Content uploaded by Paschal Sheeran
Author content
All content in this area was uploaded by Paschal Sheeran on Jan 12, 2016
Content may be subject to copyright.
Chapter
2
The health belief model
Charles Abraham and Paschal Sheeran
1 General background
In the 1950s, US public health researchers began developing psychological models designed
to enhance the effectiveness of health education programmes (Hochbaum 1958; Rosenstock
1966). Demographic characteristics such as socio-economic status, gender, ethnicity, and age
were known to be associated with preventive health-related behaviour patterns (i.e. patterns
of behaviour predictive of differences in morbidity and mortality) as well as differential use of
health services (Rosenstock 1974). Even when services were publicly financed, socio-economic
status was associated with health-related behaviour patterns. Demographic and socio-economic
characteristics could not be modified through health education but it was hypothesized that
other potentially modifiable individual characteristics associated with health-related behaviour
patterns could be changed through educational interventions, and thus shift health behaviour
patterns at population levels.
Beliefs provide a crucial link between socialization and behaviour. Beliefs are enduring indi-
vidual characteristics that shape behaviour and can be acquired through primary socialization.
Beliefs are also modifiable and can differentiate between individuals from the same background.
If persuasive techniques can be used to change behaviour-related beliefs and such interventions
result in behaviour change, this provides a theoretical and practical basis for evidence-based
health education.
The relationship between health beliefs and behaviours was conceptualized primarily in
terms of Lewin’s (1951) idea of ‘valence’. Particular beliefs were thought to make behaviours
more or less attractive. This resulted in an expectancy-value model of belief–behaviour relation-
ships in which events believed to be more or less likely were positively or negatively evaluated
by individuals. In particular, the likelihood of experiencing a health problem, the severity of the
consequences of that problem, and the perceived benefits of a preventive behaviour, in combi-
nation with its potential costs, were seen as key beliefs that shaped health-related behaviour
MHBK140-ch02_p30-69.indd 30 28/05/15 3:19 PM
HEALTH BELIEF MODEL 31
patterns. Early research found that these health beliefs were indeed correlated with differ-
ences in health-related behaviour patterns (referred to below as ‘health behaviours’ or ‘health
behaviour patterns’) and so could be used to differentiate between those who did and did not
undertake such behaviours. The model was initially applied to preventive behaviours but later
successfully extended to identify the correlates of health service usage and adherence to medi-
cal advice (Becker et al. 1977b).
Rosenstock (1974) attributed the first health belief model (HBM) research to Hochbaum’s
(1958) studies of the uptake of tuberculosis X-ray screening. Hochbaum found that perceived sus-
ceptibility to tuberculosis and the belief that people with the disease could be asymptomatic (mak-
ing screening beneficial) distinguished between those who had and had not attended for chest
X-rays. Similarly, a prospective study by Kegeles (1963) showed that perceived susceptibility to
the worst imaginable dental problems and awareness that visits to the dentist might prevent these
problems were useful predictors of the frequency of dental visits over the next three years. Haefner
and Kirscht (1970) took this research further by demonstrating that an HBM-based health educa-
tion intervention designed to increase participants’ perceived susceptibility, perceived severity,
and anticipated benefits resulted in a greater number of check-up visits to the doctor compared
with no intervention over an eight-month follow-up. Thus, by the early 1970s a series of studies
suggested that the HBM specified a series of key health beliefs that provided a useful framework
for understanding individual differences in health behaviour patterns and for designing behaviour
change interventions.
The HBM had the advantage of specifying a discrete set of common-sense beliefs that
appear to explain, or mediate, the effects of demographic variables on health behaviour pat-
terns and are amenable to change through educational intervention. The model could be applied
to a range of health behaviours and so provided a framework for shaping behaviour patterns
relevant to public health as well as training health care professionals to work from their patients’
subjective perceptions of illness and treatment. Consensus regarding the utility of the HBM was
important for public health research and, simultaneously, placed cognition modelling at the
centre of health service research.
The HBM was consolidated when Becker et al. (1977b) published a consensus statement
from the Carnegie Grant Subcommittee on Modification of Patient Behavior for Health Mainte-
nance and Disease Control. This paper considered a range of alternative approaches to under-
standing the social psychological determinants of health and illness behaviour and endorsed
the HBM framework. The components of the model were defined and further research on the
relationships between individual beliefs and health behaviours was called for.
2 Description of the model
The HBM focused on two aspects of individuals’ representations of health and health behav-
iour: threat perception and behavioural evaluation. Threat perception was construed as two
key beliefs: perceived susceptibility to illness or health problems, and anticipated severity of the
consequences of illnesses. Behavioural evaluation also consisted of two distinct sets of beliefs:
those concerning the benefits or efficacy of a recommended health behaviour, and those con-
cerning the costs of, or barriers to, enacting the behaviour. In addition, the model proposed that
cues to action can activate health behaviour when appropriate beliefs are held. These ‘cues’
MHBK140-ch02_p30-69.indd 31 28/05/15 3:19 PM
32 PREDICTING AND CHANGING HEALTH BEHAVIOUR
included a diverse range of triggers, including individual perceptions of symptoms, social influ-
ence, and health education campaigns. Finally, an individual’s general health motivation, or
‘readiness to be concerned about health matters’, was included in later versions of the model
(e.g. Becker et al. 1977b). There were therefore six distinct constructs specified by the HBM.
As Figure 2.1 indicates, there were no clear guidelines on how to operationalize the links
between perceived susceptibility, severity, and overall threat perception. Similarly, although
it was suggested that perceived benefits were ‘weighted against’ perceived barriers (Becker
et al. 1977b), no formula for creating an overall behavioural evaluation measure was developed.
Consequently, the model has usually been operationalized as a series of up to six separate inde-
pendent variables that potentially account for variance in health behaviours. Even the defini-
tion of these six constructs was left open to debate. Rosenstock (1974) and Becker and Maiman
(1975) illustrated how various researchers used somewhat different operationalizations of these
constructs and, in a meta-analysis of predictive applications of the HBM, Harrison et al. (1992)
concluded that this lack of operational homogeneity weakens the HBM’s status as a coherent
psychological model of the prerequisites of health behaviour. Nevertheless, a series of studies
has shown that these various operationalizations allowed identification of beliefs correlated
with health behaviours (e.g. Janz and Becker 1984).
3 Summary of research
3.1 Overview of HBM applications and research strategies
The HBM has been applied to the prediction of an impressively broad range of health behaviours
among a wide range of populations. Table 2.1 illustrates the range of behaviours that have been exam-
ined. Three broad areas can be identified: (1) preventive health behaviours, which include health-
promoting (e.g. diet, exercise) and health-risk (e.g. smoking) behaviours as well as vaccination and
DEMOGRAPHIC
VARIABLES
PSYCHOLOGICAL
CHARACTERISTICS
class, gender, age, etc.
Perceived susceptibility
Perceived severity
Health motivation
Perceived benefits
Perceived barriers Cues to action
Action
personality, peer
group pressure, etc.
Figure 2.1 The health belief model
MHBK140-ch02_p30-69.indd 32 28/05/15 3:19 PM
HEALTH BELIEF MODEL 33
Table 2.1 Illustrative applications of the HBM
Behaviour Researchers
Preventive behaviours
Screening
`¼ A?H?NC= Tay-Sachs trait: Becker et al. (1975); faecal occult blood: Hoogewerf et al.
(1990)
`¼ B?;FNB Hypertension: King (1984); cervical cancer: Orbell et al. (1995); colorectal
cancer: Rawl et al. (2001), Hay et al. (2003); mammography: Aiken et al.
(1994a, 1994b); STI test: Simon and Das (1984); HIV test: Dorr et al. (1999)
Risk behaviours
`¼ MGIECHA Penderson et al. (1982), Gianetti et al. (1985), Mullen et al. (1987), Stacy and
Lloyd (1990)
Exposure to environmental tobacco smoke: Li et al. (2003)
`¼ ;F=IBIF Por tnoy (1980), Beck (1981), Gottlieb and Baker (1986)
`¼ ?;NCHA¼G?;N Weitkunat et al. (2003)
Influenza vaccination Cummings et al. (1979), Oliver and Berger (1979), Rundall and Wheeler
(1979), Larson et al. (1982)
Breast self-examination Champion (1984), Calnan (1985), Owens et al. (1987), Ronis and Harel
(1989), Umeh and Rogan-Gibson (2001), Norman and Brain (2005)
Contraceptive use
(including condom use)
Eisen et al. (1985), Hester and Macrina (1985), Lowe and Radius (1987)
Abraham et al. (1992, 1996), Lollis et al. (1997), Adih and Alexander (1999),
Volk and Koopman (2001), Drayton et al. (2002), Winfield and Whaley (2002)
Diet and exercise Langlie (1977), Aho (1979a)
In relation to osteoporosis prevention: Silver Wallace (2002)
Dental behaviours
`¼ >?HN;F¼PCMCNM Kegeles (1963), Chen and Land (1986)
`¼ <LOMBCHA@FIMMCHA Chen and Tatsuoka (1984)
Others Cholera prevention: Ogionwo (1973); coronary heart disease prevention: Ali
(2002); osteoporosis prevention: Schmiege et al. (2007); safety helmet use:
Quine et al. (1998)
Sick role/adherence behaviours
Anti-hypertensive regimens Kirscht and Rosenstock (1977), Nelson et al. (1978), Taylor (1979), Hershey
et al. (1980)
Diabetic regimens Harris and Lynn (1985), Bradley et al. (1987), Brownlee-Duffeck et al. (1987),
Wdowik et al. (2001)
Renal disease regimens Heinzelmann (1962), Hartman and Becker (1978), Cummings et al. (1982)
Psychiatric regimens Kelly et al. (1987), Smith et al. (1999), Perkins (2002)
Parental adherence to
children’s regimens
Obesity: Becker et al. (1977b); rheumatic fever: Gordis et al. (1969); otitis medea:
Charney et al. (1967), Becker et al. (1972); asthma: Becker et al. (1978)
(Continued)
MHBK140-ch02_p30-69.indd 33 28/05/15 3:19 PM
34 PREDICTING AND CHANGING HEALTH BEHAVIOUR
contraceptive practices; (2) sick role behaviours, particularly adherence to recommended medical
regimens; and (3) clinic use, which includes physician visits for a variety of reasons.
Early HBM studies focused on prediction of preventive health behaviours. One of the first
reviews of research (Becker et al. 1977a) examined 20 studies, 13 of which were investigations of
preventive actions. These 13 studies examined seven distinct behaviours (X-ray screening for TB,
polio vaccination, influenza vaccination, use of safety gloves, pap test, preventive dental visits, and
screening for Tay-Sachs trait). In contrast, six of the remaining seven studies of sick role behav-
iours concerned adherence to penicillin prescriptions. When Janz and Becker reviewed the HBM
literature in 1984, smoking, alcohol use, dieting, exercise, and attendance at blood pressure screen-
ing had been added to the list of preventive behaviours examined from an HBM perspective. Studies
of sick role behaviours also increased to include adherence to regimens for hypertension, insulin-
dependent and non-insulin-dependent diabetes, end-stage renal disease, obesity, and asthma. Stud-
ies often examined a range of outcomes relevant to a particular regimen. For example, Cummings
and colleagues’ (1982) study of end-stage renal disease patients included measures of serum phos-
phorus and potassium levels, fluid intake, weight gain, and patients’ self-reports of diet and medica-
tion. Subsequent research has extended the range of behaviours examined to include contraceptive
use, including condom use, and personal dental behaviours such as teeth brushing and flossing,
as well as screening for faecal occult blood, colorectal cancer, and sexually transmitted diseases.
Many HBM predictive studies have employed cross-sectional designs, although Janz and
Becker’s (1984) review found that 40% (n = 18) of identified HBM studies were prospective.
Prospective studies are important because simultaneous measurement of health beliefs and
(especially self-reported) behaviour may be subject to memory and social desirability biases
and do not permit causal inferences (Field 2000). Most studies also have used self-report meas-
ures of behaviour but some have used physiological measures (e.g. Bradley et al. 1987), behav-
ioural observations (e.g. Alagna and Reddy 1984; Dorr et al. 1999; Hay et al. 2003) or medical
records (e.g. Orbell et al. 1995; Drayton et al. 2002) as outcome measures. While the majority of
measures of health beliefs employ self-completion questionnaires, structured face-to-face (e.g.
Cummings et al. 1982; Volk and Koopman 2001) and telephone (e.g. Grady et al. 1983) interviews
have also been employed. The use of random sampling techniques is commonplace and specific
representation of low-income and minority groups is also evident (e.g. Becker et al. 1974; Mullen
et al. 1987; Ronis and Harel 1989; Winfield and Whaley 2002).
Behaviour Researchers
Others Regimen for urinary tract infection: Reid and Christensen (1988); malaria
prophylaxis regimens: Abraham et al. (1999a), Farquharson et al. (2004)
Clinic use
Physician visits
`¼ JL?P?HNCP? Kirscht et al. (1976), Leavitt (1979), Berkanovich et al. (1981), Norman and
Conner (1993)
`¼ J;L?HN¼;H>¼=BCF> Becker et al. (1972, 1977a), Kirscht et al. (1978)
`¼ JMS=BC;NLC= Connelly et al. (1982), Connelly (1984), Rees (1986), Pan and Tantam (1989)
Table 2.1 Continued
MHBK140-ch02_p30-69.indd 34 28/05/15 3:19 PM
HEALTH BELIEF MODEL 35
Findings from research studies employing the HBM are reviewed below. We first examine
evidence for the predictive utility of the model’s four major constructs: susceptibility, severity,
benefits, and barriers. Second, we consider findings relating to cues to action and health moti-
vation, which have received more limited empirical attention. Third, we examine the issue of
combining health beliefs and the potential importance of interactions among beliefs. Finally, we
discuss the extent to which health beliefs have been successful in mediating the effects of social
structural variables and past behaviour.
3.2 Utility of perceived susceptibility, severity, benefit, and barrier constructs
Three quantitative reviews of research using the HBM with adults have been published (Janz
and Becker 1984; Harrison et al. 1992; Carpenter 2010). These reviews adopted different strate-
gies in quantifying findings from primary research studies. There is also a substantial literature
reporting applications of the HBM to children’s behaviour, which we do not discuss below (see,
for example, Gochman and Parcel 1982).
In their review, Janz and Becker (1984) employed a vote count procedure (see Cooper 1986:
36). A significance ratio was calculated ‘wherein the number of positive and statistically signifi-
cant findings for an HBM dimension are divided by the total number of studies which reported
significance levels for that dimension’. Janz and Becker’s significance ratios show the percent-
age of times each HBM construct was statistically significant in the predicted direction across
46 studies. Across all studies, the significance ratios are very supportive of HBM predictions.
Susceptibility was significant in 81% (30/37) of studies, severity in 65% (24/37), benefits in 78%
(29/37), and barriers in 89% (25/28). When prospective studies only (n = 18) were examined,
findings appeared to confirm a predictive role for these health beliefs. The ratios were 82%,
65%, 81%, and 100% for susceptibility, severity, benefits, and barriers based on 17, 17, 16, and 11
studies, respectively. Results show that barriers are the most reliable predictor of behaviour,
followed by susceptibility and benefits, and finally severity.
Figure 2.2 presents significance ratios separately for preventive, sick role, and clinic utili-
zation behaviours based in each case on the number of studies examined by Janz and Becker.1
Across 24 studies of preventive behaviours, barriers were significant predictors in 93% of hypothe-
ses, susceptibility in 86%, benefits in 74%, and severity in 50%. Barriers were also the most frequent
predictor in 19 studies of sick role behaviours (92%), with severity second (88%) followed by ben-
efits (80%) and susceptibility (77%). Janz and Becker only included three clinic use studies in their
review. Benefits were significant in all studies, susceptibility was significant in two of three, and
severity was significant in one of three. Barriers were significant in one of the two studies of clinic
use that examined this component. It is interesting to note that while severity has only a moderate
effect upon preventive behaviour or clinic utilization, it is the second most powerful predictor of
sick role behaviour. Janz and Becker suggest that these differences might be due to respondents’
difficulty in conceptualizing this component when they are asymptomatic or when the effects of
the health threat are unfamiliar or only occur in the long term.
Janz and Becker’s findings appear to provide strong support for the HBM across a range
of behaviours, although limitations of the vote count procedure suggest caution in interpreting
these results. The significance ratios only reveal how often HBM components were significantly
associated with behaviour, not how large the effects of HBM measures were on outcomes (e.g.
behaviour). Moreover, significance ratios give equal weighting to findings from studies with
MHBK140-ch02_p30-69.indd 35 28/05/15 3:19 PM
36 PREDICTING AND CHANGING HEALTH BEHAVIOUR
large and small numbers of participants, and do not differentiate between bivariate relation-
ships between an HBM construct and behaviour and multivariate associations. In addition, Janz
and Becker’s analysis did not properly control for multiple behavioural outcomes.
Harrison and colleagues’ (1992) meta-analytic review of the HBM addressed these meth-
odological issues. Harrison et al. originally identified 234 published empirical tests of the HBM.
Of these, only 16 studies (6.8%) measured all four major components and included reliability
checks. This clearly demonstrates the extent to which operationalizations of the HBM have failed
to measure all constructs or provide psychometric tests of measures (see Conner 1993). The
meta-analysis involved converting associations between HBM constructs and behaviour meas-
ures, in each study, into a common effect size, namely Pearson’s r. A weighted average of these
effect sizes was then computed for each component (see Rosenthal 1984). Figure 2.3 shows that,
across all studies, the average correlations between HBM components and behaviour were 0.15,
0.08, 0.13, and 0.21 for susceptibility, severity, benefits, and barriers, respectively. While these
correlations are statistically significant they are all small, with individual constructs account-
ing for between just 0.5% and 4% of the variance in behaviour across studies. Unlike Janz and
Becker (1984), Harrison et al. found that HBM components had different associations in cross-
sectional versus longitudinal designs. Both benefits and barriers had significantly larger effect
sizes in prospective than in retrospective studies, whereas in the case of severity, the effect size
was significantly larger in retrospective studies.
Updating Harrison and colleagues’ work, Carpenter (2010) conducted a meta-analysis of
18 studies in which HBM measures had been used to prospectively predict health behaviours,
including 12 studies published after Harrison et al. (1992) and involving 2702 respondents. The
studies reviewed focused on a range of health behaviours, including screening attendance, den-
tal care, condom use, smoking cessation, and diet and exercise; the predictive period ranged
between 2 and 365 days. As can be seen from Table 2.2, apart from perceived susceptibility,
Figure 2.2 Significance ratios for HBM constructs for preventive, sick role and clinic use behaviours (after
Janz and Becker 1984)
100
80
60
40
20
0
Susceptibility
Significance ratio (%)
Severity Benefits Barriers
Preventive
Sick role
Clinic use
MHBK140-ch02_p30-69.indd 36 28/05/15 3:19 PM
HEALTH BELIEF MODEL 37
Carpenter found somewhat higher average correlations between health beliefs and measures
of behaviour than did Harrison et al. (0.05, 0.15, 0.27, and 0.30, respectively), suggesting that
associations between benefits and barriers and health-related behaviours may be stronger than
indicated by Harrison et al. Carpenter also noted that these correlations were higher when the
time interval between belief and behavioural measurement was shorter, and that correlations
varied across health behaviours. Overall, all three of the reviews considered here confirm a con-
sistent pattern of small, significant correlations between HBM-specified beliefs and measures of
health-related behaviours varying across behavioural measures and follow-up periods.
0.2
0.1
0.0
–0.1
–0.2
–0.3
–0.4 Susceptibility
Effect size (r)
Severity Benefits Barriers
All studies
Prospective
Retrospective
Figure 2.3 Effect sizes for HBM constructs for prospective and retrospective studies (after Harrison
et al. 1992)
Table 2.2 Average correlations between HBM-specified beliefs and health behaviours
Review Susceptibility Severity Benefits Barriers
Harrison et al. (1992) 0.15 0.08 0.13 0.21
Carpenter (2010) 0.05 0.15 0.27 0.30
3.3 Utility of cues to action and health motivation constructs
Cues to action and health motivation have been relatively neglected in empirical tests of the
HBM. Janz and Becker (1984), Harrison et al. (1992), and Carpenter (2010) did not include these
components in their reviews because of the paucity of studies including these measures. One
reason for researchers’ failure to measure these components may be the lack of clear construct
definitions. Unlike the four key health beliefs included in the HBM, ‘cues to action’ can include
a wide range of experiences and contexts and so have been operationalized differently by dif-
ferent researchers. For example, Grady et al. (1983) found significant associations between
the numbers of family members with breast and other cancers and participation in a breast
MHBK140-ch02_p30-69.indd 37 28/05/15 3:19 PM
38 PREDICTING AND CHANGING HEALTH BEHAVIOUR
self-examination teaching programme. These authors did not, however, refer to these measures
as ‘cues to action’, while an almost identical measure in Keesling and Friedman’s (1987) study of
skin cancer prevention was categorized as a ‘cue to action’.
Physicians’ advice or recommendations have been found to be successful cues to action in
the contexts of smoking cessation (Weinberger et al. 1981; Stacy and Lloyd 1990) and flu vac-
cination (Cummings et al. 1979). Postcard reminders have also been successful (e.g. Larson
et al. 1982; Norman and Conner 1993), although the effect of other media cues to action is less
clear. While Ogionwo (1973) found that a radio, film, and poster campaign was successful in
attempts to prevent cholera, Bardsley and Beckman (1988) reported a negative effect of an
advert for alcoholism treatment. Mullen et al. (1987) found no effect for memory of a mass media
campaign upon smoking, while Li et al. (2003) found that reported exposure to anti-smoking
campaigns on radio, TV, and billboards was not associated with young people’s exposure to
tobacco smoke. Knowing someone who is HIV positive or has AIDS has not been predictive of
behavioural change among gay men (e.g. McCuskar et al. 1989; Wolcott et al. 1990), and Winfield
and Whaley (2002) found that a multi-item scale, including assessment of knowledge of others
with HIV/AIDS, previous discussion of HIV/AIDS, and exposure to HIV/AIDS campaigns, was not
significantly correlated with condom use among African American college students. Similarly,
Umeh and Rogan-Gibson (2001) found that a multi-item cues-to-action scale that included social
pressure, recommendations from health care professionals, family experiences, and physical
symptoms was not associated with reported breast self-examination. However, Aho (1979b)
found that knowing someone who had experienced negative side-effects from influenza vac-
cination was negatively related to inoculation behaviour. Perhaps unsurprisingly, measures of
‘internal’ cues to action, namely the presence or intensity of symptoms, have also been generally
predictive of behaviour (King 1984; Harris and Lynn 1985; Kelly et al. 1987).
Measures of ‘health motivation’ have generally been single questionnaire items, usually
expressing general ‘concern’ about health, although a few researchers have developed psycho-
metric scales (e.g. Maiman et al. 1977; Champion 1984; Umeh and Rogan-Gibson 2001). Bivariate
relationships between health motivation and health behaviour are generally small but statistically
significant (e.g. Ogionwo 1973; Berkanovich et al. 1981; Champion 1984; Casey et al. 1985; Ali
2002), with some non-significant exceptions (e.g. Harris and Guten 1979; Rayant and Sheiham
1980; Umeh and Rogan-Gibson 2001). Findings from multivariate analyses are mixed, with some
studies finding positive relationships (e.g. Portnoy 1980; Thompson et al. 1986; Ali 2002) and others
finding no association (e.g. King 1982; Wagner and Curran 1984). Health motivation has also been
used as an outcome in predictive studies using the HBM when HBM measures are used to predict
intentions (e.g. Petosa and Kirby 1991). Overall, however, the primary challenge in operationalizing
‘cues to action’ and ‘health motivation’ either as explanatory measures or intervention targets is
that their broad definitions may mean that they can refer to quite different modifiable psychologi-
cal factors across different populations, behaviour patterns, and contexts.
3.4 Combining HBM components
Failures to operationalize the HBM in its entirety may be partly due to the early suggestion
that susceptibility and severity could be combined under a single construct, that is ‘threat’, and
similarly, that benefits and barriers should be subtracted from one another rather than treated
as separate constructs (Becker and Maiman 1975). Some researchers have used a threat index
MHBK140-ch02_p30-69.indd 38 28/05/15 3:19 PM
HEALTH BELIEF MODEL 39
rather than measure susceptibility and severity separately (e.g. Kirscht et al. 1976). This appears
to violate the expectancy-value structure of the HBM and so can be seen as an inferior, and per-
haps incorrect, operationalization of the model (see Feather 1982).
While most HBM studies measure benefits and barriers, some researchers have also com-
bined benefits and barriers in a single index (e.g. Oliver and Berger 1979; Gianetti et al. 1985).
This practice raises theoretical and empirical issues. At a theoretical level, Weinstein (1988) sug-
gests that there is a qualitative difference between benefits and barriers, at least in hazardous
situations, which means that they should be treated as distinct constructs. For example, while
barriers relating to taking exercise or giving up salt are certain and concrete (e.g. time and
effort, loss of pleasure), the benefits in terms of avoiding hypertension are more hypothetical.
At an empirical level, the benefits construct may comprise distinct components, namely the
efficacy of the behaviour in achieving an outcome (response efficacy) as well as possible psy-
chosocial benefits such as social approval. Similarly, the barriers construct may comprise both
physical limitations on performing a behaviour (e.g. expense) and psychological costs associ-
ated with its performance (e.g. distress). It is unlikely that a single index could adequately rep-
resent these different outcome expectancies. An empirical approach to resolving this issue is to
employ factor and reliability analyses to assess whether, and which, benefits and barriers can be
legitimately combined, from a psychometric perspective (e.g. Abraham et al. 1992).
A separate issue concerns whether susceptibility and severity scores should combine addi-
tively or multiplicatively as the HBM’s expectancy-value structure would suggest. Rogers and
Mewborn (1976) have investigated this issue experimentally from a protection motivation theory
perspective. These researchers found no support for the predicted susceptibility × severity inter-
action (see also Weinstein 1982; Maddux and Rogers 1983; Rogers 1983; Ronis and Harel 1989). In a
rare HBM study addressing this question, Lewis (1994) noted that the severity manipulation check
in Rogers and Mewborn’s study was not successful, so their data did not represent a useful test
of the interaction hypothesis. Lewis’s data found no support for the interaction hypothesis using
parametric and non-parametric statistical tests on retrospective data. However, in a prospective
study employing a small sample, the susceptibility × severity interaction contributed a significant
proportion of unique variance (sr2 = 0.12, p < 0.05). Lewis suggests that the equation
threat = susceptibility + (susceptibility × severity)
may better represent the effects of the severity component, at least for some health behaviours,
than a simple additive model. Kruglanski and Klar (1985) and Weinstein (1988) concur, suggest-
ing that severity must reach a certain magnitude to figure in health decisions, but once that
magnitude has been reached, decisions are based solely on perceived susceptibility. The rela-
tively poor findings for the severity component in quantitative reviews appear to support these
interpretations, although further research on this issue is required.
3.5 Utility of HBM components in mediating the impact of past experience or social
structural position
In a useful review of literature on the impact of past experience of a behaviour upon its future
performance, Sutton (1994) points out that almost all health behaviours are capable of being
repeated. Janz and Becker (1984: 44) acknowledge that ‘some behaviours (e.g. cigarette smok-
ing, tooth-brushing) have a substantial habitual component obviating any ongoing psychosocial
MHBK140-ch02_p30-69.indd 39 28/05/15 3:19 PM
40 PREDICTING AND CHANGING HEALTH BEHAVIOUR
decision-making process’, but do not address the question of whether health beliefs might have
a role in breaking unhealthy habits. While the issue of whether cognitions mediate the effects of
past experience has been a central concern of researchers using the theory of reasoned action
(see Bentler and Speckart 1979), few HBM studies measure past behaviour.
Some researchers using HBM have explicitly addressed this mediation hypothesis. In a pro-
spective study, Cummings et al. (1979) found both direct and indirect effects for ‘past experi-
ence with flu shots’ upon subsequent inoculation behaviour. Perceived efficacy of vaccination
(a benefit) and the behavioural intention construct of the theory of reasoned action (Fishbein
and Ajzen 1975) were both partial mediators of the effects of experience. Similarly, Norman
and Brain (2005) found that perceived susceptibility and barriers both partially mediated the
effects of past behaviour on subsequent breast self-examination. Two studies by Otten and van
der Pligt (1992) tested whether perceived susceptibility mediated the relationship between past
and future preventive health behaviours. While past behaviour was predictive of susceptibility
assessments and a proxy measure of future behaviour (behavioural expectation; Warshaw and
Davis 1985), susceptibility was negatively associated with expectation and did not mediate the
effects of past behaviour. Otten and van der Pligt’s (1992) studies underline the need for further
longitudinal research on this issue.
Another important, and neglected, issue concerns the ability of HBM components to medi-
ate the effects of social structural position upon health behaviour. Cummings et al. (1979) found
that socio-economic status (SES) was not related to health beliefs, although both SES and
beliefs were significantly related to inoculation behaviour in bivariate analyses. Orbell et al.
(1995), on the other hand, found that perceived susceptibility and barriers entirely mediated the
effects of social class upon uptake of cervical screening. Direct effects were, however, obtained
for both marital status and sexual experience. Salloway et al. (1978) obtained both direct and
indirect effects of occupational status, sex, and income and an indirect effect of education upon
appointment-keeping at an inner-city hypertension clinic (see also Chen and Land 1990; Sullivan
et al. 2004).
Salloway et al. (1978) are critical of Rosenstock’s (1974) contention that the HBM may be
more applicable to middle-class samples because of their orientation towards the future, deliber-
ate planning, and deferment of immediate gratification. Salloway et al. (1978: 113) point out that
working-class people ‘are subject to real structural barriers and constrained by real differences
in social network structure which are not present in middle-class populations’. Further research
is needed to determine the impact of SES upon health beliefs and their relationship to behaviour,
and to discriminate between the effects of cognitions and the effects of factors such as financial
constraints, culture of poverty/network effects, and health system/provider barriers upon the
likelihood of health behaviours (Rundall and Wheeler 1979).
4 Developments
Recognizing limitations of the HBM, Rosenstock (1974) suggested that a more comprehensive
model of cognitive antecedents could reveal how health beliefs are related to other psychologi-
cal stages in decision-making and action. King (1982) demonstrated how this might be achieved
by ‘extending’ the HBM in a study of screening for hypertension. She included measures of
individuals’ causal understanding of high blood pressure derived from ‘attribution theory’
MHBK140-ch02_p30-69.indd 40 28/05/15 3:19 PM
HEALTH BELIEF MODEL 41
(Kelley 1967), which she theorized as determinants of health beliefs, which, in turn, prompted
intention formation (Fishbein and Ajzen 1975), a more immediate cognitive antecedent of action.
Using a prospective design, King found that eight measures, including intention, could correctly
classify 82% of respondents as either attenders or non-attenders. She also reported that four
measures (perceived severity, two measures of perceived benefits, and the extent to which
respondents identified one or many causes of high blood pressure) accounted for 18% of the
variance in behavioural intention, which was the best single predictor of attendance. This study
is noteworthy because it combined constructs from a number of theories (attribution theory, the
HBM, and the theory of reasoned action) and created a new model that simultaneously explored
the cognitive foundations of health beliefs and sketched a mechanism by which they might gen-
erate action. King’s research is a good example of how pathways between cognition measures
may be empirically examined to provide evidence relating to psychological processes rather
than static belief strengths and valences. Unfortunately, studies of this kind are rare in HBM
research (but see Quine et al. 1998; Abraham et al. 1999a). This failure to extend the model has
distanced it from theoretical advances in social cognition research, and later attempts to situate
health beliefs in more comprehensive models of the cognitive antecedents of action have tended
to abandon the HBM structure in favour of new conceptual frameworks (e.g. protection motiva-
tion theory; Prentice-Dunn and Rogers 1986).
By 1980, work on ‘locus of control’ by Rotter (1966) and Wallston and Wallston (1982) and,
more importantly, ‘perceived self-efficacy’ by Bandura (1977) had established perceived con-
trol as an important determinant of health behaviour. King (1982) included a measure of per-
ceived control derived from attribution theory, which was found to predict attendance. Later,
Janz and Becker (1984) also recognized the importance of perceived control but speculated
that it might be thought of as a component of perceived barriers rather than an additional
theoretical construct. Consequently, the HBM remained unmodified, whereas Ajzen added per-
ceived behavioural control to the theory of reasoned action to re-launch it as the theory of
planned behaviour (TPB; Ajzen and Madden 1986; Ajzen 1998). Two years later, Rosenstock
et al. (1988) acknowledged that Janz and Becker (1984) underestimated the importance of
self-efficacy and proposed that it be added to the HBM. Subsequent studies have tested the
predictive utility of an extended HBM, including self-efficacy, and generally confirmed that
self-efficacy is a useful additional predictor (e.g. Silver Wallace 2002; Wallace 2002; Hay et al.
2003; Norman and Brain 2005; Schmiege et al. 2007). When floor or ceiling effects are observed,
such as when participants are uniformly confident that they can take action, self-efficacy may
not provide additional discrimination (e.g. Weitkunat et al. 2003) between those who act and
those who do not. Often, however, self-efficacy is a useful additional predictor, partly because
it reflects individuals’ views of the behaviour being predicted rather perceived health risk. For
example, Schmiege et al. (2007) found that health beliefs about current preventive behaviours,
including perceived barriers and self-efficacy, were more powerful predictors of calcium con-
sumption and weight-bearing exercise, at six months follow-up, than beliefs about potential
negative health outcomes.
Unfortunately, unlike King (1982), Rosenstock et al. (1988) offered no new theoretical for-
mulation for considering self-efficacy. They suggested that self-efficacy could be added to the
other HBM constructs without elaboration of the model’s theoretical structure. This may have
been short-sighted because subsequent research indicated that key HBM constructs have indi-
rect effects on behaviour as a result of their effect on perceived control and intention, which
MHBK140-ch02_p30-69.indd 41 28/05/15 3:19 PM
42 PREDICTING AND CHANGING HEALTH BEHAVIOUR
may, therefore, be regarded as more proximal determinants of action (Schwarzer 1992; Abraham
et al. 1999a). For example, Schmiege et al. (2007) found that intention measures entirely medi-
ated the effects of health beliefs on behaviour.
A number of researchers have included HBM-specified health beliefs in more comprehen-
sive models of the cognitive antecedents of action. For example, Schwarzer’s (1992) ‘health
action process approach’ combines constructs from the HBM with those from other social cogni-
tive models (see Schwarzer and Luszczynska, Chapter 8 this volume). Susceptibility and sever-
ity beliefs are construed as antecedents of outcome expectancies and intention, and intention
and self-efficacy are identified as more proximal antecedents of action. Abraham et al. (1999a)
found that including health beliefs (concerning perceived susceptibility and perceived side-
effects) in a TPB model helped identify key cognitive antecedents of the intention to adhere to
malaria prophylaxis after returning from a malarious region. Like King, Abraham et al. found
that specific health beliefs enhanced the prediction of intention and that intention was the
strongest predictor of adherence. Intention mediated the effects of most cognitions on behav-
iour, although among a sample taking a drug known to have serious side-effects, perceived
side-effects added to the variance explained in adherence, after controlling for the effects of
intention. Jones et al. (2001) found that an HBM-derived measure of perceived threat contrib-
uted to the prediction of intention to use sunscreen in a model that also included measures of
knowledge, norms, importance of short-term negative consequences, and self-efficacy. Such
research suggests that health beliefs may be more usefully construed as cognitive antecedents
of self-efficacy and intention, rather than direct antecedents of action. However, in certain
circumstances, particularly salient beliefs about a procedure or medication (e.g. beliefs about
perceived side-effects) may enhance the prediction of behaviour, controlling for the effects of
intention and self-efficacy.
Correspondence between social cognition models has been recognized for some time. For
example, Kirscht (1983: 287) noted that HBM constructs could be ‘mapped onto’ the theory of
reasoned action. Prompted by efforts to promote HIV-preventive behaviours, leading theorists
held a workshop in 1991 to ‘identify a finite set of variables to be considered in any behavioural
analysis’ (Fishbein et al. 2001: 3). Theorists included Fishbein (theory of planned behaviour; e.g.
Fishbein and Ajzen 1975), Bandura (social cognitive theory and self-efficacy; e.g. Bandura 1977),
and Becker (health belief model; e.g. Becker et al. 1977a, 1977b). They identified eight core
variables important to explaining behaviour and promoting behaviour change. Three variables
were regarded as necessary and sufficient prerequisites, namely a strong intention, the neces-
sary skills, and the absence of environmental constraints that prevent the specified actions. In
addition, five further antecedents were identified as determinants of intention strength: self-
efficacy, the belief that advantages (e.g. benefits) outweigh disadvantages (e.g. costs), the per-
ception of greater social (normative) pressure to perform the behaviour than not to perform the
behaviour, the belief that the action is consistent with the person’s self-image, and anticipation
of a more positive than negative emotional reaction to undertaking the specified action(s). This
framework maps the main constructs from the HBM onto the more general attitude and nor-
mative constructs included in the theory of reasoned action. Beliefs about the seriousness of
a health threat, personal susceptibility to the threat, efficacy of medication, and side-effects
of medication are all construed as perceived advantages and disadvantages of action, which
are, in turn, determinants of strength of intention. This produces a logical and parsimonious
framework that incorporates the two-stage model of decision-making and action inherent in the
MHBK140-ch02_p30-69.indd 42 28/05/15 3:19 PM
HEALTH BELIEF MODEL 43
theory of reasoned action (Fishbein and Ajzen 1975) – that is, intention predicts behaviour and
is predicted by a series of other cognitive antecedents. This amalgamation of predictive models
is a helpful development for researchers wishing to apply social cognition models. However,
researchers should continue to explore and measure health beliefs because research has shown
that such beliefs can explain additional variance in intention beyond that engendered by gen-
eral measures of attitude (e.g. semantic differential measures; Fishbein et al. 2001: 11). Overall,
however, with some exceptions, it may be prudent to regard many HBM-specified beliefs as
antecedents of intention, rather than predictors of behaviour.
5 Operationalization of the model
Below we outline the steps involved in developing an HBM questionnaire. We briefly review
available instruments and analyse a study by Champion (1984) in which she developed health
belief scales to investigate the frequency of breast self-examination. Determination of reliability
and validity of scales is addressed in some depth. Finally, we identify some conceptual difficul-
ties with HBM components and briefly address problems of response bias.
5.1 Developing an HBM questionnaire
Formulating hypotheses or research questions clearly so that they translate into relationships
between variables, defining an appropriate sample, gaining access to that sample, and deciding
the mode of data collection (e.g. pencil-and-paper test or telephone interview) are generally
prerequisites of instrument development. There are two ways to determine the content of the
items of the questionnaire. The first is to conduct a literature search for previous HBM studies in
the area and determine whether previous instruments are published or available from authors.
Scales should be checked to determine whether internal reliability is satisfactory and whether
the scale has face validity (respondents believe that the scale measures what it says it does). A
scale obtained in this way might be used in its entirety but may require modification if it is to be
used with a different sample.
The HBM scales from the Standardized Compliance Questionnaire (Sackett et al. 1974) have
been modified for use in a variety of settings (e.g. Cerkoney and Hart 1980; Bollin and Hart
1982; Connelly 1984) but this instrument may be difficult to obtain and other scales have also
been employed. For example, Calnan (1984) and Hallal (1982) employed measures derived from
Stillman’s (1977) research on breast cancer. Fincham and Wertheimer (1985) used items derived
from Leavitt (1979) in their study of uptake of prescriptions, while Hoogewerf et al. (1990)
examined compliance with genetic screening using items from Halper et al. (1980). There are
also published HBM scales in the areas of compliance with hypertension regimens (Abraham
and Williams 1991), children’s obesity regimens (Maiman et al. 1977), breast self-examination
(Champion 1984), and other behaviours.
If no appropriate, previously developed HBM measures are available, researchers must
develop their own (for a general guide to scale development, see DeVellis 1991). A useful exam-
ple of this process is provided by Champion’s (1984) study of breast self-examination The first
step involves generating items that purport to measure HBM components (the item pool). Again,
previous HBM studies can be used as a guide. It is good practice, however, for researchers to
MHBK140-ch02_p30-69.indd 43 28/05/15 3:19 PM
44 PREDICTING AND CHANGING HEALTH BEHAVIOUR
conduct semi-structured interviews with 20 or 30 potential respondents in order to determine
respondents’ perceptions of the health threat and beliefs about the behaviour in an open-ended
manner. This process will ensure that questionnaire items are salient to the population of inter-
est and will provide guidance on how well respondents will understand medical terminology
and other terms. Identification of sample relevant benefits, barriers, and cues to action is likely
to provide better behavioural predictors than researcher-imposed conceptualizations. Relevant
experts can also be used to develop and select items.
Champion initially developed 20–24 items for each HBM component (excluding cues to
action) but then retained only those items that at least six of eight judges (faculty and doctoral
students knowledgeable about the HBM) agreed represented the constructs in question. Ran-
dom presentation of items to judges allowed assessment of the content (or face) validity of each
scale – that is, the extent to which the items accurately and adequately reflected the content of
HBM constructs.
The next step in developing the instrument is the pilot study. While a small number of
studies in the literature report pilots of the instruments employed in the main study (e.g. Eisen
et al. 1985; Orbell et al. 1995), these, unfortunately, are exceptions rather than the rule, and
this lack of piloting may help to explain some of the difficulties with previous research using
the HBM. Champion’s pilot questionnaires included the items judged to have good content
validity (10–12 items for each construct) and employed a 5-point Likert scale for responses
(where 5 = ‘strongly agree’ and 1 = ‘strongly disagree’). The questionnaires were posted to a
convenience sample of women together with a prepaid return envelope. Three hundred and
one women participated.
Reliability and validity analyses constitute the final step in determining scale items. When
a scale has high reliability, random measurement error is low and the items can be viewed as
indices of one underlying construct. Scale reliability can be assessed using Cronbach’s alpha
coefficient (Cortina 1993) or the Spearman–Brown formula (Rust and Golombok 1990). One
can determine error over time by correlating scores on the same scale at different time points,
for example, two weeks apart. Champion determined alpha coefficients for each HBM compo-
nent, dropping items that reduced the reliability of the scale. While coefficients for three con-
structs exceeded the generally accepted level of 0.70 (susceptibility = 0.78, severity = 0.78, and
barriers = 0.76), the reliabilities for benefits and health motivation were weaker (0.62 and 0.61,
respectively). Two weeks after the original questionnaires were distributed, these revised scales
were sent to a subsample that had agreed to take part in a further study. Correlations were
computed between scores on the scales at these two time-points. These test–retest correlations
were satisfactory (> 0.70) for four of the five components (susceptibility = 0.86, severity = 0.76,
benefits = 0.47, barriers = 0.76, health motivation = 0.81).
The construct validity of the scales (the extent to which scales measure what they are
designed to measure) was next determined by factor analysing all of the item scores. This sta-
tistical procedure sorts individual items into groupings or factors on the basis of correlations
between items. Factor analysis showed that, with one exception, items all loaded on the fac-
tors (HBM constructs) they were originally assigned to, demonstrating satisfactory construct
validity. Criterion validity was also determined by demonstrating that the HBM measures were
significantly related to previous practice of breast self-examination. Table 2.3 presents the items
used to measure the susceptibility, severity, benefits and barriers constructs, following the reli-
ability and validity checks.
MHBK140-ch02_p30-69.indd 44 28/05/15 3:19 PM
HEALTH BELIEF MODEL 45
Table 2.3 Items representing susceptibility, severity, benefits, and barriers components in a study of breast
self-examination (Champion 1984)
HBM constructs, items, and reliability
Susceptibility
1. My chances of getting breast cancer are great.
2. My physical health makes it more likely that I will get breast cancer.
3. I feel that my chances of getting breast cancer in the future are good.
4. There is a good possibility that I will get breast cancer.
5. I worry a lot about getting breast cancer.
6. Within the next year I will get breast cancer.
Cronbach’s alpha = 0.78
Severity
1. The thought of breast cancer scares me.
2. When I think about breast cancer I feel nauseous.
3. If I had breast cancer, my career would be endangered.
4. When I think about breast cancer, my heart beats faster.
5. Breast cancer would endanger my marriage (or a significant relationship).
6. Breast cancer is a hopeless disease.
7. My feelings about myself would change if I got breast cancer.
8. I am afraid to even think about breast cancer.
9. My financial security would be endangered if I got breast cancer.
10. Problems I would experience from breast cancer would last a long time.
11. If I got breast cancer, it would be more serious than other diseases.
12. If I had breast cancer, my whole life would change.
Cronbach’s alpha = 0.70
Benefits
1. Doing self-breast exams prevents future problems for me.
2. I have a lot to gain by doing self-breast exams.
3. Self-breast exams can help me find lumps in my breast.
4. If I do monthly breast exams, I may find a lump before it is discovered by regular health exams.
5. I would not be so anxious about breast cancer if I did monthly exams.
Cronbach’s alpha = 0.61
Barriers
1. It is embarrassing for me to do monthly breast exams.
2. In order for me to do monthly breast exams, I have to give up quite a bit.
3. Self-breast exams can be painful.
4. Self-breast exams are time-consuming.
5. My family would make fun of me if I did self-breast exams.
6. The practice of self-breast exams interferes with my activities.
7. Doing self-breast exams would require starting a new habit, which is difficult.
8. I am afraid I would not be able to do self-breast exams.
Cronbach’s alpha = 0.76
MHBK140-ch02_p30-69.indd 45 28/05/15 3:19 PM
46 PREDICTING AND CHANGING HEALTH BEHAVIOUR
Although there were some difficulties with Champion’s (1984) analyses, this paper provides
an example of good practice in the design of a study applying the HBM. Champion rightly con-
trasts her own study with previous research, pointing out that the validity and reliability of HBM
measures has rarely been tested, that multiple-item measures are not routinely employed, that
operational definitions vary across studies, and that the use of response options that generate
nominal or ordinal (as opposed to interval) data limits statistical exploration of relationships
between measures.
5.2 Problems of operationalization: conceptual difficulties with HBM components
Champion (1984) describes regression of a measure of breast self-examination practice on HBM
components as evidence for construct validity. We would argue that these data are indicative of
criterion validity. There is also some difficulty with interpretation of the factor analysis in that
a three-factor solution for the perceived severity component was not pursued. An item relat-
ing to having ‘relatives and friends with breast cancer’ (p. 83) was not interpreted as a cue to
action, so this component of the HBM was ignored in Champion’s analysis. Finally, further item
development on the benefits component should properly have been conducted to improve its
poor reliability. Both the methodological problems revealed by consideration of Champion’s
(1984) paper and the heterogeneity of effect sizes obtained by Harrison et al. (1992) and Car-
penter (2010) highlight difficulties with the conceptual definition of HBM constructs. A variety
of theorists have drawn attention to problematic assumptions inherent in the HBM, including
the assumption that HBM constructs are unidimensional and that relationships between HBM
constructs and behaviour are fixed and linear. In this sub-section, we briefly review theoretical
issues relevant to the conceptualization of each of the HBM constructs.
5.2.1 Susceptibility
Becker and Maiman (1975: 20) acknowledge the wide variety of operationalizations of suscep-
tibility:
Hochbaum’s questions apparently emphasized the concept of perceived possibility of
contracting the disease; Kegeles’ questions were directed at the probability of becom-
ing ill; Heinzelmann requested estimates of likelihood of recurrence, while Elling et al.
asked for similar re-susceptibility estimates from the mother concerning her child; and
Rosenstock introduced ‘self-reference’ versus ‘reference to men (women) your age’
(as well as ‘fixed-alternative’ versus ‘open-ended’ items). [original emphasis]
Tversky and Kahneman (1981) showed that even quite small changes in the wording of risk
choices have significant and predictable effects upon responses. Thus, considerable care needs
to be taken in the phrasing of items measuring perceived susceptibility, and multi-item measures
are essential.
People may employ cognitive heuristics in their susceptibility judgements. Slovic et al.
(1977) pointed out that, in general, people seem to overestimate the frequency of rare causes
of death and underestimate common causes of death. In particular, events that are dramatic or
personally relevant, and therefore easy to imagine or recall, tend to be overestimated. There is
also a tendency for people to underestimate the extent to which they are personally vulnerable
MHBK140-ch02_p30-69.indd 46 28/05/15 3:19 PM
HEALTH BELIEF MODEL 47
to health and life-threatening problems. Weinstein (1980) has termed this phenomenon ‘unrealis-
tic optimism’. This sense of unique invulnerability has been demonstrated in the context of both
relative risk comparisons of self to others (e.g. Weinstein 1984) and subjective versus objective
risk appraisals (Gerrard and Warner 1991). Cognitive factors, including perceptions of control,
egocentric bias, personal experience, and stereotypical beliefs, have been posited as explana-
tions for this tendency, as well as motivational factors, including self-esteem maintenance and
defensive coping (see Van der Pligt et al. 1993). The impact of these cognitive and motivational
processes on risk estimation may help to explain the small effect sizes obtained for associations
between perceived susceptibility and health-protective behaviours (Harrison et al. 1992).
Weinstein (1988) has also drawn attention to other difficulties with the HBM conceptualiza-
tion of susceptibility. He suggests that beliefs about susceptibility should be characterized in
terms of three stages. The first stage involves the awareness that the health threat exists. The
second stage involves determining how dangerous the threat is and how many people are likely
to be affected. This is inevitably an ambiguous question and many people will display unreal-
istic optimism at this stage. Only in the final stage, when the threat has been personalized, will
personal susceptibility be acknowledged. This processual account of risk perception implies
that susceptibility levels are likely to change over time as populations are influenced by health
education and that, consequently, the point at which susceptibility is measured may determine
the strength of its association with subsequent health behaviour.
The interpretation of correlations between perceived susceptibility and health behaviour
may also be problematic in cross-sectional studies because both positive and negative asso-
ciations between risk and behaviour are easily interpreted. For example, suppose someone
believes he or she is at risk of HIV infection and therefore decides that he or she will use a con-
dom during sex. In this case, high susceptibility leads to safer behaviour and so the correlation is
positive. The same person, having adopted consistent condom use, however, may now estimate
his or her risk of infection as low. In this case, protective behaviour leads to lowered suscep-
tibility, resulting in a negative correlation. Cross-sectional data do not allow determination of
the causal relationships between beliefs and behaviour and vice versa. In a review of this issue,
Weinstein and Nicolich (1993: 244) concluded that: ‘the correlation between perceived personal
risk and simultaneous preventive behaviors should not be used to assess the effects of percep-
tions on behavior. It is an indicator of risk perception accuracy.’ Gerrard et al. (1996) supported
this conclusion by examining four prospective studies that measured perceived susceptibility to
HIV infection and subsequent safer sexual behaviour. They found no evidence that perceived
susceptibility predicts behaviour when the effects of past behaviour were controlled. By con-
trast, they found a small but significant average weighted association (r+ = –0.11) between past
risk behaviour and perceived HIV susceptibility across 26 cross-sectional studies. Gerrard et al.
point out that sexual behaviour is social and complex and that the impact of perceived suscep-
tibility may be reduced for these reasons. Nonetheless, these findings underline the need for
longitudinal studies and analyses that control for the effects of past behaviour in modelling the
impact of perceived susceptibility (and other cognitions) on future behaviour. Such findings
also suggest that, when evaluating the impact of belief-changing interventions, it is important to
assess cognitions immediately after risk information has been received – that is, before partici-
pants have had an opportunity to change their behaviour.
Finally, individual differences may moderate the relationship between past risk behaviour
and perceived susceptibility. For example, Gerrard et al. found that this relationship was strongest
MHBK140-ch02_p30-69.indd 47 28/05/15 3:19 PM
48 PREDICTING AND CHANGING HEALTH BEHAVIOUR
among older respondents, women (vs. men), gay (vs. straight) men, and college (vs. clinic) sam-
ples. Smith et al. (1997) found that self-esteem moderated the effect of past behaviour on per-
ceived susceptibility, and Gladis et al. (1992) found that although for most participants previous
risk behaviour was positively related to perceived susceptibility, this relationship was reversed
among pupils classed as ‘repressors’. Myers (2000) further highlighted the importance of this
personality trait in studies of health beliefs and cognitions. There is also evidence that personal-
ity differences moderate the relationship between perceived risk and subsequent behaviour. For
example, Hampson et al. (2000) found that perceived risk was associated with a reduction in
cigarettes smoked indoors but only for those high in conscientiousness.
5.2.2 Severity
Severity has been conceptualized as a multidimensional construct involving both the medical
severity of a disease (pain, complications, etc.) and its psychosocial severity (e.g. the extent
to which illness might interfere with valued social roles). Unfortunately, as Haefner (1974: 96)
noted: ‘examining the literature, one becomes aware of the variation in the selection of particu-
lar dimensions of seriousness to be studied’. For example, in a study of osteoporosis prevention,
Smith Klohn and Rogers (1991) used essays to manipulate three severity dimensions: visibility
of disablement (high vs. low), time of onset (near vs. distant future), and rate of onset (gradual
vs. sudden). These researchers obtained a significant main effect for visibility and a significant
interaction between visibility and time of onset on post-test intention measures. The more vis-
ibly disabling descriptions of the effects of osteoporosis were, the stronger were intentions to
take preventive action. In addition, low visibility consequences in the distant future were associ-
ated with weaker intentions than high visibility consequences with either time of onset. These
findings underline the importance of pilot work in identifying salient dimensions of severity,
including beliefs about visibility and how quickly consequences are likely to occur.
Ronis and Harel (1989) combined elements of the HBM and subjective expected utility (SEU)
theory in a study of breast examination behaviours. Since breast examination leads to early
detection and treatment, these researchers divided the severity component into severity follow-
ing action (severity of breast cancer if treated promptly) and severity following inaction (sever-
ity of breast cancer if treated late). They found support for this distinction using confirmatory
factor analysis. Path analysis showed that severity dimensions did not directly affect behaviour.
Rather, the benefits constructs entirely mediated the effects of severity. This study offered an
interesting reconceptualization of the threat component of the HBM. Instead of directly influenc-
ing behaviour, threat appraisal is thought to contribute to the subjective utility of taking action
versus not taking action. This is reflected in Schwarzer’s (1992) health action process approach,
in which perceived threat is construed as a determinant of outcome expectancies and intention.
Further research comparing direct effects (Janz and Becker 1984), interactions (e.g. Lewis 1994),
and mediational (Ronis and Harel 1989) models of severity would be informative.
5.2.3 Benefits, barriers, cues to action, and health motivation
The remaining HBM constructs have also raised problems of multidimensionality in operation-
alizations of the model. As we have noted, the benefits construct comprises both medical and
psychosocial benefits of engaging in health-promoting behaviours. Similarly, the barriers com-
ponent comprises practical barriers to performing the behaviour (e.g. time, expense, availability,
transport, waiting time), as well as psychological costs associated with performing the behaviour
MHBK140-ch02_p30-69.indd 48 28/05/15 3:19 PM
HEALTH BELIEF MODEL 49
(pain, embarrassment, threat to well-being or lifestyle and livelihood). Later HBM formulations
(Rosenstock et al. 1988) included psychological barriers to performing the behaviour. While self-
efficacy has received considerable attention (Bandura 1986, 1997), other specific psychological
barriers might include poor understanding of complex recommendations (e.g. by a learning disa-
bled person with diabetes) or a lack of social skills (e.g. to negotiate condom use successfully).
Indeed, later work has indicated that social skills prerequisite to interpersonal negotiation may be
more important predictors of safer sexual behaviour than the beliefs specified by the HBM (Ban-
dura 1992; Abraham and Sheeran 1993, 1994).
As we have seen, the cues to action construct can encompass a variety of influences upon
behaviour, ranging from awareness and memory of mass media campaigns, through leaflets
and reminder letters, to perceived descriptive and injunctive normative influence exerted by
health care professionals and significant others. Thus the coherence of this construct has been
questioned by a number of researchers. For example, Weinstein (1988) argued that the con-
struct does not fit easily alongside the rational expectancy-value structure of the model’s major
constructs. Mattson (1999) suggested that ‘cues to action’ could include all persuasive experi-
ences, including interpersonal communication, exposure to mass media, and internal responses
to threat. Conceptualized in this way, cues to action are causally prior to beliefs and the effect of
cues on beliefs depends on the content of the persuasive communications (e.g. fear appeals vs.
self-efficacy enhancing communication). Schwarzer (1992) suggested that actual and perceived
cues should be distinguished and that cues to action might be more appropriately construed as
antecedents of intention formation and action (once other beliefs were established). Arguably,
operationalizations of cues to action could ask respondents about the presence or absence of
cues and also ask them to indicate the extent to which cues were available and influenced their
decisions (see Bagozzi 1986). Such measures may more closely represent the original concep-
tion of ‘cues’. The challenge facing researchers is to define this construct so that it can be trans-
lated into clearly defined measures that have both theoretical and psychometric coherence or,
alternatively, to divide the construct into a series of clearly defined behavioural prompts.
Multi-item measures of health motivation have included a variety of items. For example,
Chen and Land’s (1986) measure included items relating to control over health and perceived
health status, while Umeh and Rogan-Gibson (2001) included measures of past performance
of a range of health behaviours. This underlines problems with the discriminant validity of the
health motivation construct. If health motivation is to be used as a distinct measure, further
research is needed to clarify the relationship between this construct and related constructs,
including past behaviour, health locus of control (Wallston and Wallston 1982), health value
(Kristiansen 1985), and intention, as specified in the theories of reasoned action and planned
behaviour (Fishbein and Ajzen 1975).
5.3 Problems of operationalization: response bias
A final problem, common to all social cognition models, concerns social desirability. Respond-
ents may be aware of the purposes of interviews and questionnaires and so may be motivated
to exaggerate both the desirability of their beliefs and behaviours and the consistency between
the two. Unfortunately, this issue has received little attention. Prospective studies and objec-
tive outcome measures help to reduce bias, and individual difference measures of social desir-
ability may also identify participants most likely to shape their responses to present a socially
MHBK140-ch02_p30-69.indd 49 28/05/15 3:19 PM
50 PREDICTING AND CHANGING HEALTH BEHAVIOUR
desirable picture of themselves. Sheeran and Orbell (1995) found that responses to HBM meas-
ures may also be subject to bias when questionnaire items are not randomized and can easily be
‘read’ by respondents.
6 Intervention studies
We have examined the utility of the HBM as a model of modifiable cognitive determinants of
health-related behaviour and reviewed studies evaluating its utility as a predictive model of
behaviour. This is foundational because accurate prediction is an indicator of veridical expla-
nation. As Sutton (1998: 1317) observed, ‘models that do not enable us to predict behaviour
are unlikely to be useful as explanatory models’. Explanatory models are useful to behaviour
change intervention designers to the extent that they identify modifiable intrapersonal factors
that underpin the regulation of behaviours that health care professionals wish to promote or
discourage. Thus, if HBM-specified beliefs can be changed and such changes result in changes in
health behaviour patterns, such as increasing the prevalence of consistent condom use among
sexually active adolescents (e.g. Abraham et al. 1992; Abraham and Sheeran 1994; Sheeran et al.
1999), then the model provides a useful resource for behaviour change intervention designers.
The HBM specifies a series of potentially modifiable cognitions. Perceived susceptibility
and severity represent perceived illness threat, and while these beliefs may be prerequisite to
preventive motivation, they may be less strongly associated with behaviour than beliefs about
the specific behaviour, such as benefits, barriers, and self-efficacy (Sullivan et al. 2004). Other
cognition models, such as the theory of planned behaviour (Ajzen 1991, 1998) and the health
action process approach (Schwarzer 1992, 2008), have conceptualized all HBM cognitions as
determinants of the strength and stability of motivation to change behaviour, with self-efficacy
also having effects on the translation of motivation into action. However, the impact of changing
any cognition on subsequent behaviour depends on the target audience and the behaviour. For
example, even when self-efficacy is a strong correlate of a specified preventive behaviour
pattern in a particular population, changing self-efficacy is only likely to promote behav-
iour change among those with low behaviour-specific self-efficacy beliefs. The selection of cog-
nition targets for intervention design must be group specific.
Once modifiable cognitions are identified, intervention designers need to identify particular
change techniques known to effectively change those cognitions (Abraham and Michie 2008).
For example, an intervention may use the technique of ‘communicating group-specific disease
incidence’ in order to increase perceived susceptibility. However, even when one or more evi-
dence-based change techniques are identified, many design and implementation choices remain.
Disease incidence may be communicated in different ways using different media. Abraham
(2012) discusses how health care professionals can use identified change targets and techniques
to develop text-based change techniques in paper or electronic media (see also Abraham et al.
[2002] and Abraham et al. [2007] for analyses of how identified change targets have, in prac-
tice, been translated into text-based communication in nationally available, European leaflets
intended to promote condom use and reduce excessive alcohol use, respectively).
Alternatively, rather than text-based communication, an intervention designer might find
evidence suggesting that a drama-based approach to incidence communication would be more
effective than a text-based intervention. For example, in the ‘handshake transmission activity’
MHBK140-ch02_p30-69.indd 50 28/05/15 3:19 PM
HEALTH BELIEF MODEL 51
(e.g. Stephenson et al. 2004), students are instructed to shake hands with other students (or
not) and then later asked to stand depending on status. The handshake can represent sexual
intercourse and standing can represent being infected with a specific sexually transmitted infec-
tion. Some standing students represent initially infected individuals, others those who become
infected during the encounter, while those who remain sitting represent those who avoided
infection because they used condoms or avoided having sex (by not shaking hands). This mode
of delivery is more resource intensive than provision of a text-based leaflet or website but may
be more effective in changing perceived susceptibility than other delivery formats. The Inter-
vention Mapping framework (Bartholomew et al. 2011) guides intervention designers through
such decisions (for a discussion of the contextualization and use of intervention mapping in
behaviour change intervention design, evaluation, and implementation, see Denford et al. 2015).
It is important to remember that interventions differ across multiple dimensions when
attributing differences in effectiveness to particular characteristics of behaviour change inter-
ventions. For example, one HBM-based intervention targeting perceived susceptibility may be
effective while another, also targeting perceived susceptibility, may be ineffective because the
particular change techniques or mode of delivery differ and/or the implementation of the inter-
vention was more or less well suited to the recipients and context. Moreover, de Bruin et al.
(2010) demonstrated that the content of active control conditions, such as ‘usual care’, can deter-
mine the effectiveness of interventions. If an intervention is compared with optimal usual care
as opposed to poor usual care, it is less likely to appear effective. Thus investigation of associa-
tions between targeted cognitions and intervention effectiveness can provide helpful guidance
to intervention designers (for an insightful illustration, see Albarracín et al. 2005). However, all
such advice should be considered in the context of the multiple dimensions along which inter-
vention designs and evaluations differ.
Below we consider behaviour change intervention evaluations in which the HBM was identi-
fied as the theoretical or ‘logic’ model identifying potentially modifiable beliefs. Table 2.4 lists
19 HBM-based, behaviour change intervention evaluations that illustrate the diversity of this lit-
erature. We have not included intervention evaluations that used knowledge and health beliefs
as outcome measures (e.g. Booth et al. 1999; Out and Lafreniere 2001; Aoun et al. 2002).
Among this small sample of intervention studies, there is evidence of selectivity in the
choice of health behaviours targeted for intervention: Some behaviours were targeted several
times (e.g. mammogram screening), whereas other behaviours were not targeted at all. Some
interventions were derived directly from the HBM (e.g. Carmel et al. 1996; Toro-Alfonso et al.
2002; Anderson et al. 2011), whereas others drew upon HBM and other social cognition models
in order to target a broader range of cognitions (e.g. Strecher et al. 1994; Lu 2001). Some inter-
ventions took the form of educational presentations to groups in classes or workshops (e.g.
Ford et al. 1996; Abood et al. 2003; Shariatjafari et al. 2012) and/or involved the distribution of
leaflets or booklets (e.g. Carmel et al. 1996; Hawe et al. 1998), whereas others were delivered
at an individual level (referred to variously as ‘educational’ or ‘counselling’ interventions) and
often involved assessment of the recipient’s current beliefs before new information and persua-
sive arguments were presented (e.g. Cummings et al. 1981; Jones et al. 1988a, 1988b; Hegel et al.
1992; Champion 1994; Anderson et al. 2011). Such interventions are tailored to the individual’s
cognitions. Computer-generated, individually tailored letters have also been used (Strecher
et al. 1994). All of these interventions relied on information provision and verbal persuasion as
means to change HBM-specified beliefs.
MHBK140-ch02_p30-69.indd 51 28/05/15 3:19 PM
52 PREDICTING AND CHANGING HEALTH BEHAVIOUR
Table 2.4 Evaluations of behaviour change interventions based on the HBM
Behaviour Target group Investigators/effectiveness Intervention
Screening behaviours
Stroke-risk screening
Preventive behaviours
Adults with stroke risk Anderson et al. (2011) Motivational telephone call
Smoking cessation Adult patients Strecher et al. (1994) Tailored letters
Women with cardiac risk Schmitz et al. (1999) Individual educational
programme
Breast self-examination Female adolescents Ludwick and Gaczkowski
(2001)
Teaching with video role
model
Taiwanese beauticians Lu (2001) Instruction, practice, and
follow-up
Mammogram screening Elderly minority women Fox et al. (2001) Postal advice on cost
Women (40–48 years) Champion (1994) Home interview
Women (35+ years) Champion (1995) Home interview
Women (50–85 years) Champion et al. (2003) Various (five interventions)
Safer sexual practices Men who have sex with
men
Toro-Alfonso et al. (2002) Workshop
Condom use Low-cost sex workers Ford et al. (1996) Outreach group educational
programmes
Teenage contraception Adolescents Eisen et al. (1992) HBM-based sex education
Healthy diet University employees Abood et al. (2003) Eight-week worksite
intervention
Healthy women Shariatjafari et al. (2012) Three-session group
education and cooking
instruction
Sun exposure protection Elderly kibbutz members Carmel et al. (1996) Multi-component
intervention
Measles vaccination Parents (pre-vaccination) Hawe et al. (1998) Modified postal reminder
card
Adherence behaviours
Time in treatment Alcohol clinic patients Rees (1986) Weekly group meeting
Fluid restrictions Male haemodialysis
patients
Hegel et al. (1992) HBM-based counselling
Keeping appointments Emergency room
patients (11 problems)
Jones et al. (1988b) HBM-promoting interviews
Note: Studies highlighted in bold found evidence of behaviour change following the HBM-based intervention.
MHBK140-ch02_p30-69.indd 52 28/05/15 3:19 PM
HEALTH BELIEF MODEL 53
Fifteen of these 19 evaluations (79%), highlighted in bold in Table 2.4, found evidence of
behaviour change following HBM-based interventions. This is very encouraging, suggesting that
HBM-specified cognitions can provide an important basis for behaviour change intervention
design. However, because these evaluations were not systematically selected or assessed for
methodological rigour, conclusions regarding effectiveness should be examined on a study-by-
study basis. For example, some evaluations did not include a control group (e.g. Carmel et al.
1996), and weaknesses inherent in before-and-after designs mean that observed changes in such
evaluations cannot be confidently assigned to the intervention. Other evaluations employed
randomized controlled trials (RCTs) and some investigated moderator effects (see Baron and
Kenny 1986). For example, Strecher et al. (1994) found that their computer-tailored letters were
effective for moderate but not heavy smokers in two studies using random assignment to an
intervention or control group. Some evaluations also report intervention effects on hypothesized
cognitive mediators (e.g. changes in targeted health beliefs). We will highlight methodological
and theoretical issues emerging from this literature by considering four of these 19 intervention
evaluations in greater detail.
Ludwick and Gaczkowski (2001) used a pre-/post-test design without a control group to
evaluate an HBM-based intervention to increase breast self-examination (BSE) among ninety-
three 14- to 18-year-old US teenagers. The intervention was a school-based, multi-stage teaching
session delivered by an undergraduate nursing student. Fibrocystic changes and risks of con-
tracting breast cancer were explained to classes in order to increase knowledge of perceived
susceptibility and severity. Cards that could be placed in showers were distributed as cues to
action. Classes also watched a video that explained breast anatomy, showed teenagers perform-
ing BSE, and demonstrated mammography. In addition, classes watched a demonstration of
BSE on a breast model and practised BSE on breast models under supervision. The interven-
tion was evaluated by questionnaire one month after the teaching session. Self-reported BSE
increased significantly. For example, the proportion that had never performed BSE fell from
64% to 32%. The HBM components were measured using multi-item scales but the authors do
not report pre-/post-test comparisons of HBM measures. It is, therefore, unclear whether the
observed self-report behaviour change could be explained by changes in the target HBM cogni-
tions. Although the increase in reported BSE initiation is substantial, this evaluation is weak
because no control group was included and the follow-up was short-term. Thus, for example,
the results do not clarify whether the completion of BSE-related questionnaires on its own might
have prompted increased BSE (without the class), or what proportion of these teenagers were
still performing BSE at three months or a year post-intervention.
Lu (2001) assessed the effectiveness of a work-site intervention designed to promote BSE
among women who scored highly on a measure of perceived barriers to BSE. High scoring
women were allocated by place of work to a control group (n = 40) or intervention group
(n = 30). The educational programme was based on the HBM, the theory of reasoned action,
and Bandura’s (1977) social cognitive theory. A brief description indicates that the intervention
included BSE instructions and practice using breast models as well as discussion of individual
barriers to BSE performance. In addition, participants received a monthly reminder telephone
call. The intervention was evaluated using self-report questionnaires three months later. The
HBM constructs were measured using multi-item scales derived from Champion’s work (see, for
example, Table 2.3). Significant differences between the intervention and control group were
found for reported BSE, BSE accuracy, perceived susceptibility, perceived benefits and barriers,
MHBK140-ch02_p30-69.indd 53 28/05/15 3:19 PM
54 PREDICTING AND CHANGING HEALTH BEHAVIOUR
perceived competency, perceived normative influence, and intention, but not for perceived seri-
ousness at three-month follow-up. However, these analyses did not control for pre-intervention
scores (e.g. using analysis of covariance) and no mediation analysis is reported. Consequently,
it is unclear whether differences between the intervention and control group on HBM-specified
beliefs accounted for differences in reported BSE. Multiple regression analyses indicated that
perceived competency and normative influence were significant predictors of BSE frequency
(with perceived competency accounting for 13% of the variance in BSE frequency) but that
HBM-specified beliefs did not add to the variance explained in BSE. This implies that HBM-spec-
ified beliefs may not be the most important cognitive targets for BSE-promoting interventions.
The intervention evaluations reported by Lu (2001) and Ludwick and Gaczkowski (2001)
suggest that educational programmes, including BSE instruction, practice with breast models,
and follow-up reminders, are likely to promote BSE (see also Champion 1995). However, more
sophisticated intervention evaluation designs, such as RCTs, with longer-term follow-up, are
required before conclusions can be reached about evidence-based practice for health educa-
tors in this field. Although these interventions were, at least partially, inspired by the HBM, it is
unclear whether their apparent effectiveness depended on promotion of HBM-specified beliefs.
It is possible, for example, that enhanced self-efficacy, rather than changes in HBM-specified
beliefs, is crucial to the effectiveness of such BSE educational programmes.
Champion (1994) reported a more robust evaluation of an intervention designed to pro-
mote mammography attendance in women over 35 years. An RCT was used to compare four
conditions: a no-intervention control group, an information-giving intervention, an individual
counselling intervention designed to change HBM-specified beliefs, and a combined interven-
tion designed both to provide information and change health beliefs. Self-reported adherence
to mammography attendance guidelines was assessed for 301 women one year later. Control-
ling for pre-intervention compliance, the results indicated that only the combination intervention
had a significantly greater post-intervention adherence rate than the control group, with this
group being almost four times more likely to adhere. Thus, this evaluation establishes that both
information provision and belief-change interventions are required to maximize mammography
adherence (see also Champion and Huster 1995; Mandelblatt and Yabroff 1999). The belief-change
interventions resulted in greater perceived seriousness, greater benefits, and reduced barriers
but did not increase perceived susceptibility. However, no mediation analysis was reported and,
since knowledge and perceived control were also enhanced, the findings do not demonstrate con-
clusively that HBM-specified belief changes were critical to intervention effectiveness.
Jones et al. (1988b) report an RCT of an intervention designed to persuade patients using
hospital emergency services to make and keep follow-up appointments with their own doc-
tor. The sample comprised 842 patients with 11 presenting problems (chest pain, hypertension,
asthma, otitis media, diabetes, urinary tract infection, headache, urethritis [men], vaginitis
[women], low back pain, and rash) that did not require hospitalization. An intervention for indi-
vidual patients was developed. This involved assessment of patients’ HBM-specified beliefs and
delivery of protocol-based, condition-specific educational messages to target beliefs that were
not accepted by recipients. The intervention was designed to increase the patients’ perceived
susceptibility to illness complications, perceived seriousness of the complications, and benefits
of a follow-up referral appointment in terms of avoiding further complications. It was delivered
by a research nurse during necessary nursing care. Four intervention conditions were tested:
(1) a routine care, control group; (2) an individual, nurse-delivered hospital intervention; (3) the
MHBK140-ch02_p30-69.indd 54 28/05/15 3:19 PM
HEALTH BELIEF MODEL 55
hospital intervention combined with a follow-up telephone call; and (4) a follow-up telephone call
without the hospital intervention. Only 33% of the control group patients scheduled a follow-up
appointment, whereas 76% of the hospital intervention group, 85% of the telephone intervention
group, and 85% of the combined intervention group did so. Twenty-four per cent of the control
group kept a follow-up appointment compared with 59% in the hospital intervention group, 59%
in the telephone intervention group, and 68% in the combination group. Thus, the combination
intervention worked most effectively. Jones et al. did not conduct a cost-effectiveness analysis
but noted that the telephone intervention alone might be the most effective practical interven-
tion when costs such as staff training and staff time are taken into account.
Jones et al. (1988b) found that presenting problem had a moderating effect on the impact
of the intervention – that is, there were no significant differences between conditions in relation
to keeping a follow-up appointment for four of the 11 illness groups (i.e. diabetes, headache,
urethritis, and vaginitis). The results of this study were also reported separately for asthmatic
patients (Jones et al. 1987a), hypertensive patients (Jones et al. 1987b), low back pain patients
(Jones et al. 1988a), urinary tract patients (Jones et al. 1991a), and chronic versus acute patients
(Jones et al. 1991b). Mediation analysis was conducted. The researchers found that among those
patients who had scheduled a follow-up appointment, the interventions did not have an effect
on keeping an appointment. This suggests that the interventions were effective because they
prompted appointment scheduling. The availability of childcare and being older than 30 years
also made keeping a scheduled appointment more likely. These mediation and moderation anal-
yses help clarify how the intervention(s) work and for whom. However, the researchers did not
report analyses testing whether differences in pre- and post-intervention HBM-specified beliefs
could account for the effect of the intervention on scheduling follow-up appointments. Nonethe-
less, these studies indicate that a HBM-based intervention worked for a variety of patients and
that the model of delivery could enhance effectiveness.
These illustrative evaluation studies demonstrate that the HBM has inspired effective
behaviour change interventions across a range of health behaviours. However, these studies
also highlight six key shortcomings in studies evaluating HBM-inspired interventions. First,
some evaluation designs are limited due to the lack of appropriate control groups, lack of rand-
omization to condition, samples that do not support generalization or short-term follow-up. Sec-
ond, the variety of behaviours targeted and the multidimensionality of HBM constructs means
that the nature of persuasive messages may differ across behaviours and thereby undermine
the validity of cross-behaviour cognition and technique comparisons. For example, one inter-
vention may attempt to reduce perceived barriers by informing patients of available financial
support (e.g. Fox et al. 2001), while another attempts to reduce barriers by enhancing commu-
nication about risk and precautions in sexual relationships (e.g. Toro-Alfonso et al. 2002). The
HBM-specified change process is the same (perceived barriers to behaviour change) but the
content of the intervention is quite different. Third, the HBM, like other social cognition models,
does not describe how to change beliefs. It is possible to combine models like the HBM with
cognition change theories such as the elaboration likelihood model (ELM; Petty and Cacioppo
1986; for an empirical example, see Quine et al. 2001) or cognitive dissonance theory (CDT; Fes-
tinger 1957; for an empirical example, see Stone et al. 1994) but this approach is not typical of
HBM based interventions. Consequently, the selection of change techniques is often not, or not
explicitly, theory based. Fourth, interventions usually comprise a variety of change techniques,
making it unclear which particular technique (or combinations of techniques) is/are crucial to
MHBK140-ch02_p30-69.indd 55 28/05/15 3:19 PM
56 PREDICTING AND CHANGING HEALTH BEHAVIOUR
effectiveness. For example, in considering the BSE-promoting interventions by Ludwick and
Gaczkowski (2001) and Lu (2001), we might ask whether practice examination of breast models
is crucial to effectiveness or whether reminders are necessary to ensure maintenance. To iden-
tify the contribution of specific change techniques, evaluations need to test effects on changing
specified cognitions (e.g. perceived barriers) and behaviours of single techniques and combi-
nations of techniques. Fifth, to establish whether an intervention generates behaviour change
because it alters target beliefs, it is necessary both to measure cognition and behaviour pre- and
post-intervention and to relate the former to the latter including mediation analyses (Baron and
Kenny 1986). However, mediation analysis is rarely reported in HBM-based intervention evalu-
ations. Consequently, even when these interventions are effective in changing behaviour, it is
unclear whether such effects are due to changes in HBM-specified beliefs. Sixth, once an effec-
tive technique is identified, it is important to explore moderating effects such as patient type and
mode of delivery to establish for whom (and how) the intervention is most likely to be effective.
To date, only two meta-analyses have been conducted that provide useful insights into the
utility of the HBM as a model to inform the design of health behaviour change interventions.
Jones et al. (2013) conducted a systematic review of evaluations of HBM-based interventions
designed to increase adherence to medical regimens as measured by behaviour change. Eighteen
intervention evaluations were included. Fourteen targeted adults and 9 targeted patients’ adher-
ence behaviours across a range of illnesses including alcoholism, asthma, diabetes, and obstruc-
tive sleep apnoea. Sixteen used RCTs to evaluate the intervention with follow-up periods ranging
from one to twelve months. Overall, 14 of the 18 studies (77%) reported significant improvements
in adherence behaviours following an HBM-based behaviour change intervention. Effect sizes
varied from small to large (Cohen’s d ranged from 0.2 to 1.0), with six studies (33%) reporting
moderate to large effect sizes (d > 0.5). As Jones et al. note, the HBM was originally developed
to guide public health interventions, and all three studies reporting large (d > 0.8) significant
effects on adherence behaviours were aimed at primary prevention of disease. Moreover, two
of these three studies used objective measures of patient attendance. Jones et al. noted that the
interventions used a variety of behaviour change techniques but were unable to map change
techniques onto specific HBM beliefs, or to relate changes in HBM beliefs to changes in adher-
ence behaviour. Jones et al. (2013) found that interventions varied in the HBM beliefs that were
targeted, with 16 targeting perceived benefits, 15 targeting perceived susceptibility, 14 targeting
perceived barriers, 11 targeting perceived severity, 7 targeting cues to action, and 4 targeting
self-efficacy. The authors did not include health motivation in their conceptualization of the
HBM. Overall, despite the small sample of studies included, this review suggests that HBM-
based behaviour change interventions can effectively promote adherence behaviours and, in
certain cases, are surprisingly effective (e.g. achieving d values of 0.8 and above).
Meta-analyses of correlational data (e.g. Harrison et al. 1992; Carpenter 2010) and meta-
analyses focusing on changes in outcome behaviours (e.g. Jones et al. 2013) cannot assess the
extent to which beliefs specified by the HBM have a causal impact on health-related behaviours
(see our fifth shortcoming above). A recent meta-analysis explored this issue by selecting stud-
ies that (1) randomly assigned participants to intervention versus control conditions, (2) observed
a significant increase in perceived susceptibility or perceived severity due to the intervention, and
(3) compared subsequent behaviour for treatment versus control participants (Sheeran et al. 2014).
Findings showed that interventions that increased perceptions of susceptibility or severity led to
significant changes in behaviour in the predicted direction (mean d = 0.25 and 0.34, respectively).
MHBK140-ch02_p30-69.indd 56 28/05/15 3:19 PM
HEALTH BELIEF MODEL 57
There was also evidence that heightening perceptions of both susceptibility and severity led to
greater changes in behaviour compared with heightening either susceptibility or severity on its
own. Interventions that increased perceptions of susceptibility also had larger effects on behav-
iour when perceived benefits (response efficacy) or self-efficacy were simultaneously enhanced.
The largest effect on behaviour was observed when interventions were successful in increasing
perceived susceptibility, response efficacy, and self-efficacy (average d = 0.52). These results
indicate that successfully changing beliefs specified by the original HBM and by an extended HBM
effectively promote health behaviour change. Overall, then, the HBM, which has been used to
guide health behaviour change interventions since 1970 (e.g. Haefner and Kirscht 1970), appears
to have considerable potential as a basis for health behaviour change intervention design, despite
its relatively poor predictive utility. They key to effectiveness may be identification of predictive
beliefs that are missing and amenable to change within particular populations.
7 Future directions
The HBM has provided a useful theoretical framework for the identification of modifiable beliefs
predictive of health-related behaviours for more than 50 years (Rosenstock 1974). The model’s
common-sense constructs are easy for non-psychologists to assimilate and can be readily and
inexpensively operationalized in self-report questionnaires. The HBM has focused researchers’
and health care professionals’ attention on modifiable psychological prerequisites of behaviour
and provided a basis for practical interventions across a range of behaviours. Research has,
however, predominantly employed cross-sectional correlational designs, and further experi-
mental studies are required to clarify the causal impact of changing beliefs specified by the
HBM on a range of health behaviours (e.g. Sheeran et al. 2014). The proposed mediation of
socio-economic influences on health behaviour by health beliefs also remains unclear. Research
identifying which beliefs or cognitions mediate the effects of socio-economic status in relation
to particular health behaviours (e.g. Orbell et al. 1995) would be especially valuable.
The common-sense, expectancy-value framework of the HBM simplifies health-related rep-
resentational processes. Further elaboration of HBM constructs, as seen in Weinstein’s (1988)
precaution adoption process, may therefore be necessary. The model also excludes cognitions
that have been shown to be powerful predictors of behaviour. In contrast to the theory of rea-
soned action, it fails to highlight the importance of intention formation or the influence that
others’ approval may have upon our behaviour. It portrays individuals as asocial, economic
decision-makers and consequently fails to account for behaviour under social and affective con-
trol. This is evident in applications to sexual behaviour, where, despite initial optimism, it has
failed to distinguish between ‘safer’ and ‘unsafe’ behaviour patterns. The model is also limited
because it does not articulate hierarchical or temporal relationships between cognitions. Despite
King’s (1982, 1984) innovative extension, the model has not distinguished between proximal and
distal antecedents of behaviour. More recent models, such as the theory of planned behaviour
(Ajzen and Madden 1986) and protection motivation theory (Prentice-Dunn and Rogers 1986),
propose direct and indirect cognitive influences on behaviour. This facilitates a more power-
ful analysis of data and a clearer indication of how interventions might exert their effects. For
example, if a certain level of perceived severity must be reached before perceived susceptibility
becomes dominant in guiding behaviour, this would explain why severity generally has weak
MHBK140-ch02_p30-69.indd 57 28/05/15 3:19 PM
58 PREDICTING AND CHANGING HEALTH BEHAVIOUR
associations with behaviour and suggest that this variable should be regarded as a more distal
cognitive antecedent (Schwarzer 1992). Intentions and perceived self-efficacy may mediate the
effects of health beliefs on behaviour (Cummings et al. 1979; Warwick et al. 1993), confirming
Rosenstock’s (1974) suggestion that HBM constructs could be seen as ‘the setting for . . . subse-
quent responses at other stages in the decision process’ leading to action. More recent research
has focused upon specifying cognitions that distinguish between people who intend and subse-
quently undertake behaviours and people with equivalent intentions who do not act (Abraham
et al. 1999b; Gollwitzer 1999; Sheeran 2002; Sheeran and Abraham 2003). Health beliefs may,
therefore, be seen as distant from action facilitation and self-regulation processes. Nonetheless,
even if other models specify stronger predictors of behaviour, in certain instances, beliefs about
susceptibility, benefits of precautions or treatments or barriers to performing health behaviours
may remain potentially important if variability in these beliefs is key to motivation to act.
Further systematic examination of evaluations of HBM-inspired interventions could clarify
patterns of effectiveness across this literature. However, given the heterogeneity of evaluation
designs, intervention techniques, target behaviours, and populations, it is likely that reviews
focusing on interventions designed to change particular behaviours for particular populations will
be most informative (e.g. Kelley et al. 2001). For example, in a review of 63 interventions designed
to increase mammography use, Yabroff and Mandelblatt (1999) found that four theory-based
interventions drawing upon the HBM (see Aiken et al. 1994a, 1994b; Champion 1994) increased
mammography utilization, on average, by an impressive 23% compared with usual care. The
review also indicated that theory-based cognitive interventions that did not involve interper-
sonal interaction (e.g. those distributing letters or videos) were not effective. Meta-analyses of
this kind can identify types of intervention and modes of intervention delivery that are effective
in changing specified health behaviours. This information could then be used to design experi-
mental studies that isolate particular techniques and combinations of techniques and measure
potential mediators, including pre- and post-intervention beliefs. Such findings would permit
identification of techniques that are effective in changing certain beliefs that are important to
particular health behaviours and allow these techniques to be tested against one another (see,
for example, Hegel et al. 1992). Much remains to be done but, overall, evaluation studies and
meta-analyses highlight the ongoing usefulness of the HBM as a basis for behaviour change
intervention design and evaluation.
Note
1. The number of hypotheses examined for each HBM component varies across behaviour
types. The relevant numbers for vulnerability, severity, benefits, and barriers in the case of
preventive behaviours are 21, 18, 19, and 14 respectively. In the case of sick role behaviours,
the numbers of hypotheses are 13, 16, 15, and 12 respectively.
References
Abood, D.A., Black, D.R. and Feral, D. (2003) Smoking cessation in women with cardiac risk: a comparative study
of two theoretically based therapies, Journal of Nutrition Education and Behavior, 35, 260–7.
MHBK140-ch02_p30-69.indd 58 28/05/15 3:19 PM
HEALTH BELIEF MODEL 59
Abraham, C. (2012) Mapping change mechanisms and behaviour change techniques: a systematic approach to
promoting behaviour change through text, in C. Abraham and M. Kools (eds.) Writing Health Communica-
tion: An Evidence-Based Guide (pp. 83–98). London: Sage.
Abraham, C. and Michie, S (2008) A taxonomy of behavior change techniques used in interventions, Health Psy-
chology, 27, 379–87.
Abraham, C. and Sheeran, P. (1993) In search of a psychology of safer-sex promotion: beyond beliefs and texts,
Health Education Research: Theory and Practice, 8, 245–54.
Abraham, C. and Sheeran, P. (1994) Modelling and modifying young heterosexuals’ HIV-preventive behav-
iour: a review of theories, findings and educational implications, Patient Education and Counselling, 23,
173–86.
Abraham, C. and Sheeran, P. (2003) Implications of goal theories for the theories of reasoned action and planned
behaviour, Current Psychology, 22, 264–80.
Abraham, C., Clift, S. and Grabowski, P. (1999a) Cognitive predictors of adherence to malaria prophylaxis
regimens on return from a malarious region: a prospective study, Social Science and Medicine, 48,
1641–54.
Abraham, C., Krahé, B., Dominic, R. and Fritsche, I. (2002) Do health promotion messages target cognitive and
behavioural correlates of condom use? A content analysis of safer-sex promotion leaflets in two countries,
British Journal of Health Psychology, 7, 227–46.
Abraham, C., Sheeran, P., Abrams, D. and Spears, R. (1996) Health beliefs and teenage condom use: a prospective
study, Psychology and Health, 11, 641–55.
Abraham, C., Sheeran, P., Norman, P., Conner, M., de Vries, N. and Otten, W. (1999b) When good intentions are not
enough: modeling post-intention cognitive correlates of condom use, Journal of Applied Social Psychology,
29, 2591–612.
Abraham, C., Sheeran, P., Spears, R. and Abrams, D. (1992) Health beliefs and the promotion of HIV-preventive
intentions amongst teenagers: a Scottish perspective, Health Psychology, 11, 363–70.
Abraham, C., Southby, L., Quandte, S., Krahé, B. and van der Sluijs, W. (2007), What’s in a leaflet? Identifying
research-based persuasive messages in European alcohol-education leaflets, Psychology and Health, 22,
31–60.
Abraham, I.L. and Williams, B.M. (1991) Hypertensive elders’ perception and management of their disease: health
beliefs or health decisions? Journal of Applied Gerontology, 10, 444–54.
Adih, W.H. and Alexander, C.S. (1999) Determinants of condom use to prevent HIV infection among youth in
Ghana, Journal of Adolescent Health, 24, 63–72.
Aho, W.R. (1979a) Smoking, dieting and exercise: age differences in attitudes and behaviour relevant to selected
health belief model variables, Rhode Island Medical Journal, 62, 95–102.
Aho, W.R. (1979b) Participation of senior citizens in the Swine Flu inoculation programme: an analysis of health
belief model variables in preventive health behaviour, Journal of Gerontology, 34, 201–8.
Aiken, L.S., West, S.G., Woodward, C.K. and Reno, R.R. (1994a) Health beliefs and compliance with mammography-
screening: recommendations in asymptomatic women, Health Psychology, 13, 122–9.
Aiken, L.S., West, S.G., Woodward, C.K., Reno, R.R. and Reynolds, K.D. (1994b) Increasing screening mammogra-
phy in asymptomatic women: evaluation of a second generation theory-based program, Health Psychology,
13, 526–38.
Ajzen, I. (1991) The theory of planned behavior, Organizational Behavior and Human Decision Processes, 50,
179–211.
Ajzen, I. (1998) Models of human social behaviour and their application to health psychology, Psychology and
Health, 13, 735–9.
Ajzen, I. and Madden, T.J. (1986) Prediction of goal-directed behaviour: attitudes, intentions and perceived behav-
ioral control, Journal of Experimental Social Psychology, 22, 453–74.
MHBK140-ch02_p30-69.indd 59 28/05/15 3:19 PM
60 PREDICTING AND CHANGING HEALTH BEHAVIOUR
Alagna, S.W. and Reddy, D.M. (1984) Predictors of proficient technique and successful lesion detention in breast
self-examination, Health Psychology, 3, 113–27.
Albarracín, D., Gillete, J.C., Earl, A.N., Glasman, L.R. and Durantini, M.R. (2005) A test of major assumptions
about behavior change: a comprehensive look at the effects of passive and active HIV-prevention interven-
tions since the beginning of the epidemic, Psychological Bulletin, 131, 856–97.
Ali, N.S. (2002) Prediction of coronary heart disease preventive behaviors in women: a test of the Health Belief
Model, Women and Health, 35, 83–96.
Anderson, R.T., Camacho, F., Iaconi, A.I., Tegeler, C.H. and Balkrishnan, R. (2011) Enhancing the effectiveness
of community stroke risk screening: a randomised controlled trial, Journal of Stroke and Cerebrovascular
Diseases, 20, 330–5.
Aoun, S., Donovan, R.J., Johnson, L. and Egger, G. (2002) Preventive care in the context of men’s health, Journal
of Health Psychology, 7, 243–52.
Bagozzi, R.P. (1986) Attitude formation under the theory of reasoned action and a purposeful behaviour reformu-
lation, British Journal of Social Psychology, 25, 95–107.
Bandura, A. (1977) Self-efficacy: towards a unifying theory of behavioural change, Psychological Review, 84,
191–215.
Bandura, A. (1986) Social Foundations of Thought and Action. Englewood Cliffs, NJ: Erlbaum.
Bandura, A. (1992) Exercise of personal agency through the self-efficacy mechanism, in R. Schwarzer (ed.) Self-
Efficacy: Thought Control of Action (pp. 3–38). Washington, DC: Hemisphere.
Bandura, A. (1997) Self-Efficacy: The Exercise of Control. New York: Freeman.
Bardsley, P.E. and Beckman, L.J. (1988) The health belief model and entry into alcoholism treatment, Interna-
tional Journal of the Addictions, 23, 19–28.
Baron, R. and Kenny, D.A. (1986) The moderator–mediator variable distinction in social psychological research:
conceptual, strategic, and statistical considerations, Journal of Personality and Social Psychology, 51,
1173–82.
Bartholomew, L.K., Parcel, G.S., Kok, G., Gottlieb, N.H. and Fernández, M.E. (2011) Planning Health Promotion
Programs: An Intervention Mapping Approach. San Francisco, CA: Jossey-Bass.
Beck, K.H. (1981) Driving while under the influence of alcohol: relationship to attitudes and beliefs in a college
sample, American Journal of Drug and Alcohol Abuse, 8, 377–88.
Becker, M.H. and Maiman, L.A. (1975) Sociobehavioural determinants of compliance with health and medical
care recommendations, Medical Care, 13, 10–24.
Becker, M.H., Drachman, R.H. and Kirscht, J.P. (1972) Predicting mothers’ compliance with pediatric medical
regimens, Journal of Pediatrics, 81, 843–54.
Becker, M.H., Drachman, R.H. and Kirscht, J.P. (1974) A new approach to explaining sick-role behaviour in low
income populations, American Journal of Public Health, 64, 205–16.
Becker, M.H., Haefner, D.P. and Maiman, L.A. (1977b) The health belief model in the prediction of dietary compli-
ance: a field experiment, Journal of Health and Social Behaviour, 18, 348–66.
Becker, M.H., Radius, S.M. and Rosenstock, I.M. (1978) Compliance with a medical regimen for asthma: a test of
the health belief model, Public Health Reports, 93, 268–77.
Becker, M.H., Haefner D.P., Kasl, S.V., Kirscht, J.P., Maiman, L.A. and Rosenstock, I.M. (1977a) Selected psychoso-
cial models and correlates of individual health-related behaviors, Medical Care, 15, 27–46.
Becker, M.H., Kaback, M.M., Rosenstock, I.R. and Ruth, M. (1975) Some influences of public participation in a
genetic screening program, Journal of Community Health, 1, 3–14.
Bentler, P.M. and Speckart, G. (1979) Models of attitude–behaviour relations, Psychological Review, 86,
452–64.
Berkanovich, E., Telesky, C. and Reeder, S. (1981) Structural and social psychological factors on the decision to
seek medical care for symptoms, Medical Care, 19, 693–709.
MHBK140-ch02_p30-69.indd 60 28/05/15 3:19 PM
HEALTH BELIEF MODEL 61
Bollin, N.W. and Hart, L.K. (1982) The relationship of health belief motivations, health focus of control and health
valuing to dietary compliance of haemodialysis patients, American Association of Nephrology Nurses and
Technicians Journal, 9, 41–7.
Booth, R.E., Zhang, Y. and Kwiatkowski, C.F. (1999) The challenge of changing drug and sex risk behaviors of
runaway and homeless adolescents, Child Abuse and Neglect, 23, 1295–306.
Bradley, C., Gamsu, D.S. and Moses, S.L. (1987) The use of diabetes-specific perceived control and health belief
measures to predict treatment choice and efficacy in a feasibility study of continuous subcutaneous insulin
infusion pumps, Psychology and Health, 1, 133–46.
Brownlee-Duffeck, M., Peterson, L. and Simonds, J.F. (1987) The role of health beliefs in the regimen adherence
and metabolic control of adolescents and adults with diabetes mellitus, Journal of Consulting and Clinical
Psychology, 55, 139–44.
Calnan, M. (1984) The health belief model and participation in programmes for the early detection of breast can-
cer: a comparative analysis, Social Science and Medicine, 19, 823–30.
Calnan, M. (1985) An evaluation of the effectiveness of a class teaching breast self-examination, British Journal
of Medical Psychology, 53, 317–29.
Carmel, S., Shani, E. and Rosenberg, L. (1996) Skin cancer protective behaviors among the elderly: explaining
their response to a health education program using the health belief model, Educational Gerontology, 22,
651–68.
Carpenter, C.J. (2010) A meta-analysis of the effectiveness of health belief model variables in predicting behav-
iour, Health Communication, 25, 661–9.
Casey, R., Rosen, B., Glowasky, A. and Ludwig, S. (1985) An intervention to improve follow-up of patients with
otitis media, Clinical Pediatrics, 24, 149–52.
Cerkoney, K.A. and Hart, K.L. (1980) The relationship between the health belief model and compliance of persons
with diabetic regimens, Diabetes Care, 3, 594–8.
Champion, V.L. (1984) Instrument development for health belief model constructs, Advances in Nursing Science,
6, 73–85.
Champion, V.L. (1994) Strategies to increase mammography utilization, Medical Care, 32, 118–29.
Champion, V.L. (1995) Results of a nurse-delivered intervention on proficiency and nodule detection with BSE,
Oncology Nursing Forum, 22, 819–24.
Champion, V. and Huster, G. (1995) Effect of interventions on stage of mammography adoption, Journal of
Behavioral Medicine, 18, 169–87.
Champion, V., Maraj, M., Hui, S., Perkins, A.J., Tierney, W., Menon, U. et al. (2003) Comparison of tailored
interventions to increase mammography screening in nonadherent older women, Preventive Medicine, 36,
150–8.
Charney, E., Bynum, R., Eldridge, D., Frank, D., MacWhinney, J.B., McNabb, N. et al. (1967) How well do patients
take oral penicillin? A collaborative study in private practice, Journal of Pediatrics, 40, 188–95.
Chen, M. and Land, K.C. (1986) Testing the Health Belief Model: Lisrel analysis of alternative models of causal
relationships between health beliefs and preventive dental behaviour, Social Psychology Quarterly, 49, 45–60.
Chen, M. and Land, K.C. (1990) Socioeconomic status (SES) and the health belief model: LISREL analysis of uni-
dimensional versus multidimensional formulations, Journal of Social Behaviour and Personality, 5, 263–84.
Chen, M. and Tatsuoka, M. (1984) The relationship between American women’s preventive dental behaviour and
dental health beliefs, Social Science and Medicine, 19, 971–8.
Connelly, C.E. (1984) Compliance with outpatient lithium therapy, Perspectives in Psychiatric Care, 22, 44–50.
Connelly, C.E., Davenport, Y.B. and Nurnberger, J.I. (1982) Adherence to treatment regimen in a lithium carbon-
ate clinic, Archives of General Psychiatry, 39, 585–8.
Conner, M. (1993) Pros and cons of social cognition models in health behaviour, Health Psychology Update, 14,
24–30.
MHBK140-ch02_p30-69.indd 61 28/05/15 3:19 PM
62 PREDICTING AND CHANGING HEALTH BEHAVIOUR
Cooper, H.M. (1986) Integrating Research: A Guide for Literature Reviews. London: Sage.
Cortina, J.M. (1993) What is coefficient alpha? An examination of theory and applications, Journal of Applied
Psychology, 78, 98–104.
Cummings, K.M., Becker, M.H. and Kirscht, J.P. (1982) Psychosocial factors affecting adherence to medical regi-
mens in a group of haemodialysis patients, Medical Care, 20, 567–79.
Cummings, K.M., Becker, M.H., Kirscht, J.P. and Levin, N.W. (1981) Intervention strategies to promote compliance
with medical regimens by ambulatory haemodialysis patients, Journal of Behavioral Medicine, 4, 111–27.
Cummings, K.M., Jette, A.M. and Brock, B.M. (1979) Psychological determinants of immunization behaviour in a
swine influenza campaign, Medical Care, 17, 639–49.
De Bruin, M., Viechtbauer, W., Schaalma, H.P., Kok, H., Abraham, C. and Hospers, H.J. (2010) Standard care
impact on effects of highly active antiretroviral therapy adherence interventions: A meta-analysis of rand-
omized controlled trials, Archives of Internal Medicine, 170, 240–50.
Denford, S., Abraham, C., Smith, J., Lloyd, J.J., White, M., Tarrant, M. et al. (2015) Designing and evaluating
behavior change interventions to promote health, in K.J. Reynolds and N.R. Branscombe (eds.) The Psychol-
ogy of Change: Life Contexts, Experiences, and Identities. New York: Psychology Press.
DeVellis, R.F. (1991) Scale Development: Theory and Applications. Newbury Park, CA: Sage.
Dorr, N., Krueckeberg, S., Strathman, A. and Wood, M.D. (1999) Psychosocial correlates of voluntary HIV anti-
body testing in college students, AIDS Education and Prevention, 11, 14–27.
Drayton, V.L.C., Montgomery, S.B., Modeste, N.N. and Frye-Anderson, B.A. (2002) The Health Belief Model as a
predictor of repeat pregnancies among Jamaican teenage mothers, International Quarterly of Community
Health Education, 21, 67–81.
Eisen, M., Zellman, G.L. and McAlister, A.L. (1985) A health belief model approach to adolescents fertility control:
some pilot program findings, Health Education Quarterly, 12, 185–210.
Eisen, M., Zellman, G.L. and McAlister, A.L. (1992) A health belief model–social learning theory approach to
adolescents’ fertility control: findings from a controlled field trial, Health Education Quarterly, 19, 249–62.
Farquharson, L., Noble, L., Barker, C. and Behrens, R.H. (2004) Health beliefs and communication in the travel
clinic as predictors to adherence to malaria chemoprophylaxis, British Journal of Health Psychology, 9,
201–17.
Feather, N.T. (1982) Expectations and Actions: Expectancy-value Models in Psychology. Hillsdale, NJ: Erlbaum.
Festinger, L. (1957) A Theory of Cognitive Dissonance. Stanford, CA: Stanford University Press.
Field, A. (2000) Discovering Statistics: Using SPSS for Windows. London: Sage.
Fincham, J.E. and Wertheimer, A.L. (1985) Using the health belief model to predict initial drug therapy defaulting,
Journal of Psychology, 118, 101–5.
Fishbein, M. and Ajzen, I. (1975) Belief, Attitude, Intention and Behavior: An Introduction to Theory and
Research. Reading, MA: Addison-Wesley.
Fishbein, M., Triandis, H.C., Kanfer, F.H., Becker, M., Middlestadt, S.E. and Eichler, A. (2001) Factors influencing
behaviour and behaviour change, in A. Baum, T.A. Revenson and J.E. Singer (eds.) Handbook of Health Psy-
chology (pp. 3–17). Mahwah, NJ: Erlbaum.
Ford, K., Wirawan, D.N., Fajans, P., Meliawan, P., MacDonald, K. and Thorpe, L. (1996) Behavioural interventions
for reduction of sexually transmitted disease/HIV transmission among female commercial sex workers and
clients in Bali, Indonesia, AIDS, 10, 213–22.
Fox, S.A., Stein, J.A., Sockloskie, R.J. and Ory, M.G. (2001) Targeted mailed materials and the Medicare benefici-
ary: increasing mammogram screening among the elderly, American Journal of Public Health, 91, 55–61.
Gerrard, M. and Warner, T.D. (1991) Antecedents of pregnancy among women marines, Journal of the Washing-
ton Academy of Sciences, 80, 1015.
Gerrard, M., Gibbons, F.X. and Bushman, B.J. (1996) Relation between perceived vulnerability to HIV and precau-
tionary sexual behaviour, Psychological Bulletin, 119, 390–409.
MHBK140-ch02_p30-69.indd 62 28/05/15 3:19 PM
HEALTH BELIEF MODEL 63
Gianetti, V.J., Reynolds, J. and Rihen, T. (1985) Factors which differentiate smokers from ex-smokers among
cardiovascular patients: a discriminant analysis, Social Science and Medicine, 20, 241–5.
Gladis, M.M., Michela, J.L., Walter, H.J. and Vaughan, R.D. (1992) High school students’ perceptions of AIDS risk:
realistic appraisal or motivated denial? Health Psychology, 11, 307–16.
Gochman, D.S. and Parcel, G.S. (eds.) (1982) Children’s health beliefs and health behaviours, Special Issue of
Health Education Quarterly, 9, 104–270.
Gollwitzer, P.M. (1999) Implementation intentions: strong effects of simple plans, American Psychologist, 54,
493–503.
Gordis, L., Markowitz, M. and Lilienfeld, A.M. (1969) Why patients don’t follow medical advice: a study of children
on long-term antistreptococcal prophylaxis, Journal of Pediatrics, 75, 957–68.
Gottlieb, N.H. and Baker, J.A. (1986) The relative influence of health beliefs, parental and peer behaviours and
exercise program participation on smoking, alcohol use and physical activity, Social Science and Medicine,
22, 915–27.
Grady, K.E., Kegeles, S.S., Lund, A.K., Wolk, C.H. and Farber, N.J. (1983) Who volunteers for a breast self-exami-
nation program? Evaluating the bases for self-selection, Health Education Quarterly, 10, 79–94.
Haefner, D.P. (1974) The health belief model and preventive dental behavior, in M.H. Becker (ed.) The Health
Belief Model and Personal Health Behavior (pp. 93–105). Thorofare, NJ: Slack.
Haefner, D.P. and Kirscht, J.P. (1970) Motivational and behavioural effects of modifying health beliefs, Public
Health Reports, 85, 478–84.
Hallal, J.C. (1982) The relationship of health beliefs, health locus of control, and self-concept to the practice of
breast self-examination in adult women, Journal of Nursing Research, 31, 127–42.
Halper, M., Winawer, S. and Body, R. (1980) Issues of patient compliance, in S. Winawer, D. Schottenfeld and P. Sher-
lock (eds.) Colorectal Cancer: Prevention, Epidemiology and Screening (pp. 299–310). New York: Raven Press.
Hampson, S.E., Andrews, J.A., Barckley, M., Lichenstein, E. and Lee, M.E. (2000) Conscientiousness, perceived
risk and risk reduction behaviors: a preliminary study, Health Psychology, 19, 496–500.
Harris, D.M. and Guten, S. (1979) Health protective behaviour: an exploratory study, Journal of Health and
Social Behaviour, 20, 17–29.
Harris, R. and Lynn, M.W. (1985) Health beliefs, compliance and control of diabetes mellitus, Southern Medical
Journal, 2, 162–6.
Harrison, J.A., Mullen, P.D. and Green, L.W. (1992) A meta-analysis of studies of the health belief model with
adults, Health Education Research, 7, 107–16.
Hartman, P.E. and Becker, M.H. (1978) Non-compliance with prescribed regimen among chronic haemodialysis
patients, Journal of Dialysis and Transplantation, 7, 978–86.
Hawe, P., McKenzie, N. and Scurry, R. (1998) Randomised controlled trial of the use of a modified postal reminder
card in the uptake of measles vaccination, Archives of Disease in Childhood, 79, 136–40.
Hay, J.L., Ford, J.S., Klein, D., Primavera, L.H., Buckley, T.R. Stein, T.R. et al. (2003) Adherence to colorectal
cancer screening in mammography-adherent older women, Journal of Behavioral Medicine, 26, 553–76.
Hegel, M.T., Ayllon, T., Thiel, G. and Oulton, B. (1992) Improving adherence to fluid restrictions in male hemodi-
alysis patients: a comparison of cognitive and behavioural approaches, Health Psychology, 11, 324–30.
Heinzelmann, F. (1962) Factors in prophylaxis behaviour in treating rheumatoid fever: an exploratory study,
Journal of Health and Human Behaviour, 3, 73.
Hershey, J.C., Morton, B.G., Davis, J.R. and Reichgolt, M.J. (1980) Patient compliance with antihypertensive
medication, American Journal of Public Health, 70, 1081–9.
Hester, N.R. and Macrina, D.M. (1985) The health belief model and the contraceptive behaviour of college women:
implications for health education, Journal of American College Health, 33, 245–52.
Hochbaum, G.M. (1958) Public Participation in Medical Screening Programs: A Socio-Psychological Study.
Public Health Service Publication #572. Washington, DC: US Government Printing Office.
MHBK140-ch02_p30-69.indd 63 28/05/15 3:19 PM
64 PREDICTING AND CHANGING HEALTH BEHAVIOUR
Hoogewerf, P.E., Hislop, T.G., Morrison, B.J., Burns, S.D. and Sitzo, R. (1990) Health belief and compliance with
screening for faecal occult blood, Social Science and Medicine, 30, 721–6.
Janz, N. and Becker, M.H. (1984) The health belief model: a decade later, Health Education Quarterly, 11, 1–47.
Jones, C.J., Smith, H. and Llewellyn, C. (2013) Evaluating the effectiveness of health belief model interventions in
improving adherence: a systematic review, Health Psychology Review, 8, 253–69.
Jones, F., Abraham, C., Harris, P., Schulz, J. and Chrispin, C. (2001) From knowledge to action regulation: modeling
the cognitive prerequisites of sunscreen use in Australian and UK samples, Psychology and Health, 16, 191–206.
Jones, P.K., Jones, S.L. and Katz, J. (1987a) Improving compliance for asthmatic patients visiting the emergency
department using a health belief model intervention, Journal of Asthma, 24, 199–206.
Jones, P.K., Jones, S.L. and Katz, J. (1987b) Improving follow-up among hypertensive patients using a health
belief model intervention, Archives of Internal Medicine, 147, 1557–60.
Jones, S.L., Jones, P.K. and Katz, J. (1988a) Compliance for low-back pain patients in the emergency department:
a randomized trail, Spine, 13, 553–6.
Jones, S.L., Jones, P.K. and Katz, J. (1988b) Health belief model intervention to increase compliance with emer-
gency department patients, Medical Care, 26, 1172–84.
Jones, S.L., Jones, P.K. and Katz, J. (1991a) A randomized trial to improve compliance in urinary tract patients in
the emergency department, Annals of Emergency Medicine, 19, 16–20.
Jones, S.L., Jones, P.K. and Katz, J. (1991b) Compliance in acute and chronic patients receiving a health belief
model intervention in the emergency department, Social Science and Medicine, 32, 1183–9.
Keesling, B. and Friedman, H.S. (1987) Psychological factors in sunbathing and sunscreen use, Health Psychol-
ogy, 6, 477–93.
Kegeles, S.S. (1963) Why people seek dental care: a test of a conceptual formulation, Journal of Health and
Human Behaviour, 4, 166–73.
Kelley, H.H. (1967) Attribution theory in social psychology, in D. Levine (ed.) Nebraska Symposium on Motiva-
tion (pp. 192–241). Lincoln, NB: University of Nebraska Press.
Kelley, K., Bond, R. and Abraham, C. (2001) Effective approaches to persuading pregnant women to quit smoking:
a meta-analysis of intervention evaluation studies, British Journal of Health Psychology, 6, 207–28.
Kelly, G.R., Mamon, J.A. and Scott, J.E. (1987) Utility of the health belief model in examining medication compli-
ance among psychiatric outpatients, Social Science and Medicine, 25, 1205–11.
King, J.B. (1982) The impact of patients’ perceptions of high blood pressure on attendance at screening: an exten-
sion of the health belief model, Social Science and Medicine, 16, 1079–91.
King, J.B. (1984) Illness attributions and the health belief model, Health Education Quarterly, 10, 287–312.
Kirscht, J.P. (1983) Preventive health behaviour: a review of research and issues, Health Psychology, 2, 277–301.
Kirscht, J.P. and Rosenstock, I.M. (1977) Patient adherence of antihypertensive medical regimens, Journal of
Community Health, 3, 115–24.
Kirscht, J.P., Becker, M.H. and Eveland, P. (1976) Psychological and social factors as predictors of medical behav-
iour, Journal of Medical Care, 14, 422–31.
Kirscht, J.P., Becker, M.H., Haefner, D.P. and Maiman, L.A. (1978) Effects of threatening communications and
mothers’ health beliefs on weight change in obese children, Journal of Behavioural Medicine, 1, 147–57.
Kristiansen, C.M. (1985) Value correlates of preventive health behaviour, Journal of Personality and Social Psy-
chology, 49, 748–58.
Kruglanski, A.W. and Klar, Y. (1985) Knowing what to do: on the epistemology of actions, in J. Kuhl and J. Beck-
mann (eds.) Action Control: From Cognition to Behaviour (pp. 41–60). Berlin: Springer.
Langlie, J.K. (1977) Social networks, health beliefs and preventive health behaviour, Journal of Health and Social
Behaviour, 18, 244–60.
Larson, E.B., Bergman, J. and Heidrich, F. (1982) Do postcard reminders improve influenza vaccination compli-
ance? Journal of Medical Care, 20, 639–48.
MHBK140-ch02_p30-69.indd 64 28/05/15 3:19 PM
HEALTH BELIEF MODEL 65
Leavitt, F. (1979) The health belief model and utilization of ambulatory care services, Social Science and Medi-
cine, 13, 105–12.
Lewin, R.W. (1951) Field Theory in Social Science. New York: Harper.
Lewis, K.S. (1994) An examination of the health belief model when applied to diabetes mellitus. Unpublished
doctoral dissertation, University of Sheffield.
Li, C., Unger, J.B., Schuster, D., Rohrbach, L.A., Howard-Pitney, B. and Norman, G. (2003) Youths’ exposure to
environmental tobacco smoke (ETS): associations with health beliefs and social pressure, Addictive Behav-
iors, 28, 39–53.
Lollis, C.M., Johnson, E.H. and Antoni, M.H. (1997) The efficacy of the health belief model for predicting condom
usage and risky sexual practices in university students, AIDS Education and Prevention, 9, 551–63.
Lowe, C.S. and Radius, S.M. (1987) Young adults’ contraceptive practices: an investigation of influences, Adoles-
cence, 22, 291–304.
Lu, Z.J. (2001) Effectiveness of breast self-examination nursing interventions for Taiwanese community target
groups, Journal of Advanced Nursing, 34, 163–70.
Ludwick, R. and Gaczkowski, T. (2001) Breast self-exams by teenagers: outcome of a teaching program, Cancer
Nursing, 24, 315–19.
Maddux, J.E. and Rogers, R.W. (1983) Protection motivation and self-efficacy: a revised theory of fear appeals
and attitude change, Journal of Experimental Social Psychology, 19, 469–79.
Maiman, L.A., Becker, M.H., Kirscht, J.P., Haefner, D.P. and Drachman, R.H. (1977) Scales for measuring health
belief model dimensions: a test of predictive value, internal consistency and relationships among beliefs,
Health Education Quarterly, 4, 215–31.
Mandelblatt, J.S and Yabroff, K.R. (1999) Effectiveness of interventions designed to increase mammography use:
a meta-analysis of provider-targeted strategies, Cancer Epidemiology, Biomarkers and Prevention, 8, 759–67.
Mattson, M. (1999) Towards a reconceptualization of communication cues to action in the health belief model:
HIV test counselling, Communication Monographs, 66, 240–65.
McCuskar, J., Stoddard, A.M., Zapka, J.G., Zorn, M. and Mayer, K.H. (1989) Predictors of AIDS-preventive behav-
iour among homosexually active men: a longitudinal analysis, AIDS, 3, 443–8.
Mullen, P.D., Hersey, J.C. and Iversen, D.C. (1987) Health behaviour compared, Social Science and Medicine, 24,
973–81.
Myers, L.B. (2000) Identifying repressors: a methodological issue for health psychology, Psychology and Health,
15, 205–14.
Nelson, E.C., Stason, W.B. and Neutra, R.R. (1978) Impact of patients’ perceptions on compliance with treatment
for hypertension, Journal of Medical Care, 16, 893–906.
Norman, P. and Brain, K. (2005) An application of the extended health belief model to the prediction of breast self-
examination among women with a family history of breast cancer, British Journal of Health Psychology, 10, 1–16.
Norman, P. and Conner, M. (1993) The role of social cognition models in predicting attendance at health checks,
Psychology and Health, 8, 447–62.
Ogionwo, W. (1973) Socio-psychological factors in health behaviour: an experimental study of methods and atti-
tude change, International Journal of Health Education, 16 (suppl.), 1–14.
Oliver, R.L. and Berger, P.K. (1979) A path analysis of preventive care decision models, Journal of Consumer
Research, 6, 113–22.
Orbell, S., Crombie, I. and Johnston, G. (1995) Social cognition and social structure in the prediction of cervical
screening uptake, British Journal of Health Psychology, 1, 35–50.
Otten, W. and van der Pligt, J. (1992) Risk and behaviour: the mediating role of risk appraisal, Acta Psychologia,
80, 325–46.
Out, J.W. and Lafreniere, K. (2001) Baby think it over: using role play to prevent teen pregnancy, Adolescence, 36,
571–82.
MHBK140-ch02_p30-69.indd 65 28/05/15 3:19 PM
66 PREDICTING AND CHANGING HEALTH BEHAVIOUR
Owens, R.G., Daly, J., Heron, K. and Lemster, S.J. (1987) Psychological and social characteristics of attenders for
breast screening, Psychology and Health, 1, 303–13.
Pan, P. and Tantam, D. (1989) Clinical characteristics, health beliefs and compliance with maintenance treatment:
a comparison between regular and irregular attenders at a depot clinic, Acta Psychiatrica Scandinavica, 79,
564–70.
Penderson, L.L., Wanklin, J.M. and Baskerville, J.C. (1982) Multivariate statistical models for predicting change in
smoking behaviour following physician advice to stop smoking, Journal of Preventive Medicine, 11, 536–49.
Perkins, D.O. (2002) Predictors of noncompliance in patients with schizophrenia, Journal of Clinical Psychiatry,
63, 1121–8.
Petosa, R. and Kirby, J. (1991). Using the health belief model to predict safer sex intentions among adolescents,
Health Education Quarterly, 18, 463–76.
Petty, R.E. and Cacioppo, J.T. (1986) The elaboration likelihood model of persuasion, in L. Berkowitz (ed.)
Advances in Experimental Social Psychology (Vol. 19, pp. 123–205). New York: Academic Press.
Portnoy, B. (1980) Effects of a controlled-usage alcohol education program based on the health belief model,
Journal of Drug Education, 10, 181–95.
Prentice-Dunn, S. and Rogers, R.W. (1986) Protection motivation theory and preventive health: beyond the health
belief model, Health Education Research: Theory and Practice, 3, 153–61.
Quine, L., Rutter, D.R. and Arnold, L. (1998) Predicting and understanding safety helmet use among schoolboy
cyclists: a comparison of the theory of planned behaviour and the health belief model, Psychology and Health,
13, 251–69.
Quine, L., Rutter, D.R. and Arnold, L. (2001) Persuading school-age cyclists to use safety helmets: effectiveness
of an intervention based on the theory of planned behaviour, British Journal of Health Psychology, 6, 327–45.
Rawl, S., Champion, V., Menon, U., Loehrer, P.J., Vance, G.H. and Skinner, C.S. (2001) Validation of scales to
measure benefits of and barriers to colorectal cancer screening, Journal of Psychosocial Oncology, 19, 47–63.
Rayant, G.A. and Sheiham, A. (1980) An analysis of factors affecting compliance with tooth-cleaning recommen-
dations, Journal of Clinical Periodontology, 7, 289–99.
Rees, D.W. (1986) Changing patients’ health beliefs to improve compliance with alcohol treatment: a controlled
trial, Journal of Studies on Alcoholism, 47, 436–9.
Reid, L.D. and Christensen, D.B. (1988) A psychosocial perspective in the explanation of patients’ drug-taking
behaviour, Social Science and Medicine, 27, 277–85.
Rogers, R.W. (1983) Cognitive and physiological processes in fear appeals and attitude change: a revised theory
of protection motivation, in J. Cacioppo and R. Petty (eds.) Social Psychophysiology (pp. 153–76). New York:
Guilford Press.
Rogers, R.W. and Mewborn, C.R. (1976) Fear appeals and attitude change: effects of a threat’s noxiousness, prob-
ability of occurrence and the efficacy of coping responses, Journal of Personality and Social Psychology, 34,
54–61.
Ronis, D.L. and Harel, Y. (1989) Health beliefs and breast examination behaviours: analysis of linear structural
relations, Journal of Psychology and Health, 3, 259–85.
Rosenstock, I.M. (1966) Why people use health services, Milbank Memorial Fund Quarterly, 44, 94–124.
Rosenstock, I.M. (1974) Historical origins of the health belief model, Health Education Monographs, 2, 328–335.
Rosenstock, I.M., Strecher, V.J. and Becker, M.H. (1988) Social learning theory and the health belief model, Health
Education Quarterly, 15, 175–83.
Rosenthal, R. (1984) Meta-analysis Procedures for Social Research. Beverly Hills, CA: Sage.
Rotter, J.B. (1966) Generalized expectancies for internal versus external control of reinforcement, Psychological
Monographs, 80 (whole no. 609), 1–28.
Rundall, T.G. and Wheeler, J.R. (1979) The effect of income on use of preventive care: an evaluation of alternative
explanations, Journal of Health and Social Behaviour, 20, 397–406.
MHBK140-ch02_p30-69.indd 66 28/05/15 3:19 PM
HEALTH BELIEF MODEL 67
Rust, J. and Golombok, S. (1990) Modern Psychometrics. London: Routledge.
Sackett, D.L., Becker, M.H. and MacPherson, A.S. (1974) The Standardized Compliance Questionnaire. Hamilton,
Ontario: McMaster University.
Salloway, J.C., Pletcher, W.R. and Collins, J.J. (1978) Sociological and social psychological models of compliance
with prescribed regimen: in search of a synthesis, Sociological Symposium, 23, 100–21.
Schmiege, S.J., Aiken, L.S., Sander, J.L. and Gerend, M.A. (2007) Osteoporosis prevention among young women:
psychological models of calcium consumption and weight bearing exercise, Health Psychology, 26, 577–87.
Schmitz, J.M., Spiga, R., Rhoades, H.M., Fuentes, F. and Grabowski, J. (1999) Nutrition education worksite inter-
vention for university staff: application of the health belief model, Nicotine and Tobacco Research, 1, 87–94.
Schwarzer, R. (1992) Self-efficacy in the adoption and maintenance of health behaviours: theoretical approaches
and a new model, in R. Schwarzer (ed.) Self-Efficacy: Thought Control of Action (pp. 217–42). Washington,
DC: Hemisphere.
Schwarzer, R. (2008) Modeling health behavior change: how to predict and modify the adoption and maintenance
of health behaviors, Applied Psychology, 57, 1–29.
Shariatjafari, S., Omidvar, N., Shakibazadeh, E., Majdzadeh, R., Minaei, M. and Gholamzade, M. (2012) Effective-
ness of community-based intervention to promote Iran’s food-based dietary guidelines. International Journal
of Preventive Medicine, 4, 249–61.
Sheeran, P. (2002) Intention–behavior relations: a conceptual and empirical review, in W. Stroebe and M. Hew-
stone (eds.) European Review of Social Psychology (Vol. 12, pp. 1–36). Chichester: Wiley.
Sheeran, P. and Abraham, C. (2003) The importance of temporal stability of intention relative to other moderators
of the intention–behavior relation, Personality and Social Psychology Bulletin, 29, 205–15.
Sheeran, P. and Orbell, S. (1995) How confidently can we infer health beliefs from questionnaire responses?
Psychology and Health, 11, 273–90.
Sheeran, P., Abraham, C. and Orbell, S. (1999) Psychosocial correlates of condom use: a meta-analysis, Psychologi-
cal Bulletin, 125, 90–132.
Sheeran, P., Harris, P. and Epton, T. (2014) Does heightening risk appraisals change people’s intentions and
behavior? A meta-analytic review of the experimental evidence, Psychological Bulletin, 140, 511–43.
Silver Wallace, L. (2002) Osteoporosis prevention in college women: application of the expanded health belief
model, American Journal of Health Behavior, 26, 163–72.
Simon, K.J. and Das, A. (1984) An application of the health belief model toward educational diagnosis for VD
education, Health Education Quarterly, 11, 403–18.
Slovic, P., Fischoff, B. and Lichtenstein, S. (1977) Behavioral decision theory, Annual Review of Psychology, 28,
1–39.
Smith, G.E., Gerrard, M. and Gibbons, F.X. (1997) Self-esteem and the relation between risk behaviour and per-
ceptions of vulnerability to unplanned pregnancy in college women, Health Psychology, 16, 137–46.
Smith, J.A., Hughes, I.C.T. and Budd, R.J. (1999) Non-compliance with anti-psychotic medication: users’ views on
advantages and disadvantages, Journal of Mental Health, 8, 287–96.
Smith Klohn, L. and Rogers, R.W. (1991) Dimensions of severity of health threat: the persuasive effects of visibility,
time of onset and rate of onset on young women’s intentions to prevent osteoporosis, Health Psychology, 10, 323–9.
Stacy, R.D. and Lloyd, B.H. (1990) An investigation of beliefs about smoking among diabetes patients: informa-
tion for improving cessation efforts, Journal of Patient Education and Counselling, 15, 181–9.
Stephenson, J.M., Strange, V., Forrest, S., Oakley, A., Copas, A., Allen, E. et al. (2004) Pupil-led sex education in
England (RIPPLE study): cluster-randomised intervention trial, Lancet, 364, 338–46.
Stillman, M. (1977) Women’s health beliefs about breast cancer and breast self-examination, Nursing Research,
26, 121–7.
Stone, J., Aronson, E., Crain, A.L., Winslow, M.P. and Fried, C.B. (1994) Inducing hypocrisy as a means of encour-
aging young adults to use condoms, Personality and Social Psychology Bulletin, 20, 116–28.
MHBK140-ch02_p30-69.indd 67 28/05/15 3:19 PM
68 PREDICTING AND CHANGING HEALTH BEHAVIOUR
Strecher, V.J., Kreuter, M., DenBoer, D.J., Kobrin, S., Hospers, H.J. and Skinner, C.S. (1994) The effects of compu-
ter-tailored smoking cessation messages in family practice settings, Journal of Family Practice, 39, 262–70.
Sullivan, K., Pasch, L.A., Cornelius, T. and Cirigliano, E. (2004) Predicting participation in premarital prevention
programs: the health belief model and social norms, Family Process, 43, 175–93.
Sutton, S. (1998) Predicting and explaining intentions and behavior: how well are we doing? Journal of Applied
Social Psychology, 15, 1317–38.
Sutton, S.R. (1994) The past predicts the future: interpreting behaviour–behaviour relationships in social-
psychological models of health behaviours, in D.R. Rutter and L. Quine (eds.) Social Psychology and Health:
European Perspectives (pp. 71–88). Aldershot: Avebury Press.
Taylor, D.W. (1979) A test of the health belief model in hypertension, in R.B. Haynes, D.W. Taylor and K.L. Sackett
(eds.) Compliance in Health Care (pp. 103–9). Baltimore, MD: Johns Hopkins University Press.
Thompson, R.S., Michnich, M.E., Gray, J., Friedlander, L. and Gilson, B. (1986) Maximizing compliance with
hemoccult screening for colon cancer in clinical practice, Medical Care, 24, 904–14.
Toro-Alfonso, J., Varas-Dias, N. and Andujar-Bello, I. (2002) Evaluation of an HIV/AIDS prevention intervention
targeting latino gay men and men who have sex with men in Puerto Rico, AIDS Education and Prevention,
14, 445–56.
Tversky, A. and Kahneman, D. (1981) The framing of decisions and the psychology of choice, Science, 211, 453–8.
Umeh, K. and Rogan-Gibson, J. (2001) Perceptions of threat, benefits, and barriers in breast self-examination
amongst young asymptomatic women, British Journal of Health Psychology, 6, 361–72.
Van der Pligt, J., Otten, W., Richard, R. and Van der Velde, F. (1993) Perceived risk of AIDS: unrealistic opti-
mism and self-protective action, in J.B. Prio and G.D. Reeder (eds.) The Social Psychology of HIV Infection
(pp. 39–58). Hillsdale, NJ: Erlbaum.
Volk, J.E. and Koopman, C. (2001) Factors associated with condom use in Kenya: a test of the health belief model,
AIDS Education and Prevention, 13, 495–508.
Wagner, P.J. and Curran, P. (1984) Health beliefs and physician identified ‘worried well’, Health Psychology, 3,
459–74.
Wallace, L.S. (2002) Osteoporosis prevention in college women: application of the expanded health belief model,
American Journal of Health Behavior, 26, 163–72.
Wallston, K.A. and Wallston, B.S. (1982) Who is responsible for your health? The construct of health locus of
control, in G.S. Sanders and J. Suls (eds.) The Social Psychology of Health and Illness (pp. 65–95). Hillsdale,
NJ: Erlbaum.
Warshaw, P.R. and Davis, F.D. (1985) Disentangling behavioral intention and behavioral expectation, Journal of
Experimental Social Psychology, 21, 213–28.
Warwick, P., Terry, D. and Gallois, C. (1993) Extending the theory of reasoned action: the role of health beliefs,
in D.J. Terry, C. Gallois and M. McCamish (eds.) The Theory of Reasoned Action: Its Application to AIDS-
Preventive Behaviour (pp. 117–34). Oxford: Pergamon Press.
Wdowik, M.J., Kendall, P.A., Harris, M.A. and Auld, G. (2001) Expanded health belief model predicts diabetes self-
management in college students, Journal of Nutrition Education, 33, 17–23.
Weinberger, M., Green, J.Y. and Mandin, J.J. (1981) Health beliefs and smoking behaviour, American Journal of
Public Health, 71, 1253–5.
Weinstein, N.D. (1980) Unrealistic optimism about future life events, Journal of Personality and Social Psychol-
ogy, 39, 806–20.
Weinstein, N.D. (1982) Unrealistic optimism about susceptibility to health problems, Journal of Behavioral Medi-
cine, 5, 441–60.
Weinstein, N.D. (1984) Why it won’t happen to me: perceptions of risk factors and illness susceptibility, Health
Psychology, 3, 431–57.
Weinstein, N.D. (1988) The precaution adoption process, Health Psychology, 7, 355–86.
MHBK140-ch02_p30-69.indd 68 28/05/15 3:19 PM
HEALTH BELIEF MODEL 69
Weinstein, N.D. and Nicolich, M. (1993) Correct and incorrect interpretations of correlations between risk percep-
tions and risk behaviours, Health Psychology, 12, 235–45.
Weitkunat, R., Pottgiesser, C., Meyer, N., Crispin, A., Fischer, R., Schotten, K. et al. (2003) Perceived risk of bovine
spongiform encephalopathy and dietary behaviour, Journal of Health Psychology, 8, 373–81.
Winfield, E.B. and Whaley, A.L. (2002) A comprehensive test of the health belief model in the prediction of con-
dom use among African American college students, Journal of Black Psychology, 28, 330–46.
Wolcott, D.L., Sullivan, G. and Klein, D. (1990) Longitudinal change in HIV transmission risk behaviours by gay
male physicians, Journal of Psychosomatics, 31, 159–67.
Yabroff, K.R. and Mandelblatt, J.S. (1999) Interventions targeted towards patients to increase mammography use,
Cancer Epidemiology, Biomarkers and Prevention, 8, 749–57.
MHBK140-ch02_p30-69.indd 69 28/05/15 3:19 PM
Predicting and changing
health behaviour
Research and Practice with
Social Cognition Models
Third edition
Edited by
Mark Conner and Paul Norman
MHBK140-FM.indd 3 28/05/15 3:46 PM
Open University Press
McGraw-Hill Education
McGraw-Hill House
Shoppenhangers Road
Maidenhead
Berkshire
England
SL6 2QL
email: enquiries@openup.co.uk
world wide web: www.openup.co.uk
and Two Penn Plaza, New York, NY 10121-2289, USA
First published 1995
Second edition published 2005
First published in this third edition 2015
Copyright © Mark Conner and Paul Norman, 2015
All rights reserved. Except for the quotation of short passages for the purposes of criticism and review,
no part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form
or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written
permission of the publisher or a licence from the Copyright Licensing Agency Limited. Details of such
licences (for reprographic reproduction) may be obtained from the Copyright Licensing Agency Ltd of
Saffron House, 6–10 Kirby Street, London, EC1N 8TS.
A catalogue record of this book is available from the British Library
ISBN-13: 978-0-335-26378-3
ISBN-10: 0-335-26378-X
eISBN: 978-0-335-26379-0
Library of Congress Cataloging-in-Publication Data
CIP data applied for
Typeset by Aptara, Inc.
Fictitious names of companies, products, people, characters and/or data that may be used herein (in case
studies or in examples) are not intended to represent any real individual, company, product or event.
MHBK140-FM.indd 4 28/05/15 3:46 PM
Contents
List of tables xi
List of figures xii
Contributors xiii
Preface xv
Abbreviations xvii
1 Predicting and changing health behaviour: a social cognition approach 1
Mark Conner and Paul Norman
2 The health belief model 30
Charles Abraham and Paschal Sheeran
3 Protection motivation theory 70
Paul Norman, Henk Boer, Erwin R. Seydel and Barbara Mullan
4 Self-determination theory 107
Martin S. Hagger and Nikos L.D. Chatzisarantis
5 The theory of planned behaviour and the reasoned action approach 142
Mark Conner and Paul Sparks
6 The prototype/willingness model 189
Frederick X. Gibbons, Meg Gerrard, Michelle L. Stock and Stephanie D. Finneran
7 Social cognitive theory 225
Aleksandra Luszczynska and Ralf Schwarzer
8 Health action process approach 252
Ralf Schwarzer and Aleksandra Luszczynska
9 Stage theories 279
Stephen Sutton
MHBK140-FM.indd 9 28/05/15 3:46 PM
x CONTENTS
10 Implementation intentions 321
Andrew Prestwich, Paschal Sheeran, Thomas L. Webb and Peter M. Gollwitzer
11 Health behaviour change techniques 358
Susan Michie and Caroline E. Wood
12 Predicting and changing health behaviour: future directions 390
Paul Norman and Mark Conner
Index 431
MHBK140-FM.indd 10 28/05/15 3:46 PM