PROVIDING INFORMATION AND USING FEAR APPEALS
Not all health promotion interventions, whether they are simple leaflets or more
complex programmes, are developed using the intervention mapping approach. This
can lead to mismatches between population needs and the content of interventions.
Abraham et al.(2002) surveyed widely available leaflets promoting condom use in
the UK and Germany (gathering leaflets from general practitioners’ surgeries, clinics
for the treatment of sexually transmitted infections and from the largest publishers of
such materials). The authors conducted a content analysis of the leaflets. This involved
reading and re-reading the leaflets in order to identify distinct messages appearing across
leaflet texts. Clearly defined, non-overlapping categories of message were defined and
checked to see whether they comprehensively reflect the content of the leaflets. These
category definitions were tested to see whether independent readers could assign the
same message to the same category across leaflets (that is whether they could be used
reliably). Such a set of text-derived categories can be reliably used to compre hensively
describe the content of texts (in this case leaflets) and distinguish between the content
of different texts can allow identification of intervention content that is or is not
evidence based. Evidence-based messages (or more generally intervention components)
can then be recommended for inclusion in subsequent interventions. This ‘content
analysis approach to theory-specified persuasive educational communication’ has been
used subsequently to improve the content of leaflets used in public health campaigns
(e.g. Lake et al., 2015; see Holsti, 1969 and Weber, 1990 for further guidance on
content analyses of text).
These analyses found few differences in content between UK and German leaflets.
In general, leaflets devoted most content to providing information on the transmission
of sexually transmitted infections (STIs), people’s risks of acquiring STIs, the effect -
iveness of condom use and on encouraging professional contact. This was disappointing
because the apparent change targets of these leaflets, for example, knowledge and
perceived susceptibility are not the strongest determinants of condom use. Sheeran
et al.(1999) report average weighted correlations with condom use of 0.06 for both
measures of knowledge and perceived susceptibility. Even perceived condom
effectiveness was found to have an average correlation of only 0.10 with condom use.
Thus the cognitions targeted most frequently by the majority of these leaflets were
not those found to be those most strongly associated with condom use. By contrast,
cognitions found to be stronger correlates of condom use such as those specified
by the theory of planned behaviour were targeted less frequently by the leaflets (see
Figure 9.4). The researchers identified 20 core messages that corresponded to the
cognitive determinants significantly associated with condom use in previous studies.
Seventy-five per cent of leaflets included less than half of these determinant-matched
messages. However, a small number of illustrative leaflets were identified that included
between 15 and 18 of the 20 core messages, demonstrating that safer sex promotion
leaflets can include a range of messages matched to the cognitive determinants of
condom use.
This study (and see too Abraham et al., 2007) emphasizes two points. First, careful
content analyses of intervention content can help identify change techniques (in this
case messages included in text) that are more or less well matched to the findings of
previous elicitation research. This is important in stages 2 and 3 of the intervention
202 MOTIVATION AND BEHAVIOUR