Measuring illness cognitions

Page 51
Black blue
ILLNESS COGNITIONS 51
been made previously in response to descriptions of ‘your most recent illness’. They
reported that the subjects’ piles of categories reflected the dimensions of identity
(diagnosis/symptoms), consequences (the possible effects), time line (how long it
will last), cause (what caused the illness) and cure/control (how and whether it can be
treated).
A series of experimental studies by Bishop and colleagues also provided support for
these dimensions. For example, Bishop and Converse (1986) presented subjects with
brief descriptions of patients who were experiencing six different symptoms. Subjects
were randomly allocated to one of two sets of descriptions: high prototype in which all
six symptoms had been previously rated as associated with the same disease, or low
prototype in which only two of the six symptoms had been previously rated as associated
with the same disease. The results showed that subjects in the high prototype condition
labelled the disease more easily and accurately than subjects in the low prototype condition. The authors argued that this provides support for the role of the identity dimension
(diagnosis and symptoms) of illness representations and also suggested that there is some
consistency in people’s concept of the identity of illnesses. In addition, subjects were
asked to describe in their own words ‘what else do you think may be associated with this
person’s situation?’. They reported that 91 per cent of the given associations fell into
the dimensions of illness representations as described by Leventhal and his colleagues.
However, they also reported that the dimensions consequences (the possible effects) and
time line (how long it will last) were the least frequently mentioned.
There is also some evidence for a similar structure of illness representations in other
cultures. Weller (1984) examined models of illness in English-speaking Americans and
Spanish-speaking Guatemalans. The results indicated that illness was predominantly
conceptualized in terms of contagion and severity. Lau (1995) argued that contagion is
a version of the cause dimension (i.e. the illness is caused by a virus) and severity is a
combination of the magnitude of the perceived consequences and beliefs about time line
(i.e. how will the illness effect my life and how long will it last) – dimensions which
support those described by Leventhal and his colleagues. Hagger and Orbell (2003)
carried out a meta analysis of 45 empirical studies which used Leventhal’s model of
illness cognitions. They concluded from their analysis that there was consistent support
for the different illness cognition dimensions and that the different cognitions showed a
logical pattern across different illness types.
Measuring illness cognitions
Leventhal and colleagues originally used qualitative methods to assess people’s illness
cognitions. Since this time other forms of measurement have been used. These will be
described in terms of questionnaires that have been developed and methodological issues
surrounding measurement.
The use of questionnaires
Although it has been argued that the preferred method to access illness cognitions
is through interview, interviews are time consuming and can only involve a limited
Page 51
Black blue
Page 52
Black blue
52 HEALTH PSYCHOLOGY
number of subjects. In order to research further into individuals’ beliefs about illness,
researchers in New Zealand and the UK have developed the ‘Illness Perception Questionnaire’ (IPQ) (Weinman et al. 1996). This questionnaire asks subjects to rate a series of
statements about their illness. These statements reflect the dimensions of identity (e.g. a
set of symptoms such as pain, tiredness), consequences (e.g. ‘My illness has had major
consequences on my life’), time line (e.g. ‘My illness will last a short time’), cause (e.g.
‘Stress was a major factor in causing my illness’) and cure/control (e.g. ‘There is a lot
I can do to control my symptoms’). This questionnaire has been used to examine beliefs
about illnesses such as chronic fatigue syndrome, diabetes and arthritis and provides
further support for the dimensions of illness cognitions (Weinman and Petrie 1997).
Recently a revised version of the IPQ has been published – IPQR (Moss-Morris et al.
2002) which had better psychometric properties than the original IPQ and included
three additional subscales: cyclical timeline perceptions, illness coherence and emotional
representations. However, people have beliefs not only about their illness but also about
their treatment, whether it is medication, surgery or behaviour change. In line with this
Horne (1997; Horne et al. 1999) developed a questionnaire to assess beliefs about medicine which was conceptualized along four dimensions: two of these are specific to the
medication being taken: ‘specific necessity’ (to reflect whether their medicine is seen as
important) and ‘specific concerns’ (to reflect whether the individual is concerned about
side effects) and two are these are general beliefs about all medicines: ‘general-overuse’
(to reflect doctors over use of medicines) and ‘general-harm’ (to reflect the damage that
medicines can do).
Measurement issues
Beliefs about illness can be assessed using a range measures. Some research has used
interviews (e.g. Leventhal et al. 1980; Leventhal and Lorenz 1985; Schmidt and
Frohling 2000), some has used formal questionnaires (e.g. Horne and Weinman 2002;
Llewellyn et al. 2003), some have used vignette studies (e.g. French et al. 2002) and
others have used a repertory grid method (e.g. Walton and Eves 2001). French and
colleagues asked whether the form of method used to elicit beliefs about illness influenced the types of beliefs reported. In one study French et al. (2002) compared the
impact of eliciting beliefs using either a questionnaire or a vignette. Participants were
asked either simply to rate a series of causes for heart attack (the questionnaire) or to
read a vignette about a man and to estimate his chances of having a heart attack. The
results showed that the two different methods resulted in different beliefs about the
causes of heart attack and different importance placed upon these causes. Specifically,
when using the questionnaire smoking and stress came out as more important causes
than family history, whereas when using the vignette smoking and family history came
out as more important causes than stress. In a similar vein French et al. (2001) carried
out a systematic review of studies involving attributions for causes of heart attack and
compared these causes according to method used. The results showed stressors, fate or
luck were more common beliefs about causes when using interval rating scales (i.e. 1–5)
than when studies used dichotomous answers (i.e. yes/no).
In summary, it appears that individuals may show consistent beliefs about illness
Page 52
Black blue