A decision aid to support informed choices about bowel cancer screening among adults with low education: randomised controlled trial
BMJ 2010; 341 doi: https://doi.org/10.1136/bmj.c5370 (Published 26 October 2010) Cite this as: BMJ 2010;341:c5370
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
The article by Smith and colleagues [1] is a timely and well-
conducted evaluation of a highly detailed decision-aid for colorectal
cancer decisions in low education/literacy groups. However, there are
some significant concerns regarding the measures employed and their
relationship to the informed choice framework used to structure this
research. According to the framework proposed by Marteau and colleagues
[2], the categorisation of people into eight separate 'informed choice'
groups is based on differences in knowledge (good/poor), attitudes
(positive/negative) and behaviour (participate/don't). The categorisation
of positive/negative attitudes is normally based on the median-split
method [2, 3] (as occurred in the present paper) or can also be based on
the median of the scale [4].
Smith et al [1] report that 51% of participant's in intervention
groups and 65% in the control group have positive attitudes toward
screening. However, the median-split scores for these two groups are
extraordinarily high - 27.3 out of 30 for the standard information group
and 26.4 out of 30 for the intervention groups (attitude scale range 6 to
30). This means that any participants who scored less than 89% in the
standard information group, or less than 85% in the intervention groups
were classified as having negative attitudes according to the theory [1,
2]. If the median of the scale was used, it seems unlikely anyone would
have been 'negative' about screening at all. A person in the standard
information group could circle '5' for three attitude items (very
positive) and '4' for the three remaining items (somewhat positive) and
still be categorised as having 'negative' attitudes towards screening.
This makes very little conceptual sense and does question validity of the
attitude scale for use in this particular field of research. Although the
high median attitude scores likely reflect the overwhelmingly positive
attitudes that the public hold in regards to cancer screening [5],
classifying people as negative based on 'not almost 100% in favour of the
test' suggests that people are being arbitrarily categorised for ease of
analysis, rather than reflecting their actual values or attitudes towards
screening.
Possibly of greater concern is the classification of 'adequate
knowledge' based on the knowledge scale used in the present paper. The
authors report that the decision aid significantly increased adults'
knowledge in comparison to the standard information group. However, a
careful examination of the overall knowledge score reveals that the
difference in knowledge between the intervention groups and the control
group seems exclusively based on 'numerical' knowledge (where even the
decision aid groups failed to score above 50%). 'Conceptual' knowledge
(possibility of a false-positive or false-negative result) was understood
equally as well between the intervention and control groups (no
significant difference reported). However, the very specific 'numerical'
knowledge meant that 72% of the standard information group did not have
adequate knowledge (Table 4), and hence, could not be classified as making
an 'informed choice'. Although it is essential that people do have a
greater understanding of screening to enable informed decisions to be made
about participating or not in screening, do health professionals expect
the population to be as statistically or numerically competent as the
experts? Given that the decision aid groups only averaged a score of 3
out of 8 'numerical' component of the knowledge scale, apparently even
those making an informed choice were displaying a selective understanding
of the knowledge items presented to them.
The authors suggest that the 16% reduction in screening participation
for the intervention groups "may have resulted from [participants]
increasing their knowledge about the low personal benefit of screening".
Unfortunately, there is little in the article to reinforce this claim with
the exception of the higher knowledge score and a slightly reduced
attitudes score for the intervention groups, although it is expected that
the qualitative component of the research that will be published will
illuminate this conclusion. The authors do provide a comprehensive
evaluation of participants' perceptions of the information materials, with
the one exception being in regards to 'fair and balanced'. 48% of the
intervention groups and 41% of the control group reported that the
information was 'fair and balanced'. However, it is uncertain whether the
remaining participants felt that the decision-aid was discouraging
screening, or the control group felt their information encouraged
screening - a potential source of the differences in participation.
This field of research desperately needs a accepted theoretical basis
that allows a comprehensive evaluation of how people make decisions based
on high-quality information. However, arbitrarily categorising people
based on an unrepresentative attitude scale and requiring lay people to
have expert numerical understanding of particular factors associated with
the marginal individual benefits of screening does not achieve this goal -
indeed, if anything, it provides ammunition to those who wish to reduce
the amount of high-quality information to people invited to screening in
favour of simply providing limited information about only the benefits of
participating.
Refs
1. Smith SK, Trevena L, Simpson JM, Barratt A, Nutbeam D, etc. A
decision aid to support informed choices about bowel cancer screening
among adults with low education: randomised controlled trial. BMJ
2. Marteau T, Dormandy E, Mitchie S. A measure of informed choice.
Health Expect 2001; 4: 99-108.
3. Dormandy E, Hooper R, Michie S, Marteau TM. Informed choice to
undergo prenatal screening: a comparison of two hospitals conducting
testing either as part of routine visit or requiring a separate visit. J
Med Screen 2002; 9: 109-114.
4. Dormandy E, Michie S, Hooper R, Marteau TM. Informed choice in
antenatal Down syndrome screening: a cluster-randomised trial of combined
versus separate visit teaching. Pat Edu Counsel 2006; 61: 56-64.
5. Schwatz LM, Woloshin S, Fowler FJ, Welch HG. Enthusiasm for
cancer screening in the United States. JAMA 2004; 291: 71-78.
Competing interests: No competing interests
Re:A question of validity
Paul Hewitson raises important issues around determining whether
people have made an 'informed choice' about screening.
1. Categorisation of positive/negative attitudes
Classifying attitudes as 'positive' or 'negative' is a difficult
issue and one which we carefully considered when designing the trial. We
identified 3 possible ways to deal with the attitudes construct. These
were as follows:
(1) To use the median of the scale in which positive and negative
attitudes are defined by the midpoint of the scale [1 2].
(2) To use the median split as used by the authors of the construct
[3] and the method applied in the current study. This leads to the
categorisation of people as 'positive' or 'negative' based on the median
of the overall sample;
(3) To choose a threshold (on the scale) over which we believe
attitudes are 'positive' vs 'negative'.
As Hewitson notes, method 1 may be inappropriate in the context of
screening as people's attitudes towards screening tend to be highly
positive, and few participants would fall into the 'negative' attitude
category. In the current study, we chose to use method 2, although we do
recognise that there are some limitations to this approach. We have also
applied method 3 to the trial data. This method has been applied in our
previously published decision aid work [4]. We reanalysed the trial data
using scores of 4 or above out of 5, on each item of the attitude scale
and categorised these scores as 'positive'. Score below this threshold
were categorised as 'negative'. Using this alternative approach made no
difference to our findings; 39% of decision aid participants made an
informed choice, compared to 16% in the control groups (P<0.001) a 23%
difference (very similar to our finding of a 22% difference using method
2).
A 4th approach is to determine people's values about screening using
the values clarity subscale on the decisional conflict scale. This
approach has been used by our colleagues in a mammography screening
decision aid trial [5]. Applying this approach to our trial data again did
not change our findings: the proportion of participants making an informed
choice remained approximately the same (35% decision aid groups vs 14%
control group, 21% difference). The reason we did not use the values
clarity subscale was because we felt that the attitude measure developed
by Marteau and colleagues was more appropriate in assessing people's
attitudes towards actually doing the bowel cancer screening test.
Finally, it could be that the problems associated with categorising
people as 'positive' and 'negative' about screening may lie more in how we
label attitudes. It may be more appropriate to use labels such as
'positive' vs 'less positive' or 'highly positive' vs 'moderately
positive' in the case of screening, rather than categorising people in
such as absolute terms (i.e. 'positive' and 'negative').
2. Classification of 'adequate' knowledge
Participants' overall knowledge score was based on the combined
scores for the conceptual and numeric items. Thus participants were
considered to have 'adequate knowledge' if their total score (conceptual
and numeric items combined) was 6 or more out of 12 rather than based on
proportions scoring correctly on individual items. This is a very common
practice in knowledge assessment. It was decided a priori to set the pass
mark for knowledge at 50%, however, we also examined the effects of
increasing or decreasing the knowledge pass mark on informed choice. As
described in our paper, increasing and decreasing the pass mark for
'adequate' knowledge had little effect on the difference (%) of
participants making an informed choice between the decision aid and
control groups. For example, increasing the pass mark to 75% correct (i.e.
9 or more out of 12), the difference in the percentage of participants
making an informed choice remained approximately the same (i.e. 20%)
between the decision aid and control groups.
We did not expect participants to be 'statistically or numerically
competent as the experts'. Rather, we wanted to help participants have a
better understanding of the approximate number of people who may
experience the different outcomes of screening. As described in our paper,
the marking scheme was designed to award participants who gave a ballpark
answer to the numeric questions. Thus, participants were not penalised if
they did not know the precise/exact numeric answer.
3. Reason for reduction in screening participation and participants'
perceptions of materials
Participants in the decision aid group (compared to control
participants) generally had a better understanding of their baseline risk
of bowel cancer and the absolute risk reduction (or number of lives saved)
by screening. In turn, participants' knowledge about the absolute benefit
of screening may help to explain differences in screening behaviour
between the two groups. The qualitative study (as Hewitson points out) is
indeed helping us to better understand the reduced uptake in screening
among decision aid participants. We hope to publish the findings from this
study shortly.
With regard to participants' perceptions of the information
materials, we found that 52% of the decision aid participants and 59% of
the control participants felt the information was encouraging bowel cancer
screening. None of the participants thought the decision aid or control
booklet was discouraging screening. This is perhaps not surprising given
that all participants were sent a bowel screening test kit, which some
participants may have interpreted as a sign of encouragement. In addition,
we know from our previous work that people often expect information about
screening to be in favour of screening [6 7]. Together, these factors may
have contributed to the small difference we observed in the way people
perceived the two resources.
We agree that more theoretical work is needed to examine how people
make screening decisions using high quality information, however we have
neither 'arbitrarily categorised' people nor expected lay people to have
'expert numerical understanding'.
1. Dormandy E, Hooper R, Michie S, Marteau TM. Informed choice to
undergo prenatal screening: a comparison of two hospitals conducting
testing either as part of a routine visit or requiring a separate visit. J
Med Screen 2002;9(3):109-14.
2. Dormandy E, Michie S, Hooper R, Marteau TM. Low uptake of prenatal
screening for Down syndrome in minority ethnic groups and socially
deprived groups: a reflection of women's attitudes or a failure to
facilitate informed choices? Int. J. Epidemiol. 2005;34(2):346-52.
3. Marteau T, Dormandy E, Michie S. A measure of informed choice.
Health Expectations 2001;4(2):99 - 108.
4. McCaffery KJ, Irwig L, Turner R, Chan SF, Macaskill P, Lewicka M,
et al. Psychosocial outcomes of three triage methods for the management of
borderline abnormal cervical smears: an open randomised trial. BMJ
2010;340.
5. Mathieu E, Barratt A, Davey H, McGeechan K, Howard K, Houssami N.
Informed Choice in Mammography Screening: A Randomized Trial of a Decision
Aid for 70-Year-Old Women. Archives of Internal Medicine 2007;167(19):2039
- 46.
6. Smith SK, Trevena L, Nutbeam D, Barratt A, McCaffery KJ.
Information needs and preferences of low and high literacy consumers for
decisions about colorectal cancer screening: utilizing a linguistic model.
Health Expect 2008;11(2):123-36.
7. Smith SK, Trevena L, Barratt A, Dixon A, Nutbeam D, Simpson JM, et
al. Development and preliminary evaluation of a bowel cancer screening
decision aid for adults with lower literacy. Patient Education and
Counseling 2009;75(3):358-67.
Competing interests: No competing interests