Introduction

Primary Prevention of Cancer

Prevention and screening are the two main strategies for reducing the burden of cancer. Primary prevention of cancer in general and breast cancer (BC) in particular is based on awareness of and thus consequent identification of risk factors (RF).

The French National Cancer Institute (INCa) works to raise awareness regarding exposure to individual BCRFs, including hormone therapy, tobacco consumption, alcohol consumption, and overweight [1]. In the context of the nationwide BC screening program in France, the question of adapting screening according to individual BCRFs is currently under debate [2]. The widely implemented program in France recommends a mammography every 2 years for all women in the average-risk population aged 50–74 years. However, the latest public health guidelines published in 2014 [2] also defined populations at high and very high risk of BC, irrespective of age, including women with a personal history of BC, benign breast pathology (in situ carcinoma, hyperplasia), or high-dose thoracic medical irradiation, as well as those with a family history of cancer. Specific screening strategies are recommended for this population, screening examinations based on imaging techniques other than X-Ray (echography or magnetic resonance imaging), and initiated earlier or repeated more frequently. Oncogenetic testing is also used.

The EDIFICE Surveys

EDIFICE surveys aim to improve insight into the behavior of the French population with regard to cancer prevention and screening uptake. Since 2005, they have addressed various issues regarding participation in population-based national screening programs (breast, colorectal) and individual attitudes towards screening for prostate, cervical, or lung cancer. Other health-related issues have also been addressed, including tobacco, e-cigarette or alcohol consumption, and the assessment of individual risks of cancer. Factors likely to impede or trigger screening uptake have also been studied.

Effective prevention strategies such as improved understanding and thus identification of RFs among the general population are useful from two standpoints. By encouraging a change in individual behavior, with a potentially positive impact on individual risk, they may also incite greater compliance to screening programs at the same time as boosting individual decision-making initiatives.

Although this topic has previously been studied in different countries [3,4,5,6], our analysis provides current-day insight into the knowledge and awareness of BCRFs not only in the population at large but also among physicians. We compared RFs as perceived by the laypopulation with those that are scientifically acknowledged.

Methods

Methodology of the EDIFICE Surveys

This fourth nationwide observational survey, EDIFICE 4, was conducted by phone interviews, using the quota method [7]. Survey questionnaires were administered by experienced independent interviewers (Kantar Health) using the computer-assisted telephone interview (CATI) technique. They collected information on sociodemographic characteristics (age, area of residence, town size) and socioprofessional factors. The present work focuses on the breakdown of spontaneous replies to the question “In your opinion, what are the five main RFs that increase the risk of breast cancer?”. When interviewees cited “tobacco,” they were asked to specify whether it was passive or active smoking. Similarly, details were requested when diet was cited.

Sample Populations

An initial representative sample of 1602 individuals (age 40–75 years) was interviewed between June 12 and July 10, 2014. Population statistics were adjusted for sex, age, and profession and stratified by territory and size of urban areas based on data from the French Employment Survey conducted in 2009 and updated in 2012 by the French National Institute for Statistics and Economic Studies (Institut National de Statistiques et d’Etudes Economiques [INSEE]) [8].

A mirror survey on a representative sample of 301 physicians was conducted between July 9 and August 8, 2014. This population comprised 201 general practitioners (GPs) and 100 oncologists. Data on the former were adjusted for age and geographical area and on the latter, for the type of healthcare institution and geographical area.

Statistics

Student’s ttest for quantitative data was used for all comparisons between two populations and the Ztest and the χ2test were used for comparisons of categorical data. Differences were considered statistically significant when the P value was < 0.05 (two-sided test) (SAS software, version 8.2 SAS Institute Inc., Cary, NC, USA).

Results

Cited Risk Factors

The present analysis is based on two subpopulations: 737 laywomen with no history of cancer and 105 female physicians. When asked to cite five BCRFs, laywomen cited an average of 2.2 (standard deviation [SD] 1.3) and physicians 3.9 (SD 1.2). All physicians cited at least one RF whereas 22% of laywomen did not answer the question.

Heredity/Family history of breast cancer was the most widely recognized RF in both populations (Table 1). In second place, after this non-behavioral RF, both populations cited behavioral RFs, namely Drug therapy for physicians and Unhealthy lifestyle for laywomen. Lifestyle RFs cited by similar proportions of physicians and laywomen included Smoking and significantly less frequently, Alcohol consumption.

Table 1 Breakdown of answers to the question “In your opinion, what are the five main risk factors that increase the risk of breast cancer?” Comparison between female physicians and laywomen with no history of cancer. NC, not cited
Full size table

Late/No childbearing was rated in the middle range by physicians and rarely cited by laywomen. Overweight/obesity, No breastfeeding, Age, and Menopause were all ranked at similar levels by physicians though at much lower levels by laywomen.

Surprisingly, a number of established RFs were very rarely cited by either physicians or laywomen, e.g., Exposure to medical radiation or even not at all, i.e., Benign breast conditions and a Personal history of breast cancer. Other RFs were cited by less than one in ten laywomen.

Medical Population

The breakdown of answers from GPs and oncologists shows that physicians all ranked heredity/family history first, drug therapy second, and an unhealthy life style third, as risk factors for breast cancer (Table 2). However, compared to GPs, oncologists were more likely to cite age (34.3 vs. 12.9%, P = 0.02), overweight/obesity (34.3 vs 15.7%, P ≤ 0.05), and menopause (37.1 vs. 15.7%, P = 0.03) but less likely to cite tobacco use (8.6 vs. 24.3%, P ≤ 0.05).

Table 2 Breakdown of answers to the question “In your opinion, what are the five main risk factors that increase the risk of breast cancer?” Comparison between female general practitioners (GPs) and oncologists
Full size table

Discussion

Our analysis suggests that both physicians and laywomen are largely aware of Heredity/Family history as the major RFs for BC. Other recognized RFs such as Exposure to medical radiation or Alcohol were however cited at surprisingly low levels. Moreover, and unexpectedly, Benign mastopathy and a Personal history of breast cancer were either omitted or not acknowledged, notably by physicians. This may be due to the inherent potential bias of declarative surveys where participants are not asked to select a response but to provide spontaneous replies, with the obvious risk of memorization bias or social desirability bias. Our survey does however have the advantage of a robust, validated methodology [8].

Risk Factor Ranking

Ranking BCRFs according to their relative impact on morbimortality is a key issue for targeting prevention. However, it becomes critical in the absence of any absolute scale [9, 10]. The relative risk for a given RF and the population-attributable fraction of cancer incidence and/or death are generally used to provide a quantitative, relative evaluation of the impact of a given RF.

Obesity, smoking, alcohol, and lack of physical activity are nowadays recognized as the most prevalent lifestyle RFs [11, 12•, 13, 14]. A US prospective cohort study showed that 15% of BC and 45% of BC deaths could be prevented in the USA through lifestyle changes [15].

However, prevention strategies should also take into consideration the fact that the absolute level of risk is likely to be the result of a combination of RFs, which notably is closely related to hormonal status [14, 16, 17, 18•] or genetic susceptibility [19•].

One attempt to rate BCRFs [10] reached the conclusion that factors associated with personal or familial history of BC were correlated with the highest relative risks, ranging from 3 to 200 (for BRCA1 and BRCA2 mutations). Exposure to medical radiation was associated with a relative risk of 5.2, whereas lower relative risks (< 2) were associated with hormone use, alcohol consumption, or even obesity.

Heredity and Family History

Our analysis highlights Heredity/Family history as the primary RF recognized by both physicians and laywomen. Interestingly, Genetic factors ranked lower but may have been implicitly included with the more general Heredity/Family history item for most interviewees. Previous investigations have reported conflicting opinions on this subject [4, 5]. Extensive media coverage of cases such as that of Angelina Jolie, which without question increased public awareness regarding BC, may well account for this observation [20]. One can legitimately extrapolate this observation and hypothesize that media coverage also increased awareness of the hereditary component of BC.

Little Known Risk Factors

In contrast, Exposure to medical radiation was very seldom mentioned in our study, and Benign mastopathy along with Personal history of breast cancer was not spontaneously cited by either physicians or laywomen. In other investigations, Personal history of breast cancer was indeed acknowledged but only when suggested by the interviewer [21]. However, because both are conditions requiring specific follow-up and screening procedures [2], these findings raise concerns regarding compliance with the national recommendations.

Awareness of Actual Risk Factors

It is nowadays widely acknowledged that reproductive factors and drug therapies (hormone replacement or oral contraception) have contributed to the increased incidence of BC [12•, 22, 23]. Therefore, as expected, these criteria were rated in the middle range of our survey, although less frequently acknowledged by laywomen.

A number of modifiable RFs related to specific at-risk behavior or exposure are individual components of the risk of BC. The most widely acknowledged BCRFs are alcohol consumption [12•, 24,25,26,27,28,29], overweight/obesity [12•, 30, 31], lack of physical activity [12•, 32, 16], and diet [12•, 33, 34]. The link with tobacco consumption is still under debate and the inconsistency of results is recognized in the literature [35,36,37,38,39,40]. High body mass index (BMI) has been shown to be the highest ranking BCRF in postmenopausal women, followed by alcohol consumption and lack of physical activity, with tobacco in the last position [18•]. Tobacco was however not ranked at all in other studies [12•, 10].

Our results suggest that the impact of tobacco was over-estimated by both healthcare professionals (GPs) and laywomen in our survey. Laywomen and GPs under-estimate the risks associated with a high BMI, and there is little awareness of the relationship between alcohol consumption and breast cancer. This has already been reported [41, 42], and ignorance of the cancer risk in countries where alcohol consumption is closely linked to cultural aspects has been suggested [42].

Conclusion

Heredity/Family history is widely recognized by both the laypopulation and physicians as a major RF for BC, no doubt partly due to recent extensive media coverage of high-profile cases. The limited awareness of the risk related to Exposure to medical radiation, Benign mastopathy, or Personal history highlights the issue of compliance with the national guidelines, which recommend specific screening procedures for high-risk populations. A number of misconceptions with regard to behavioral RFs have also been highlighted, including the relative impact of alcohol and tobacco consumption and the importance of menopausal status. Future breast-cancer awareness-raising campaigns should include messages on at-risk behavior that is largely unknown to the general population. Physicians should be encouraged to focus strongly on these issues in their discussions with patients.