We invited all 2369 members of a panel of the national Dutch CVD patients’ association (Harteraad) to take part in an online survey. This panel consists of CVD patients, individuals who are, in some way, related to CVD patients (partners, parents/legal guardians and caregivers) and individuals at increased risk of CVD (self-reported, e.g. those with diabetes, hypertension or overweight). The panel regularly receives surveys from healthcare professionals, researchers and policymakers on their experiences with cardiovascular health.
On 15 November 2018, the panel received an email invitation with a hyperlink to an anonymous online survey (in Dutch) that was available for 7 weeks. Completion of the survey was voluntary. No personal reminder could be sent because the survey was anonymous and a general reminder would pose too great a burden on the panel at large. As the Dutch Medical Research Involving Human Subjects Act did not apply to this study, explicit ethical approval was not needed.
The survey included instructions and a brief introduction into the topic, followed by general questions on sociodemographic variables (age, country of birth, parents’ country of birth, sex, gender identity [personal perception of one’s own gender: man, woman, non-binary, gender fluid, transgender, transsexual, other], and educational level) and history of CVD. As a measure for priority setting, we asked respondents to choose a maximum of 3 out of 16 conventional risk factors (mentioned on the website of the Dutch Heart Foundation [11
]; see Tab. 1 in Electronic Supplementary Material) on which they would like to spend a fictional research budget for new research within their gender group (‘Imagine that you can spend one million euros on more research on risk factors in your gender group. On which studies would you spend this money? You can choose a maximum of three’). Respondents could explain their choice in an open-text field (‘If you wish, you can elaborate on your choice of risk factor(s) below’).
As a measure of perceived relevance of new research on their chosen risk factors, we asked respondents to rate the importance of their budget allocation on a scale from 1 to 10 (’How important do you think it is to spend this money on new research on at least one of the three chosen risk factors? With ‘1’, you indicate that you think it is not important at all new research will be initiated, and with ‘10’, you indicate that you think it is very important new research will be initiated’).
We asked the same questions for 19 sex- and gender-related risk factors (Tab. 2 in Electronic Supplementary Material). These were characteristics, behaviours or conditions that were not considered conventional, such as hypertension and stress [11
], and that were chosen based on previous research [3
Finally, to identify any other factors that might be related to CVD risk according to the respondents, they could suggest additional factors in an open-text field. All questions were closed-ended, except the 3 open-text field questions.
Data and statistical analyses
The survey was anonymous and we did not track whether invitees filled in the form. Since it was not possible to check whether respondents who did not complete the survey filled in a complete second version at a later time, we used questionnaires with fully completed closed-ended questions in our data analyses (n = 980).
We analysed sociodemographic data and history of CVD descriptively, depicting them as percentage and mean ± standard deviation (SD). Gender group was based on the individual’s sex at birth combined with self-reported gender identity, and classified into cisgender men (those identifying as men or who were assigned the male sex at birth), cisgender women (those identifying as women or who were assigned the female sex at birth) and other gender groups. Ethnicity was defined using the validated country of birth indicator, by combining the individual’s country of birth with the parental countries of birth. We distinguished between a Dutch, non-Western migration and Western migration background [21
]. Educational level was classified into high level (higher vocational schooling or university) and other.
All data on conventional and sex- and gender-related risk factors were descriptively analysed as mean ± SD, stratified by gender group. All analyses were performed in IBM SPSS Statistics 25. The open-text field questions were summarised, but not further analysed.
Post hoc sample
After the first survey, we distributed a shortened version of it to a convenience sample of 52 women with a non-Western migration background, who were recruited at local markets in Amsterdam, the Netherlands. We studied this group of women specifically because previous research has shown that they are a potentially overlooked high-risk group for CVD [23
]. Using this post hoc sample, we were also able to include a larger and more heterogeneous group of respondents with a non-Western migration background.