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

Learn More →

A randomised controlled trial of an intervention to promote early presentation of breast cancer in older women: effect on breast cancer awareness

A randomised controlled trial of an intervention to promote early presentation of breast cancer... British Journal of Cancer (2009) 101, S40 – S48 & 2009 Cancer Research UK All rights reserved 0007 – 0920/09 $32.00 www.bjcancer.com Full Paper A randomised controlled trial of an intervention to promote early presentation of breast cancer in older women: effect on breast cancer awareness 1 ,1 1 1 1 1 1 L Linsell , LJL Forbes , M Kapari , C Burgess , L Omar , L Tucker and AJ Ramirez King’s College London, Cancer Research UK Promoting Early Presentation Group, Institute of Psychiatry, St Thomas’ Hospital, London SE1 7EH, UK BACKGROUND: There is virtually no evidence for the effectiveness of interventions to promote early presentation in breast cancer. We aimed to test the efficacy of an intervention to equip older women with the knowledge, skills, confidence and motivation to detect symptoms and seek help promptly, with the aim of promoting early presentation with breast cancer symptoms. METHODS: We randomised 867 women aged 67–70 years attending for their final routine appointment on the UK NHS Breast Screening Programme to receive: a scripted 10-min interaction with a radiographer plus a booklet, a booklet alone or usual care. The primary outcome was whether or not a woman was breast cancer aware based on knowledge of breast cancer symptoms and age- related risk, and reported breast checking. RESULTS: At 1 month, the intervention increased the proportion who were breast cancer aware compared with usual care (interaction arm: 32.8% vs 4.1%; odds ratio (OR): 24.0, 95% confidence interval (CI): 7.7–73.7; booklet arm: 12.7% vs 4.1%; OR: 4.4, 95% CI: 1.6–12.0). At 1 year, the effects of the interaction plus booklet, and the booklet, on breast cancer awareness were largely sustained, although the interaction plus booklet remained much more effective. CONCLUSION: An intervention to equip older women with the knowledge, skills, confidence and motivation to detect breast cancer symptoms and seek help promptly increases breast cancer awareness at 1 year. Future research will evaluate whether the intervention promotes early presentation and reduces breast cancer mortality. British Journal of Cancer (2009) 101, S40 – S48. doi:10.1038/sj.bjc.6605389 www.bjcancer.com & 2009 Cancer Research UK Keywords: aged; awareness; breast neoplasms; complex intervention; delayed presentation; randomised controlled trials Women in the United Kingdom have poorer survival from We have built two versions of an intervention to promote early breast cancer than many other Western European countries presentation of breast cancer in older women (Burgess et al, 2008) (Berrino et al, 2007), and differences in stage at diagnosis are aiming to equip them with breast cancer awareness: a booklet largely responsible (Sant et al, 2003). Late stage at diagnosis is containing health-promoting messages and a scripted one-to-one almost certainly due to late presentation by some women and interaction with a radiographer, supported by the booklet, both delays in onward referral by some general practitioners. In the designed to be delivered during the final routine appointment on developed world, 17–35% of women with breast cancer delay the NHS Breast Screening Programme, a setting giving access to presenting for 43 months, and 9–20% delay 46 months most English women aged 67–70 years. We targeted older women (Westcombe et al, 1999; Arndt, 2002). We have found no because they are at higher risk of breast cancer (Cancer Research high quality evidence of effectiveness of interventions to promote UK, 2008), are more likely to delay presentation (Ramirez et al, early presentation in breast, or any other cancer (Austoker et al, 1999) and have poorer survival (Office for National Statistics, 2009). 2008). The positive predictive value of breast symptoms for breast Risk factors for delay in presentation in breast cancer include cancer is higher in older than younger women (Nichols et al, 1981). older age, less education and presenting with non-lump sym- We based the design of the intervention on a theoretical ptoms (Ramirez et al, 1999). Older women have particularly framework for delayed presentation (Bish et al, 2005), and poor knowledge of non-lump symptoms and the increase in incorporated techniques to maximise the probability of behaviour breast cancer risk with age (Grunfeld et al, 2002; Linsell et al, change (Rollnick and Miller, 1995; Gollwitzer, 1999; Jepson, 2000; 2008). About 20% of older women report that they never look Wardle et al, 2003). at or feel their breasts (Linsell et al, 2008). We hypothesise that We have developed and evaluated the intervention in line with breast cancer awareness, which might include breast cancer the Medical Research Council guidance on complex interventions knowledge and the confidence, skills and motivation to detect (Medical Research Council, 2008). We have previously shown in a symptoms and seek help appropriately, will reduce delays in before-and-after exploratory trial that the intervention increased presentation. breast cancer awareness in older women at 6 months (Burgess et al, 2009). We report here the 1 month and 1 year results of a randomised controlled trial (RCT) of efficacy of the 10-min interaction with a radiographer supported by a booklet vs the *Correspondence: Dr LJL Forbes; E-mail: [email protected] booklet alone vs usual care alone. Intervention to promote early presentation of breast cancer L Linsell et al S41 MATERIALS AND METHODS five radiographers and two research psychologists delivered the interactions, and received ongoing performance feedback through- Participants and setting out. All interactions were video recorded and quality of each was assessed by an independent rater using a trial-specific quality Participants were a consecutive series of women aged 67–70 checklist assessing content and style of delivery. Quality scores attending final routine appointments on the NHS Breast Screening were standardised on a scale of 0–100. Programme. All women attending were eligible, unless they had a significant disorder that may have affected their ability to consent or participate, insufficient English or other language difficulties, or Outcomes were going overseas during the subsequent 6 months. The women We measured outcomes at baseline and 1, 6 and 12 months after were recruited from seven breast screening units in London and randomisation using a trial-specific questionnaire, an earlier Surrey. version of which we have used in a survey (Linsell et al, 2008) and an exploratory trial (Burgess et al, 2009). We have demon- Procedure strated good test–retest reliability (Linsell, 2006) and sensitivity to change (Burgess et al, 2009). The screening units sent an invitation letter and information sheet The primary outcome was proportion of women achieving to each potential participant 2 weeks before her appointment. On breast cancer awareness at 1 month, measured using a breast the day of attendance, a trained radiographer assessed whether the cancer awareness score. The score was a composite of responses to woman was eligible and obtained written informed consent. After three questions from the questionnaire, relating to knowledge of completing a baseline questionnaire, women were randomly symptoms, knowledge of age-related risk and reported breast allocated to: usual care, booklet alone or the 10-min one-to-one checking: interaction supported by the booklet, in addition to usual care. We Knowledge of breast cancer symptoms:‘Do you know any of the sent trial-specific postal questionnaires at 1 and 12 months after warning signs of breast cancer? If yes, please circle the signs you randomisation to collect outcome data. know below’. Women circled symptoms on a scattered list of 11 symptoms (two lump and nine non-lump). To score one point, The intervention the woman had to identify at least five non-lump symptoms, that is, over half; Usual care The screening unit receptionist informed each woman Knowledge of age-related risk:‘In the next year who is most who had received her final routine mammogram that she was no likely to get breast cancer?’ Response categories: a 30-year-old longer eligible for routine screening, advised her that she might woman, a 50-year-old woman, a 70-year-old woman, a woman continue to be screened every 3 years on request, and provided a of any age. To score one point, the woman had to identify that a card with contact details and a suggested date for contact. 70-year-old woman was most likely to get breast cancer; Breast checking:‘How often do you check your breasts?’ Response The booklet In addition to usual care, a radiographer gave a categories: rarely or never, at least every 6 months, at least once booklet to each woman who had received her final routine a month, at least once a week. To score one point, the woman had mammogram. The booklet conveyed key breast cancer awareness to report checking her breasts at least once a month or at least messages, including: once a week. A list of breast cancer symptoms; Each item was given equal weighting and contributed one point Age-related and absolute risk of developing breast cancer; to the total score (range: 0–3). To achieve breast cancer awareness, How to detect a breast change; the woman had to respond correctly to all three items. What to do on discovering a breast change; We also collected data on relationship status, education and A strong direct recommendation to seek medical attention ethnicity at baseline. To estimate socio-economic status, we used immediately on discovering a breast change, outlining the the Index of Multiple Deprivation (Communities and Local benefits of prompt help seeking and suggestions for overcoming Government, 2007) based on area of residence. This is a measure barriers such as embarrassment and fear; of deprivation at the small area level (32 482 areas in England) A direct recommendation to tell someone close in the event of based on seven dimensions: income, employment, health, educa- discovering a breast change; tion, housing and services, living environment and crime. Every An action plan to be completed by the woman about how she area in England is ranked from 1 (most deprived) to 32 482 (least will be breast aware and what she will do on discovering a breast deprived); median rank is 16 241. We assigned each woman a rank change; of Index of Multiple Deprivation according to the rank of her area A series of statements describing possible positive feelings of residence. (relief, reassurance, satisfaction), resulting from seeking help immediately with a breast change; Sample size A reminder that she might request further breast screening. We estimated that 2% of women would be breast cancer aware at The interaction plus booklet In addition to usual care, women baseline (Burgess et al, 2009) and that there would be a 12% received a scripted 10-min one-to-one interaction with a radio- difference between trial arms. Incorporating a design effect to take grapher or research psychologist. During this, the radiographer/ account of clustering by centre and radiographer (Lee and psychologist verbally delivered all the messages in the booklet in a Thompson, 2005) (assuming an intracluster correlation coefficient positive, collaborative and motivational style, referred to the of 0.08 and 14 participants per centre–radiographer cluster) and booklet throughout and gave it to the woman to take home. She allowing for 70% response, we required 238 women per arm (total also showed photographs of breast cancer signs and demonstrated 714) with a significance level of 5% (two-sided) and power of 80%. and rehearsed breast checking using a silicone breast. The radiographer was able to tailor the key messages by checking the Randomisation woman’s understanding and answering any questions. We randomised women individually on the day of attendance, with Quality assurance and quality control for the interaction plus equal probability of assignment to each arm. The trial statistician booklet After receiving training and being assessed as competent, computer generated the allocation sequence using stratified block & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S40 – S48 Intervention to promote early presentation of breast cancer L Linsell et al S42 1945 targeted with mailing 1209 (62.2%) attended and assessed for eligibility 342 (28.3%) excluded Reasons: 176 chose not to participate 70 insufficient English 48 missed for logistic reasons 28 significant disorder 20 going overseas 867 (71.7%) randomly allocated 287 (33.1%) allocated 294 (33.9%) allocated 286 (33.0%) allocated to usual care to usual care plus to usual care plus booklet booklet and interaction 286 received usual care 291 received booklet 279 received interaction Questionnaire Questionnaire Questionnaire completed at completed at completed at Baseline: n =287 Baseline: n =292 Baseline: n =284 1 month: n =251 (88%) 1 month: n =261 (91%) 1 month: n =257 (88%) 1 year: n =237 (83%) 1 year: n =240 (84%) 1 year: n =245 (83%) Lost to follow-up: Lost to follow-up: Lost to follow-up: n =10 (4%) n =7 (2%) n =8 (3%) 288 (98.0%) in main 282 (98.6%) in main 281 (97.9%) in main analysis analysis analysis 6 excluded from main 4 excluded from main 6 excluded from main analyses (no data) analyses (no data) analyses (no data) Figure 1 Flow of participants through trial. randomisation with centre and radiographer as stratification between repeated observations from the same individual. To variables (block sizes of three, six and nine). To ensure examine the intervention effect, we tested the interaction between concealment, assignments were enclosed in sequentially num- intervention group and time in each model, and presented odds bered, opaque, sealed envelopes and stored by the trial coordi- ratios (OR) with 95% confidence intervals (CIs). We also analysed nator before randomisation. The radiographer recorded each the data adjusting for stratification variables, relationship status, participant’s trial identification number on the envelope before education, ethnicity and Index of Multiple Deprivation, fitting opening it. categorical variables as binary variables and Index of Multiple Deprivation on a continuous scale (0–100). We calculated the intracluster correlation coefficient for radiographer and centre Statistical analysis using one-way analysis of variance, adjusting for unequal cluster We analysed the data by intention to treat: all participants were size (Fleiss, 1981; Armitage and Berry, 1994). analysed in the groups to which they were allocated. We summarised continuous variables using means, standard devia- tions, medians and ranges, and categorical variables using counts RESULTS and percentages. We used two-sided significance tests, taking P¼ 0.05 as significant. All analyses were performed using Stata Flow of participants version 10.0. The primary comparative analyses for all outcomes examined the difference between baseline and 1, 6 and 12 months Between August 2007 and May 2008, we randomised 867 women to for each pair of intervention groups. We used robust generalised one of three arms: usual care (n¼ 287), the booklet in addition to estimating equations (Zeger and Liang, 1986) with unstructured usual care (n¼ 294) and the interaction supported by the booklet correlation structure using a logit link and binomial distribution in addition to usual care (n¼ 286) (Figure 1). Only 15% (176 out of for the outcomes. This method takes account of the correlation 1209) of women who were assessed for eligibility chose not to British Journal of Cancer (2009) 101(S2), S40 – S48 & 2009 Cancer Research UK Follow-up Allocation Enrolment Targeted Intervention to promote early presentation of breast cancer L Linsell et al S43 Table 1 Baseline characteristics of participants Usual care (n¼ 287) Booklet (n¼ 292) Interaction plus booklet (n¼ 284) Relationship status, n (%) (n ¼ 285) (n¼ 283) (n¼ 282) Married or cohabiting 162 (56.8) 151 (53.4) 174 (61.7) Widowed 56 (19.7) 61 (21.6) 50 (17.7) Single 22 (7.7) 28 (9.9) 17 (6.0) Divorced or separated 45 (15.8) 43 (15.2) 41 (14.5) Education, n (%) (n ¼ 266) (n¼ 269) (n¼ 263) No formal qualifications 93 (35.0) 109 (40.5) 116 (44.1) O level or school certificate 89 (33.5) 81 (30.1) 77 (29.3) A level or higher school certificate 34 (12.8) 31 (11.5) 29 (11.0) Degree or above 50 (18.8) 48 (17.8) 41 (15.6) Ethnic group, n (%) (n ¼ 281) (n¼ 284) (n¼ 280) White British 187 (66.6) 196 (69.0) 186 (66.4) White other 23 (8.2) 23 (8.1) 32 (11.4) Black-Caribbean 36 (12.8) 38 (13.4) 37 (13.2) Other 35 (12.5) 27 (9.5) 25 (8.9) Index of multiple deprivation, median rank (IQR) (n ¼ 286) (n¼ 292) (n¼ 284) (1 (most) to 32 482 (least) deprived) 14 557 (8222 – 5989) 16 511 (8809 – 6184) 15 375 (8575 – 5729) participate. Eleven women did not receive the allocated interven- although estimated differences between arms were mostly slightly tion and 25 were lost to follow-up (20 withdrew consent, three larger (data not shown). moved with no forwarding address, one for medical reasons and one died). We included women with data for the primary outcome Knowledge of breast cancer symptoms on at least one occasion in the main analysis (n¼ 851). We were unable to measure any outcomes for four women who did not Forty-two per cent of women were able to identify five or more complete any questionnaires (two booklet arm, two interaction non-lump symptoms at baseline. At 1 month, the interaction plus arm). We received breast cancer awareness questionnaires from booklet increased the proportion of women able to identify five or 89% of those randomised at 1 month and 83% at 1 year; response more non-lump symptoms compared with usual care (78.9% vs rates were similar in each arm (Figure 1). 54.2%; difference 24.7%; OR: 2.5, 95% CI: 1.7–3.6) but the booklet alone did not (61.6% vs 54.2%; difference 7.4%; OR: 1.1, 95% CI: 0.8–1.5). The increase in knowledge of symptoms associated with Baseline socio-demographic characteristics the interaction plus booklet was maintained at 1 year. Before Socio-demographic characteristics were well balanced across the receiving the intervention, the women were able to identify a arms (Table 1), except for a slight difference in the proportion of median of four non-lump symptoms from the list of nine (IQR: women with no educational qualifications (35% usual care, 41% 2–6), and most recognised a lump in the breast or armpit as booklet and 44% interaction arm). Women in the booklet arm had symptoms. At 1 year, this increased to a median of six symptoms a similar deprivation score to the English median, but women in (IQR: 4–9) among those receiving the booklet only, and to seven the usual care and interaction plus booklet arms lived in somewhat (IQR: 4–9) among those receiving the interaction plus booklet. more deprived areas. For women invited for screening but not The intervention had most impact on the two least recognised recruited, the only characteristic we had data was postcode (and, symptoms, redness of skin and nipple rash (Figure 3). therefore, rank of Index of Multiple Deprivation). The median rank for the 1078 women not recruited was 9068 (interquartile range (IQR): 4784–15 802) compared with 15 664 (IQR: 8589–25 989) for Knowledge of age-related risk the 867 women in the trial, so women who did not take part lived Only 11.4% of the women knew that a 70-year-old woman was in more deprived areas than those who did. most at risk of breast cancer at baseline. At 1 month, the inter- action plus booklet increased the proportion knowing that a Breast cancer awareness 70-year-old woman was at most risk of breast cancer compared with usual care (44.7% vs 8.7%; difference 36.0%: OR: 9.5, 95% CI: Table 2 shows the main results for breast cancer awareness score 5.1–17.6), as did the booklet alone (24.9% vs 8.7%; difference and its components for baseline, 1 month and 1 year (6 months 16.2%; OR: 3.2, 95% CI: 1.8–5.8). At 1 year, the effect of the inter- data are not presented). Figure 2 illustrates breast cancer action plus booklet and the booklet alone remained significant. awareness and components of the score over the 12-month period. Overall, only 2.7% of women were breast cancer aware at baseline. At 1 month, the interaction plus booklet increased the proportion Breast checking who were breast cancer aware compared with usual care (32.8% vs 4.1%; OR: 24.0, 95% CI: 7.7–73.7), as did the booklet, although the About half of the women reported checking their breasts at least effect of the booklet was much less striking (12.7% vs 4.1%; OR: once a month at baseline. At 1 month, the interaction plus booklet 4.4, 95% CI: 1.6–12.0). At 1 year, the effect of the interaction plus increased the proportion of women checking their breasts at least booklet, and the booklet alone, on breast cancer awareness monthly compared with usual care (77.7% vs 62.5%; difference remained significant, with the interaction plus booklet remaining 15.2%; OR: 2.0, 95% CI: 1.4–2.8), but the booklet alone did more effective. The results of the adjusted analysis were similar, not (61.3% vs 62.5%; difference 1.2%; OR: 1.2, 95% CI: 0.9–1.6). & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S40 – S48 Intervention to promote early presentation of breast cancer L Linsell et al S44 Table 2 Breast cancer awareness and component items at 1 month and 1 year after randomisation Baseline 1 month 1 year Interaction Interaction Interaction Usual care Booklet plus booklet Usual care Booklet plus booklet Usual care Booklet plus booklet Breast cancer awareness Number (%) breast cancer aware 9/267 (3.4) 8/275 (2.9) 5/272 (1.8) 10/244 (4.1) 30/237 (12.7) 75/229 (32.8) 9/229 (3.9) 26/227 (11.5) 53/225 (23.6) Odds ratio (95% CI), 1.0 4.4 (1.6 – 2.0) 24.0 (7.7 – 73.7) 1.0 3.5 (1.2 – 10.5) 15.2 (4.8 – 47.8) P-value (vs usual care) P¼ 0.004 Po0.001 P¼ 0.025 Po0.001 Knowledge of breast cancer symptoms Number (%) identifying 111/284 (39.1) 126/286 (44.1) 122/280 (43.6) 136/251 (54.2) 151/245 (61.6) 187/237 (78.9) 142/233 (60.9) 167/236 (70.8) 170/230 (73.9) X5 non-lump symptoms Odds ratio (95% CI), 1.0 1.1 (0.8 – 1.5) 2.5 (1.7 – 3.6) 1.0 1.3 (0.9 – 1.9) 1.7 (1.1 – 2.4) P-value (vs usual care) P¼ 0.61 Po0.001 P¼ 0.23 P ¼ 0.01 Knowledge of age-related risk Number (%) identifying 30/269 (11.2) 36/282 (12.8) 28/276 (10.1) 22/254 (8.7) 62/249 (24.9) 109/244 (44.7) 16/234 (6.8) 53/237 (22.4) 78/234 (33.3) a 70-year-old woman as most likely to get breast cancer Odds ratio (95% CI), 1.0 3.2 (1.8 – 5.8) 9.5 (5.1 – 17.6) 1.0 3.4 (1.8 – 6.7) 7.4 (3.7 – 14.7) P-value (vs usual care) Po0.001 Po0.001 Po0.001 Po0.001 Breast checking Number (%) reporting 152/285 (53.3) 139/288 (48.3) 154/284 (54.2) 163/261 (62.5) 157/256 (61.3) 192/247 (77.7) 171/239 (71.6) 169/243 (70.0) 180/234 (76.9) breast checking at least once a month Odds ratio (95% CI), 1.0 1.2 (0.9 – 1.6) 2.0 (1.4 – 2.8) 1.0 1.1 (0.8 – 1.6) 1.3 (0.9 – 1.8) P-value (vs usual care) P¼ 0.25 Po0.001 P¼ 0.47 P ¼ 0.23 A woman scored three points on the breast cancer awareness score if she: identified at least five non-lump symptoms (one point), identified that a 70-year-old woman is most at risk of breast cancer (one point) and reported checking her breasts at least once a month (one point). The effect of the interaction plus booklet was no longer significant Only 3% of the women were breast cancer aware (as defined at 1 year. prospectively for this study) at baseline. This may explain why so many women delay presenting with breast cancer symptoms and have poor survival as a result. Knowledge of age-related risk was Intervention delivery particularly poor (only 11% were aware that a 70-year-old woman was at higher risk than a 30-year-old woman or a 50-year-old A total of 279 interactions were conducted and 82% were video woman), perhaps because of heavy media coverage of younger recorded and assessed by an independent rater (38 were not usable women with breast cancer and the current upper age limits on the due to technical faults, five were incomplete, four were audio taped NHS Breast Screening Programme. only, two women refused recording and one tape was lost). Quality The efficacy of the booklet alone was limited. It was important to of intervention delivery was high: median content score was 96 test the booklet alone, as it would be cheaper to deliver on the NHS (range: 68–100) (only five interactions (2%) scored o80); median than the one-to-one interaction. The interaction plus booklet was style score was 85 (range: 37–100) (79 interactions (35%) scored probably more effective because it incorporated features thought o80). One per cent of the total variability in the breast cancer to promote behaviour change: a direct recommendation from a awareness score at 1 month was attributable to radiographer, and health professional (Jepson, 2000), tailoring (Jepson, 2000) and 0.6% to centre. positive motivational style and verbal persuasion (Rollnick and Miller, 1995; Wardle et al, 2003). Knowledge of non-lump symptoms and reported breast check- DISCUSSION ing increased quite markedly in the women who received usual The intervention increased breast cancer awareness among older care alone over the 12-month follow-up. This may be due to what women compared with usual care at 1 month, with the interaction has been called the ‘mere measurement’ effect (Godin et al, 2008): supplemented by a booklet having a greater effect than the booklet either the questionnaire itself increased awareness, or women alone. Thirty-three per cent of those receiving the interaction plus started to guess the correct, or most appropriate, answers because booklet and 13% of those receiving the booklet alone were breast they were repeatedly asked the same question. cancer aware compared with 4% of women receiving usual care. A systematic review of interventions to promote cancer These improvements in breast cancer awareness were sustained at awareness found very limited evidence of effectiveness of any 12 months although were somewhat less marked (24% of the interventions. It found only five RCTs of interventions to promote interaction plus booklet group and 12% of the booklet only group cancer awareness aimed at individuals, of moderate to good compared with 4% of the usual care group). Of the three quality (Austoker et al, 2009). All found more modest effects on components of breast cancer awareness, the interaction plus cancer awareness than we achieved. The trial finding the greatest booklet and the booklet alone had the most marked effect on effect was of an intensive intervention (tailored written informa- knowledge of age-related risk. The interaction plus booklet was tion with a reinforcing newsletter at 12 months plus two telephone also associated with a statistically significant increase in knowledge counselling sessions) primarily aiming to increase breast screening of breast cancer symptoms. uptake. It increased the proportion who gave the correct answer to British Journal of Cancer (2009) 101(S2), S40 – S48 & 2009 Cancer Research UK Intervention to promote early presentation of breast cancer L Linsell et al S45 Breast cancer awareness Knowledge of breast cancer symptoms 100 100 90 90 Usual care Booklet Interaction Usual care Booklet Interaction 70 70 60 60 40 40 30 30 0 0 Baseline 1 month 1 year Baseline 1 month 1 year Knowledge of age-related risk 100 Breast checking Usual care Booklet Interaction Usual care Booklet Interaction 40 40 30 30 20 20 10 10 0 0 Baseline 1 month 1 year Baseline 1 month 1 year Figure 2 Breast cancer awareness and component items at baseline, 1 month and 1-year after randomisation. *A woman scored three points on the breast cancer awareness score if she: identified at least five non-lump symptoms (one point), identified that a 70-year-old woman is most at risk of breast cancer (one point) and reported checking her breasts at least once a month (one point). a question about age-related risk by 12% compared with usual care Screening Programme, 2009). Women who take up breast screen- after 2 years (Rimer et al, 2002). In our study, the interaction plus ing live in less deprived areas than those who do not (Banks et al, booklet, which increased the proportion correctly identifying a 2002; Maheswaran et al, 2006), although the women in our study 70-year-old woman as at higher risk than a 50-year-old woman or lived in slightly more deprived areas than the English average. 30-year-old woman by 27% at 12 months, compares well with Should the intervention be implemented across more affluent this. Less intensive interventions such as mailed information populations, it is likely that women receiving it would be of higher (de Nooijer et al, 2004) and interactive computer programmes socio-economic status than those not receiving it. However, these (Glazebrook et al, 2006) increased cancer awareness more women are at higher risk of breast cancer (Threlfall and modestly. The effects of the interaction plus booklet on reported Woodman, 2001; Shack et al, 2008) so it is appropriate to target them in this setting. breast checking at least monthly (77% vs 72% after 12 months) are similar to those found in trials to promote breast self-examination: We developed a score for measuring breast cancer awareness a RCT in the United States found that a 45-min class increased the to be used in surveys and trials. Currently, there is no univer- proportion who reported monthly breast self-examination from sally accepted measure of breast cancer awareness (although the 51% to 62% after 6 months (Strickland et al, 1997). breast cancer module of the Cancer Awareness Measure is being Our intervention is not a tutorial in breast self-examination. The developed) and no published agreement on what the con- evidence to support systematic, regular breast self-examination cept means. We argue that it is not a single construct, so, in is weak: a Cochrane systematic review found that breast self- developing our measure, we included three constructs that we felt examination did not reduce mortality and increased investigations encompass the minimum information women need to be able to (Kosters and Gotzsche, 2003). However, both trials included in present promptly with breast cancer symptoms: why to look for the review recruited women under the age of 67 years; whether them (magnitude of risk), what to look for (the range of breast checking, or even breast self-examination, would increase symptoms) and how to look for them (to look at and feel their detection rates and reduce mortality in older women is unknown. breasts). What seems highly unlikely is that women who never look at or Health professional-delivered complex interventions such as touch their breasts (20% of older women (Linsell et al, 2008)) will ours are prone to variability in the quality of delivery because they detect symptoms early; our intervention is designed to encourage are made up of many components and are operator dependent, simply looking and touching. and this may influence whether they work or not. In psychother- Strengths of our trial were the high level of participation (84% of apy trials, a significant amount of the variability in participant eligible women were randomised) and the high response to follow- outcome has been shown to be attributable to variable delivery by up (83% at 12 months). The usual care arm had slightly higher different therapists (Crits-Christoph et al, 1991; Okiishi et al, 2003; levels of education than the interaction arm; however, adjusting for Wampold and Brown, 2005). In our trial, we found no evidence of baseline characteristics did not significantly change the size of the important variation in the quality of delivery of the interaction estimates. between those delivering it. The NHS Breast Screening Programme is an efficient setting for We recognise that some of the symptoms included in the recruiting large numbers of healthy older women: uptake of breast intervention and the questionnaire (e.g. lump in armpit) are likely screening in women aged 65–70 years is over 70% (NHS Breast to indicate disease of a worse prognosis. We included these & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S40 – S48 % identifying a 70-year-old as most likely to get breast cancer % breast cancer aware* % reporting breast checking at % identifying 5 or more non-lump least once a month symptoms Intervention to promote early presentation of breast cancer L Linsell et al S46 It could be argued that inviting older women for screening Baseline 1 month 1 year on the NHS Breast Screening Programme might be a better way Usual care of promoting earlier diagnosis in breast cancer (currently the upper age limit for routine invitation is 70; this is soon to increase Lump in breast 92 to 73). However, breast screening is expensive and the cost- Lump under armpit 87 effectiveness of inviting older women for screening is not clear. Raising awareness is likely to compare favourably in cost- Discharge from nipple 75 effectiveness terms, and may have a more prolonged effect Change in shape 69 than a further round of screening. In addition, research in the Pain in breast/armpit 71 71 United Kingdom has consistently shown low public awareness of Change in nipple position the early warning signs of breast cancer (Brunswick et al, 2001; Pulling-in of nipple 62 Wardle et al, 2001; Grunfeld et al, 2002; Adlard and Hume, 2003; Moser et al, 2007) particularly among older women (Linsell et al, Puckering/dimpling 2008). This is thought to be more likely to explain survival Change in size differences between the United Kingdom and other countries Redness of skin than differences in availability or uptake of breast screening Nipple rash (Richards, 2009). Another reason to evaluate interventions to promote breast cancer awareness is that it may have implications for increasing cancer awareness and survival 0 20 40 60 80 100 Percentage in other cancers, for which screening programmes may not be available. Booklet We have shown that our intervention, a 10-min interaction with a health professional plus a booklet, promotes breast Lump in breast 95 cancer awareness in older women after 12 months. This trial Lump under armpit 87 was not designed to show whether it will promote early Discharge from nipple 80 presentation in breast cancer and thereby improve survival, although we do plan to evaluate the effect on screening uptake Change in shape 3 years after randomisation. There is an established associa- Pain in breast/armpit 71 tion between delayed presentation and survival (Richards et al, Change in nipple position 67 73 1999), but the evidence linking cancer awareness and early Pulling-in of nipple 69 presentation is less strong. There is evidence that poor knowledge of non-lump symptoms is associated with delay in presenta- Puckering/dimpling 64 44 tion (Ramirez et al, 1999). Other evidence is indirect: women Change in size 64 belonging to populations most likely to have delayed diagnosis Redness of skin 37 of breast cancer (Ramirez et al, 1999), also have lower cancer Nipple rash 32 awareness, including older women, women of lower socio- economic status and black women (Grunfeld et al, 2002; Scanlon 0 20 40 60 80 100 and Wood, 2005; Linsell et al, 2008). We acknowledge that this Percentage trial does not provide evidence that our intervention will pro- mote early presentation, although that is its ultimate aim. We Interaction plus booklet plan further trials to examine whether the intervention reduces delay in presentation, but these, much larger, studies would not Lump in breast 92 be possible without first building the evidence for its effect on Lump under armpit 87 breast cancer awareness, which we hypothesise is on the causal Discharge from nipple 85 pathway. Delay in presentation in breast cancer is an important Change in shape 80 57 public health problem. We estimate that 7000–12 000 women Pain in breast/armpit 83 delay presentation for 43 months in England each year (Richards Change in nipple position 76 et al, 1999; Office for National Statistics, 2008). Women who Pulling-in of nipple 75 delay presenting for 3 to 6 months have 7% lower 5-year survival Puckering/dimpling 72 than those with shorter delays (Richards et al, 1999). If only 7000 women per year in England delay presentation for 43 months, Change in size 72 15 about 500 will die as a result (assuming a 5-year breast cancer Redness of skin 57 survival of 80% in women who delay o3 months and 73% in Nipple rash 47 those who delay 43 months). If we find that the interven- tion reduces breast cancer mortality, it could be one of the key 0 20 40 60 80 100 elements of a programme to bring UK breast cancer survival Percentage closer to the standards obtained in similar countries. It may also Figure 3 Proportions of women identifying symptoms of breast cancer deliver other benefits, to women themselves and the NHS, as at baseline, 1 month and 1-year after randomisation. a result of less intensive breast cancer treatment given at an earlier stage. symptoms because even in women with more advanced disease ACKNOWLEDGEMENTS than stage I, earlier detection may improve prognosis. In addition, we felt it important to include all symptoms of breast cancer in the We thank the staff and participating women from the South list as their absence might give a misleading message. East London, South West London and Surrey Breast Screening British Journal of Cancer (2009) 101(S2), S40 – S48 & 2009 Cancer Research UK Intervention to promote early presentation of breast cancer L Linsell et al S47 Services for their support of the trial; Liz Kedge, Rachel Baxter, (Ref No: 05/Q0703/51). This study was funded by Cancer Ragini Jhalia and Jo Sippitt for delivering the intervention; Research UK. Angela Thurnham for the analysis of the 1 year data and Matthew Hotopf, Rebecca Walwyn and Gunna Dietrich of the Data Monitoring Committee. Trial registration was by National Cancer Conflict of interest Research Network and ISRCTN (31994827). Ethics approval was by Kings College Hospital Research Ethics Committee 2007 The authors declare no conflict of interest. REFERENCES Adlard JW, Hume MJ (2003) Cancer knowledge of the general public in the Kosters JP, Gotzsche PC (2003) Regular self-examination or clinical United Kingdom: survey in a primary care setting and review of the examination for early detection of breast cancer. Cochrane literature. Clin Oncol 15: 174–180 Database of Systematic Reviews, Issue 2. Art. No.: CD003373. DOI: Armitage P, Berry G (1994) Statistical Methods in Medical Research. 10.1002/14651858.CD003373. Blackwell: Oxford Lee KJ, Thompson SG (2005) Clustering by health professional in Arndt V (2002) Patient delay and stage of diagnosis among breast individually randomised trials. BMJ 330: 142–144 cancer patients in Germany – a population based study. Br J Cancer 86: Linsell L (2006) Breast cancer awareness in older women: analysis report. 1034– 1040 http://www.iop.kcl.ac.uk/iopweb/blob/downloads/locator/l_370_Age_ Austoker J, Bankhead C, Forbes LJL, Atkins L, Martin F, Robb K, Concern_Report.pdf (accessed 2 April 2009) Wardle J, Ramirez AJ (2009) Interventions to promote cancer awareness Linsell L, Burgess CC, Ramirez AJ (2008) Breast cancer awareness among and early presentation: systematic review. Br J Cancer 101(Suppl 2): S31–S39 older women. Br J Cancer 99: 1221–1225 Banks E, Beral V, Cameron R, Hogg A, Langley N, Barnes I, Bull D, Maheswaran R, Pearson T, Jordan H, Black D (2006) Socioeconomic Reeves G, English R, Taylor S, Elliman J, Lole HC (2002) Comparison of deprivation, travel distance, location of service, and uptake of breast various characteristics of women who do and do not attend for breast cancer screening in North Derbyshire, UK. J Epidemiol Community cancer screening. Breast Cancer Res 4: R1 Health 60: 208–212 Berrino F, De Angelis R, Sant M, Rosso S, Lasota MB, Coebergh JW, Medical Research Council (2008) Developing and Evaluating Complex Santaquilani M (2007) Survival for eight major cancers and all cancers Interventions: New Guidance. Medical Research Council: London combined for European adults diagnosed in 1995– 1999: results of the Moser K, Patnick J, Beral V (2007) Do women know that the risk of breast EUROCARE-4 study. Lancet Oncol 8: 773–783 cancer increases with age? Br J Gen Pract 57: 404–406 Bish A, Ramirez A, Burgess C, Hunter M (2005) Understanding why women NHS Breast Screening Programme (2009) Breast Screening Programme, delay in seeking help for breast cancer symptoms. J Psychosom Res 58: England 2007–2008. The Information Centre for Health and Social Care: 321–326 London Brunswick N, Wardle J, Jarvis MJ (2001) Public awareness of warning signs Nichols S, Waters WE, Fraser JD, Wheeller MJ, Ingham SK (1981) Delay in for cancer in Britain. Cancer Causes Control 12: 33–37 the presentation of breast symptoms for consultant investigation. Burgess CC, Bish AM, Hunter HS, Salkovskis P, Michell M, Whelehan P, J Public Health 3: 217–225 Ramirez AJ (2008) Promoting early presentation of breast cancer: Office for National Statistics (2008) Cancer Statistics: Registrations of development of a psycho-educational intervention. Chronic Illness 4: Cancers Diagnosed in 2005. Series MB1. 36. Office for National Statistics: 13–27 Newport Burgess CC, Linsell L, Kapari M, Omar L, Michell M, Whelehan P, Richards Okiishi J, Lambert MJ, Nielsen SL, Ogles BM (2003) Waiting for MA, Ramirez AJ (2009) Promoting early presentation of breast cancer by Supershrink: an empirical analysis of therapist effects. Clin Psychol older women: a preliminary evaluation of a one-to-one health profes- Psychother 10: 361 –373 sional-delivered intervention. J Psychosom Res 67(5): 377–387 Ramirez AJ, Westcombe AM, Burgess CC, Sutton S, Littlejohns P, Richards Cancer Research UK (2008) Breast cancer incidence statistics by age. MA (1999) Factors predicting delayed presentation of symptomatic Cancer Research UK. 17-10-2008. http://info.cancerresearchuk.org/ breast cancer: a systematic review. Lancet 353: 1127–1131 cancerstats/types/breast/incidence/ (accessed 27 March 2009) Richards MA (2009) The size of the prize for earlier diagnosis of cancer in Communities and Local Government (2007) The Indices of Multiple Depri- England. Br J Cancer 101(Suppl 2): S125–S129 vation, http://www.communities.gov.uk/communities/neighbourhoodrenewal/ Richards MA, Westcombe AM, Love SB, Littlejohns P, Ramirez AJ (1999) deprivation/deprivation07/ (accessed 1 April 2009) Influence of delay on survival in patients with breast cancer: a systematic Crits-Christoph P, Baranackie K, Kurcias J, Beck A, Carroll K, Perry K, review. Lancet 353: 1119– 1126 Luborsky L, McLellan A, Woody G, Thompson L, Gallagher D, Zitron C Rimer BK, Halabi S, Sugg Skinner C, Lipkus IM, Strigo TS, Kaplan EB, (1991) Meta-analysis of therapist effects in psychotherapy outcome Samsa GP (2002) Effects of a mammography decision-making interven- studies. Psychother Res 1: 81–91 tion at 12 and 24 months. Am J Prev Med 22: 247– 257 de Nooijer J, Lechner L, Candel M, de VH (2004) Short- and long-term Rollnick SR, Miller WR (1995) What is motivational interviewing? Behav effects of tailored information versus general information on determi- Cogn Psychother 23: 325–334 nants and intentions related to early detection of cancer. Prev Med 38: Sant M, Allemani C, Capocaccia R, Hakulinen T, Aareleid T, Coebergh JW, 694–703 Coleman MP, Grosclaude P, Martinez C, Bell J, Youngson J, Berrino F, Fleiss JL (1981) Statistical Methods for Rates and Proportions. John Wiley EUROCARE Working Group (2003) Stage at diagnosis is a key and Son: New York explanation of differences in breast cancer survival across Europe. Int Glazebrook C, Garrud P, Avery A, Coupland C, Williams H (2006) Impact J Cancer 106: 416–422 of a multimedia intervention ‘Skinsafe’ on patients; knowledge and Scanlon K, Wood A (2005) Breast cancer awareness in Britain: are protective behaviors. Prev Med 42: 449–454 there differences based on ethnicity? Diversity Health Social Care 2: Godin G, Sheeran P, Conner M, Germain M (2008) Asking questions 211–221 changes behavior: mere measurement effects on frequency of blood Shack L, Jordan C, Thomson CS, Mak V, Moller H, UK Association of donation. Health Psychol 27: 179–184 Cancer Registries (2008) Variation in incidence of breast, lung and Gollwitzer PM (1999) Implementation intentions: strong effects of simple cervical cancer and malignant melanoma of skin by socioeconomic plans. Am Psychol 54: 493–503 group in England. BMC Cancer 8: 271 Grunfeld EA, Ramirez AJ, Hunter MS, Richards MA (2002) Strickland CJ, Feigl P, Upchurch C, King DK, Pierce HI, Grevstad PK, Women’s knowledge and beliefs regarding breast cancer. Br J Cancer Bearden III JD, Dawson M, Loewen WC, Meyskens Jr FL (1997) 86: 1373–1378 Improving breast self-examination compliance: a Southwest Oncology Jepson R (2000) The effectiveness of interventions to change health-related Group randomized trial of three interventions. Prev Med 26: 320–332 behaviours: a review of reviews. MRC Social and Public Health Sciences Threlfall AG, Woodman CB (2001) Risk of breast cancer in women who Unit Occasional Paper No 3. Glasgow attend the NHS breast screening programme: cohort study. BMJ 323: 140 & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S40 – S48 Intervention to promote early presentation of breast cancer L Linsell et al S48 Wampold BE, Brown GS (2005) Estimating variability in outcomes the efficacy of a mailed, psychoeducational intervention in a community attributable to therapists: a naturalistic study of outcomes in managed sample of older adults. Health Psychol 22: 99–105 care. J Consult Clin Psychol 73: 914–923 Westcombe AM, Richards MA, Ramirez AJ, Love SB, Sutton S, Burgess C, Wardle J, Waller J, Brunswick N, Jarvis MJ (2001) Awareness of risk factors Littlejohns P (1999) A Systematic Review of the Delay in Diagnosis/Treatment for cancer among British adults. Public Health 115: 173– 174 of Symptomatic Breast Cancer. NHS Research and Development: London Wardle J, Williamson S, McCaffery K, Sutton S, Taylor T, Edwards R, Atkin Zeger SL, Liang K-Y (1986) Longitudinal data analysis for discrete and W (2003) Increasing attendance at colorectal cancer screening: testing continuous outcomes. Biometrics 42: 121 –130 British Journal of Cancer (2009) 101(S2), S40 – S48 & 2009 Cancer Research UK http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png British Journal of Cancer Springer Journals

A randomised controlled trial of an intervention to promote early presentation of breast cancer in older women: effect on breast cancer awareness

Download PDF
9 pages

Loading next page...
 
/lp/springer-journals/a-randomised-controlled-trial-of-an-intervention-to-promote-early-mG2X704Pxg

References (100)

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.6605389
Publisher site
See Article on Publisher Site

Abstract

British Journal of Cancer (2009) 101, S40 – S48 & 2009 Cancer Research UK All rights reserved 0007 – 0920/09 $32.00 www.bjcancer.com Full Paper A randomised controlled trial of an intervention to promote early presentation of breast cancer in older women: effect on breast cancer awareness 1 ,1 1 1 1 1 1 L Linsell , LJL Forbes , M Kapari , C Burgess , L Omar , L Tucker and AJ Ramirez King’s College London, Cancer Research UK Promoting Early Presentation Group, Institute of Psychiatry, St Thomas’ Hospital, London SE1 7EH, UK BACKGROUND: There is virtually no evidence for the effectiveness of interventions to promote early presentation in breast cancer. We aimed to test the efficacy of an intervention to equip older women with the knowledge, skills, confidence and motivation to detect symptoms and seek help promptly, with the aim of promoting early presentation with breast cancer symptoms. METHODS: We randomised 867 women aged 67–70 years attending for their final routine appointment on the UK NHS Breast Screening Programme to receive: a scripted 10-min interaction with a radiographer plus a booklet, a booklet alone or usual care. The primary outcome was whether or not a woman was breast cancer aware based on knowledge of breast cancer symptoms and age- related risk, and reported breast checking. RESULTS: At 1 month, the intervention increased the proportion who were breast cancer aware compared with usual care (interaction arm: 32.8% vs 4.1%; odds ratio (OR): 24.0, 95% confidence interval (CI): 7.7–73.7; booklet arm: 12.7% vs 4.1%; OR: 4.4, 95% CI: 1.6–12.0). At 1 year, the effects of the interaction plus booklet, and the booklet, on breast cancer awareness were largely sustained, although the interaction plus booklet remained much more effective. CONCLUSION: An intervention to equip older women with the knowledge, skills, confidence and motivation to detect breast cancer symptoms and seek help promptly increases breast cancer awareness at 1 year. Future research will evaluate whether the intervention promotes early presentation and reduces breast cancer mortality. British Journal of Cancer (2009) 101, S40 – S48. doi:10.1038/sj.bjc.6605389 www.bjcancer.com & 2009 Cancer Research UK Keywords: aged; awareness; breast neoplasms; complex intervention; delayed presentation; randomised controlled trials Women in the United Kingdom have poorer survival from We have built two versions of an intervention to promote early breast cancer than many other Western European countries presentation of breast cancer in older women (Burgess et al, 2008) (Berrino et al, 2007), and differences in stage at diagnosis are aiming to equip them with breast cancer awareness: a booklet largely responsible (Sant et al, 2003). Late stage at diagnosis is containing health-promoting messages and a scripted one-to-one almost certainly due to late presentation by some women and interaction with a radiographer, supported by the booklet, both delays in onward referral by some general practitioners. In the designed to be delivered during the final routine appointment on developed world, 17–35% of women with breast cancer delay the NHS Breast Screening Programme, a setting giving access to presenting for 43 months, and 9–20% delay 46 months most English women aged 67–70 years. We targeted older women (Westcombe et al, 1999; Arndt, 2002). We have found no because they are at higher risk of breast cancer (Cancer Research high quality evidence of effectiveness of interventions to promote UK, 2008), are more likely to delay presentation (Ramirez et al, early presentation in breast, or any other cancer (Austoker et al, 1999) and have poorer survival (Office for National Statistics, 2009). 2008). The positive predictive value of breast symptoms for breast Risk factors for delay in presentation in breast cancer include cancer is higher in older than younger women (Nichols et al, 1981). older age, less education and presenting with non-lump sym- We based the design of the intervention on a theoretical ptoms (Ramirez et al, 1999). Older women have particularly framework for delayed presentation (Bish et al, 2005), and poor knowledge of non-lump symptoms and the increase in incorporated techniques to maximise the probability of behaviour breast cancer risk with age (Grunfeld et al, 2002; Linsell et al, change (Rollnick and Miller, 1995; Gollwitzer, 1999; Jepson, 2000; 2008). About 20% of older women report that they never look Wardle et al, 2003). at or feel their breasts (Linsell et al, 2008). We hypothesise that We have developed and evaluated the intervention in line with breast cancer awareness, which might include breast cancer the Medical Research Council guidance on complex interventions knowledge and the confidence, skills and motivation to detect (Medical Research Council, 2008). We have previously shown in a symptoms and seek help appropriately, will reduce delays in before-and-after exploratory trial that the intervention increased presentation. breast cancer awareness in older women at 6 months (Burgess et al, 2009). We report here the 1 month and 1 year results of a randomised controlled trial (RCT) of efficacy of the 10-min interaction with a radiographer supported by a booklet vs the *Correspondence: Dr LJL Forbes; E-mail: [email protected] booklet alone vs usual care alone. Intervention to promote early presentation of breast cancer L Linsell et al S41 MATERIALS AND METHODS five radiographers and two research psychologists delivered the interactions, and received ongoing performance feedback through- Participants and setting out. All interactions were video recorded and quality of each was assessed by an independent rater using a trial-specific quality Participants were a consecutive series of women aged 67–70 checklist assessing content and style of delivery. Quality scores attending final routine appointments on the NHS Breast Screening were standardised on a scale of 0–100. Programme. All women attending were eligible, unless they had a significant disorder that may have affected their ability to consent or participate, insufficient English or other language difficulties, or Outcomes were going overseas during the subsequent 6 months. The women We measured outcomes at baseline and 1, 6 and 12 months after were recruited from seven breast screening units in London and randomisation using a trial-specific questionnaire, an earlier Surrey. version of which we have used in a survey (Linsell et al, 2008) and an exploratory trial (Burgess et al, 2009). We have demon- Procedure strated good test–retest reliability (Linsell, 2006) and sensitivity to change (Burgess et al, 2009). The screening units sent an invitation letter and information sheet The primary outcome was proportion of women achieving to each potential participant 2 weeks before her appointment. On breast cancer awareness at 1 month, measured using a breast the day of attendance, a trained radiographer assessed whether the cancer awareness score. The score was a composite of responses to woman was eligible and obtained written informed consent. After three questions from the questionnaire, relating to knowledge of completing a baseline questionnaire, women were randomly symptoms, knowledge of age-related risk and reported breast allocated to: usual care, booklet alone or the 10-min one-to-one checking: interaction supported by the booklet, in addition to usual care. We Knowledge of breast cancer symptoms:‘Do you know any of the sent trial-specific postal questionnaires at 1 and 12 months after warning signs of breast cancer? If yes, please circle the signs you randomisation to collect outcome data. know below’. Women circled symptoms on a scattered list of 11 symptoms (two lump and nine non-lump). To score one point, The intervention the woman had to identify at least five non-lump symptoms, that is, over half; Usual care The screening unit receptionist informed each woman Knowledge of age-related risk:‘In the next year who is most who had received her final routine mammogram that she was no likely to get breast cancer?’ Response categories: a 30-year-old longer eligible for routine screening, advised her that she might woman, a 50-year-old woman, a 70-year-old woman, a woman continue to be screened every 3 years on request, and provided a of any age. To score one point, the woman had to identify that a card with contact details and a suggested date for contact. 70-year-old woman was most likely to get breast cancer; Breast checking:‘How often do you check your breasts?’ Response The booklet In addition to usual care, a radiographer gave a categories: rarely or never, at least every 6 months, at least once booklet to each woman who had received her final routine a month, at least once a week. To score one point, the woman had mammogram. The booklet conveyed key breast cancer awareness to report checking her breasts at least once a month or at least messages, including: once a week. A list of breast cancer symptoms; Each item was given equal weighting and contributed one point Age-related and absolute risk of developing breast cancer; to the total score (range: 0–3). To achieve breast cancer awareness, How to detect a breast change; the woman had to respond correctly to all three items. What to do on discovering a breast change; We also collected data on relationship status, education and A strong direct recommendation to seek medical attention ethnicity at baseline. To estimate socio-economic status, we used immediately on discovering a breast change, outlining the the Index of Multiple Deprivation (Communities and Local benefits of prompt help seeking and suggestions for overcoming Government, 2007) based on area of residence. This is a measure barriers such as embarrassment and fear; of deprivation at the small area level (32 482 areas in England) A direct recommendation to tell someone close in the event of based on seven dimensions: income, employment, health, educa- discovering a breast change; tion, housing and services, living environment and crime. Every An action plan to be completed by the woman about how she area in England is ranked from 1 (most deprived) to 32 482 (least will be breast aware and what she will do on discovering a breast deprived); median rank is 16 241. We assigned each woman a rank change; of Index of Multiple Deprivation according to the rank of her area A series of statements describing possible positive feelings of residence. (relief, reassurance, satisfaction), resulting from seeking help immediately with a breast change; Sample size A reminder that she might request further breast screening. We estimated that 2% of women would be breast cancer aware at The interaction plus booklet In addition to usual care, women baseline (Burgess et al, 2009) and that there would be a 12% received a scripted 10-min one-to-one interaction with a radio- difference between trial arms. Incorporating a design effect to take grapher or research psychologist. During this, the radiographer/ account of clustering by centre and radiographer (Lee and psychologist verbally delivered all the messages in the booklet in a Thompson, 2005) (assuming an intracluster correlation coefficient positive, collaborative and motivational style, referred to the of 0.08 and 14 participants per centre–radiographer cluster) and booklet throughout and gave it to the woman to take home. She allowing for 70% response, we required 238 women per arm (total also showed photographs of breast cancer signs and demonstrated 714) with a significance level of 5% (two-sided) and power of 80%. and rehearsed breast checking using a silicone breast. The radiographer was able to tailor the key messages by checking the Randomisation woman’s understanding and answering any questions. We randomised women individually on the day of attendance, with Quality assurance and quality control for the interaction plus equal probability of assignment to each arm. The trial statistician booklet After receiving training and being assessed as competent, computer generated the allocation sequence using stratified block & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S40 – S48 Intervention to promote early presentation of breast cancer L Linsell et al S42 1945 targeted with mailing 1209 (62.2%) attended and assessed for eligibility 342 (28.3%) excluded Reasons: 176 chose not to participate 70 insufficient English 48 missed for logistic reasons 28 significant disorder 20 going overseas 867 (71.7%) randomly allocated 287 (33.1%) allocated 294 (33.9%) allocated 286 (33.0%) allocated to usual care to usual care plus to usual care plus booklet booklet and interaction 286 received usual care 291 received booklet 279 received interaction Questionnaire Questionnaire Questionnaire completed at completed at completed at Baseline: n =287 Baseline: n =292 Baseline: n =284 1 month: n =251 (88%) 1 month: n =261 (91%) 1 month: n =257 (88%) 1 year: n =237 (83%) 1 year: n =240 (84%) 1 year: n =245 (83%) Lost to follow-up: Lost to follow-up: Lost to follow-up: n =10 (4%) n =7 (2%) n =8 (3%) 288 (98.0%) in main 282 (98.6%) in main 281 (97.9%) in main analysis analysis analysis 6 excluded from main 4 excluded from main 6 excluded from main analyses (no data) analyses (no data) analyses (no data) Figure 1 Flow of participants through trial. randomisation with centre and radiographer as stratification between repeated observations from the same individual. To variables (block sizes of three, six and nine). To ensure examine the intervention effect, we tested the interaction between concealment, assignments were enclosed in sequentially num- intervention group and time in each model, and presented odds bered, opaque, sealed envelopes and stored by the trial coordi- ratios (OR) with 95% confidence intervals (CIs). We also analysed nator before randomisation. The radiographer recorded each the data adjusting for stratification variables, relationship status, participant’s trial identification number on the envelope before education, ethnicity and Index of Multiple Deprivation, fitting opening it. categorical variables as binary variables and Index of Multiple Deprivation on a continuous scale (0–100). We calculated the intracluster correlation coefficient for radiographer and centre Statistical analysis using one-way analysis of variance, adjusting for unequal cluster We analysed the data by intention to treat: all participants were size (Fleiss, 1981; Armitage and Berry, 1994). analysed in the groups to which they were allocated. We summarised continuous variables using means, standard devia- tions, medians and ranges, and categorical variables using counts RESULTS and percentages. We used two-sided significance tests, taking P¼ 0.05 as significant. All analyses were performed using Stata Flow of participants version 10.0. The primary comparative analyses for all outcomes examined the difference between baseline and 1, 6 and 12 months Between August 2007 and May 2008, we randomised 867 women to for each pair of intervention groups. We used robust generalised one of three arms: usual care (n¼ 287), the booklet in addition to estimating equations (Zeger and Liang, 1986) with unstructured usual care (n¼ 294) and the interaction supported by the booklet correlation structure using a logit link and binomial distribution in addition to usual care (n¼ 286) (Figure 1). Only 15% (176 out of for the outcomes. This method takes account of the correlation 1209) of women who were assessed for eligibility chose not to British Journal of Cancer (2009) 101(S2), S40 – S48 & 2009 Cancer Research UK Follow-up Allocation Enrolment Targeted Intervention to promote early presentation of breast cancer L Linsell et al S43 Table 1 Baseline characteristics of participants Usual care (n¼ 287) Booklet (n¼ 292) Interaction plus booklet (n¼ 284) Relationship status, n (%) (n ¼ 285) (n¼ 283) (n¼ 282) Married or cohabiting 162 (56.8) 151 (53.4) 174 (61.7) Widowed 56 (19.7) 61 (21.6) 50 (17.7) Single 22 (7.7) 28 (9.9) 17 (6.0) Divorced or separated 45 (15.8) 43 (15.2) 41 (14.5) Education, n (%) (n ¼ 266) (n¼ 269) (n¼ 263) No formal qualifications 93 (35.0) 109 (40.5) 116 (44.1) O level or school certificate 89 (33.5) 81 (30.1) 77 (29.3) A level or higher school certificate 34 (12.8) 31 (11.5) 29 (11.0) Degree or above 50 (18.8) 48 (17.8) 41 (15.6) Ethnic group, n (%) (n ¼ 281) (n¼ 284) (n¼ 280) White British 187 (66.6) 196 (69.0) 186 (66.4) White other 23 (8.2) 23 (8.1) 32 (11.4) Black-Caribbean 36 (12.8) 38 (13.4) 37 (13.2) Other 35 (12.5) 27 (9.5) 25 (8.9) Index of multiple deprivation, median rank (IQR) (n ¼ 286) (n¼ 292) (n¼ 284) (1 (most) to 32 482 (least) deprived) 14 557 (8222 – 5989) 16 511 (8809 – 6184) 15 375 (8575 – 5729) participate. Eleven women did not receive the allocated interven- although estimated differences between arms were mostly slightly tion and 25 were lost to follow-up (20 withdrew consent, three larger (data not shown). moved with no forwarding address, one for medical reasons and one died). We included women with data for the primary outcome Knowledge of breast cancer symptoms on at least one occasion in the main analysis (n¼ 851). We were unable to measure any outcomes for four women who did not Forty-two per cent of women were able to identify five or more complete any questionnaires (two booklet arm, two interaction non-lump symptoms at baseline. At 1 month, the interaction plus arm). We received breast cancer awareness questionnaires from booklet increased the proportion of women able to identify five or 89% of those randomised at 1 month and 83% at 1 year; response more non-lump symptoms compared with usual care (78.9% vs rates were similar in each arm (Figure 1). 54.2%; difference 24.7%; OR: 2.5, 95% CI: 1.7–3.6) but the booklet alone did not (61.6% vs 54.2%; difference 7.4%; OR: 1.1, 95% CI: 0.8–1.5). The increase in knowledge of symptoms associated with Baseline socio-demographic characteristics the interaction plus booklet was maintained at 1 year. Before Socio-demographic characteristics were well balanced across the receiving the intervention, the women were able to identify a arms (Table 1), except for a slight difference in the proportion of median of four non-lump symptoms from the list of nine (IQR: women with no educational qualifications (35% usual care, 41% 2–6), and most recognised a lump in the breast or armpit as booklet and 44% interaction arm). Women in the booklet arm had symptoms. At 1 year, this increased to a median of six symptoms a similar deprivation score to the English median, but women in (IQR: 4–9) among those receiving the booklet only, and to seven the usual care and interaction plus booklet arms lived in somewhat (IQR: 4–9) among those receiving the interaction plus booklet. more deprived areas. For women invited for screening but not The intervention had most impact on the two least recognised recruited, the only characteristic we had data was postcode (and, symptoms, redness of skin and nipple rash (Figure 3). therefore, rank of Index of Multiple Deprivation). The median rank for the 1078 women not recruited was 9068 (interquartile range (IQR): 4784–15 802) compared with 15 664 (IQR: 8589–25 989) for Knowledge of age-related risk the 867 women in the trial, so women who did not take part lived Only 11.4% of the women knew that a 70-year-old woman was in more deprived areas than those who did. most at risk of breast cancer at baseline. At 1 month, the inter- action plus booklet increased the proportion knowing that a Breast cancer awareness 70-year-old woman was at most risk of breast cancer compared with usual care (44.7% vs 8.7%; difference 36.0%: OR: 9.5, 95% CI: Table 2 shows the main results for breast cancer awareness score 5.1–17.6), as did the booklet alone (24.9% vs 8.7%; difference and its components for baseline, 1 month and 1 year (6 months 16.2%; OR: 3.2, 95% CI: 1.8–5.8). At 1 year, the effect of the inter- data are not presented). Figure 2 illustrates breast cancer action plus booklet and the booklet alone remained significant. awareness and components of the score over the 12-month period. Overall, only 2.7% of women were breast cancer aware at baseline. At 1 month, the interaction plus booklet increased the proportion Breast checking who were breast cancer aware compared with usual care (32.8% vs 4.1%; OR: 24.0, 95% CI: 7.7–73.7), as did the booklet, although the About half of the women reported checking their breasts at least effect of the booklet was much less striking (12.7% vs 4.1%; OR: once a month at baseline. At 1 month, the interaction plus booklet 4.4, 95% CI: 1.6–12.0). At 1 year, the effect of the interaction plus increased the proportion of women checking their breasts at least booklet, and the booklet alone, on breast cancer awareness monthly compared with usual care (77.7% vs 62.5%; difference remained significant, with the interaction plus booklet remaining 15.2%; OR: 2.0, 95% CI: 1.4–2.8), but the booklet alone did more effective. The results of the adjusted analysis were similar, not (61.3% vs 62.5%; difference 1.2%; OR: 1.2, 95% CI: 0.9–1.6). & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S40 – S48 Intervention to promote early presentation of breast cancer L Linsell et al S44 Table 2 Breast cancer awareness and component items at 1 month and 1 year after randomisation Baseline 1 month 1 year Interaction Interaction Interaction Usual care Booklet plus booklet Usual care Booklet plus booklet Usual care Booklet plus booklet Breast cancer awareness Number (%) breast cancer aware 9/267 (3.4) 8/275 (2.9) 5/272 (1.8) 10/244 (4.1) 30/237 (12.7) 75/229 (32.8) 9/229 (3.9) 26/227 (11.5) 53/225 (23.6) Odds ratio (95% CI), 1.0 4.4 (1.6 – 2.0) 24.0 (7.7 – 73.7) 1.0 3.5 (1.2 – 10.5) 15.2 (4.8 – 47.8) P-value (vs usual care) P¼ 0.004 Po0.001 P¼ 0.025 Po0.001 Knowledge of breast cancer symptoms Number (%) identifying 111/284 (39.1) 126/286 (44.1) 122/280 (43.6) 136/251 (54.2) 151/245 (61.6) 187/237 (78.9) 142/233 (60.9) 167/236 (70.8) 170/230 (73.9) X5 non-lump symptoms Odds ratio (95% CI), 1.0 1.1 (0.8 – 1.5) 2.5 (1.7 – 3.6) 1.0 1.3 (0.9 – 1.9) 1.7 (1.1 – 2.4) P-value (vs usual care) P¼ 0.61 Po0.001 P¼ 0.23 P ¼ 0.01 Knowledge of age-related risk Number (%) identifying 30/269 (11.2) 36/282 (12.8) 28/276 (10.1) 22/254 (8.7) 62/249 (24.9) 109/244 (44.7) 16/234 (6.8) 53/237 (22.4) 78/234 (33.3) a 70-year-old woman as most likely to get breast cancer Odds ratio (95% CI), 1.0 3.2 (1.8 – 5.8) 9.5 (5.1 – 17.6) 1.0 3.4 (1.8 – 6.7) 7.4 (3.7 – 14.7) P-value (vs usual care) Po0.001 Po0.001 Po0.001 Po0.001 Breast checking Number (%) reporting 152/285 (53.3) 139/288 (48.3) 154/284 (54.2) 163/261 (62.5) 157/256 (61.3) 192/247 (77.7) 171/239 (71.6) 169/243 (70.0) 180/234 (76.9) breast checking at least once a month Odds ratio (95% CI), 1.0 1.2 (0.9 – 1.6) 2.0 (1.4 – 2.8) 1.0 1.1 (0.8 – 1.6) 1.3 (0.9 – 1.8) P-value (vs usual care) P¼ 0.25 Po0.001 P¼ 0.47 P ¼ 0.23 A woman scored three points on the breast cancer awareness score if she: identified at least five non-lump symptoms (one point), identified that a 70-year-old woman is most at risk of breast cancer (one point) and reported checking her breasts at least once a month (one point). The effect of the interaction plus booklet was no longer significant Only 3% of the women were breast cancer aware (as defined at 1 year. prospectively for this study) at baseline. This may explain why so many women delay presenting with breast cancer symptoms and have poor survival as a result. Knowledge of age-related risk was Intervention delivery particularly poor (only 11% were aware that a 70-year-old woman was at higher risk than a 30-year-old woman or a 50-year-old A total of 279 interactions were conducted and 82% were video woman), perhaps because of heavy media coverage of younger recorded and assessed by an independent rater (38 were not usable women with breast cancer and the current upper age limits on the due to technical faults, five were incomplete, four were audio taped NHS Breast Screening Programme. only, two women refused recording and one tape was lost). Quality The efficacy of the booklet alone was limited. It was important to of intervention delivery was high: median content score was 96 test the booklet alone, as it would be cheaper to deliver on the NHS (range: 68–100) (only five interactions (2%) scored o80); median than the one-to-one interaction. The interaction plus booklet was style score was 85 (range: 37–100) (79 interactions (35%) scored probably more effective because it incorporated features thought o80). One per cent of the total variability in the breast cancer to promote behaviour change: a direct recommendation from a awareness score at 1 month was attributable to radiographer, and health professional (Jepson, 2000), tailoring (Jepson, 2000) and 0.6% to centre. positive motivational style and verbal persuasion (Rollnick and Miller, 1995; Wardle et al, 2003). Knowledge of non-lump symptoms and reported breast check- DISCUSSION ing increased quite markedly in the women who received usual The intervention increased breast cancer awareness among older care alone over the 12-month follow-up. This may be due to what women compared with usual care at 1 month, with the interaction has been called the ‘mere measurement’ effect (Godin et al, 2008): supplemented by a booklet having a greater effect than the booklet either the questionnaire itself increased awareness, or women alone. Thirty-three per cent of those receiving the interaction plus started to guess the correct, or most appropriate, answers because booklet and 13% of those receiving the booklet alone were breast they were repeatedly asked the same question. cancer aware compared with 4% of women receiving usual care. A systematic review of interventions to promote cancer These improvements in breast cancer awareness were sustained at awareness found very limited evidence of effectiveness of any 12 months although were somewhat less marked (24% of the interventions. It found only five RCTs of interventions to promote interaction plus booklet group and 12% of the booklet only group cancer awareness aimed at individuals, of moderate to good compared with 4% of the usual care group). Of the three quality (Austoker et al, 2009). All found more modest effects on components of breast cancer awareness, the interaction plus cancer awareness than we achieved. The trial finding the greatest booklet and the booklet alone had the most marked effect on effect was of an intensive intervention (tailored written informa- knowledge of age-related risk. The interaction plus booklet was tion with a reinforcing newsletter at 12 months plus two telephone also associated with a statistically significant increase in knowledge counselling sessions) primarily aiming to increase breast screening of breast cancer symptoms. uptake. It increased the proportion who gave the correct answer to British Journal of Cancer (2009) 101(S2), S40 – S48 & 2009 Cancer Research UK Intervention to promote early presentation of breast cancer L Linsell et al S45 Breast cancer awareness Knowledge of breast cancer symptoms 100 100 90 90 Usual care Booklet Interaction Usual care Booklet Interaction 70 70 60 60 40 40 30 30 0 0 Baseline 1 month 1 year Baseline 1 month 1 year Knowledge of age-related risk 100 Breast checking Usual care Booklet Interaction Usual care Booklet Interaction 40 40 30 30 20 20 10 10 0 0 Baseline 1 month 1 year Baseline 1 month 1 year Figure 2 Breast cancer awareness and component items at baseline, 1 month and 1-year after randomisation. *A woman scored three points on the breast cancer awareness score if she: identified at least five non-lump symptoms (one point), identified that a 70-year-old woman is most at risk of breast cancer (one point) and reported checking her breasts at least once a month (one point). a question about age-related risk by 12% compared with usual care Screening Programme, 2009). Women who take up breast screen- after 2 years (Rimer et al, 2002). In our study, the interaction plus ing live in less deprived areas than those who do not (Banks et al, booklet, which increased the proportion correctly identifying a 2002; Maheswaran et al, 2006), although the women in our study 70-year-old woman as at higher risk than a 50-year-old woman or lived in slightly more deprived areas than the English average. 30-year-old woman by 27% at 12 months, compares well with Should the intervention be implemented across more affluent this. Less intensive interventions such as mailed information populations, it is likely that women receiving it would be of higher (de Nooijer et al, 2004) and interactive computer programmes socio-economic status than those not receiving it. However, these (Glazebrook et al, 2006) increased cancer awareness more women are at higher risk of breast cancer (Threlfall and modestly. The effects of the interaction plus booklet on reported Woodman, 2001; Shack et al, 2008) so it is appropriate to target them in this setting. breast checking at least monthly (77% vs 72% after 12 months) are similar to those found in trials to promote breast self-examination: We developed a score for measuring breast cancer awareness a RCT in the United States found that a 45-min class increased the to be used in surveys and trials. Currently, there is no univer- proportion who reported monthly breast self-examination from sally accepted measure of breast cancer awareness (although the 51% to 62% after 6 months (Strickland et al, 1997). breast cancer module of the Cancer Awareness Measure is being Our intervention is not a tutorial in breast self-examination. The developed) and no published agreement on what the con- evidence to support systematic, regular breast self-examination cept means. We argue that it is not a single construct, so, in is weak: a Cochrane systematic review found that breast self- developing our measure, we included three constructs that we felt examination did not reduce mortality and increased investigations encompass the minimum information women need to be able to (Kosters and Gotzsche, 2003). However, both trials included in present promptly with breast cancer symptoms: why to look for the review recruited women under the age of 67 years; whether them (magnitude of risk), what to look for (the range of breast checking, or even breast self-examination, would increase symptoms) and how to look for them (to look at and feel their detection rates and reduce mortality in older women is unknown. breasts). What seems highly unlikely is that women who never look at or Health professional-delivered complex interventions such as touch their breasts (20% of older women (Linsell et al, 2008)) will ours are prone to variability in the quality of delivery because they detect symptoms early; our intervention is designed to encourage are made up of many components and are operator dependent, simply looking and touching. and this may influence whether they work or not. In psychother- Strengths of our trial were the high level of participation (84% of apy trials, a significant amount of the variability in participant eligible women were randomised) and the high response to follow- outcome has been shown to be attributable to variable delivery by up (83% at 12 months). The usual care arm had slightly higher different therapists (Crits-Christoph et al, 1991; Okiishi et al, 2003; levels of education than the interaction arm; however, adjusting for Wampold and Brown, 2005). In our trial, we found no evidence of baseline characteristics did not significantly change the size of the important variation in the quality of delivery of the interaction estimates. between those delivering it. The NHS Breast Screening Programme is an efficient setting for We recognise that some of the symptoms included in the recruiting large numbers of healthy older women: uptake of breast intervention and the questionnaire (e.g. lump in armpit) are likely screening in women aged 65–70 years is over 70% (NHS Breast to indicate disease of a worse prognosis. We included these & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S40 – S48 % identifying a 70-year-old as most likely to get breast cancer % breast cancer aware* % reporting breast checking at % identifying 5 or more non-lump least once a month symptoms Intervention to promote early presentation of breast cancer L Linsell et al S46 It could be argued that inviting older women for screening Baseline 1 month 1 year on the NHS Breast Screening Programme might be a better way Usual care of promoting earlier diagnosis in breast cancer (currently the upper age limit for routine invitation is 70; this is soon to increase Lump in breast 92 to 73). However, breast screening is expensive and the cost- Lump under armpit 87 effectiveness of inviting older women for screening is not clear. Raising awareness is likely to compare favourably in cost- Discharge from nipple 75 effectiveness terms, and may have a more prolonged effect Change in shape 69 than a further round of screening. In addition, research in the Pain in breast/armpit 71 71 United Kingdom has consistently shown low public awareness of Change in nipple position the early warning signs of breast cancer (Brunswick et al, 2001; Pulling-in of nipple 62 Wardle et al, 2001; Grunfeld et al, 2002; Adlard and Hume, 2003; Moser et al, 2007) particularly among older women (Linsell et al, Puckering/dimpling 2008). This is thought to be more likely to explain survival Change in size differences between the United Kingdom and other countries Redness of skin than differences in availability or uptake of breast screening Nipple rash (Richards, 2009). Another reason to evaluate interventions to promote breast cancer awareness is that it may have implications for increasing cancer awareness and survival 0 20 40 60 80 100 Percentage in other cancers, for which screening programmes may not be available. Booklet We have shown that our intervention, a 10-min interaction with a health professional plus a booklet, promotes breast Lump in breast 95 cancer awareness in older women after 12 months. This trial Lump under armpit 87 was not designed to show whether it will promote early Discharge from nipple 80 presentation in breast cancer and thereby improve survival, although we do plan to evaluate the effect on screening uptake Change in shape 3 years after randomisation. There is an established associa- Pain in breast/armpit 71 tion between delayed presentation and survival (Richards et al, Change in nipple position 67 73 1999), but the evidence linking cancer awareness and early Pulling-in of nipple 69 presentation is less strong. There is evidence that poor knowledge of non-lump symptoms is associated with delay in presenta- Puckering/dimpling 64 44 tion (Ramirez et al, 1999). Other evidence is indirect: women Change in size 64 belonging to populations most likely to have delayed diagnosis Redness of skin 37 of breast cancer (Ramirez et al, 1999), also have lower cancer Nipple rash 32 awareness, including older women, women of lower socio- economic status and black women (Grunfeld et al, 2002; Scanlon 0 20 40 60 80 100 and Wood, 2005; Linsell et al, 2008). We acknowledge that this Percentage trial does not provide evidence that our intervention will pro- mote early presentation, although that is its ultimate aim. We Interaction plus booklet plan further trials to examine whether the intervention reduces delay in presentation, but these, much larger, studies would not Lump in breast 92 be possible without first building the evidence for its effect on Lump under armpit 87 breast cancer awareness, which we hypothesise is on the causal Discharge from nipple 85 pathway. Delay in presentation in breast cancer is an important Change in shape 80 57 public health problem. We estimate that 7000–12 000 women Pain in breast/armpit 83 delay presentation for 43 months in England each year (Richards Change in nipple position 76 et al, 1999; Office for National Statistics, 2008). Women who Pulling-in of nipple 75 delay presenting for 3 to 6 months have 7% lower 5-year survival Puckering/dimpling 72 than those with shorter delays (Richards et al, 1999). If only 7000 women per year in England delay presentation for 43 months, Change in size 72 15 about 500 will die as a result (assuming a 5-year breast cancer Redness of skin 57 survival of 80% in women who delay o3 months and 73% in Nipple rash 47 those who delay 43 months). If we find that the interven- tion reduces breast cancer mortality, it could be one of the key 0 20 40 60 80 100 elements of a programme to bring UK breast cancer survival Percentage closer to the standards obtained in similar countries. It may also Figure 3 Proportions of women identifying symptoms of breast cancer deliver other benefits, to women themselves and the NHS, as at baseline, 1 month and 1-year after randomisation. a result of less intensive breast cancer treatment given at an earlier stage. symptoms because even in women with more advanced disease ACKNOWLEDGEMENTS than stage I, earlier detection may improve prognosis. In addition, we felt it important to include all symptoms of breast cancer in the We thank the staff and participating women from the South list as their absence might give a misleading message. East London, South West London and Surrey Breast Screening British Journal of Cancer (2009) 101(S2), S40 – S48 & 2009 Cancer Research UK Intervention to promote early presentation of breast cancer L Linsell et al S47 Services for their support of the trial; Liz Kedge, Rachel Baxter, (Ref No: 05/Q0703/51). This study was funded by Cancer Ragini Jhalia and Jo Sippitt for delivering the intervention; Research UK. Angela Thurnham for the analysis of the 1 year data and Matthew Hotopf, Rebecca Walwyn and Gunna Dietrich of the Data Monitoring Committee. Trial registration was by National Cancer Conflict of interest Research Network and ISRCTN (31994827). Ethics approval was by Kings College Hospital Research Ethics Committee 2007 The authors declare no conflict of interest. REFERENCES Adlard JW, Hume MJ (2003) Cancer knowledge of the general public in the Kosters JP, Gotzsche PC (2003) Regular self-examination or clinical United Kingdom: survey in a primary care setting and review of the examination for early detection of breast cancer. Cochrane literature. Clin Oncol 15: 174–180 Database of Systematic Reviews, Issue 2. Art. No.: CD003373. DOI: Armitage P, Berry G (1994) Statistical Methods in Medical Research. 10.1002/14651858.CD003373. Blackwell: Oxford Lee KJ, Thompson SG (2005) Clustering by health professional in Arndt V (2002) Patient delay and stage of diagnosis among breast individually randomised trials. BMJ 330: 142–144 cancer patients in Germany – a population based study. Br J Cancer 86: Linsell L (2006) Breast cancer awareness in older women: analysis report. 1034– 1040 http://www.iop.kcl.ac.uk/iopweb/blob/downloads/locator/l_370_Age_ Austoker J, Bankhead C, Forbes LJL, Atkins L, Martin F, Robb K, Concern_Report.pdf (accessed 2 April 2009) Wardle J, Ramirez AJ (2009) Interventions to promote cancer awareness Linsell L, Burgess CC, Ramirez AJ (2008) Breast cancer awareness among and early presentation: systematic review. Br J Cancer 101(Suppl 2): S31–S39 older women. Br J Cancer 99: 1221–1225 Banks E, Beral V, Cameron R, Hogg A, Langley N, Barnes I, Bull D, Maheswaran R, Pearson T, Jordan H, Black D (2006) Socioeconomic Reeves G, English R, Taylor S, Elliman J, Lole HC (2002) Comparison of deprivation, travel distance, location of service, and uptake of breast various characteristics of women who do and do not attend for breast cancer screening in North Derbyshire, UK. J Epidemiol Community cancer screening. Breast Cancer Res 4: R1 Health 60: 208–212 Berrino F, De Angelis R, Sant M, Rosso S, Lasota MB, Coebergh JW, Medical Research Council (2008) Developing and Evaluating Complex Santaquilani M (2007) Survival for eight major cancers and all cancers Interventions: New Guidance. Medical Research Council: London combined for European adults diagnosed in 1995– 1999: results of the Moser K, Patnick J, Beral V (2007) Do women know that the risk of breast EUROCARE-4 study. Lancet Oncol 8: 773–783 cancer increases with age? Br J Gen Pract 57: 404–406 Bish A, Ramirez A, Burgess C, Hunter M (2005) Understanding why women NHS Breast Screening Programme (2009) Breast Screening Programme, delay in seeking help for breast cancer symptoms. J Psychosom Res 58: England 2007–2008. The Information Centre for Health and Social Care: 321–326 London Brunswick N, Wardle J, Jarvis MJ (2001) Public awareness of warning signs Nichols S, Waters WE, Fraser JD, Wheeller MJ, Ingham SK (1981) Delay in for cancer in Britain. Cancer Causes Control 12: 33–37 the presentation of breast symptoms for consultant investigation. Burgess CC, Bish AM, Hunter HS, Salkovskis P, Michell M, Whelehan P, J Public Health 3: 217–225 Ramirez AJ (2008) Promoting early presentation of breast cancer: Office for National Statistics (2008) Cancer Statistics: Registrations of development of a psycho-educational intervention. Chronic Illness 4: Cancers Diagnosed in 2005. Series MB1. 36. Office for National Statistics: 13–27 Newport Burgess CC, Linsell L, Kapari M, Omar L, Michell M, Whelehan P, Richards Okiishi J, Lambert MJ, Nielsen SL, Ogles BM (2003) Waiting for MA, Ramirez AJ (2009) Promoting early presentation of breast cancer by Supershrink: an empirical analysis of therapist effects. Clin Psychol older women: a preliminary evaluation of a one-to-one health profes- Psychother 10: 361 –373 sional-delivered intervention. J Psychosom Res 67(5): 377–387 Ramirez AJ, Westcombe AM, Burgess CC, Sutton S, Littlejohns P, Richards Cancer Research UK (2008) Breast cancer incidence statistics by age. MA (1999) Factors predicting delayed presentation of symptomatic Cancer Research UK. 17-10-2008. http://info.cancerresearchuk.org/ breast cancer: a systematic review. Lancet 353: 1127–1131 cancerstats/types/breast/incidence/ (accessed 27 March 2009) Richards MA (2009) The size of the prize for earlier diagnosis of cancer in Communities and Local Government (2007) The Indices of Multiple Depri- England. Br J Cancer 101(Suppl 2): S125–S129 vation, http://www.communities.gov.uk/communities/neighbourhoodrenewal/ Richards MA, Westcombe AM, Love SB, Littlejohns P, Ramirez AJ (1999) deprivation/deprivation07/ (accessed 1 April 2009) Influence of delay on survival in patients with breast cancer: a systematic Crits-Christoph P, Baranackie K, Kurcias J, Beck A, Carroll K, Perry K, review. Lancet 353: 1119– 1126 Luborsky L, McLellan A, Woody G, Thompson L, Gallagher D, Zitron C Rimer BK, Halabi S, Sugg Skinner C, Lipkus IM, Strigo TS, Kaplan EB, (1991) Meta-analysis of therapist effects in psychotherapy outcome Samsa GP (2002) Effects of a mammography decision-making interven- studies. Psychother Res 1: 81–91 tion at 12 and 24 months. Am J Prev Med 22: 247– 257 de Nooijer J, Lechner L, Candel M, de VH (2004) Short- and long-term Rollnick SR, Miller WR (1995) What is motivational interviewing? Behav effects of tailored information versus general information on determi- Cogn Psychother 23: 325–334 nants and intentions related to early detection of cancer. Prev Med 38: Sant M, Allemani C, Capocaccia R, Hakulinen T, Aareleid T, Coebergh JW, 694–703 Coleman MP, Grosclaude P, Martinez C, Bell J, Youngson J, Berrino F, Fleiss JL (1981) Statistical Methods for Rates and Proportions. John Wiley EUROCARE Working Group (2003) Stage at diagnosis is a key and Son: New York explanation of differences in breast cancer survival across Europe. Int Glazebrook C, Garrud P, Avery A, Coupland C, Williams H (2006) Impact J Cancer 106: 416–422 of a multimedia intervention ‘Skinsafe’ on patients; knowledge and Scanlon K, Wood A (2005) Breast cancer awareness in Britain: are protective behaviors. Prev Med 42: 449–454 there differences based on ethnicity? Diversity Health Social Care 2: Godin G, Sheeran P, Conner M, Germain M (2008) Asking questions 211–221 changes behavior: mere measurement effects on frequency of blood Shack L, Jordan C, Thomson CS, Mak V, Moller H, UK Association of donation. Health Psychol 27: 179–184 Cancer Registries (2008) Variation in incidence of breast, lung and Gollwitzer PM (1999) Implementation intentions: strong effects of simple cervical cancer and malignant melanoma of skin by socioeconomic plans. Am Psychol 54: 493–503 group in England. BMC Cancer 8: 271 Grunfeld EA, Ramirez AJ, Hunter MS, Richards MA (2002) Strickland CJ, Feigl P, Upchurch C, King DK, Pierce HI, Grevstad PK, Women’s knowledge and beliefs regarding breast cancer. Br J Cancer Bearden III JD, Dawson M, Loewen WC, Meyskens Jr FL (1997) 86: 1373–1378 Improving breast self-examination compliance: a Southwest Oncology Jepson R (2000) The effectiveness of interventions to change health-related Group randomized trial of three interventions. Prev Med 26: 320–332 behaviours: a review of reviews. MRC Social and Public Health Sciences Threlfall AG, Woodman CB (2001) Risk of breast cancer in women who Unit Occasional Paper No 3. Glasgow attend the NHS breast screening programme: cohort study. BMJ 323: 140 & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S40 – S48 Intervention to promote early presentation of breast cancer L Linsell et al S48 Wampold BE, Brown GS (2005) Estimating variability in outcomes the efficacy of a mailed, psychoeducational intervention in a community attributable to therapists: a naturalistic study of outcomes in managed sample of older adults. Health Psychol 22: 99–105 care. J Consult Clin Psychol 73: 914–923 Westcombe AM, Richards MA, Ramirez AJ, Love SB, Sutton S, Burgess C, Wardle J, Waller J, Brunswick N, Jarvis MJ (2001) Awareness of risk factors Littlejohns P (1999) A Systematic Review of the Delay in Diagnosis/Treatment for cancer among British adults. Public Health 115: 173– 174 of Symptomatic Breast Cancer. NHS Research and Development: London Wardle J, Williamson S, McCaffery K, Sutton S, Taylor T, Edwards R, Atkin Zeger SL, Liang K-Y (1986) Longitudinal data analysis for discrete and W (2003) Increasing attendance at colorectal cancer screening: testing continuous outcomes. Biometrics 42: 121 –130 British Journal of Cancer (2009) 101(S2), S40 – S48 & 2009 Cancer Research UK

Journal

British Journal of Cancer – Springer Journals

Published: Dec 3, 2009

There are no references for this article.