319

British Journal of Health Psychology (2003), 8, 319333

2003 The British Psychological Society
www.bps.org.uk

Help-seeking intentions for breast-cancer

symptoms: A comparison of the self-regulation
model and the theory of planned behaviour

Myra S. Hunter1 , Elizabeth A. Grunfeld2 * and

Amanda J. Ramirez1
1
ICRF Psychosocial Oncology Group, Guys, Kings and St Thomas Medical School,
St Thomas Hospital, London, UK
2
Unit of Psychology, Guys, Kings and St Thomass Medical School, Guys Hospital,
London, UK

Purpose. Delays in seeking help for symptoms have been found to be associated with
poorer outcome in breast-cancer patients. This study explores symptom perceptions
and health beliefs as predictors of intentions to seek medical help in a general female
population. The utility of the self-regulation model of illness cognition and the theory of
planned behaviour were examined in predicting help-seeking intentions for potential
symptoms of breast cancer in a general population sample.
Methods. A general population sample of 546 women completed a postal question-
naire comprising items examining components of the self-regulation model and the
theory of planned behaviour. Help-seeking intention was determined by asking
participants to rate the likelihood of visiting their GP for a range of breast symptoms.
Results. Hierarchical multiple regression analysis revealed that the cognitive com-
ponent of the self-regulation model accounted for approximately 22% of the variance in
help-seeking intention. Identity (b = 0.45, p < .001) emerged as a signi cant predictor of
intention to seek help. Inclusion of the components of the theory of planned behaviour
accounted for an additional 7% of the variance; the signi cant predictors were attitude
to help-seeking (b = 0.19, p < .001) and perceived behavioural control (b = 0.12,
p < .01).
Conclusions. Intention to seek medical help for a potential breast-cancer symptom
may be mediated, partly, by cognitive representations of the identity and consequences
of breast cancer and by attitudes towards help-seeking and perceived behavioural
control. Although less than one-third of the variance was accounted for, these results

* Requests for reprints should be addressed to Elizabeth Grunfeld, Unit of Psychology, 5th Floor, Thomas Guy House, Guys
Hospital, London SE1 9RT, UK (e-mail: beth.grunfeld@kcl.ac.uk).
320 Myra S. Hunter et al.
have important implications for future research (in terms of identifying which variables
should be examined) and for the development of a model of help-seeking behaviour in
women with breast-cancer symptoms.

Breast cancer is a signicant health risk to women in the UK. Approximately 35,000 new
cases of breast cancer are diagnosed each year, and the UK has one of the highest
mortality rates in the world. Health education and screening programmes are aimed at
increasing early detection, but there is concern that, following self-discovery of a breast
symptom, approximately one-third of women delay for 12 weeks or more prior to
presentation to a health care provider (Richards, Westcombe, Love, Littlejohns, &
Ramirez, 1999a). This is important, as over 75% of patients present after symptoms have
occurred rather than through the National Health Service screening programme
(Richards et al., 1999a). Furthermore, evidence suggests that longer delay intervals
are associated with a larger tumour size, more advanced disease, and poorer survival
(Richards, Smith, Ramirez, Fentiman, & Rubens, 1999). To encourage changes in help-
seeking behaviour, it is useful to understand not only cognitive representations of
disease and associated symptoms but also beliefs regarding the treatment for breast
cancer and the implications of seeking help.
The self-regulation model proposes that people construct cognitive representations
of an illness or disease in order to understand and cope with it (Leventhal, Nerenz, &
Steele, 1984). These cognitive representations are seen to determine emotional
responses and to guide coping responses. The illness-related cognitions, or illness
representations as they are also known, comprise ve main components or dimensions:
identity (interpretation of the symptoms associated with the illness and of the labels
attached to the illness), cause (likely causes of the illness, which may relate to the
persons beliefs about personal risk), time-line component (the likely duration of the
illness), consequences (the perceived severity of the illness and the potential impact on
physical, psychological, and social functioning), and cure/control (the extent to which
the illness can be successfully controlled or cured). Evidence from the eld of breast
cancer suggests that patient delay behaviour may be inuenced, at least in part, by
experience of a non-lump breast-cancer symptom and by attribution of symptoms to
causes other than cancer (Burgess, Ramirez, Richards, & Love, 1998; Ramirez et al.,
1999). In addition, it has been demonstrated that decisions to seek medical help are
inuenced not only by the presence of an atypical symptom but also by the perceived
severity and consequences of the symptom (Cameron, Leventhal, & Leventhal, 1993).
Examination of the ve components of Leventhals illness representations may help
explain womens appraisal of symptoms of breast cancer and their decisions regarding
the necessity of seeking medical help.
The theory of planned behaviour states that the intention to perform certain
behaviours is determined by two forms of beliefs (Ajzen & Madden, 1986): behavioural
beliefs (attitudes towards a particular behaviour) and normative beliefs (beliefs con-
cerning the likely approval or disapproval of key referents towards performing a certain
behaviour). There is some evidence from qualitative studies to suggest that signicant
others may inuence a womans decision to seek help for symptoms of breast cancer
(Burgess et al., 1998; Facione, 1993). The theory of planned behaviour also takes
account of a persons perceived behavioural control (how easy or difcult it is to
perform that behaviour). Previous work employing the theory of reasoned action (the
precursor to the theory of planned behaviour that does not include perceived
behavioural control) has demonstrated that intention to delay in seeking help for a
 Help-seeking intentions for breast cancer 321
breast symptom (among healthy women) is associated with holding positive attitudes
towards delay and with perceived social pressure to delay (Timko, 1987). The type of
favourable attitudes that Timko found to be important included beliefs that delaying
would allow one to maintain some control over, and would prevent disruption to, ones
own or signicant others lives. As this study was based on the theory of reasoned
action, it did not examine the inuence of perceived behavioural control with regard to
intention to seek help.
The theory of planned behaviour has not previously been employed to examine
actual help-seeking behaviour in women with breast cancer but has been used to
examine breast-cancer screening behaviour (Rutter, 2000). Attendance at screening is a
different behaviour to help-seeking for symptoms, as it does not require the individual to
interpret and attribute symptoms, and the individual is provided with information on
how, when, and where to implement their intention. However, Rutters (2000) large
sample study does provide a useful comparison. The study found that attitude,
perceived behavioural control, and subjective norm were all predictors of intention
to attend screening, although only attitude and subjective norm were predictors of
actual attendance behaviour.
Additional variables, outside the two models, may also inuence intention to seek
help. Previous work examining intentions to take hormone-replacement therapy
extended the theory of planned behaviour to include similar prior behaviour and age
(Quine & Rubin, 1997). Both these additional variables and the components of the
theory of planned behaviour were found to predict intention. Similar prior behaviour in
relation to help-seeking for potential breast-cancer symptoms could include breast self-
examination behaviour, and the present study included this variable. The aim of the
present study was to apply the self-regulation model and the theory of planned
behaviour to explore womens interpretations and representations of potential
breast-cancer symptoms and to identify their beliefs regarding help-seeking and treat-
ment for breast cancer. The ultimate aim was to determine which aspects of these two
theories best predict intention to seek help promptly for potential breast-cancer
symptoms.

Methods
Development of the questionnaire
The questionnaire items were developed from previous quantitative and qualitative
research with breast-cancer patients (Burgess et al., 1998; Burgess, Hunter, & Ramirez,
2001). For example, this previous work demonstrated that type of symptom and
misattribution of symptoms were key factors associated with delay among breast-
cancer patients, so a range of breast changes were examined in the current study.
The previous work also identied unwillingness to prioritize oneself over others,
concern about talking to GPs, and being able to talk to others about symptoms as
being inuential in the delay process. Questionnaire items were generated from this
previous work and were structured so as to represent the components of the self-
regulation model (Leventhal & Nerenz, 1985) and the components of the theory of
planned behaviour (Ajzen & Madden, 1986). All items were piloted on a sample of
healthy women who were not involved with the study or with the health service (N = 8)
to assess comprehensibility and suitability (i.e. that they did not cause embarrassment or
distress). The questionnaire was easily understood by all the participants (in the pilot
322 Myra S. Hunter et al.
group), and there were no items that needed to be eliminated on the grounds of poor
comprehensibility or unsuitability. The questionnaire comprised two sections.

Self-regulation model
The items in the rst section were based on the format of the Illness Perceptions
Questionnaire (Weinman, Petrie, Moss-Morris, & Horne, 1996), and the ve scales
related to the components of Lenventhals illness representations (Leventhal & Nerenz,
1985). The identity scale included 12 symptoms, eight of which were potential
symptoms of breast cancer and four of which were general, non-breast-cancer-related
symptoms. Participants were required to state whether each of the symptoms was a
potential symptom of breast cancer on a 7-point scale from `denitely to `denitely not.
The identity scale was scored by assigning 1 to all symptoms that were rated as
symptoms of breast cancer (1 4 on the scale) and 0 to all symptoms rated as non-
symptoms (57 on the scale). The scores for the eight symptoms of breast cancer were
summed to give an identity scale score (range 18). Cronbachs a for the summed scale
was .71.
The four remaining scales comprised attitudinal statements and required partici-
pants to rate their responses on a 7-point scale from `strongly agree to `strongly
disagree. The cause scale comprised 14 items adapted from the Breast Cancer Risk
Assessment Scale (Royak-Schaler, Cheuvront, Wilson, & Williams, 1996). Seven of the
items were established risk factors for breast cancer, and seven were factors that people
sometimes associate with breast cancer but that are not established risk factors. Each
item on the cause scale was seen to represent a specic causal belief, and therefore each
item was examined individually (Weinman et al., 1996).
The time-line scale comprised three statements about the potential duration of
breast cancer following treatment. The internal consistency for this subscale (Cron-
bachs a .54) was lower than that reported for the original Illness Perceptions
Questionnaire (Weinman et al., 1996). The lower the score, the shorter was the
associated time line. The consequences scale comprised six statements regarding the
physical, social, and psychological consequences of breast cancer (Cronbachs a .74).
The higher the score, the more positive was the persons attitude toward the outcomes
associated with breast cancer. The cure/control scale comprised ve questions related
to the management of breast cancer. The higher the score, the more positive was the
persons attitude toward the controllability of the disease. The internal consistency for
this subscale was rather low (Cronbachs a .50).

Theory of planned behaviour

The items in the second section of the questionnaire were designed to measure the
constructs contained within the theory of planned behaviour and were based on the
format reported by Ajzen and Madden (1986) and as detailed in Conner and Norman
(1995).

Attitudes towards help-seeking

Five semantic differential scales (good/bad, benecial/harmful, pleasant/unpleasant,
wise/foolish, necessary/unnecessary) were used to assess responses towards the
question `Making an appointment to see my doctor for a symptom that might be
breast cancer would be . . .. The scores obtained for the ve semantic scales were
 Help-seeking intentions for breast cancer 323
summed to produce a direct attitude to help-seeking score. Belief-based measures of
attitude towards help-seeking were obtained by asking participants to rate (on a 7-point
scale from `strongly agree to `strongly disagree) eight statements relating to the
consequences of seeking help for a potential breast-cancer symptom (i.e. `I would
nd it easy to talk to my doctor about a breast symptom). The scores on the eight items
were averaged to produce a belief-based attitude to help-seeking score. The higher the
score, the more positive was the attitude toward help-seeking. Cronbachs a for the
summed scale was .57.

Subjective norm
The subjective norm measurement was based on the normative beliefs concerning three
referents: family, spouse/partner, and close friends. The strength of each normative
belief (e.g. `My family would encourage/discourage me to go to the doctor for a symptom
of breast cancer) was assessed on a 7-point scale (`encourage to `discourage). The
respondents motivation to comply with each referent (e.g. `In general, when making
decisions I am inuenced by my familys opinions) was assessed on a 7-point scale (from
`agree to `disagree). Each normative belief was multiplied by the motivation to comply
with the referent and the sum of the products constituted the belief-based measure of
subjective norm (Cronbachs a .68). The higher the score, the greater was the perceived
social pressure to seek help.

Perceived behavioural control

Perceived behavioural control was assessed with four items relating to the degree of
perceived control over help-seeking for a symptom of breast cancer (e.g. `There is
nothing I could do to make sure that I got help for a breast-cancer symptom).
Responses were recorded on a 7-point scale (from `agree to `disagree), and the
product of the four responses served as the measure of perceived behavioural control
(Cronbachs a .74). The higher the score, the greater was the perceived behavioural
control.

Intention
Intention to seek medical help for breast-cancer symptoms was assessed in two
ways, because (1) there was an interest in seeking help for symptoms that might
signify breast-cancer, and (2) there was an interest in the intention to seek help
promptly.

Symptom-based measure. Participants rated (on a 7-point scale from `denitely to

`denitely not) how likely they would be to visit their doctor for each of 12
symptoms, eight of which were potential symptoms of breast cancer and four
that were general non-breast-cancer symptoms. The intention score was calculated
by assigning 1 to all symptoms where the participant indicated that they would
denitely visit their doctor (14 on the scale) and 0 to all symptoms where the
participant indicated that they would denitely not visit their doctor (57 on the
scale). The scores for the eight symptoms of breast cancer were summed to give
an intention score (range 18). This score was used as the independent variable
for multiple-regression analysis.
324 Myra S. Hunter et al.
Temporal-based measure. Temporal information regarding help-seeking intentions of the
participants was elicited by asking, `If you developed a symptom that you thought
might be breast cancer, how soon would you visit the doctor?. Responses were pro-
vided on an 8-point scale from `immediately to `never. To examine differences
between women who would promptly seek help and those that would delay for a
period of time, all participants choosing `immediately were classied as prompt help-
seekers, and all other respondents were classi ed as potential delayers.

Additional variables included in the analysis

It was hypothesized that attitudes to treatment for breast cancer might inuence a
womans intention to seek help for a symptom of breast cancer. To measure attitudes to
treatment, the same ve semantic differential scales employed for the attitude to help-
seeking measure were used to assess attitudes towards undergoing each of four
treatments for breast cancer (surgery, chemotherapy, radiotherapy, and hormone
therapy). The scores obtained for the ve semantic scales were summed to produce a
score for the direct attitude to each form of treatment. The scores for the four treatments
were then averaged to produce a composite score for attitudes towards treatment.
Belief-based measures of attitude towards treatment were obtained by asking partici-
pants to rate (on a 7-point scale from `strongly agree to `strongly disagree) nine
statements relating to the consequences of treatments for breast cancer (i.e. `The
treatment for breast cancer is worse than the disease itself ). The scores for the nine
items were averaged to produce a belief-based attitude to treatment score. Cronbachs a
for the summed scale was .59.
Participants were also asked about their personal and familial history of breast cancer
and about their practice of breast self-examination, as it was hypothesized that these
may inuence intention to seek help for a symptom of breast cancer.

Participants
The sample was recruited from a previous study examining womens knowledge of
breast cancer (Grunfeld, Ramirez, Hunter, & Richards, 2001), and the sample obtained
was representative of the geographical, socio-economic and age distribution of the UK
population. From this original study, 781 participants provided their consent to
participate in the current study. Of these, 546 respondents completed and returned
the questionnaire (a response rate of 70%). The mean age of the sample was 47 years
(range 1686). Thirty-ve per cent of the sample was classiable as professional or
intermediate workers, 45% as skilled or partly skilled, and 21% as unskilled or
unemployed. The study did not examine ethnic differences in responses, as there
were too few women within each particular group to draw meaningful conclusions.
Ninety-three per cent of the sample were classied as White, 3% as Black African and
Black Caribbean, 2% as Indian, Pakistani, or Bangladeshi, 1% as Chinese, and the
remaining 1% were assigned to other ethnic groups.

Analysis
The means and standard deviations for each sub-scale and the inter-correlations between
the sub-scales for each section of the questionnaire are reported. Analysis of variance
was used to examine differences between prompt help-seekers and potential delayers
(classied according to responses to the temporal-based intention measure) on each of
 Help-seeking intentions for breast cancer 325
the subscales. The symptom-based intention measure was used for all correlational
analyses. Hierarchical multiple regression was used to identify the best predictors of
the symptom-based measure of intention. To evaluate whether the addition of the
components of the self-regulation model and the components of the theory of
planned behaviour accounted for signicantly more variance, the formula proposed
by Tabachnik and Fidell (1996, p. 144) was applied.

Results
Responses to the temporal measure of intention to seek help revealed that 58.6% of
respondents would seek immediate help for a breast symptom, and this group was
classied as prompt help-seekers. The remaining participants were classied as poten-
tial delayers; 29.7% indicated that they would seek help within 1 week, 8.5% would seek
help within 1 month, and 3.2% would wait for 2 months or more. Using the main
intention variable (based on the sum of intentions to seek help for individual
symptoms (M = 5.34, SD = 0.85) univariate regression analyses were carried out
between intention and age and intention and socio-economic status. Age was
found to be a signicant predictor of intention to seek help, F(1, 543) = 36.09,
p < .001, whereas socio-economic group was not, F(1, 518) = 0.29, p = .591.

Table 1. Mean scores (SD) for each of the sub-scales from the self-regulation model as a function of
help-seeking intention

Potential delayers Prompt help-seekers

Scale (N = 222) (N = 318)

Identity 6.5 (1.5) 6.8 (1.4)

Timeline 4.0 (1.1) 4.0 (1.3)
Consequences 4.6 (1.1) 4.7 (1.3)
Cure/control 5.7 (0.9) 5.7 (0.9)

Self-regulation model
The mean scores obtained for prompt help-seekers and potential delayers on each of the
sub-scales are presented in Table 1. Potential delayers score signicantly lower on the
identity scale than prompt help-seekers, F(1, 542) = 5.08, p < .05, suggesting that
prompt help-seekers accurately identied more symptoms of breast cancer. The mean
scores for each of the cause items are shown separately in Table 2. It can be seen that the
two groups were similar in their attributions of the risk factors for breast cancer. Family
history was the most strongly endorsed cause of breast cancer, closely followed by
previous breast cancer and smoking. The least-favoured attributions were a previous
breast problem, late onset of menses, and excess body weight. There were no signicant
differences between the two groups in terms of their responses on the time-line,
consequences, and control/cure scales.
The inter-correlations between the self-regulation model scales are shown in Table 3
and logical (weak to moderate) relationships between the scales were apparent.
Respondents who reported a longer time course for breast cancer also reported more
negative attitudes towards the consequences of breast cancer and the controllability of
326 Myra S. Hunter et al.
Table 2. Mean scores (SD) for the cause items from the self-regulation model (range of potential scores
17) (the higher the score, the stronger the causal belief )

Attribution/cause Potential delayers Prompt help-seekers

Family history 5.9 (1.5) 6.0 (1.5)

Previous breast-cancer 5.3 (1.5) 5.3 (1.6)
Bumping or bruising the breast 3.0 (1.8) 3.0 (1.9)
Smoking 5.1 (1.9) 5.4 (1.9)
High-fat diet 4.1 (1.9) 4.1 (2.0)
Older age 4.9 (1.8) 4.8 (1.9)
Alcohol consumption 3.5 (1.7) 3.6 (1.9)
Age at time of rst full pregnancy 5.0 (1.8) 4.8 (1.9)
Late onset of menses 2.7 (1.5) 2.8 (1.8)
Excess body weight 2.8 (1.6) 2.8 (1.8)
Late menopause 5.0 (1.5) 4.9 (1.7)
Previous breast problem 2.3 (1.3) 2.2 (1.4)
Oral contraceptives 3.9 (1.8) 4.0 (1.9)
Stress 4.3 (1.9) 4.2 (2.0)

the disease. Participants who reported a more positive attitude toward the conse-
quences of breast cancer were also more likely to report a greater belief in the
controllability of breast cancer. However, it was shown that participants with the
greatest awareness of breast-cancer symptoms (higher identity score) were more likely
to report negative beliefs about the consequences of breast cancer.

Theory of planned behaviour

The mean scores obtained for prompt help-seekers and potential delayers on each of
the sub-scales of the theory of planned behaviour are presented in Table 4. Prompt
help-seekers scored signicantly higher on both the belief-based, F(1, 542) = 35.18,
p < .001, and direct, F(1, 542) = 13.94, p < .001, measures of attitude toward
help-seeking. Prompt help-seekers also reported greater perceived behavioural con-
trol regarding obtaining help for a breast cancer symptom, F(1, 542) = 10.87, p < .01.
There were no group differences in the responses on the subjective norms scale.
The intercorrelations between the subscales can be found in Table 3. Logical
relationships were again apparent between the scales. As expected, the belief-based
and direct measures of attitude were signicantly correlated. Because of this, and the
fact that there were stronger correlations of intention with the belief-based measures of
attitude than with the direct measures, the direct measures of attitude were excluded
from further analysis, and therefore references to attitude correspond to the belief-based
measures. The subjective norm scores were negatively correlated with attitudes to help-
seeking, suggesting that participants with more positive attitudes to help-seeking were
less likely to feel social pressure to seek help. Examination of the correlation coefcient,
however, revealed that this was a weak relationship.

Multiple-regression analysis
Hierarchical multiple-regression analysis was performed to examine the combined
sufciency of the self-regulation model and the theory of planned behaviour to explain
Table 3. Intercorrelations between the subscales of the self-regulation model, theory of planned behaviour subscales and intention to seek help (N = 546)

Attitudes to Attitudes to Perceived

help-seeking help-seeking Subjective behavioural
Scales Identity Time-line Consequences Cure/control (belief based) (direct measure) norm control

Identity
Time-line .05
Consequences .12** .40**

Cure/control .02 .29** .38**

Attitude to help-seeking (belief-based) .07 .18** .37** .31**

Attitude to help-seeking (direct measure) .02 .12** .20** .16** .38**

Subjective norm .02 .01 .02 .02 .09* .06*

Perceived behavioural control .03 .13** .16** .28** .33** .20** .11*

Intention .43** .09* .08 .16** .30** .28** .07 .19**

* p < .05; ** p < .01.
Help-seeking intentions for breast cancer
327
328 Myra S. Hunter et al.
Table 4. Mean scores (SD) for each of the subscales within the theory of planned behaviour
questionnaire as a function of help-seeking intention

Potential delayers Prompt help-seekers

Scale (N = 222) (N = 318)

Attitude to help-seeking (belief-based) 5.4 (0.9) 5.8 (0.9)

Attitude to help-seeking (direct measure) 6.0 (0.6) 6.2 (0.6)
Subjective norm 92.3 (38.1) 83.9 (43.5)
Perceived behavioural control 6.4 (0.8) 6.6 (0.7)

intention to seek help for a potential breast-cancer symptom. As there were no prior
expectations as to which theory would account for the majority of variance, the theories
were entered in a logical temporal sequence based on the assumption that a woman
would rst draw upon her beliefs and knowledge of breast cancer (as outlined in the
SRM) before drawing upon her beliefs about help-seeking behaviour (outlined in the
TPB). The components of the self-regulation model (identity, time-line, cure/control,
and consequences) were entered on the rst step, and the components of the theory
of planned behaviour (attitude to help-seeking, subjective norm, and perceived
behavioural control) on the second step. As shown in Table 5, the components of
the self-regulation model signicantly increased the explained variance to 22.1%,
F(4, 536) = 23.11, p < .001. The explained variance was again signicantly increased
to 29.2% by the addition of the components of the theory of planned behaviour,
F(2, 500) = 6.70, p < .05. The strongest predictor of intention to seek help was
the perceived identity of symptoms, followed by attitude toward help-seeking and
perceived behavioural control (Table 5).

Table 5. Hierarchical multiple regression analysis of intention to seek help for a symptom of
breast-cancer

b in nal
Step/predictor R2 Adjusted R2 F equation

1. Self-regulation model .227 .221 39.41

Identity .454**
Time-line .058

Cure/control .021
Consequences .024
2. Theory of planned behaviour .301 .292 30.89
Attitude (help-seeking) .187**
Subjective norm .040

Perceived behavioural control .122*

* p < .01; ** p < .001.

Additional potential in uences of help-seeking behaviour

It was hypothesized that personal and familial history of breast cancer, previous breast-
examination behaviour, and attitudes towards treatment for breast cancer might also be
 Help-seeking intentions for breast cancer 329
related to intention to seek help for a potential breast-cancer symptom. Point-biserial
correlations were carried out to examine the relationship between help-seeking
intention and personal and familial history (answered as dichotomous yes/no
responses). There were no signicant relationships between help-seeking intention
and personal history of breast problems (rpb = .042, N = 545, p = .328), help-seeking
for breast problems (rpb = .069, N = 112, p = .145), or familial history of breast

cancer (rpb = .014, N = 546, p = .750). Bivariate linear correlations were used to
examine the relationship between help-seeking intention and frequency of breast self-
examination and attitudes towards breast cancer. There was not a signicant relation-
ship between help-seeking intention and frequency of breast self-examination (r = .038,
N = 445, p = .427). However, there was a weak signicant negative relationship
between intention to seek help for a symptom of breast cancer and attitude towards
treatment for breast cancer (r = .129, N = 546, p < .01).

Discussion
The study examined the utility of the self-regulation model and the theory of planned
behaviour in explaining help-seeking intention for symptoms of breast cancer. The
results of the hierarchical multiple regression revealed that individual components of
both the self-regulation model and the theory of planned behaviour were able to predict
intention to seek help for symptoms of breast cancer but that the total variance
explained by the two models was moderate, being approximately 30%.
The proportion of estimated prompt help-seekers was higher than that found in
other studies of clinical populations (Richards et al., 1999b), and this might be
explained by the absence of the emotional response to possible breast cancer and the
inuence of competing demands which could not be assessed in the design of this
study of well women. Therefore, a higher proportion of the variance in help-seeking
behaviour might be expected in future studies of clinical samples.

Self-regulation model
Using the self-regulation model, differences in illness representations were identied
between women intending to seek help for potential breast-cancer symptoms and
between prompt and potential delayed help-seekers. Identication of symptoms as
potential signs of breast cancer (identity) was the variable within this model that most
strongly predicted intention to seek help for breast-cancer symptoms. This supports one
of the main assumptions of the self-regulation model, namely that symptoms are key
factors in the cognitive representations of health threats and in the initiation of the self-
regulatory process (Cameron et al., 1993).
There has been little research examining womens knowledge of breast-cancer
symptoms, despite the fact that misattribution of symptoms is a well-established
predictor of delay behaviour in medical situations (Cameron, Leventhal, & Leventhal,
1995; Jones, 1990; Stoller & Forster, 1994). In particular, a patient with a non-lump
breast symptom is over four times more likely to delay seeking medical help than an
individual with a breast lump (Burgess et al., 1998; Ramirez et al., 1999). A painless
breast lump is the most frequently recognised symptom of breast cancer (Grunfeld,
Ramirez, Hunter, & Richards, 2001). It may be that limited knowledge of other symptoms
of breast cancer leads to the misattribution of these symptoms to a benign process,
resulting in delay in seeking medical help (Facione & Dodd 1995). In the present study,
330 Myra S. Hunter et al.
even prompt help-seekers did not identify all potential symptoms of breast cancer. These
results have implications not only for the inclusion of symptom information in health
education campaigns but also for ensuring that the array of potential breast symptoms
becomes part of a common `cancer knowledge. It is important to encourage people to
be aware of non-lump symptoms (i.e. nipple retraction) and that the ability to correctly
identify such symptoms may have survival implications.
Time-line, consequences, and control/cure beliefs were not found to be inuential
predictors of help-seeking intention. This study was conducted with a sample of well
women, and it may be unlikely that they would consider the timescale of a disease when
making decisions regarding help-seeking for hypothetical symptoms. Beliefs regarding
the time course of breast cancer might be more likely to inuence the affective response
to the disease, as has been suggested for other illnesses (Moss-Morris, Petrie, &
Weinman, 1996). As the study was hypothetical in nature, it was not possible to
measure affective response to illness within this design. In addition, the a values for
both the time-line and the cure/control scale were rather low, and this would need to be
addressed in future studies based upon these subscales.

Theory of planned behaviour

Holding a positive attitude towards help-seeking for breast cancer was a strong predictor
of intention to seek help for breast-cancer symptoms and prompt help-seeking. These
results are in line with previous research that has demonstrated an association between
less favourable attitudes to the medical profession and increased likelihood to delay
seeking medical help (Battistella, 1971). Previous qualitative research highlighted that
some women may be hesitant to `bother their GP for symptoms of breast cancer,
particularly if the symptoms are ambiguous (Burgess et al., 2001). These ndings have
important implications for ensuring that high-risk women, in particular, are encouraged
to seek help for ambiguous symptoms.
Intention to seek help was also inuenced by beliefs regarding personal condence
in the ability to seek help for a symptom of breast cancer. The present study did not
examine in detail the inuences on perceived behavioural control under these condi-
tions, but there is evidence to suggest that some women may not prioritize their own
health over perceived competing demands (Burgess et al., 2001). Future research is
necessary to elicit reasons why some women lack the condence to seek help for
symptoms of breast cancer.

Relative suf ciency of the models

The models provided a clear framework for examining variables that may inuence help-
seeking intentions for breast symptoms. Combined, the models accounted for almost
one-third of the total variance in intention to seek help for breast cancer symptoms, and
this is in line with the variance explained using social-cognition models for other health
behaviours (i.e. Hill, Gardner, & Rassaby, 1985). However, to explain even this relatively
moderate degree of variance provides valuable insight into the reasons why women may
or may not seek help for breast-cancer symptoms. This in turn could have a signicant
impact in terms of health education and informing help-seeking behaviour.
The current study did not examine the emotional component of the self-regulation
model, and it is probable that more variance would have been explained had this
component been included in a sample of women with breast-cancer symptoms
 Help-seeking intentions for breast cancer 331

Figure 1. Hybrid model incorporating components of the self-regulation model and theory of planned
behaviour shown to be important in predicting intentions to seek help for breast-cancer symptoms.
The heavy lines show the variables that emerged as signi cant predictors of intention following
hierarchical multiple-regression analysis (b values are shown). The dashed lines represent signi cant
correlations between variables (the correlation coef cients are shown).

(Leventhal & Nerenz, 1984). Furthermore, unlike previous research examining screen-
ing and treatment intentions (Quine & Rubin, 1997; Rutter, 2000), subjective norms
were not found to be a signicant predictor of intention to seek help among this sample.
Previous research has suggested that subjective norms are primarily of importance in
situations where the behaviour will directly affect the health of signicant others, for
example the decision to terminate a pregnancy (Smetana & Adler, 1980), or where the
behaviour is performed publicly, for example wearing a seat belt while driving
(Wittenbraker, Gibbs, & Kahle, 1983). Since help-seeking for a potential, possibly
ambiguous, symptom of breast cancer does not fall into either of these categories, it
could be argued that subjective norms would not be a key inuential variable in a
decision to seek help under these circumstances. However, the importance of social
networks in help-seeking for breast symptoms has been demonstrated previously
(Burgess et al., 1998). This work demonstrated that not disclosing the discovery of a
symptom to someone within a few days was associated with increased delay and
that being prompted to seek help by someone else was associated with prompt
help-seeking.
The current study forms part of a larger programme of research examining delay
behaviour for breast cancer. The study aimed to draw upon the elements of two models
in order to maximize understanding of the process of help-seeking. The results suggest
the possibility of a two-component process whereby a woman appraises breast changes
and, having interpreted the change as a possible breast-cancer symptom, cognitively
processes the advantages and disadvantages of seeking help, drawing upon her health
beliefs and self-efcacy beliefs. Figure 1 depicts a hybrid model drawing on the
components of the self-regulatory model and the theory of planned behaviour that
were found to be of signicance in predicting intention to seek help for breast
symptoms. The results suggest that, although both models are useful as an initial
starting point in the examination of intentions to perform health behaviours, different
types of behaviour even within the same patient group (i.e. screening, starting
treatment, and help-seeking for symptoms) may be inuenced by separate components
of the models. Reasons for these discrepancies have already been discussed. The
research highlights the value of utilizing multiple models in the initial stages of research
332 Myra S. Hunter et al.
to identify key components that can be used to inform both research and policy
developments.
Inevitably, additional variables, outside the remit of the two models, will contribute
to help-seeking intentions. For example, it is likely that there is a complex, non-linear
relationship between age and health beliefs regarding breast cancer. Research is needed
to examine the usefulness of these models across different age groupings, to see which
variables exert most inuence at different life stages. The addition of previous behaviour
has been shown to increase the explained variance for some health-related behaviours
(Quine & Rubin, 1997). In the present study, however, breast self-examination and
previous experience of seeking help for a breast symptom were not signicantly
correlated with help-seeking intention. The results do suggest that some women,
who hold negative perceptions of breast-cancer treatments, may delay seeking medical
help for symptoms of breast cancer. However, this conclusion is drawn from a relatively
weak correlation, and therefore further research is needed to explain the remaining
variance not accounted for by the models used in this study. Additional variables that
could be examined in future research include the emotional component of the self-
regulation model and situational factors (e.g. competing demands). A future study is
being planned, drawing upon these ndings with a clinical sample of women who have
recently sought medical help for breast-cancer symptoms, with the aim of developing a
health-promotion intervention.

Acknowledgements
This study was supported by a Project Grant from The Breast Cancer Campaign (charity number
299758, grant reference number 1999/96). The authors would like to thank Barzan Rahman for his
help with data collection.

References
Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behaviour: Attitudes, intentions and
perceived behavioural control. Journal of Experimental and Social Psychology, 22, 453474.
Battistella, R. M. (1971). Factors associated with delay in the initiation of physicians care among
late adulthood persons. American Journal of Public Health, 6(7), 1348 1361.
Burgess, C. C., Hunter, M. S., & Ramirez, A. J. (2001). A qualitative study of delay among women
reporting symptoms of breast cancer. British Journal of General Practice, 51, 967971.
Burgess, C. C., Ramirez, A. J., Richards, M. A., & Love, S. B. (1998). Who and what inuences
delayed presentation in breast cancer? British Journal of Cancer, 77(8), 1343 1348.
Cameron, L., Leventhal, E. A., & Leventhal, H. L. (1993). Symptom representations and affect as
determinants of care seeking in a community-dwelling, adult sample population. Health
Psychology, 12(3), 171179.
Cameron, L., Leventhal, E. A., & Leventhal, H. L. (1995). Seeking medical care in response to
symptoms and life stress. Psychosomatic Medicine, 57, 3747.
Conner, M., & Norman, P. (1995). Predicting health behaviour (pp. 121162). Buckingham, UK:
Open University Press.
Facione, N. C. (1993). Delay versus help seeking for breast-cancer symptoms: A critical review of
the literature on patient and provider delay. Social Science and Medicine, 36, 1521 1534.
Facione, N. C., & Dodd, M. J. (1995). Womens narratives of help-seeking for breast cancer. Cancer
Practice, 3(4), 219225.
Grunfeld, E. A., Ramirez, A. J., Hunter, M. S., & Richards, M. A. (2001). Womens knowledge and
beliefs regarding breast cancer. British Journal of Cancer, 86, 1373 1378.
 Help-seeking intentions for breast cancer 333
Hill, D., Gardner, G., & Rassaby, J. (1985). Factors predisposing women to take precautions against
breast and cervix cancer. Journal of Applied Social Psychology, 15, 59 79.
Jones, R. A. (1990). Expectations and delay in seeking medical care. Journal of Social Issues,
46(2), 8195.
Leventhal, H., & Nerenz, D. (1985). The assessment of illness cognition. In P. Karoly (Ed.),
Measurement strategies in health psychology (pp. 517555). New York: Wiley.
Leventhal, H., Nerenz, D. R., & Steele, D. J. (1984). Illness representations and coping with health
threats. In A. Baum, S. E. Taylor, & J. E. Singer (Eds.), Handbook of psychology and health.
Vol IV: Social psychological aspects of health (pp. 219252). Hillsdale, NJ: Erlbaum.
Moss-Morris, R., Petrie, K. J., & Weinman, J. (1996). Functioning in chronic fatigue syndrome:
Do illness perceptions play a regulatory role? British Journal of Health Psychology, 1(1),
1525.
Quine, L., & Rubin, R. (1997). Attitude, subjective norm and perceived behavioural control as
predictors of womens intentions to take hormone replacement therapy. British Journal of
Health Psychology, 2, 199216.
Ramirez, A. J., Westcombe, A. M., Burgess, C. C., Sutton, S., Littlejohns, P., & Richards, M. A. (1999).
Factors predicting delayed presentation of symptomatic breast cancer: A systematic review.
Lancet, 353, 1127 1131.
Richards, M. A., Smith, P., Ramirez, A. J., Fentiman, I. S., & Rubens, R. D. (1999b). The inuence on
survival of delay in the presentation and treatment of symptomatic breast cancer. British
Journal of Cancer, 79(5/6), 858864.
Richards, M. A., Westcombe, A. M., Love, S. B., Littlejohns, P., & Ramirez, A. J. (1999a). Inuence of
delay on survival in patients with breast cancer: A systematic review of the literature. The
Lancet, 353, 1119 1126.
Royak-Schaler, R., Cheuvront, B., Wilson, K. R., & Williams, C. M. (1996). Addressing womens
breast cancer risk and perceptions of control in medical settings. Journal of Clinical
Psychology in Medical Settings, 3(3), 185199.
Rutter, D. R. (2000). Attendance and reattendance for breast cancer screening: A prospective
3-year test of the Theory of Planned Behaviour. British Journal of Health Psychology, 2,
199216.
Safer, M. A., Tharps, Q. J., Jackson, T. C., & Leventhal, H. (1979). Determinants of three stages of
delay in seeking care at a medical setting. Medical Care, 17(1), 1129.
Smetana, J. G., & Adler, N. E. (1980). Fishbeins value expectancy model: An examination of
some assumptions. Personality and Social Psychology Bulletin, 6, 8996.
Stoller, E. P., & Forster, L. E. (1994). The impact of symptom interpretation on physician utilization.
Journal of Aging and Health, 6(4), 507534.
Tabachnik, B. G., & Fidell, L. S. (1996). Using multivariate statistics (4th ed., pp. 144). Boston:
Allyn & Bacon.
Timko, C. (1987). Seeking medical care for a breast cancer symptom: determinants of intention to
engage in prompt or delay behavior. Health Psychology, 6, 305328.
Weinman, J., Petrie, K. J., Moss-Morris, R., & Horne, R. (1996). The Illness Perception
Questionnaire: A new method for assessing the cognitive representation of illness. Psychology
and Health, 11, 431445.
Wilcox S., & Stefanick, M. L. (1999). Knowledge and perceived risk of major diseases in middle-
aged and older women. Health Psychology, 18(4), 346353.
Wittenbreaker, J., Gibbs, B. L., & Kahle, L. R. (1983). Seat belt attitudes, habits and behaviours: An
adaptive amendment to the Fishbein model. Journal of Applied Social Psychology, 13,
406421.

Received 9 July 2001; revised version received 25 April 2002