Improving cancer control through a community-based cancer awareness initiative.

OBJECTIVE: To assess the impact of the Cancer Research UK Cancer Awareness Roadshow on intentions to change health behaviours and use local health services related to cancer.
METHOD: Feedback forms from visitors to three Roadshows collected data on anticipated lifestyle changes and health service use following their visit to the Roadshow. Demographic predictors of intentions were investigated.
RESULTS: A total of 6009 individuals completed a feedback form. On average, respondents intended to make between two and three (2.55; SD=1.77) lifestyle changes, and use between none and one (0.59; SD=0.77) local health services following their visit. Multivariable analysis showed that age (p=0.001), ethnicity (p=0.006), and occupation (p=0.043) were significant predictors of anticipated lifestyle changes. Anticipated health service use was higher among men (p=0.001), younger groups (p<0.001), and smokers (p<0.001). Overall effects of ethnicity (p=0.001) and occupation (p<0.001) on anticipated health service use were also observed. Post-hoc analyses indicated stronger effects of the Roadshow among disadvantaged groups.
CONCLUSION: High levels of anticipated health behaviour change and health service use were observed among Roadshow visitors. Disadvantaged groups such as lower socioeconomic groups, ethnic minorities, and smokers showed particularly high levels of intention. A more in-depth evaluation of the Roadshow is warranted.

Introduction

Over 40% of cancers in the UK are attributable to lifestyle and environmental risk factors (Parkin et al., 2011). A large proportion of adults in England do not meet recommendations for key behaviours that influence cancer risk, including alcohol consumption, diet, smoking and physical activity, and this is particularly apparent among disadvantaged groups (Craig and Mindell, 2012; Hamer et al., 2012; Stringhini et al., 2011; West and Brown, 2012). Lower socioeconomic status groups also demonstrate more fatalistic attitudes towards cancer which could prevent timely help-seeking (Beeken et al., 2011).

Various avenues have been used to inform the public about cancer prevention and the importance of early diagnosis. However, traditional channels such as printed information disproportionately reach those with higher literacy levels who tend to be from more affluent backgrounds (Berkman et al., 2011; Boxell et al., 2012). This health literacy discrepancy compounds existing inequalities in access to and the understanding of cancer control information (Viswanath, 2005).

Cancer-related information is often accessed by the public through incidental face-to-face interactions (Niederdeppe et al., 2007). Community engagement activities take advantage of this, providing an opportunity to reach a broad range of people with motivational communications that aim to improve knowledge, attitudes, and behaviour (Resnicow et al., 2002). Although there is little evidence on the impact of community-based interventions, they may be an effective way of informing the public about cancer (Foster et al., 2010).

This study aims to assess the impact of a community-based mobile Roadshow on anticipated behaviour in terms of lifestyle changes and use of local health services.

Method

Procedure

This study was based on survey data from adults (n = 6009) attending the Cancer Research UK Cancer Awareness Roadshow in 2009. The Roadshow is a multi-component community intervention that aims to increase awareness and encourage behaviour change. It focuses on cancer prevention, screening, early diagnosis and access to health services and operates in deprived areas of the UK.

The Roadshow enables members of the public to talk to a specially trained cancer awareness nurse in an opportunistic setting. The nurse can answer questions and provide tailored information. There are interactive resources on display to help engage visitors, the option to have a BMI test or waist measurement, and leaflets on a range of cancer-related topics. Since 2006, Roadshow staff has interacted with over 350,000 visitors.

Adults attending one of three Roadshows in the Midlands, and Northwest and Northeast England were approached opportunistically after their visit to complete a brief questionnaire about their visit. Not all attendees were approached and no quotas were used.

Measures

Respondents were asked: how useful they found the Roadshow on a four-point scale ranging from ‘very useful’ to ‘not useful at all’; whether they knew of more ways to reduce the risk of cancer (‘yes’ or ‘no’); about any anticipated plans related to behaviour change and use of local health services following their visit. Respondent characteristics included gender, age, occupation, ethnicity and smoking status.

Analysis

A total health behaviour score was calculated by summing all anticipated changes an individual expected to make and dividing this by the total number of relevant behaviours to account for smokers being asked an additional question. The same approach was used for health service use.

Missing data were minimal (< 4%) for gender, age and ethnicity, and were deleted pairwise. Missing data for smoking status (25.27%) and occupation (12.00%) were ‘missing not at random’ and separate categories created. Missing data for the dependent variables could not be determined as respondents were asked to only tick a response if they intended to perform that action.

Multivariable between-subjects ANCOVAs determined independent predictors of intentions to change health behaviour and use health services. Partial-eta squared (ηρ2) effect sizes are reported and post-hoc comparisons were carried out using the Bonferroni correction. Significance levels were set at p < 0.05. Analyses were performed in SPSS v21.

Results

Individuals (n = 6009) aged 16 and over completed a questionnaire following their visit to a Roadshow mobile unit in the Midlands (n = 2355), the Northwest (n = 1279) or the Northeast (n = 2375). The sample was mixed in terms of gender, age, ethnicity and occupation (see Table 2). The Roadshow sample was well represented by lower socioeconomic groups as assessed by occupation (17.44% unemployed; 9.69% manual workers; 7.66% administrative).

Table 2

Sample characteristics and predictors of anticipated health behaviour changes and health service use.

		Health behaviour changes		Health service use
	n (valid %)	Mean (SD)	F-value	Mean (SD)	F-value
Gender
Male	2444 (42.16)	2.56 (1.75)	(F(1, 5701) = 1.56, p = 0.212, ηρ2 = 0.000)	0.61 (0.80)	(F(1, 5701) = 11.24, p = 0.001, ηρ2 = 0.002)
Female	3353 (57.84)	2.52 (1.77)		0.56 (0.73)
Age
16–24	1220 (20.48)	2.28 (1.67)	(F(1, 5701) = 11.52, p = 0.001, ηρ2 = 0.002)	0.66 (0.79)	(F(1, 5701) = 29.12, p < 0.001, ηρ2 = 0.005)
25–34	912 (15.31)	2.49 (1.76)		0.66 (0.82)
35–44	991 (16.63)	2.65 (1.75)		0.62 (0.78)
45–54	1028 (17.25)	2.78 (1.78)		0.61 (0.77)
55–64	980 (16.45)	2.69 (1.76)		0.56 (0.76)
65 +	827 (13.88)	2.48 (1.84)		0.42 (0.66)
Ethnicity
White	5100 (85.69)	2.54 (1.77)	(F(1, 5701) = 7.47, p = 0.006, ηρ2 = 0.001)	0.58 (0.76)	(F(1, 5701) = 11.65, p = 0.001, ηρ2 = 0.002)
South Asian	419 (7.04)	2.58 (1.73)		0.69 (0.83)
Black	253 (4.25)	2.96 (1.69)		0.73 (0.86)
Other	180 (3.02)	2.47 (1.66)		0.63 (0.83)
Occupation
Managerial	317 (5.28)	2.62 (1.80)	(F(1, 5701) = 4.09, p = 0.043, ηρ2 = 0.001)	0.49 (0.75)	(F(1, 5701) = 35.63, p < 0.001, ηρ2 = 0.006)
Professional	573 (9.54)	2.56 (1.64)		0.54 (0.77)
Key worker	157 (2.61)	2.40 (1.79)		0.68 (0.85)
Administration	460 (7.66)	2.56 (1.69)		0.51 (0.66)
Manual	582 (9.69)	2.55 (1.73)		0.58 (0.72)
Unemployed	1048 (17.44)	2.57 (1.83)		0.77 (0.85)
Other	240 (3.99)	2.66 (1.65)		0.71 (0.83)
Student	657 (10.93)	2.29 (1.70)		0.60 (0.76)
Retired	1331 (22.15)	2.56 (1.81)		0.47 (0.70)
Missing	544 (10.72)	2.69 (1.82)		0.67 (0.82)
Smoking status
No	3053 (50.81)	2.73 (1.78)	(F(1, 5701) = 0.327, p = 0.568, ηρ2 = 0.000)	0.47 (0.75)	(F(1, 5701) = 201.95, p < 0.001, ηρ2 = 0.034)
Yes	1744 (29.02)	2.37 (1.64)		0.87 (0.72)
Missing	1212 (20.17)	2.36 (1.85)		0.50 (0.80)

Most (93.21%) individuals felt they knew of more ways to reduce their risk of cancer and, on average, respondents anticipated making between two and three lifestyle changes (2.55; SD = 1.77). They were particularly likely to say they were going to be more aware of the signs/symptoms of cancer, and to intend to change energy balance behaviours (see Table 1). Few respondents indicated that they were going to reduce their alcohol consumption. A high proportion of smokers intended to visit the NHS stop smoking clinics and over a fifth of the sample intended to visit their General Practitioner.

Table 1

Proportion of respondents indicating that they anticipate making changes to their health behaviours and health service use following their visit⁎.

	n	%
Health behaviours
Maintain healthy weight	2719	45.25
Eat healthier diet	2569	42.75
Be more aware of signs/symptoms of cancer	2555	42.52
Be more physically active	2312	38.48
Quit smoking⁎⁎	1354	22.53
Protect from sunburn	1306	21.73
Reduce alcohol consumption	905	15.06
Health service use
NHS stop smoking services⁎⁎	942	54.01
General Practitioner	1350	22.47
Local health and fitness group	426	7.09
Local weight loss group	336	5.59

Each interaction is personalised and therefore not all topics are covered with every visitor.

Data includes individuals who responded ‘yes’ to the smoking status question only (n = 1744).

As shown in Table 2, age (p = 0.001), ethnicity (p = 0.006), and occupation (p = 0.043) were significant predictors of anticipated health behaviour change. Black respondents (vs. all ethnicities; all ps < 0.001) were significantly more likely to anticipate changing their behaviours, while those aged 16–24 (vs. 35–44, 45–54 and 55–64 age groups; all ps < 0.001) were significantly less likely.

Respondents anticipated using an average of 0.59 (SD = 0.77) local health services following their visit. As shown in Table 2, gender (p = 0.001), age (p < 0.001), ethnicity (p = 0.001), occupation (p < 0.001) and smoking status (p < 0.001) were significant predictors of anticipated health service use. Respondents who were unemployed (vs. administration, students, managerial, manual, professional and retired, all ps < 0.001) and smokers (vs. non-smokers, ps < 0.001) were significantly more likely to anticipate using local health services after visiting the Roadshow. Fewer respondents who were 65 + (vs. all ages, all ps < 0.01), white (vs. south Asian and Black, all ps < 0.05) and retired (vs. students, key workers, other, and unemployed all ps < 0.05) anticipated using local health services.

Discussion

These data from adults attending the Cancer Research UK Cancer Awareness Roadshow demonstrate the success of the initiative in attracting people from a lower socioeconomic background to engage in discussions about cancer control. Such groups are notoriously hard to access (Alcaraz et al., 2011; Yancey et al., 2006) and tend to have less exposure to quality health information sources (Askelson et al., 2011). It was therefore reassuring that several ‘hard to reach’ groups were particularly well represented. For example, in comparison with national data, respondents were more likely to be unemployed (17.4% vs. 7.8%), and were more likely to smoke (29.0% vs. 21.0%) (Office for National Statistics, 2013; West and Brown, 2012). The proportion of Black (4.3% vs. 3.3%) and Asian (7.0% vs. 7.5%) groups were comparable to national averages (Office for National Statistics, 2012).

On average, respondents anticipated making between two and three lifestyle changes following their visit, of which weight control, diet, physical activity and increasing awareness of cancer symptoms were the most common. Alcohol consumption was a noticeably difficult behaviour to influence. On average, respondents anticipated making use of between none and one local health services following their visit, with smoking cessation or visiting the GP the most popular. Particularly high levels of intentions to make lifestyle changes and/or use local health services were noted among smokers, ethnic minorities and lower socioeconomic groups.

Considering that the majority of individuals act on their intentions (Sheeran, 2002), these findings suggest the Roadshow may be a useful channel through which to encourage behaviour change. However, the absence of a comparison group that did not attend the Roadshow limits the extent to which the initiative can be considered responsible for the high levels of intentions reported. The study was also limited by self-reported data that assessed anticipated rather than actual behaviour change. It is possible that the sample were more motivated to find out about cancer than the general population as they not only attended the Roadshow, but also agreed to complete a questionnaire. These preliminary data do however provide support for the development of a larger and more in-depth evaluation of the Roadshow. This may help to further demonstrate the value of community-based initiatives in improving cancer control behaviours among ‘hard to reach’ groups (Alcaraz et al., 2011; Foster et al., 2010).

Conflict of interest

Smith was funded by Cancer Research UK as an academic advisor on this project. The work was initiated by Cancer Research UK, analysed by Smith and interpreted and verified by all authors. Rendell, George and Power are employed by Cancer Research UK and Power has an honorary research contract at UCL.

Ethics approval

Cancer Research UK is a Market Research Society company partner and all research is carried out according to the MRS Code of Conduct. This study used anonymised records and datasets available from the Cancer Awareness Roadshow team at Cancer Research UK who had already acquired appropriate permissions from Roadshow visitors.