 1.1 If I had cancer it would be difficult for me to talk
about it with others.
public 2.4 0.9 People with cancer do not seem to avoid going out in public.
want talk 2.0 1.0 If I were diagnosed with cancer, I would want to talk with
people I know.
uneasy 3.1 1.3 I would feel uneasy if I found out that someone I knew had
cancer.
others 2.3 1.1 If someone I knew had cancer, I would worry that it might happen
to me.
vulnerability
f sd
avoid act 3.2 1.3 I avoid doing certain things because they may cause cancer.
avoid place 2.1 1.0 I avoid certain places because they may cause cancer.
avoid cancer 3.7 0.9 There are many things that can be done to avoid cancer.
experience variables (0 = no)
f sd
exp1 0.11 0.31 Have you ever been diagnosed with any form of cancer by a
doctor?
exp2 0.64 0.49 Do you have any close friends who have ever been diagnosed with
cancer?
exp3 0.51 0.50 Have you ever had a friend die of cancer?
exp4 0.80 0.41 Have any members of your close family ever been diagnosed with
cancer?
exp5 0.64 0.49 Have you ever had a close family member die of cancer?
relative risk perception
f sd
chance 2.5 1.0 What do you think your chances are of getting cancer at some
time in life?
person 2.4 0.7 What is the average person's chance for getting cancer at some
time?
odds -0.15 1.0 = person - chance. (0 = neutral neg = pessimistic pos
= optimistic)
Table 2. Evaluation of exploratory factor analysis. Principal components
extraction, varimax rotation, loadings under .40 blanked.
Factor 1 Factor 2 Factor 3 Factor 4 Factor 5
concern .79
thoughts .78
worry .78
others .60 .49
uneasy .47
want talk .80
talk is hard .78
too late .52
invade .80
other diseases .63
survival .40 .58
avoid cancer .48 -.47
painful .74
word .50 .50
public .42 .44
avoid place .74
avoid act .69
media .55
Eigenvalue 4.1 2.7 1.7 1.3 1.2
Variance 22.7 14.9 9.7 7.1 6.7
Table 3. Reliability and correlational analysis of all indices used.
Index
Reliability
Correlations
a1
a 2
r
gca
fac1
fac2
ser
soc
wor
vul
stai
wi
odds
GCA f = 46 sd = 8
(18 questions)
.75
.78
.78
-
.76**
.60**
.73**
.76**
.69**
.14
.27*
.14
-.30**
Factor 1 f = 14 sd = 4 (concern, thoughts, worry, others, uneasy)
.77
.59
.71
-
.16
.31**
.50**
.91**
.14
.05
.10
-.23*
Factor 2 f = 7 sd = 2
(want talk, talk hard, too late)
.68
.57
.75
-
.62**
.80**
.07
-.11
.29*
.13
-.22
Serious f = 10 sd = 3
(invade, painful, survival, too late)
.75
.74
.73
-
.53**
.24*
-.21
.44**
.25*
-.29*
Social f = 12 sd = 3
(talk hard, want talk, public)
.66
.66
70
-
.29*
.00
.24*
.09
-.20
Worry f = 11 sd = 4
(worry, thoughts, concern, word)
.75
.74
.70
-
.14
.05
.08
-.21
Vulnerability f = 9 sd = 2
(avoid act, place, avoid cancer)
.52
.48
.35
-
-.30**
-.12
.00
STAI f = 41 sd = 12
(20 questions from trait index)
.91
-
-
-
.49**
-.14
Whitley Index f = 23 sd = 6
(14 questions)
.82
-
-
-
.03
* p < .05 ** p < .01
[--- Pict Graphic Goes Here ---]
References.
Barsky A. J., Wyshak, G., and Klerman G. L. (1986a). Hyponchondriasis: an
evaluation of the DSM-III criteria in medical outpatients. Archives of General
Psychiatry, 43:493-500.
Barsky, A. J., Wyshak, G. and Klerman, G. L. (1986b). Medical and psychiatric
determinants of outpatients medical utilization. Medical care, 24: 548-560.
Beaber, R. J., and Rodney, W. M. (1984). Underdiagnosis of hypochondriasis in
family practice. Psychosomatics, 25: 39-46.
Berman, S. H., and Wandersman, A. (1990). Fear of cancer and knowledge of
cancer: A review and proposed relevance to hazardous waste sites. Social Science
of Medicine, 31(1): 81-90.
Brown, R. S. and Lees-Haley, P. R. (1992). Fear of future illness, chemical
AIDS, and cancerphobia: A review. Psychological Reports, 71: 187-207.
Dent, O. W., and Goulston, K. (1982). Community attitudes to cancer. Journal of
Biosocial Science, 14: 359-372.
Donato, F., Bollani, A., Spiazzi, R., Soldo, M., Pasquale, L., Monarca, S.,
Lucini, L., and Nardi, G. (1991). Factors associated with nonparticipation of
women in a breast cancer screening programme in a town in northern Italy.
Journal of Epidemiology & Community Health, 45: 59-64.
Dunne, M. P., Burnett, P., Lawton, J., and Raphael, B. (1990). The health
effects of chemical waste in an urban community. Medical Journal of Australia,
152: 592-97.
Easerling, D. V., and Leventhal, H. (1989). Contributions of concrete cognition
to emotion: neutral symptoms as elicitors of worry about cancer. Journal of
Applied Psychology, 74: 787-796.
Gram, I. T. and Slenker, S. E. (1992). Cancer anxiety and attitudes toward
mammography among screening attenders, nonattenders, and women never invited.
American Journal of Public Health, 82(2): 249-251.
Gutteling, J. M., Seydel, E. R., and Weigman, O. (1986). Perceptions of cancer.
Journal of Psychosocial Oncology, 4(3): 77-91.
Hadjistavropoulos, H. D., Craig, K. D. and Hadjistavropoulos, T. (1998).
Cognitive and behavioral responses to illness information: the role of health
anxiety. Behavior Research and Therapy, 36: 149-164.
Holland, R., and Cullen, A. (1986). New insights and attitudes. In A. I. Holleb
(Ed)., The American Cancer Society Book: prevention, detection, diagnosis,
treatment, rehabilitation, cure. New York: Doubleday.
Ilfeld, F. W. (1976). Further validation of a psychiatric symptom index in a
normal population. Psychological Reports, 39: 1215-1228.
Jepson, C., and Chaiken, S. (1990). Chronic issue-specific fear inhibits
systematic processing of persuasive communications. Journal of Social Behavior
and Personality, 5(2): 61-84.
Johnson, D. J. & Meischke, H. (1992). Differences in evaluations of
communication channels for cancer-related information. Journal of Behavioral
Medicine 15 (5): 429-445.
Kash, K. M., Holland, J. C., Halper, M. S. and Miller, D. M. (1992).
Psychological distress and surveillance behaviors of women with a family history
of breast cancer. Journal of the National Cancer Institute, 84(1): 24-30.
Katz, I. R., Hass, G., Parisi, N., Astone, J., McEvaddy, D., & Lucido, D. J.
(1987). Lay people's and health care personnel's perceptions of cancer, AIDS,
cardiac, and diabetic patients. Psychological Report 60: 615-629.
Katz, R. C., Meyers, K. and Walls, J. (1995). Cancer awareness and
self-examination practices in young men and women. Journal of Behavioral
Medicine, 18(4): 377-384.
Kellner, R., Abbott, P., Winslow, W. W. and Pathak, D. (1987). Fears, beliefs,
and attitudes in DSM-III hypochondriasis. The Journal of Nervous and Mental
Disease, 175(1): 20-25.
Love, N. (1991). Why patients delay seeking care for cancer symptoms, what you
can do about it. Postgraduate Medicine, 89: 151-152.
Lucock, M. P. and Morley, S. (1996). The health anxiety questionnaire. British
Journal of Health Psychology, 1(2): 137-150.
McCaul, K. D., Schroeder, D. M. and Reid, P. A. (1996). Breast cancer worry and
screening: Some prospective data. Health Psychology, 15(6): 430-433.
Okun, A., Stein, R. E., Bauman, L. J. and Silver, E. J. (1996). Content validity
of the Psychiatric Symptom Index, CES-Depression Scale, and State-Trait Anxiety
inventory from the perspective of the DSM-IV. Psychological Reports, 79:
1059-1069.
Pilowsky, I. (1967). Dimensions of hypochondriasis. British Journal of
Psychiatry, 113: 89-93.
Schwartz, S. P., White, P. E., and Hughes, R. G. (1985, March). Environmental
threats, communities, and hysteria. Journal of Public Health Policy 6 (1):
58-77.
Speckens, A. E. M., Spinhoven, P., Sloekers, P. P. A., Bolk, J. H., and van
Hemert, A. M. (1996a). A validation study of the Whitley Index, the Illness
Attitude Scales, and the Somatosensory Amplification Scale in general medical
and general practice patients. Journal of Psychosomatic Research, 40(1): 95-104.
Speckens, A. E. M., van Hemert, A. M., Spinhoven, P., and Bolk, J. H. (1996b).
The diagnostic and prognostic significance of the Whitley Index, the Illness
Attitude Scales and the Somatosensory Amplification Scale. Psychological
Medicine, 26: 1085-1090.
Spielberger, C. D. (1983). State-Trait Anxiety Inventory for Adults. Palo Alto,
CA: Mind Garden.
Steptoe, A., Horti, J. and Stanton, S. (1986). Concern about cancer in women
undergoing elective gynecological surgery. Social Science and Medicine, 23(11):
1139-1145.
Trumbo, C. W. (1999). Heuristic-systematic information processing and risk
judgment. Risk Analysis (in press).
|
