Women’s perception, attitudes, and intended behavior towards predictive epigenetic risk testing for female cancers in 5 European countries: a cross-sectional online survey

The study was set up as a cross-sectional population-based online survey with women from five European countries that represent Northern, Eastern, Southern, Western, and Central Europe and the nationalities of members of the FORECEE consortium. Based on these criteria the countries chosen were the Czech Republic, Germany, United Kingdom (UK), Italy, and Sweden. The design and content of the study was developed by the authors, discussed in detail with country-specific clinical partners of the FORECEE consortium (https://​forecee.​eu), and revised after feedback. To execute the study online, the survey sampling company Harris Interactive (Hamburg, Germany) was contracted in order to conducted the online study—including an online version of a leaflet and an online questionnaire—by using their online panels and the online panels of Toluna, respectively. Both online panels—comprising about 78,000 and 275,000 online panelist, respectively—are representative of the general population in these countries.

The study was performed in accordance with relevant guidelines and regulations, and informed consent was obtained from all participants prior to the study. The study was approved by the independent Institutional Ethics Review Board of the Max Planck Institute for Human Development (Germany).

The goal of the online study was to survey national samples of women who will likely be the target group for future predictive epigenetic tests for female cancer risks (40 to 75 years) in the Czech Republic, Germany, UK, Italy, and Sweden. To reduce nonrespondent bias and to better reflect the general population of women of the target group in each country, the samples were stratified based on the official combined distribution of age and education per country at the point of survey completion. Quotas per country were drawn from EUROSTAT (European Statistical System) [11] and calculated on the grounds of three levels of education (low, medium, high) as categorized by the International Standard Classification of Education (ISCED) and for four age groups (40–49 years, 50–59 years, 60–69 years, 70–75 years). We calculated that a sample size of about 300 participants per country was required to detect differences in perceptions and intended behavior (2-sided alpha of .05) of 10% or higher with a 90% power between the national samples. To allow for nonresponse and ineligibility, in January 2017, Harris Interactive sent invitation links by email to those of their female online panelists who matched the selection criteria in age and nationality. In case, the invitation caught the interest of an online panelist, she clicked on a link provided in the invitation email that directed her to the online portal of Harris Interactive. After providing informed consent, women entered the online study—entirely hosted by Harris Interactive—which included both the online survey questionnaire and the online leaflet. After study completion, the authors of this study received a completely anonymized data sheet from Harris Interactive. Thus, all respondents of this study are fully anonymous to the authors.

Of 3629 women contacted by Harris Interactive, 848 did not respond, 197 were not eligible (duplicate listing, other strata than originally coded), and 492 entered the online study after respective quotas were filled. Of the 2092 who responded, 417 did not finish the online study, resulting in 1675 completed online surveys. Using the AAPOR (American Association for Public Opinion Research) response rate calculator, [12] which incorporates a default method for estimating e (estimated proportion of cases of unknown eligibility that is eligible), the study yielded a response rate of 61.4% (1675/[1675 + 417 + e (848)]) and a cooperation rate of 80.1% (1675/[1675 + 417]).

Of the 1675 online surveys collected in total, 356 were completed in the Czech Republic, 335 in Germany, 323 in the UK, 338 in Italy, and 323 in Sweden. Across countries, the distribution of age groups was 40–49 years (29.6%), 50–59 years (29.3%), 60–69 years (29.5%), and 70–75 years (11.7%) and the distribution of educational levels was low (27.8%), medium (48.8%), and high (23.5%). 157 women reported a personal history of any cancer in the past and 104 specifically of female cancers. Table 1 provides country-specific details on these characteristics.

Respondents in each of the national online samples were similar to their respective general population in terms of the distribution of age and education, with the following exceptions: For women from the Czech Republic and Sweden presenting with low education, the age group “60 to 69 years” were overrepresented and “70 to 75 years” were underrepresented (Additional file 2: Table S1).

To ensure that all women were equipped with sufficient background information on predictive epigenetic testing for female cancer risk, the study was designed as such that the women entering the online study were first presented with an online leaflet—whose content they were advised to familiarize themselves with—before being subsequently presented with the online survey questionnaire. While then working through the online survey questionnaire, women could easily re-approach the content of the online leaflet at any time. The online leaflet (see Additional file 2) included details on the following aspects:

The development of the online survey questionnaire—designed to elicit women’s perception, attitudes, and intended behavior towards predictive epigenetic testing for female cancer risk—was informed by an assessment of findings published on these aspects in the field of genetic testing, in-depth discussions with the clinical partners and (epi-)genetic experts of the FORECEE consortium group, and by outcomes of focus group discussions with 25 women from Germany and 12 women from the UK. The survey questionnaire (for the exact wording, see Additional file 2) first asked four questions on the core principles of predictive epigenetic testing to ensure that women had a sufficient basic understanding of the cancer types targeted by the test, the potential impact of different test outcomes, and the interplay between external factors and the epigenome. For each of these four questions, women were presented with a choice of three answers of which one was correct. Women’s perception of the benefit-to-harm ratio of predictive epigenetic testing for female cancer risks was measured by a 5-point-Likert scale reaching from “harms clearly outweigh potential benefits” to “benefits clearly outweigh potential harms.” Their desire to learn about their cancer-specific risk for each of the four cancer types and their intention to take predictive epigenetic testing for female cancer risks if the test were freely available were measured by a binary choice (yes/no). To explore reasons in favor or disfavor of the test, we presented women with lists of items in favor (e.g., motivate the adoption of healthier lifestyle) and disfavor (e.g., induce unnecessary worry) of predictive epigenetic testing that were derived from focus group discussions with women in Germany and the UK. Women were asked to tick the reasons that would personally matter to them and assign ranks of importance to each (assignment of equal ranks to different reasons was possible). After women had made their choices on reasons in favor and disfavor, each woman was again presented with her chosen reasons and asked if one of these reasons was so strong that it would outweigh all the other reasons in favor or in disfavor of the test.

All study materials (online survey questionnaire, online survey leaflet) were translated into country-specific languages by a professional translation office and checked for correctness and completeness by country-specific members of the FORECEE consortium. The survey sampling company Harris Interactive (Hamburg, Germany) programmed the online version of the survey.

The online version of the questionnaire did not allow for item nonresponse, thus all 1675 questionnaires were completed. Primary outcome measures were (i) women’s evaluation of the benefit-to-harm ratio of a predictive epigenetic test for female cancer risks, (ii) women’s desire to know their cancer risk per cancer site, (iii) women’s reasons in favor of having predictive epigenetic testing for female cancer risk, (iv) women’s reasons in disfavor of having predictive epigenetic testing for female cancer risk, and (v) women’s intention to use predictive epigenetic testing for female cancer risks. All results on the primary outcomes were calculated as absolute proportions with 95% confidence intervals.

Binary and multivariate logistic regression models were used to investigate whether differences between countries for each of the primary outcomes were confounded by covariates such as women’s knowledge of the core principles of predictive epigenetic testing, their age, their education, or their personal cancer history. If not reported otherwise, none of these variables significantly influenced country-specific differences. Bonferroni corrections were employed to correct for multiple testing, and α was set at 0.05/5 = .01. All data were stored and analyzed with IBM SPSS Statistics 24 (New York City, USA). To facilitate clear and unbiased reporting of our study results, the STROBE (Strengthening the Reporting of Observational studies in Epidemiology) reporting guideline was applied (Additonal file 1: STROBE checklist). 

Overall, between 57.4% (55.0, 59.8) and 94.6% (95% confidence interval [CI]: 93.4, 95.6) of women provided correct responses to each of the four questions on the core principles of the predictive epigenetic test for female cancer risk (Table 2). Women were least likely to correctly respond to the question on how a lower-than-average risk test result may guide their cancer screening and prevention management. Women from the Czech Republic (p < .001) and women with lower education (p < .001) were more likely to provide lower proportions of correct responses on core principles than women from the other European countries or women with higher education (Additional file 2: Table S2). After adjusting for education in multivariate analyses and multiple testing, the significant association between country and women’s knowledge of core principles (p = < .001) remained.

More than half (56.6, 95% CI: 54.2, 59.0) of all women felt that the potential benefits of a predictive epigenetic test for female cancer risks would somewhat or clearly outweigh its potential harms. This evaluation was shared by the majority of women within each of the studied European countries, except by women from the Czech Republic where the majority thought that the test’s harms equal its benefits (p < .001) (Fig. 1). After adjusting for covariates (knowledge of core principles, education, personal female cancer history, age) and multiple testing, compared to women from the Czech Republic, women from the other four countries were on average only about half as likely to consider the harms of a predictive epigenetic test for female cancer risks to equal or to somewhat/clearly outweigh the benefits (Additional file 2: Table S2).

Also, women with limited knowledge of the test’s core principles (≤50% correct) evaluated the test more negatively in univariate analyses (p < .001). Multivariate analysis showed, however, that the association between women’s benefit-to-harm evaluation and their knowledge of the test’s core principles was confounded by their level of education (χ² (1) = 22.48; p < .001). After adjusting for covariates and multiple testing, women presenting with limited knowledge of the test’s core principles were seen to be nearly 2 times as likely (odds ratio [OR]: 1.79, 95% CI: 1.24–2.25, p < .001) to consider the harms of the predictive epigenetic test to equal or somewhat/clearly outweigh the benefits than women presenting with higher degrees of knowledge (> 50%) (Additional file 2: Table S2).

The majority of women (63.6%) said they want to know their individual risk for each of the four female cancers (breast, ovarian, cervical, endometrial) (Table 3). Within each country, women’s desire to know the risk for one female cancer over the other cancers did not differ significantly. Women’s desire differed between countries, however: Compared to women from the Czech Republic, women from the UK, Italy, and Sweden were about 2 to 3 times more likely to desire risk information on three or all female cancers in multivariate analysis (Additional file 2: Table S2), whereas women from Germany did not differ from the Czech Republic (p = .122). Also, women’s age was associated with a higher desire to learn about the risk for female cancers in univariate and multivariate analysis, with women aged 40 to 59 years being significantly more likely to request such predictive information on their cancer risk than women older than that (Additional file 2: Table S2).

When women were presented with the five potential reasons in favor of taking this epigenetic test, 67.2% of all women chose all five reasons, 3.9% chose four reasons, 6.4 and 7.3% chose three and two reasons, respectively, and 15.3% chose one reason. Figure 2a shows women’s ranking of their perceived importance of each of the reasons in favor of the test. When subsequently presented with all individually ticked reasons, 817 (48.8%) out of 1675 women indicated that one of these reasons was so strong that it outweighed all the others. Among these, the test’s potential to guide the personal cancer prevention strategy was the most decisive reason for women in Germany (38.8%), the Czech Republic (29.1%), and Italy (30.9%), whereas for women in the UK (37.0%) and Sweden (30.4%), it was its potential to motivate the adoption of a healthier lifestyle.

When presented with the five reasons that might speak in disfavor of taking the predictive epigenetic test, 61.1% of all women chose all five reasons, 3.5% chose four reasons, 5.6 and 8.6% chose three and two reasons, respectively, and 21.3% chose one reason. Figure 2b shows women’s ranking of their perceived importance of each of the reasons in disfavor of the test. When again subsequently presented with all individually ticked reasons in disfavor of the test, 884 (52.8%) out of 1675 women indicated that one of these reasons was so strong that it outweighed all the others. That the result of the test may induce unnecessary worry and reduce quality of life was the most decisive reason against taking the test for women in the UK (46.7%) and Sweden (33.5%), whereas for women in Germany (37.1%) and Italy (30.9%) it was the test’s potential to cause a woman to permanently expect the onset of cancer. Women in the Czech Republic (39.4%), in contrast, were most worried about the fact that the test result may pressure them into adopting a healthier lifestyle or to undergo more cancer screening.

Overall, 75% of women responded that they would definitely or probably take this epigenetic test if it were freely available. However, women’s preparedness to take the test depended on the country they came from (p < .001) and their initial evaluation of the test’s benefit-to-harm ratio (p < .001). After adjusting for covariates and multiple testing, women from the Czech Republic, UK, Italy, and Sweden were about 2 to 3.5 times more likely to consider taking the test if it were freely available than women from Germany (Additional file 2: Table S2). Results were confounded, however, by women’s evaluation of the test’s benefit-to-harm (χ² (1) = 92,67; p < .001), with women evaluating the test rather negatively being less prepared to take the test (p < .001). In accordance with this finding, we found that women who intended to take the test were also more likely to tick four or all five reasons in favor of the test than women who said they would definitely or probably not take the test: 73.7% (71.1, 76.1) versus 63.3% (58.4, 67.9). Likewise, women who indicated that they would not take the test were more likely to tick four or all reasons in disfavor of the test than women who indicated they would take the test: 71.9% (67.3, 76.2) versus 62.5% (59.8, 65.2).