Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Development of a measurement tool to assess public awareness of cancer

Development of a measurement tool to assess public awareness of cancer British Journal of Cancer (2009) 101, S13 – S17 & 2009 Cancer Research UK All rights reserved 0007 – 0920/09 $32.00 www.bjcancer.com Full Paper Development of a measurement tool to assess public awareness of cancer 1 1 1 2 3 4 5 ,1 S Stubbings , K Robb , J Waller , A Ramirez , J Austoker , U Macleod , S Hiom and J Wardle Cancer Research UK Health Behaviour Research Centre, Department of Epidemiology and Public Health, UCL, Gower Street, London WC1E 6BT, UK; 2 3 Cancer Research UK Promoting Early Presentation Group, Institute of Psychiatry, King’s College London, St Thomas’ Hospital, London, UK; Cancer Research UK Primary Care Education Research Group, Cancer Epidemiology Unit, University of Oxford, Richard Doll Building, Roosevelt Drive, Oxford, UK; General Practice and Primary Care, Division of Community Based Sciences, Faculty of Medicine, University of Glasgow, 1 Horselethill Road, Glasgow, UK; Health Information Department, Cancer Research UK, 61 Lincoln’s Inn Fields, London, UK BACKGROUND: We aimed to develop and validate a measurement tool to assess cancer awareness in the general population: the cancer awareness measure (CAM). METHODS: Items assessing awareness of cancer warning signs, risk factors, incidence, screening programmes and attitudes towards help seeking were extracted from the literature or generated by expert groups. To determine reliability, the CAM was administered to a university participant panel (n¼ 148), with a sub-sample (n¼ 94) completing it again 2 weeks later. To establish construct validity, CAM scores of cancer experts (n¼ 12) were compared with those of non-medical academics (n¼ 21). Finally, university students (n¼ 49) were randomly assigned to read either a cancer information leaflet or a leaflet with control information before completing the measure, to ensure the CAM was sensitive to change. RESULTS: Cognitive interviewing indicated that the CAM was being interpreted as intended. Internal reliability (Cronbach’s a¼ 0.77) and test–retest reliability (r¼ 0.81) were high. Scores for cancer experts were significantly higher than those for non-medical academics (t(31)¼ 6.8, Po0.001). CAM scores were higher among students who received an intervention leaflet than the control leaflet (t(47)¼ 4.8, Po0.001). CONCLUSIONS: These studies show the psychometric properties of the CAM and support its validity as a measure of cancer awareness in the general population. British Journal of Cancer (2009) 101, S13 – S17. doi:10.1038/sj.bjc.6605385 www.bjcancer.com & 2009 Cancer Research UK Keywords: cancer awareness; measurement; psychometrics Cancer is a major burden worldwide (Parkin et al, 2005) and Grunfeld et al, 2002; McCaffery et al, 2003; Rudberg et al, 2005; a leading cause of mortality in the United Kingdom, claiming West et al, 2006; FitzGerald et al, 2008). over 150 000 lives a year (Cancer Research UK, 2009). Around European comparisons show the United Kingdom to have lower 289 000 new cases are diagnosed annually in the United Kingdom, than average cancer survival rates (Coleman et al, 2003), part of with one in three people developing cancer in their lifetime which is due to patients in the United Kingdom having more (Cancer Research UK, 2008a). Despite progress in reducing advanced stage of disease at diagnosis. Later presentation of cancer mortality rates, changes in the age distribution of the population symptoms in the United Kingdom than in other European will mean that cancer incidence will continue to rise (Boyle countries may contribute to this. Raising public awareness of and Ferlay, 2005). In recent years, the UK Governments have warning signs and promoting prompt presentation could reduce developed strategies aimed at reducing cancer incidence and patient-attributable delay and result in diagnosis at an earlier stage mortality; the most recent for England is the NHS Cancer Reform (Ramirez et al, 1999; Richards et al, 1999; Macdonald et al, 2006). Strategy (Department of Health, 2007), which emphasises the Awareness of cancer risk factors except for tobacco use is also importance of raising public awareness of early warning signs and low (Wardle et al, 2001; Grunfeld et al, 2002; McCaffery et al, 2003; risk factors. Marlow et al, 2007; Redeker et al, 2009), particularly in relation Existing evidence indicates that public awareness of warning to weight, alcohol consumption, high-fat diet, exercise and older signs is poor. In a population-based survey in England, fewer than age. It is estimated that around half of all cancers could be 1 in 10 respondents could recall Europe Against Cancer’s seven prevented by modification of lifestyle risk factors (Cancer warning signs for cancer (Brunswick et al, 2001). These findings Research UK, 2008b). Although awareness alone may not be are not unique to this study nor limited to generic warning signs; sufficient to motivate change, it is unrealistic to expect changes in awareness is also low for a range of cancers (Wardle et al, 2001; behaviour if people are not at least aware of the risk factors (Viswanath et al, 2006). There is currently no validated measure of general public *Correspondence: Professor J Wardle; E-mail: [email protected] awareness of cancer, although several questionnaires have been Cancer awareness measure S Stubbings et al S14 developed to assess awareness of specific cancers (Stager, 1993; and nine recognition items); (ii) nine items on anticipated time to Rees et al, 2003; Green and Kelly, 2004). In a review of the seek medical advice (asking about each of the warning signs); (iii) literature, Adlard and Hume (2003) also identified lack of 10 items on barriers to seeking medical advice (covering a range of agreement over the best way to measure cancer awareness. As a practical, service delivery and emotional barriers); (iv) 13 items on result, different question formats are often used. Questions asked awareness of risk factors (one open-ended question, 11 recognition in a prompted (recognition) format can elicit higher apparent items and one asking participants to rank the importance of levels of cancer awareness than those asked in an unprompted different types of risk factor); (v) seven items on cancer incidence (recall) format (Waller et al, 2004). Such variations in responses (one asking about overall cancer incidence and six asking about make it difficult to establish current levels of awareness or make the three most common cancers for men and women) and (vi) six comparisons across studies. items on awareness of NHS screening programmes (asking about Education campaigns designed to improve awareness of cancer awareness of the cervical, breast and bowel screening programmes in the general population have been carried out, but without and the age from which screening is offered for each). validated instruments, it is difficult to evaluate effectiveness. This highlights the need for a measure that will enable both researchers Validating the CAM and campaigning groups to evaluate the impact of their activities. Validating the CAM was a three-stage process, with each stage The aim of this research was therefore to develop and validate a assessing a different aspect of reliability or validity. The aims of standardised measurement tool to assess cancer awareness. A good stage one were to establish internal reliability and test–retest measure should have face validity, the questions should be reliability and carry out item analyses; stage two was designed to interpreted as intended by the target audience, and it should give establish construct validity and stage three to ensure that the stable results across two occasions between which knowledge has measure was sensitive to increases in levels of awareness. not changed (test–retest reliability). Good reliability is essential Data were analysed using SPSS 14.0. Parametric statistics (e.g. for obtaining precise estimates of knowledge and for giving the Pearson’s correlation, t-tests) were performed to analyse the best statistical power to detect change. A good measure also needs reliability and validity of the measure. Descriptive statistics were to be valid, that is, groups who by general consensus have a higher used to examine the characteristics of the samples. standing on the relevant constructs should score higher; in this case, we compared cancer experts and equivalently educated non- experts to test the instrument’s sensitivity to knowledge. Finally, if RESULTS it is to have value in assessing the impact of public health interventions, it should be sensitive to change; we tested this by Stage 1: Internal reliability and test–retest reliability comparing scores before and after exposure to a brief educational intervention. Sample and methods Five hundred and fifty-one e-mails were sent to a research ‘participant panel’ with an invitation to complete the CAM anonymously online. The panel consisted of members of the general public who had previously indicated that they were MATERIALS AND METHODS willing to participate in research. The questionnaire was completed in the available time by 148 (27%) panel members. The majority of Generation of items respondents were women (76%), white (86%) and educated to Following a review of the literature, existing awareness ques- degree level (68%) (Table 1). Two weeks later, respondents were tionnaires were examined and relevant items extracted. This was asked to complete the CAM a second time. Two weeks was judged supplemented with a review of the ‘grey’ literature (i.e. unpub- as an adequate period of time for respondents neither to recall lished surveys carried out by cancer charities and other organisa- precisely their original answers, nor to be likely to have had any tions) to include items not published in academic journals. major changes in cancer awareness. Ninety-four participants Following this review, an item pool consisting of 137 items was (63%) completed the questionnaire again. Data were matched created. These covered a range of topics including awareness of using e-mail addresses. warning signs and risk factors, cancer incidence and awareness of Internal reliability assesses the extent to which all the national screening programmes. Items were then excluded if they questionnaire items measure the same underlying construct (Kline, were poorly worded, used terminology not frequently used in the 2000). It is assessed using Cronbach’s a and a minimum score of United Kingdom (e.g. Pap test) or were attitudinal in nature (e.g. ‘I 0.7 should be obtained for a questionnaire to be considered believe there are no early symptoms of cancer’). Items relating to reliable (Bland and Altman, 1997). To assess the stability of a awareness of the purpose of screening, the benefits of early questionnaire over time, a measure of test–retest reliability must detection and cancer survival rates were also omitted from the be calculated (Kline, 2000). Pearson’s correlations are computed measure because the primary focus was symptom recognition. In using scores from two time points. It is important to identify addition, the research team generated several items specifically for whether respondents find items too easy or too difficult to answer, the instrument that had not been used in previous questionnaires. and it is recommended that items are excluded if they are Once consensus over the items had been reached, a first version answered correctly by 480% or o20% of participants (Kline, of the cancer awareness measure (CAM) was circulated to a panel 2000). Item discrimination reveals the ability of an individual item of experts (n¼ 16) including academic researchers, cancer charity to discriminate between those who have high or low overall representatives, general practitioners, oncologists and experts in knowledge scores, and items should be discarded if an item-to- the field of questionnaire design, to ensure content validity and total correlation of o0.2 is yielded (Streiner and Norman, 1995). face validity. In addition, cognitive interviews were conducted with the general public. These encourage respondents to verbalise their Analysis and results cognitions, making it possible to identify areas where interpreta- tion of the questions is ambiguous (Collins, 2003). Cognitive Internal reliability A Cronbach’s a of 0.77 was achieved for the interviews were conducted with a small sample of participants whole questionnaire, with the following a values obtained for each (n¼ 6) aged between 23 and 70 years. Minor modifications were sub-section: warning signs (9 recognition items) 0.77; anticipated made to the phrasing of several items as a result. time to seek medical advice (9 items) 0.90; barriers to seeking The final version of the CAM consisted of the following: (i) 10 medical advice (10 items) 0.73; risk factors (11 recognition items) items on awareness of warning signs (one open-ended question 0.79; NHS screening programmes (6 items) 0.54. Despite obtaining British Journal of Cancer (2009) 101(S2), S13 – S17 & 2009 Cancer Research UK Cancer awareness measure S Stubbings et al S15 Table 1 Demographic characteristics of the online study sample Table 2 Test–retest reliability of the CAM (n¼ 94) (n¼ 148) Awareness section Test–retest reliability N % Warning signs 0.73 Anticipated time to seek advice 0.86 Gender Barriers to seeking advice 0.72 Male 36 24.3 Risk factors 0.73 Female 112 75.7 Incidence per 100 people 0.78 Incidence – common cancers 0.33 Age (years) Screening programmes 0.75 18 – 24 40 27.0 Screening programmes – age at first invitation 0.77 25 – 34 33 22.3 Total 0.81 35 – 44 9 6.1 45 – 54 11 7.4 Abbreviation: CAM¼ cancer awareness measure. 55 – 64 39 26.4 65 and over 16 10.8 Ethnic origin White 126 85.1 Table 3 Demographic characteristics of the two samples (n¼ 33) Other 20 13.5 Non-medical academics Cancer experts Employment status (n¼ 21) (n¼ 12) Employed full-time 37 25.0 Employed part-time 11 7.4 Student 62 41.9 n % n % Other 38 25.7 Age (years) 25 – 34 2 9.5 7 58.3 Highest qualification obtained 35 – 44 8 38.1 4 33.3 No qualifications 2 1.4 45 – 54 9 42.9 — — O level/GCSE 5 3.4 55 – 64 2 9.5 1 8.3 A level 14 9.5 Degree or above 101 68.2 Still studying 23 15.5 Gender Male 9 42.9 1 8.3 Marital status Female 12 57.1 11 91.7 Single 58 39.2 Married/cohabiting 75 50.7 Highest qualification Divorced/widowed 14 9.5 Degree — — 2 16.7 Masters 2 9.5 3 25.0 Some variables do not add up to 100% due to missing data. PhD or equivalent 19 90.5 5 41.7 Other — — 2 16.7 an a below the minimum cutoff, the decision was made to retain awareness of NHS screening programmes on the grounds of content validity. (n¼ 21) recruited from a range of departments in the university. Participants were invited to take part by e-mail and could either Test–retest reliability With the exception of incidence of complete the questionnaire electronically or print it out and common cancers, high correlations over time were found for all complete a paper copy. The demographic characteristics of the sections (Table 2), all of which reached statistical significance samples can be seen in Table 3. (Po0.001). Analyses and results Item difficulty The majority of items matched the criterion of being answered correctly byo80% and420% of participants. The The cancer experts scored consistently higher than the non- few that did not were retained on the basis of face validity (e.g. medical academics (Table 4) and, with the exception of incidence, smoking being a risk factor for cancer, a lump being a warning this reached statistical significance for each awareness section. sign for cancer), because respondents might be surprised if they Although differences in incidence scores were not statistically were not included. significant, the standard deviations varied considerably (0.8 for the cancer experts and 12.5 for the non-medical group), suggesting Item discrimination Analyses revealed item-to-total correlations that experts were more consistent at answering the questions. 40.2 for each item, suggesting all items in the CAM should be retained. Stage 3: Sensitivity to change: brief educational intervention Stage 2: Construct validity: cancer ‘experts’ vs non-medical Sample and methods A convenience sample of 49 undergraduate academics and postgraduate students was recruited to participate in this Sample and methods The ‘known-groups’ method was used to stage. They were randomised to receive one of two leaflets to read establish construct validity. If the scores of two groups known to before completing a paper copy of the CAM. One was an differ in levels of cancer awareness are significantly different then educational leaflet (‘Cancer: the facts’), which included informa- the validity of the questionnaire is supported (DeVellis, 2003). tion on the aetiology of cancer, incidence rates, warning signs, risk Cancer experts (n¼ 12) were recruited from a large cancer charity, factors and available NHS screening programmes. The other was a while the ‘non-expert’ group comprised non-medical academics control leaflet (‘Recycle to save the environment’). & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S13 – S17 Cancer awareness measure S Stubbings et al S16 Table 4 Differences in awareness scores between non-medical reached statistical significance for all awareness sections (Table 6). academics and cancer experts (n¼ 33) Not surprisingly for such a brief intervention there was comparatively small mean difference between the education and Non-medical Cancer control groups, but there was also differentiated variability with academics experts standard deviations of 3.4 (education) and 11.5 (control). (n¼ 21) (n¼ 12) Awareness section (max score) Mean s.d. Mean s.d. t (31) DISCUSSION Warning signs (9) 7.3 2.4 8.8 0.4 2.9* Risk factors (55) 42.3 4.1 50.8 4.7 5.4** There has been a good deal of interest in public awareness of Incidence per 100 people 34.0 12.5 33.3 0.8 0.2 cancer, but in most studies measures are developed on an ad hoc Incidence – common cancers (6) 2.3 1.2 3.5 1.4 2.7* basis, often with a specific population in mind, and rarely Screening programmes (3) 1.9 0.4 3.0 0.0 9.4** address psychometric properties. This not only limits general- Screening programmes – age at 0.7 0.6 2.2 1.1 4.2** isation to other groups, but also precludes comparisons between first invitation (3) studies or groups. Total (77) 57.2 5.9 69.3 5.6 6.8** The CAM was developed to provide a validated measure of *Po0.05. **Po0.001. Participants were given a score of ‘1’ if their answer was awareness of early warning signs and risk factors, and barriers to correct (correct range, 30 – 36). seeking medical advice. Reliability of the measure was high, with a total Cronbach’s a of 0.77 and all but one of the sub-scales reaching the recommended cutoff of 0.7. Test–retest reliability was Table 5 Demographic characteristics of the two samples (n¼ 49) also good, with all sub-scales except awareness of the incidence of common cancers being 40.7. The low reliability of the incidence Control (n¼ 24) Intervention (n¼ 25) items probably reflects the fact that people had little idea and were just guessing the answers; and not making the same guess on n % n % different occasions. This is understandable because although people may be able to identify the most common cancer in men Age (years) (prostate) and women (breast), the second and third most 18 – 24 9 62.5 12 48.0 25 – 34 15 37.5 12 48.0 common cancers are rarely publicised as such. However, given 35+ — — 1 4.0 that awareness of the most common cancers could improve, these items were retained. Gender Construct validity was established using a ‘known-groups’ Male 9 37.5 9 36.0 design. Overall, those with expertise in cancer achieved signifi- Female 15 62.5 16 64.0 cantly higher mean scores than non-medical academics, showing the CAM has the ability to distinguish between groups with Ethnic origin established differences in levels of awareness. In addition, the CAM White 17 70.8 16 64.0 Other 7 29.2 9 36.0 was shown to be sensitive to increases in awareness following a brief educational intervention. Table 6 Differences in awareness scores between control and Limitations intervention participants (n¼ 49) The limitations to the validation of the measure should be noted. Control Intervention The mode of administration varied across different stages of (n¼ 24) (n¼ 25) piloting, including online and with paper-and-pencil versions. Awareness section Concerns have been raised about the quality of survey data (max score) Mean s.d. Mean s.d. t (47) collected through different modes (Bowling, 2005), although the mode of administration was consistent within each stage of Warning signs (9) 6.0 1.9 7.3 0.9 2.7* piloting so the potential impact would have been kept to a Risk factors (55) 39.8 4.6 43.7 5.0 2.8* Incidence per 100 people 29.2 11.5 32.8 3.4 1.5 minimum. For use in population surveys, we suggest that the CAM Screening programmes (3) 1.8 0.7 2.7 0.7 4.2** should ideally be given in an interview format to prevent Screening programmes – age at 0.6 0.6 1.8 0.9 5.2** participants from changing their previous answers in response to first invitation (3) the prompted format of subsequent questions. The pilot samples Total (71) 48.7 5.8 56.3 5.7 4.6** tended to be female and educated to degree level, which limits generalisation to other populations. However, given that the CAM *Po0.05. **Po0.001. Participants were given a score of ‘1’ if their answer was was sensitive to differences in levels of awareness in a highly correct (correct range, 30 – 36). This total score does not include all components of the awareness measure. The intervention leaflet did not contain information on educated academic sample, it is likely that differences would be timely presentation, barriers to seeking medical advice or incidence of most common equally marked in less educated populations. To date, the CAM is cancers, so these items have been omitted from the analysis. only available in English, but we anticipate that it will be translated into a range of languages and available for use in the future. Demographic characteristics of the two groups were similar (Table 5), with no significant differences in age, gender or ethnic Future work origin being observed. This paper describes the development of the generic CAM. Work is now underway to develop tumour-specific versions of the CAM, Analyses and results including breast, prostate, bowel, cervical and ovarian, and more The cancer education group scored consistently higher than the tumour-specific versions are planned. It is anticipated that the control group and, excluding results for cancer incidence, this CAM will evolve over time to reflect advances in knowledge, and to British Journal of Cancer (2009) 101(S2), S13 – S17 & 2009 Cancer Research UK Cancer awareness measure S Stubbings et al S17 include items on cancer beliefs that may help us to better tions and to assess the impact of interventions designed to target understand cancer-related behaviour. gaps in public awareness of cancer either in whole populations or specific sub-groups. CONCLUSION ACKNOWLEDGEMENTS The CAM is a reliable and valid measure of cancer awareness and can be used to provide a comprehensive assessment of cancer The Cancer Awareness Measure was funded by Cancer Research awareness. The CAM has now been administered in a large-scale, UK and the Department of Health. population-based, British sample with a substantial ethnic boost sample, using a face-to-face, home-based interview methodology, so standardised population data for the United Kingdom are Conflict of interest available for reference (Robb et al; Waller et al). In addition, the CAM can be used by researchers to develop informed interven- The authors declare no conflict of interest. REFERENCES Adlard JW, Hume MJ (2003) Cancer knowledge of the general public in the Marlow LAV, Waller J, Wardle J (2007) Public awareness that HPV is a risk United Kingdom: survey in a primary care setting and review of the factor for cervical cancer. Br J Cancer 97: 691–694 literature. Clin Oncol 15: 174–180 McCaffery K, Wardle J, Waller J (2003) Knowledge, attitudes, and Bland JM, Altman DG (1997) Cronbach’s alpha. BMJ 314: 572 behavioral intentions in relation to the early detection of colorectal Bowling A (2005) Mode of questionnaire administration can have serious cancer in the United Kingdom. Prev Med 36: 525–535 effects on data quality. J Public Health 27: 281–291 Parkin DM, Bray F, Ferlay J, Pisani P (2005) Global cancer statistics, 2002. Boyle P, Ferlay J (2005) Cancer incidence and mortality in Europe, 2004. CA Cancer J Clin 55: 74–108 Ann Oncol 16: 481–488 Ramirez AJ, Westcombe AM, Burgess CC, Sutton S, Littlejohns P, Richards Brunswick N, Wardle J, Jarvis MJ (2001) Public awareness of warning signs MA (1999) Factors predicting delayed presentation of symptomatic for cancer in Britain. Cancer Causes Control 12: 33–37 breast cancer: a systematic review. Lancet 353: 1127–1131 Cancer Research UK (2008a) Cancer Incidence. Cancer Research UK. http:// Redeker C, Wardle J, Wilder D, Hiom S, Miles A (2009) The launch of info.cancerresearchuk.org/cancerstats/incidence/ (accessed 28 November Cancer Research UK’s ‘Reduce the Risk’ campaign: baseline measure- 2008) ments of public awareness of cancer risk factors in 2004. Eur J Cancer 45: Cancer Research UK (2008b) Lifestyle and Cancer. Cancer Research UK. 827–836 http://info.cancerresearchuk.org/cancerstats/causes/lifestyle/ (accessed 21 July Rees C, Abed R, Sheard C (2003) Development of a reliable and valid 2008) questionnaire to test the prostate cancer knowledge of men with the Cancer Research UK (2009) Cancer Mortality. Cancer Research UK. disease. Patient Educ Couns 51: 285–292 http://info.cancerresearchuk.org/cancerstats/mortality/timetrends/?a¼ 5441 Richards MA, Westcombe AM, Love SB, Littlejohns P, Ramirez AJ (1999) (accessed 15 January 2009) Influence of delay on survival in patients with breast cancer: a systematic Coleman MP, Gatta G, Verdecchia A, Esteve J, Sant M, Storm H, Allemani review. Lancet 353: 1119– 1126 C, Ciccolallo L, Santaquilani M, Berrino F (2003) EUROCARE-3 Robb K, Stubbings S, Ramirez A, Macleod U, Austoker J, Waller J, Hiom S, summary: cancer survival in Europe at the end of the 20th century. Wardle J (2009) Public awareness of cancer in Britain: a population- Ann Oncol 14(Suppl 5): v128–v149 based survey of adults. Br J Cancer 101(Suppl 2): S18–S23 Collins D (2003) Pretesting survey instruments: an overview of cognitive Rudberg L, Nilsson S, Wikblad K, Carlsson M (2005) Testicular cancer and methods. Qual Life Res 12: 229–238 testicular self-examination—knowledge and attitudes of adolescent Department of Health (2007) Cancer Reform Strategy. http://www.dh.gov.uk/ Swedish men. Cancer Nurs 28: 256–262 en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/ Stager JL (1993) The Comprehensive Breast-Cancer Knowledge Test— DH_081006 (accessed 3 December 2007) validity and reliability. J Adv Nurs 18: 1133–1140 DeVellis RF (2003) Scale Development: Theory and Applications. Sage Streiner DL, Norman GR (1995) Health Measurement Scales: A Practical Publications Inc.: Thousand Oaks, California Guide to Their Development and Use. Oxford University Press: New York FitzGerald SC, Al Sahaf M, Furlong H, Pennycooke K, Healy C, Walsh TN Viswanath K, Breen N, Meissner H, Moser RP, Hesse B, Steele WR, (2008) Lack of awareness of oesophageal carcinoma among the public in Rakowski W (2006) Cancer knowledge and disparities in the information Ireland. Ir J Med Sci 177: 151– 154 age. J Health Commun 11: 1–17 Green PM, Kelly BA (2004) Colorectal cancer knowledge, perceptions, and Waller J, McCaffery K, Wardle J (2004) Beliefs about the risk factors for behaviors in African Americans. Cancer Nurs 27: 206–215 cervical cancer in a British population sample. Prev Med 38: 745–753 Grunfeld EA, Ramirez AJ, Hunter MS, Richards MA (2002) Women’s Waller J, Robb K, Stubbings S, Ramirez A, Macleod U, Austoker J, Hiom S, knowledge and beliefs regarding breast cancer. Br J Cancer 86: Wardle J. Awareness of cancer symptoms and anticipated help seeking 1373– 1378 among ethnic minority groups in England. Br J Cancer 101(Suppl 2): Kline P (2000) The Handbook of Psychological Testing. Routledge: London S24–S30 Macdonald S, Macleod U, Campbell NC, Weller D, Mitchell E (2006) Wardle J, Waller J, Brunswick N, Jarvis MJ (2001) Awareness of risk factors Systematic review of factors influencing patient and practitioner for cancer among British adults. Public Health 115: 173– 174 delay in diagnosis of upper gastrointestinal cancer. Br J Cancer 94: West R, Alkhatib MN, McNeill A, Bedi R (2006) Awareness of mouth cancer 1272– 1280 in Great Britain. Br Dent J 200: 167–169 & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S13 – S17 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png British Journal of Cancer Springer Journals

Development of a measurement tool to assess public awareness of cancer

Download PDF
5 pages

Loading next page...
 
/lp/springer-journals/development-of-a-measurement-tool-to-assess-public-awareness-of-cancer-sq4ushrRPA

References (68)

Publisher
Springer Journals
Copyright
Copyright © 2009 by The Author(s)
Subject
Biomedicine; Biomedicine, general; Cancer Research; Epidemiology; Molecular Medicine; Oncology; Drug Resistance
ISSN
0007-0920
eISSN
1532-1827
DOI
10.1038/sj.bjc.6605385
Publisher site
See Article on Publisher Site

Abstract

British Journal of Cancer (2009) 101, S13 – S17 & 2009 Cancer Research UK All rights reserved 0007 – 0920/09 $32.00 www.bjcancer.com Full Paper Development of a measurement tool to assess public awareness of cancer 1 1 1 2 3 4 5 ,1 S Stubbings , K Robb , J Waller , A Ramirez , J Austoker , U Macleod , S Hiom and J Wardle Cancer Research UK Health Behaviour Research Centre, Department of Epidemiology and Public Health, UCL, Gower Street, London WC1E 6BT, UK; 2 3 Cancer Research UK Promoting Early Presentation Group, Institute of Psychiatry, King’s College London, St Thomas’ Hospital, London, UK; Cancer Research UK Primary Care Education Research Group, Cancer Epidemiology Unit, University of Oxford, Richard Doll Building, Roosevelt Drive, Oxford, UK; General Practice and Primary Care, Division of Community Based Sciences, Faculty of Medicine, University of Glasgow, 1 Horselethill Road, Glasgow, UK; Health Information Department, Cancer Research UK, 61 Lincoln’s Inn Fields, London, UK BACKGROUND: We aimed to develop and validate a measurement tool to assess cancer awareness in the general population: the cancer awareness measure (CAM). METHODS: Items assessing awareness of cancer warning signs, risk factors, incidence, screening programmes and attitudes towards help seeking were extracted from the literature or generated by expert groups. To determine reliability, the CAM was administered to a university participant panel (n¼ 148), with a sub-sample (n¼ 94) completing it again 2 weeks later. To establish construct validity, CAM scores of cancer experts (n¼ 12) were compared with those of non-medical academics (n¼ 21). Finally, university students (n¼ 49) were randomly assigned to read either a cancer information leaflet or a leaflet with control information before completing the measure, to ensure the CAM was sensitive to change. RESULTS: Cognitive interviewing indicated that the CAM was being interpreted as intended. Internal reliability (Cronbach’s a¼ 0.77) and test–retest reliability (r¼ 0.81) were high. Scores for cancer experts were significantly higher than those for non-medical academics (t(31)¼ 6.8, Po0.001). CAM scores were higher among students who received an intervention leaflet than the control leaflet (t(47)¼ 4.8, Po0.001). CONCLUSIONS: These studies show the psychometric properties of the CAM and support its validity as a measure of cancer awareness in the general population. British Journal of Cancer (2009) 101, S13 – S17. doi:10.1038/sj.bjc.6605385 www.bjcancer.com & 2009 Cancer Research UK Keywords: cancer awareness; measurement; psychometrics Cancer is a major burden worldwide (Parkin et al, 2005) and Grunfeld et al, 2002; McCaffery et al, 2003; Rudberg et al, 2005; a leading cause of mortality in the United Kingdom, claiming West et al, 2006; FitzGerald et al, 2008). over 150 000 lives a year (Cancer Research UK, 2009). Around European comparisons show the United Kingdom to have lower 289 000 new cases are diagnosed annually in the United Kingdom, than average cancer survival rates (Coleman et al, 2003), part of with one in three people developing cancer in their lifetime which is due to patients in the United Kingdom having more (Cancer Research UK, 2008a). Despite progress in reducing advanced stage of disease at diagnosis. Later presentation of cancer mortality rates, changes in the age distribution of the population symptoms in the United Kingdom than in other European will mean that cancer incidence will continue to rise (Boyle countries may contribute to this. Raising public awareness of and Ferlay, 2005). In recent years, the UK Governments have warning signs and promoting prompt presentation could reduce developed strategies aimed at reducing cancer incidence and patient-attributable delay and result in diagnosis at an earlier stage mortality; the most recent for England is the NHS Cancer Reform (Ramirez et al, 1999; Richards et al, 1999; Macdonald et al, 2006). Strategy (Department of Health, 2007), which emphasises the Awareness of cancer risk factors except for tobacco use is also importance of raising public awareness of early warning signs and low (Wardle et al, 2001; Grunfeld et al, 2002; McCaffery et al, 2003; risk factors. Marlow et al, 2007; Redeker et al, 2009), particularly in relation Existing evidence indicates that public awareness of warning to weight, alcohol consumption, high-fat diet, exercise and older signs is poor. In a population-based survey in England, fewer than age. It is estimated that around half of all cancers could be 1 in 10 respondents could recall Europe Against Cancer’s seven prevented by modification of lifestyle risk factors (Cancer warning signs for cancer (Brunswick et al, 2001). These findings Research UK, 2008b). Although awareness alone may not be are not unique to this study nor limited to generic warning signs; sufficient to motivate change, it is unrealistic to expect changes in awareness is also low for a range of cancers (Wardle et al, 2001; behaviour if people are not at least aware of the risk factors (Viswanath et al, 2006). There is currently no validated measure of general public *Correspondence: Professor J Wardle; E-mail: [email protected] awareness of cancer, although several questionnaires have been Cancer awareness measure S Stubbings et al S14 developed to assess awareness of specific cancers (Stager, 1993; and nine recognition items); (ii) nine items on anticipated time to Rees et al, 2003; Green and Kelly, 2004). In a review of the seek medical advice (asking about each of the warning signs); (iii) literature, Adlard and Hume (2003) also identified lack of 10 items on barriers to seeking medical advice (covering a range of agreement over the best way to measure cancer awareness. As a practical, service delivery and emotional barriers); (iv) 13 items on result, different question formats are often used. Questions asked awareness of risk factors (one open-ended question, 11 recognition in a prompted (recognition) format can elicit higher apparent items and one asking participants to rank the importance of levels of cancer awareness than those asked in an unprompted different types of risk factor); (v) seven items on cancer incidence (recall) format (Waller et al, 2004). Such variations in responses (one asking about overall cancer incidence and six asking about make it difficult to establish current levels of awareness or make the three most common cancers for men and women) and (vi) six comparisons across studies. items on awareness of NHS screening programmes (asking about Education campaigns designed to improve awareness of cancer awareness of the cervical, breast and bowel screening programmes in the general population have been carried out, but without and the age from which screening is offered for each). validated instruments, it is difficult to evaluate effectiveness. This highlights the need for a measure that will enable both researchers Validating the CAM and campaigning groups to evaluate the impact of their activities. Validating the CAM was a three-stage process, with each stage The aim of this research was therefore to develop and validate a assessing a different aspect of reliability or validity. The aims of standardised measurement tool to assess cancer awareness. A good stage one were to establish internal reliability and test–retest measure should have face validity, the questions should be reliability and carry out item analyses; stage two was designed to interpreted as intended by the target audience, and it should give establish construct validity and stage three to ensure that the stable results across two occasions between which knowledge has measure was sensitive to increases in levels of awareness. not changed (test–retest reliability). Good reliability is essential Data were analysed using SPSS 14.0. Parametric statistics (e.g. for obtaining precise estimates of knowledge and for giving the Pearson’s correlation, t-tests) were performed to analyse the best statistical power to detect change. A good measure also needs reliability and validity of the measure. Descriptive statistics were to be valid, that is, groups who by general consensus have a higher used to examine the characteristics of the samples. standing on the relevant constructs should score higher; in this case, we compared cancer experts and equivalently educated non- experts to test the instrument’s sensitivity to knowledge. Finally, if RESULTS it is to have value in assessing the impact of public health interventions, it should be sensitive to change; we tested this by Stage 1: Internal reliability and test–retest reliability comparing scores before and after exposure to a brief educational intervention. Sample and methods Five hundred and fifty-one e-mails were sent to a research ‘participant panel’ with an invitation to complete the CAM anonymously online. The panel consisted of members of the general public who had previously indicated that they were MATERIALS AND METHODS willing to participate in research. The questionnaire was completed in the available time by 148 (27%) panel members. The majority of Generation of items respondents were women (76%), white (86%) and educated to Following a review of the literature, existing awareness ques- degree level (68%) (Table 1). Two weeks later, respondents were tionnaires were examined and relevant items extracted. This was asked to complete the CAM a second time. Two weeks was judged supplemented with a review of the ‘grey’ literature (i.e. unpub- as an adequate period of time for respondents neither to recall lished surveys carried out by cancer charities and other organisa- precisely their original answers, nor to be likely to have had any tions) to include items not published in academic journals. major changes in cancer awareness. Ninety-four participants Following this review, an item pool consisting of 137 items was (63%) completed the questionnaire again. Data were matched created. These covered a range of topics including awareness of using e-mail addresses. warning signs and risk factors, cancer incidence and awareness of Internal reliability assesses the extent to which all the national screening programmes. Items were then excluded if they questionnaire items measure the same underlying construct (Kline, were poorly worded, used terminology not frequently used in the 2000). It is assessed using Cronbach’s a and a minimum score of United Kingdom (e.g. Pap test) or were attitudinal in nature (e.g. ‘I 0.7 should be obtained for a questionnaire to be considered believe there are no early symptoms of cancer’). Items relating to reliable (Bland and Altman, 1997). To assess the stability of a awareness of the purpose of screening, the benefits of early questionnaire over time, a measure of test–retest reliability must detection and cancer survival rates were also omitted from the be calculated (Kline, 2000). Pearson’s correlations are computed measure because the primary focus was symptom recognition. In using scores from two time points. It is important to identify addition, the research team generated several items specifically for whether respondents find items too easy or too difficult to answer, the instrument that had not been used in previous questionnaires. and it is recommended that items are excluded if they are Once consensus over the items had been reached, a first version answered correctly by 480% or o20% of participants (Kline, of the cancer awareness measure (CAM) was circulated to a panel 2000). Item discrimination reveals the ability of an individual item of experts (n¼ 16) including academic researchers, cancer charity to discriminate between those who have high or low overall representatives, general practitioners, oncologists and experts in knowledge scores, and items should be discarded if an item-to- the field of questionnaire design, to ensure content validity and total correlation of o0.2 is yielded (Streiner and Norman, 1995). face validity. In addition, cognitive interviews were conducted with the general public. These encourage respondents to verbalise their Analysis and results cognitions, making it possible to identify areas where interpreta- tion of the questions is ambiguous (Collins, 2003). Cognitive Internal reliability A Cronbach’s a of 0.77 was achieved for the interviews were conducted with a small sample of participants whole questionnaire, with the following a values obtained for each (n¼ 6) aged between 23 and 70 years. Minor modifications were sub-section: warning signs (9 recognition items) 0.77; anticipated made to the phrasing of several items as a result. time to seek medical advice (9 items) 0.90; barriers to seeking The final version of the CAM consisted of the following: (i) 10 medical advice (10 items) 0.73; risk factors (11 recognition items) items on awareness of warning signs (one open-ended question 0.79; NHS screening programmes (6 items) 0.54. Despite obtaining British Journal of Cancer (2009) 101(S2), S13 – S17 & 2009 Cancer Research UK Cancer awareness measure S Stubbings et al S15 Table 1 Demographic characteristics of the online study sample Table 2 Test–retest reliability of the CAM (n¼ 94) (n¼ 148) Awareness section Test–retest reliability N % Warning signs 0.73 Anticipated time to seek advice 0.86 Gender Barriers to seeking advice 0.72 Male 36 24.3 Risk factors 0.73 Female 112 75.7 Incidence per 100 people 0.78 Incidence – common cancers 0.33 Age (years) Screening programmes 0.75 18 – 24 40 27.0 Screening programmes – age at first invitation 0.77 25 – 34 33 22.3 Total 0.81 35 – 44 9 6.1 45 – 54 11 7.4 Abbreviation: CAM¼ cancer awareness measure. 55 – 64 39 26.4 65 and over 16 10.8 Ethnic origin White 126 85.1 Table 3 Demographic characteristics of the two samples (n¼ 33) Other 20 13.5 Non-medical academics Cancer experts Employment status (n¼ 21) (n¼ 12) Employed full-time 37 25.0 Employed part-time 11 7.4 Student 62 41.9 n % n % Other 38 25.7 Age (years) 25 – 34 2 9.5 7 58.3 Highest qualification obtained 35 – 44 8 38.1 4 33.3 No qualifications 2 1.4 45 – 54 9 42.9 — — O level/GCSE 5 3.4 55 – 64 2 9.5 1 8.3 A level 14 9.5 Degree or above 101 68.2 Still studying 23 15.5 Gender Male 9 42.9 1 8.3 Marital status Female 12 57.1 11 91.7 Single 58 39.2 Married/cohabiting 75 50.7 Highest qualification Divorced/widowed 14 9.5 Degree — — 2 16.7 Masters 2 9.5 3 25.0 Some variables do not add up to 100% due to missing data. PhD or equivalent 19 90.5 5 41.7 Other — — 2 16.7 an a below the minimum cutoff, the decision was made to retain awareness of NHS screening programmes on the grounds of content validity. (n¼ 21) recruited from a range of departments in the university. Participants were invited to take part by e-mail and could either Test–retest reliability With the exception of incidence of complete the questionnaire electronically or print it out and common cancers, high correlations over time were found for all complete a paper copy. The demographic characteristics of the sections (Table 2), all of which reached statistical significance samples can be seen in Table 3. (Po0.001). Analyses and results Item difficulty The majority of items matched the criterion of being answered correctly byo80% and420% of participants. The The cancer experts scored consistently higher than the non- few that did not were retained on the basis of face validity (e.g. medical academics (Table 4) and, with the exception of incidence, smoking being a risk factor for cancer, a lump being a warning this reached statistical significance for each awareness section. sign for cancer), because respondents might be surprised if they Although differences in incidence scores were not statistically were not included. significant, the standard deviations varied considerably (0.8 for the cancer experts and 12.5 for the non-medical group), suggesting Item discrimination Analyses revealed item-to-total correlations that experts were more consistent at answering the questions. 40.2 for each item, suggesting all items in the CAM should be retained. Stage 3: Sensitivity to change: brief educational intervention Stage 2: Construct validity: cancer ‘experts’ vs non-medical Sample and methods A convenience sample of 49 undergraduate academics and postgraduate students was recruited to participate in this Sample and methods The ‘known-groups’ method was used to stage. They were randomised to receive one of two leaflets to read establish construct validity. If the scores of two groups known to before completing a paper copy of the CAM. One was an differ in levels of cancer awareness are significantly different then educational leaflet (‘Cancer: the facts’), which included informa- the validity of the questionnaire is supported (DeVellis, 2003). tion on the aetiology of cancer, incidence rates, warning signs, risk Cancer experts (n¼ 12) were recruited from a large cancer charity, factors and available NHS screening programmes. The other was a while the ‘non-expert’ group comprised non-medical academics control leaflet (‘Recycle to save the environment’). & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S13 – S17 Cancer awareness measure S Stubbings et al S16 Table 4 Differences in awareness scores between non-medical reached statistical significance for all awareness sections (Table 6). academics and cancer experts (n¼ 33) Not surprisingly for such a brief intervention there was comparatively small mean difference between the education and Non-medical Cancer control groups, but there was also differentiated variability with academics experts standard deviations of 3.4 (education) and 11.5 (control). (n¼ 21) (n¼ 12) Awareness section (max score) Mean s.d. Mean s.d. t (31) DISCUSSION Warning signs (9) 7.3 2.4 8.8 0.4 2.9* Risk factors (55) 42.3 4.1 50.8 4.7 5.4** There has been a good deal of interest in public awareness of Incidence per 100 people 34.0 12.5 33.3 0.8 0.2 cancer, but in most studies measures are developed on an ad hoc Incidence – common cancers (6) 2.3 1.2 3.5 1.4 2.7* basis, often with a specific population in mind, and rarely Screening programmes (3) 1.9 0.4 3.0 0.0 9.4** address psychometric properties. This not only limits general- Screening programmes – age at 0.7 0.6 2.2 1.1 4.2** isation to other groups, but also precludes comparisons between first invitation (3) studies or groups. Total (77) 57.2 5.9 69.3 5.6 6.8** The CAM was developed to provide a validated measure of *Po0.05. **Po0.001. Participants were given a score of ‘1’ if their answer was awareness of early warning signs and risk factors, and barriers to correct (correct range, 30 – 36). seeking medical advice. Reliability of the measure was high, with a total Cronbach’s a of 0.77 and all but one of the sub-scales reaching the recommended cutoff of 0.7. Test–retest reliability was Table 5 Demographic characteristics of the two samples (n¼ 49) also good, with all sub-scales except awareness of the incidence of common cancers being 40.7. The low reliability of the incidence Control (n¼ 24) Intervention (n¼ 25) items probably reflects the fact that people had little idea and were just guessing the answers; and not making the same guess on n % n % different occasions. This is understandable because although people may be able to identify the most common cancer in men Age (years) (prostate) and women (breast), the second and third most 18 – 24 9 62.5 12 48.0 25 – 34 15 37.5 12 48.0 common cancers are rarely publicised as such. However, given 35+ — — 1 4.0 that awareness of the most common cancers could improve, these items were retained. Gender Construct validity was established using a ‘known-groups’ Male 9 37.5 9 36.0 design. Overall, those with expertise in cancer achieved signifi- Female 15 62.5 16 64.0 cantly higher mean scores than non-medical academics, showing the CAM has the ability to distinguish between groups with Ethnic origin established differences in levels of awareness. In addition, the CAM White 17 70.8 16 64.0 Other 7 29.2 9 36.0 was shown to be sensitive to increases in awareness following a brief educational intervention. Table 6 Differences in awareness scores between control and Limitations intervention participants (n¼ 49) The limitations to the validation of the measure should be noted. Control Intervention The mode of administration varied across different stages of (n¼ 24) (n¼ 25) piloting, including online and with paper-and-pencil versions. Awareness section Concerns have been raised about the quality of survey data (max score) Mean s.d. Mean s.d. t (47) collected through different modes (Bowling, 2005), although the mode of administration was consistent within each stage of Warning signs (9) 6.0 1.9 7.3 0.9 2.7* piloting so the potential impact would have been kept to a Risk factors (55) 39.8 4.6 43.7 5.0 2.8* Incidence per 100 people 29.2 11.5 32.8 3.4 1.5 minimum. For use in population surveys, we suggest that the CAM Screening programmes (3) 1.8 0.7 2.7 0.7 4.2** should ideally be given in an interview format to prevent Screening programmes – age at 0.6 0.6 1.8 0.9 5.2** participants from changing their previous answers in response to first invitation (3) the prompted format of subsequent questions. The pilot samples Total (71) 48.7 5.8 56.3 5.7 4.6** tended to be female and educated to degree level, which limits generalisation to other populations. However, given that the CAM *Po0.05. **Po0.001. Participants were given a score of ‘1’ if their answer was was sensitive to differences in levels of awareness in a highly correct (correct range, 30 – 36). This total score does not include all components of the awareness measure. The intervention leaflet did not contain information on educated academic sample, it is likely that differences would be timely presentation, barriers to seeking medical advice or incidence of most common equally marked in less educated populations. To date, the CAM is cancers, so these items have been omitted from the analysis. only available in English, but we anticipate that it will be translated into a range of languages and available for use in the future. Demographic characteristics of the two groups were similar (Table 5), with no significant differences in age, gender or ethnic Future work origin being observed. This paper describes the development of the generic CAM. Work is now underway to develop tumour-specific versions of the CAM, Analyses and results including breast, prostate, bowel, cervical and ovarian, and more The cancer education group scored consistently higher than the tumour-specific versions are planned. It is anticipated that the control group and, excluding results for cancer incidence, this CAM will evolve over time to reflect advances in knowledge, and to British Journal of Cancer (2009) 101(S2), S13 – S17 & 2009 Cancer Research UK Cancer awareness measure S Stubbings et al S17 include items on cancer beliefs that may help us to better tions and to assess the impact of interventions designed to target understand cancer-related behaviour. gaps in public awareness of cancer either in whole populations or specific sub-groups. CONCLUSION ACKNOWLEDGEMENTS The CAM is a reliable and valid measure of cancer awareness and can be used to provide a comprehensive assessment of cancer The Cancer Awareness Measure was funded by Cancer Research awareness. The CAM has now been administered in a large-scale, UK and the Department of Health. population-based, British sample with a substantial ethnic boost sample, using a face-to-face, home-based interview methodology, so standardised population data for the United Kingdom are Conflict of interest available for reference (Robb et al; Waller et al). In addition, the CAM can be used by researchers to develop informed interven- The authors declare no conflict of interest. REFERENCES Adlard JW, Hume MJ (2003) Cancer knowledge of the general public in the Marlow LAV, Waller J, Wardle J (2007) Public awareness that HPV is a risk United Kingdom: survey in a primary care setting and review of the factor for cervical cancer. Br J Cancer 97: 691–694 literature. Clin Oncol 15: 174–180 McCaffery K, Wardle J, Waller J (2003) Knowledge, attitudes, and Bland JM, Altman DG (1997) Cronbach’s alpha. BMJ 314: 572 behavioral intentions in relation to the early detection of colorectal Bowling A (2005) Mode of questionnaire administration can have serious cancer in the United Kingdom. Prev Med 36: 525–535 effects on data quality. J Public Health 27: 281–291 Parkin DM, Bray F, Ferlay J, Pisani P (2005) Global cancer statistics, 2002. Boyle P, Ferlay J (2005) Cancer incidence and mortality in Europe, 2004. CA Cancer J Clin 55: 74–108 Ann Oncol 16: 481–488 Ramirez AJ, Westcombe AM, Burgess CC, Sutton S, Littlejohns P, Richards Brunswick N, Wardle J, Jarvis MJ (2001) Public awareness of warning signs MA (1999) Factors predicting delayed presentation of symptomatic for cancer in Britain. Cancer Causes Control 12: 33–37 breast cancer: a systematic review. Lancet 353: 1127–1131 Cancer Research UK (2008a) Cancer Incidence. Cancer Research UK. http:// Redeker C, Wardle J, Wilder D, Hiom S, Miles A (2009) The launch of info.cancerresearchuk.org/cancerstats/incidence/ (accessed 28 November Cancer Research UK’s ‘Reduce the Risk’ campaign: baseline measure- 2008) ments of public awareness of cancer risk factors in 2004. Eur J Cancer 45: Cancer Research UK (2008b) Lifestyle and Cancer. Cancer Research UK. 827–836 http://info.cancerresearchuk.org/cancerstats/causes/lifestyle/ (accessed 21 July Rees C, Abed R, Sheard C (2003) Development of a reliable and valid 2008) questionnaire to test the prostate cancer knowledge of men with the Cancer Research UK (2009) Cancer Mortality. Cancer Research UK. disease. Patient Educ Couns 51: 285–292 http://info.cancerresearchuk.org/cancerstats/mortality/timetrends/?a¼ 5441 Richards MA, Westcombe AM, Love SB, Littlejohns P, Ramirez AJ (1999) (accessed 15 January 2009) Influence of delay on survival in patients with breast cancer: a systematic Coleman MP, Gatta G, Verdecchia A, Esteve J, Sant M, Storm H, Allemani review. Lancet 353: 1119– 1126 C, Ciccolallo L, Santaquilani M, Berrino F (2003) EUROCARE-3 Robb K, Stubbings S, Ramirez A, Macleod U, Austoker J, Waller J, Hiom S, summary: cancer survival in Europe at the end of the 20th century. Wardle J (2009) Public awareness of cancer in Britain: a population- Ann Oncol 14(Suppl 5): v128–v149 based survey of adults. Br J Cancer 101(Suppl 2): S18–S23 Collins D (2003) Pretesting survey instruments: an overview of cognitive Rudberg L, Nilsson S, Wikblad K, Carlsson M (2005) Testicular cancer and methods. Qual Life Res 12: 229–238 testicular self-examination—knowledge and attitudes of adolescent Department of Health (2007) Cancer Reform Strategy. http://www.dh.gov.uk/ Swedish men. Cancer Nurs 28: 256–262 en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/ Stager JL (1993) The Comprehensive Breast-Cancer Knowledge Test— DH_081006 (accessed 3 December 2007) validity and reliability. J Adv Nurs 18: 1133–1140 DeVellis RF (2003) Scale Development: Theory and Applications. Sage Streiner DL, Norman GR (1995) Health Measurement Scales: A Practical Publications Inc.: Thousand Oaks, California Guide to Their Development and Use. Oxford University Press: New York FitzGerald SC, Al Sahaf M, Furlong H, Pennycooke K, Healy C, Walsh TN Viswanath K, Breen N, Meissner H, Moser RP, Hesse B, Steele WR, (2008) Lack of awareness of oesophageal carcinoma among the public in Rakowski W (2006) Cancer knowledge and disparities in the information Ireland. Ir J Med Sci 177: 151– 154 age. J Health Commun 11: 1–17 Green PM, Kelly BA (2004) Colorectal cancer knowledge, perceptions, and Waller J, McCaffery K, Wardle J (2004) Beliefs about the risk factors for behaviors in African Americans. Cancer Nurs 27: 206–215 cervical cancer in a British population sample. Prev Med 38: 745–753 Grunfeld EA, Ramirez AJ, Hunter MS, Richards MA (2002) Women’s Waller J, Robb K, Stubbings S, Ramirez A, Macleod U, Austoker J, Hiom S, knowledge and beliefs regarding breast cancer. Br J Cancer 86: Wardle J. Awareness of cancer symptoms and anticipated help seeking 1373– 1378 among ethnic minority groups in England. Br J Cancer 101(Suppl 2): Kline P (2000) The Handbook of Psychological Testing. Routledge: London S24–S30 Macdonald S, Macleod U, Campbell NC, Weller D, Mitchell E (2006) Wardle J, Waller J, Brunswick N, Jarvis MJ (2001) Awareness of risk factors Systematic review of factors influencing patient and practitioner for cancer among British adults. Public Health 115: 173– 174 delay in diagnosis of upper gastrointestinal cancer. Br J Cancer 94: West R, Alkhatib MN, McNeill A, Bedi R (2006) Awareness of mouth cancer 1272– 1280 in Great Britain. Br Dent J 200: 167–169 & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S13 – S17

Journal

British Journal of Cancer – Springer Journals

Published: Dec 3, 2009

There are no references for this article.