Anticipated health behaviour changes and perceived control in response to disclosure of genetic risk of breast and ovarian cancer: a quantitative survey study among women in the UK

Background Genetic risk assessment for breast cancer and ovarian cancer (BCOC) is expected to make major inroads into mainstream clinical practice. It is important to evaluate the potential impact on women ahead of its implementation in order to maximise health benefits, as predictive genetic testing without adequate support could lead to adverse psychological and behavioural responses to risk disclosure. Objective To examine anticipated health behaviour changes and perceived control to disclosure of genetic risk for BCOC and establish demographic and person-specific correlates of adverse anticipated responses in a population-based sample of women. Design Cross-sectional quantitative survey study carried out by the UK Office for National Statistics in January and March 2014. Setting Face-to-face computer-assisted interviews conducted by trained researchers in participants’ homes. Participants 837 women randomly chosen from households across the UK identified from the Royal Mail’s Postcode Address File. Outcome measures Anticipated health behaviour change and perceived control to disclosure of BCOC risk. Results In response to a genetic test result, most women (72%) indicated ‘I would try harder to have a healthy lifestyle’, and over half (55%) felt ‘it would give me more control over my life’. These associations were independent of demographic factors or perceived risk of BCOC in Bonferroni-corrected multivariate analyses. However, a minority of women (14%) felt ‘it isn’t worth making lifestyle changes’ and that ‘I would feel less free to make choices in my life’ (16%) in response to BCOC risk disclosure. The former belief was more likely to be held by women who were educated below university degree level (P<0.001) after adjusting for other demographic and person-specific correlates. Conclusion These findings indicate that women in the UK largely anticipate that they would engage in positive health behaviour changes in response to BCOC risk disclosure.

1. Page 11, para 2, lines 52-53: 'which would by definition' -not true as it would depend on what had been decided / agreed in terms of returning unanticipated findings. See ACMG guidance and subsequent controversy regarding return of unanticipated findings. See Green et al. (2013) Genet Med, 15, pp. 565-574;Burke et al. (2013) Genet Med 15(11), doi:10.1038/gim.2013ACMG (2015), Genet Med 17(1): 68-69. This needs a bit of amending to clarify. 2. Page 15, para 3, lines 52-53: replace 'in' with 'than' after 'BCOC' 3. Although there is a set of bullet points on strengths and limitations of the study at the start of the manuscript, I am used to seeing a section in the Discussion part of all research articles highlighting strengths and limitations. Is it a policy of BMJ Open for this just to be included as bullet pints at the start? If not, I'd prefer to see this included in the Discussion section, as the bullet points don't highlight the following limitations: (i) the lack of ethnic diversity in the sample and (ii) data were not collected on what anticipated lifestyle changes respondents would make.

REVIEWER
Jennifer Taber Kent State University, USA REVIEW RETURNED 07-Jul-2017

GENERAL COMMENTS
This manuscript reports data from a cross-sectional survey in which a representative sample of UK women were asked about their expected responses to receipt of genetic information about their risk of breast and ovarian cancer. Most women were in favor of receiving this information and three-quarters thought they would try to improve their health in response. I appreciate the attempt to anticipate where genetic testing is headed and to assess attitudes about testing in the general population, and the sampling seems well done. My concerns, described below, center around the four items included to assess anticipated responses. 1. The manuscript is framed as assessing women's "anticipated responses." This is a vague term that could encompass affective (e.g., distress, depression), cognitive, motivational, and behavioral responses. The authors assessed a very specific type of responsewhether women expected to change their health behaviors and whether they would have control over their life. The more narrow scope of these questions should be made clearer throughout, including in the title, abstract, and introduction. Further, on the first page of the discussion, the survey is justified in part by saying that there are concerns about psychological adjustment following receipt of genetic test results, but the survey does not assess psychological outcomes (psychological outcomes are again emphasized on page 6, line 34 and again on the first line of page 7). 2. The survey questions are vague and without knowing how participants interpreted them, it is difficult to know how to interpret the results. First, the authors acknowledge (and state as a limitation) that the questions ask women to anticipate responses to learning of risk independent of whether they learn of high or low risk. It is possible that likelihood of changing behavior, or feeling like one has choices, would differ dramatically according to the risk conferred. Second, the questions refer to having a "healthy lifestyle" and "lifestyle changes" and participants may have interpreted these phrases in a variety of ways. Are diet and exercise medically advisable for women at high risk of breast cancer or ovarian cancer? Or are screening and perhaps prophylactic surgery all that would be recommended? This issue is not covered in the introduction or discussion. Without knowing how participants interpreted these phrases, or whether lifestyle changes would be advised (which would likely depend also on degree of risk conferred), interpreting the results is difficult. Participants may have selected a response because "unsure" was not an option but these responses may not meaningfully reflect their beliefs. What (if any) was the label for the midpoint of the scales for the anticipated response items that ranged from strongly disagree to strongly agree? It may be useful to report the proportion that selected the middle option for each question versus disagreement.
3. The term "personal predictors" is vague and not very informative.
4. In the introduction, it would be useful to have a sense of the degree of risk conferred by moderate-and lower-risk markers.
5. Why treat perceived risk as categorical for some analyses rather than conducting a test in which it is treated as continuous?
6. On page 13 in the discussion, the statement that "women in the UK do not seem to hold particularly strong views about" whether whole genome sequencing should be included in routine new-born screening is an extrapolation that is not supported by the data.
7. Table 2 should include statistics in addition to the p values (chi square values and DF).
8. In Table 2, why is the full sample size shown as 925 when it is stated as 837 elsewhere?
9. The abstract should include the year the data were collected.

Sarah Knerr
University of Washington, United States REVIEW RETURNED 07-Jul-2017

GENERAL COMMENTS
This manuscript describes results from a cross-sectional in person survey of adult women in the UK that asked women about genetic testing to provider breast and ovarian cancer risk information. The study's strength is the relatively large and generalizable sample of female respondents. Additional weaknesses beyond those highlighted in the manuscript are use of single questions with unknown psychometric properties/measurement characteristics as outcome measures and lack conceptual or theoretical framework(s) guiding the study design, analysis, and interpretation. Also addressing the large proportion of women who responded that they were unsure about the timing of genetic testing and what that means for the study's conclusions about overall "support for genetic testing".

Major suggested revisions:
-Additional specificity about the relationship between the study aims, outcomes of interest, and the questions used to measures those outcomes (and their limitations) needed throughout. Specifically, I don't think that the paper makes a strong enough case that a question about the timing of a genetic test is a robust indicator of whether women are supportive of the idea of genetic testing. This outcome (support of genetic testing) isn't included in the study aims, but leads off the results and discussion sections. The results and discussion also pay a good deal of attention to specific results about the timing of a genetic test, a topic that is not discussed in the introduction. Describing the question as measuring two things (support for genetic testing and also perceptions about timing of genetic testing) is confusing.
-The describing the outcome measures/questions about hypothetical behavioral and psychological responses to genetic information as "positive anticipated responses" and "deterministic attitudes" are confusing given women's potential response options, the presentation of results, and also the implication (without a corresponding theoretical justification) that endorsing these attitudes is a good/bad in some way. I would suggest keeping the description of the outcome as close to the wording of the question as possible or referring to them as hypothetical behavioral and psychological responses.
-Variable coding needs to be justified and remain consistent throughout all analyses and tables. Add in justification for dichotomizing outcomes and using logistic regression when the outcome variable is categorical. Were outcomes dichotomized in the univariate analysis as well?
-Mention descriptive statistics (Table 1) in methods section and include sample sizes for modeling in Table 3.
-Language in the results section: replace "effects" with "associations" and "survived" with "remained. Replace "finding was maintained" with "association remained".
Minor suggested revisions: -Use "variant" and/or "pathogenic variant" instead of mutation or hereditary gene mutation.
-If introducing the concept of health care disparities in relationship to the current study, more background and connecting logic needs to be provided.
-How are address stratified during sampling? Does this impact the analysis? -Replace "had stayed in education for longer" with "more years of formal education" -Concluding that women in the UK don't hold strong views about the timing of genetic testing based on the question that was asked seems misleading.
-Define genetic determinism -Education does not necessarily equate with information literacy, which wasn't measured in the study.
-Why is the sample size in 3. Page 8-9: It would be helpful in the section on 'Measures' to cross-refer to the survey form, provided as either a figure or supplementary material.
Response: Thank you for this comment. We have now added the relevant questions from the survey form as supplementary material.
4. Page 10, line 12-13: Please justify why 50 years was used as the cut-point to investigate generational differences.
Response: We thank the reviewer for this comment. We have now included the following: We chose 50 years of age as the cut-off point since public interest in the UK for genetic testing has been reported as being greatest for people aged 46-60 years and we felt that the age of 50, being close to the average age of menopause in the UK (which is 51), was a reasonable mid-point to assess generational effects.
Results: 1. Page 12, para 2: It would have been helpful to know how respondents anticipated changing their lifestyles in response to genetic risk information. This could have been included in the survey, and perhaps ought to be mentioned in the 'Limitations' section (which seems to be missing from the Discussion section).
Response: We thank the reviewer for this comment. We have now added the following in the limitations section: Secondly, because of survey constraints, we focused our investigation on a limited set of potential enablers and barriers of genetic testing, namely perceived benefits and deterministic attitudes, and were unable to further investigate how exactly respondents anticipated changing their behaviour. Please leave your comments for the authors below This manuscript reports data from a cross-sectional survey in which a representative sample of UK women were asked about their expected responses to receipt of genetic information about their risk of breast and ovarian cancer. Most women were in favor of receiving this information and three-quarters thought they would try to improve their health in response. I appreciate the attempt to anticipate where genetic testing is headed and to assess attitudes about testing in the general population, and the sampling seems well done. My concerns, described below, center around the four items included to assess anticipated responses.
1. The manuscript is framed as assessing women's "anticipated responses." This is a vague term that could encompass affective (e.g., distress, depression), cognitive, motivational, and behavioral responses. The authors assessed a very specific type of response-whether women expected to change their health behaviors and whether they would have control over their life. The more narrow scope of these questions should be made clearer throughout, including in the title, abstract, and introduction. Further, on the first page of the discussion, the survey is justified in part by saying that there are concerns about psychological adjustment following receipt of genetic test results, but the survey does not assess psychological outcomes (psychological outcomes are again emphasized on page 6, line 34 and again on the first line of page 7).
2. The survey questions are vague and without knowing how participants interpreted them, it is difficult to know how to interpret the results. First, the authors acknowledge (and state as a limitation) that the questions ask women to anticipate responses to learning of risk independent of whether they learn of high or low risk. It is possible that likelihood of changing behavior, or feeling like one has choices, would differ dramatically according to the risk conferred. Second, the questions refer to having a "healthy lifestyle" and "lifestyle changes" and participants may have interpreted these phrases in a variety of ways. Are diet and exercise medically advisable for women at high risk of breast cancer or ovarian cancer? Or are screening and perhaps prophylactic surgery all that would be recommended? This issue is not covered in the introduction or discussion. Without knowing how participants interpreted these phrases, or whether lifestyle changes would be advised (which would likely depend also on degree of risk conferred), interpreting the results is difficult. Participants may have selected a response because "unsure" was not an option but these responses may not meaningfully reflect their beliefs. What (if any) was the label for the midpoint of the scales for the anticipated response items that ranged from strongly disagree to strongly agree? It may be useful to report the proportion that selected the middle option for each question versus disagreement.
Response: We thank the reviewer for this comment. We agree that the survey questions were open to interpretation by participants, but we were unable to include more fine-grained questions because of the scope of the survey. We have now highlighted this limitation in greater detail in the discussion section of the manuscript, and also included this as a major limitation in the 'bullet points'. We have also included a full breakdown of responses for the outcome variables in the supplementary material.
3. The term "personal predictors" is vague and not very informative.
Response: We have now changed this to person-specific in the manuscript.
4. In the introduction, it would be useful to have a sense of the degree of risk conferred by moderateand lower-risk markers.
Response: We have now added the following sentence to the 1st paragraph in the Introduction: 'While high penetrance genes like BRCA1 and BRCA2 confer a lifetime risk for cancer of 50% or greater, the lifetime risk for moderately penetrant genes ranges from 20% to 50% and low penetrance genes have a limited effect.' 5. Why treat perceived risk as categorical for some analyses rather than conducting a test in which it is treated as continuous?
Response: We thank the reviewer for this comment. Given that individual categories of much lower/much higher perceived risk were very small, we chose to collapse categories as shown, and use a categorical model of analysis. We have now included this justification in the manuscript.
6. On page 13 in the discussion, the statement that "women in the UK do not seem to hold particularly strong views about" whether whole genome sequencing should be included in routine new-born screening is an extrapolation that is not supported by the data.
Response: We thank the reviewer for this comment. We have now removed this section.
7. Table 2 should include statistics in addition to the p values (chi square values and DF).
Response: We have originally not included these values in the interest of space. We have now included these. This manuscript describes results from a cross-sectional in person survey of adult women in the UK that asked women about genetic testing to provider breast and ovarian cancer risk information. The study's strength is the relatively large and generalizable sample of female respondents. Additional weaknesses beyond those highlighted in the manuscript are use of single questions with unknown psychometric properties/measurement characteristics as outcome measures and lack conceptual or theoretical framework(s) guiding the study design, analysis, and interpretation. Also addressing the large proportion of women who responded that they were unsure about the timing of genetic testing and what that means for the study's conclusions about overall "support for genetic testing".

Major suggested revisions:
Comment: Additional specificity about the relationship between the study aims, outcomes of interest, and the questions used to measures those outcomes (and their limitations) needed throughout. Specifically, I don't think that the paper makes a strong enough case that a question about the timing of a genetic test is a robust indicator of whether women are supportive of the idea of genetic testing. This outcome (support of genetic testing) isn't included in the study aims, but leads off the results and discussion sections. The results and discussion also pay a good deal of attention to specific results about the timing of a genetic test, a topic that is not discussed in the introduction. Describing the question as measuring two things (support for genetic testing and also perceptions about timing of genetic testing) is confusing.
Response: Thank you for this critique. We agree and have removed this section entirely from the manuscript.
Comment: The describing the outcome measures/questions about hypothetical behavioral and psychological responses to genetic information as "positive anticipated responses" and "deterministic attitudes" are confusing given women's potential response options, the presentation of results, and also the implication (without a corresponding theoretical justification) that endorsing these attitudes is a good/bad in some way. I would suggest keeping the description of the outcome as close to the wording of the question as possible or referring to them as hypothetical behavioral and psychological responses.
Response: We thank the reviewer for this comment. We have now made every effort to stick closely to the question wording rather than extrapolating from it.