Ethical Aspects of Genotype Disclosure: Perceptions of Participants in a Nutrigenetic Study in Finland

The ethical issues surrounding disclosure of genotype information following genetic testing of an individual have been widely discussed in the literature [1-3]. The heritability and the permanence of genotype information can be considered as the major reasons behind these ethical issues. Due to the heritability, information such as genetic susceptibility to diseases may be relevant also to those who are genetically related to the tested individual. Following genetic testing, fear of passing on disease-associated genetic polymorphisms to the next generation may affect a tested individual’s reproductive decisions. Feelings of guilt may result from knowing that one’s child may have inherited a disease-associated genetic polymorphism. Due to the permanence of genotype information, anxiety or depression may result in a healthy individual from discovering his/her genetic predisposition to diseases. Improper disclosure of an individual’s genetic predisposition to diseases may cause discrimination at workplace or in obtaining health insurance [4]. Despite these issues, however, use of genotype information has already proved to be useful in diagnosing, preventing, and treating certain diseases and shows promise in further advancement of healthcare [5, 6]. In the post-genomic era, use of genotype information in biomedical research has become unavoidable, thereby necessitating discussions on ethically sound ways of disclosing genotype information to the participants. Under these circumstances, understanding the perceptions of the research participants regarding genetic information can be very valuable.

Nutrigenetics is one of the fields using genotype information, with potential for prevention of several complex diseases (e.g., cardiovascular diseases [CVDs] and cancers) [7]. Nutrigenetics deals with understanding how genetic variation affects dietary responses and thereby human health [8]. The APOE gene, which has 3 major alleles, ε2, ε3, and ε4, is involved in cholesterol metabolism through its product, apolipoprotein E. Of the 3 alleles, APOE ε4 has been found to be strongly associated (homozygous stronger than heterozygous) with the incidence of CVDs and late-onset Alzheimer’s disease (AD) [9-12]. Presence of APOE ε4, however, is not an absolute necessity to develop these diseases, nor does its presence confirm the development of the diseases in the lifetime of an individual [13]. The disease phenotype can also be modified by lifestyle changes, especially cessation of smoking, reduction of fat and increase of fruits and vegetables in diet, and increase of physical exercise [14]. Among the carriers of various allelic forms of APOE, those with ε4 have also been shown to benefit the most from these lifestyle changes [15, 16]. The frequency of APOE ε4 has been reported to be considerably higher in the Finnish population than in most other populations [17, 18].

Research participants are one of the most important stakeholders in biomedical research. According to the Declaration of Helsinki [19], well-being of a participant comes at the forefront of an ethically well-conducted research and is to be prioritized over social or scientific advancement. Adequate provision of information and professional support in the form of genetic counselling are therefore important ethical considerations in research where genotype is disclosed [20]. Obtaining participants’ views on the concerns, and other issues discussed in the literature, also empowers them to collaborate in the ongoing ethical discussions [1]. In the literature, the psychological effects of APOE genotype disclosure in healthy participants have been reported by some studies using self-report inventories or self-developed questionnaires [21, 22]. However, there is a lack of studies where perception of participants of a nutrigenetic study (with an opportunity to receive their APOE allele status) has been studied with respect to the ethical aspects of genotype disclosure.

In this study, we have developed a close-ended questionnaire to gain views of the participants in a nutrigenetic study in Finland, in which the effect of disclosing the APOE allele status on lifestyle changes was pursued. Within the nutrigenetic study, the participants were provided information on study-related genetics and basics of genetics in written articles and through lectures. In addition, they had access to genetic counselling and multiple discussions with relevant professionals. As provision of information is regarded as an important ethical requirement in genotype disclosure after susceptibility testing [20], assessment of study-related knowledge was also carried out through the questionnaire.

“Effects of ApoE4 genotype information and intervention intensity on the fulfilment of lifestyle changes and sensory preferences” (ApoE4mot) is a collaborative project between the University of Turku, the University of Eastern Finland, the Seinäjoki Central Hospital, and the Natural Resources Institute, Finland (Luke). The ApoE4mot project in South Ostrobothnia, Finland, includes a nutrigenetic intervention study, which aimed at assessing the effects of disclosing genetic risk information (the APOE allele status) on lifestyle changes (Fig. 1). This is a follow-up to a previous 2010–2011 intervention study “Healthier food choices, tailored models for eating and exercises” (TERVAS) [21]. Participants from the TERVAS study were also invited to take part in the ApoE4mot project. Recruitment started at the beginning of the year 2017. Invitation for participation was sent through personal letters to the TERVAS participants and through recruiting advertisements in local newspapers, notice boards, and internal communication channels of some major employers in Seinäjoki (principal town in the region). Those who were interested to participate were sent an information letter containing the study protocol and an overview of the APOE gene. The inclusion criteria have been mentioned in detail in another paper about ApoE4mot project [23].

A compulsory lecture on genetics and an opportunity for genetic counselling by a clinical geneticist were arranged for all those who expressed interest to continue. A period of 2 weeks was given to reconsider the decision to participate before consenting (during which 10 people originally interested decided against participation). After this period, the signed informed consent form was returned by the consenting participants at the first contact meeting. Those who returned the signed informed consent form underwent genetic testing for the APOE gene and other health measurements. The participants were interviewed about their medical history, medication and dietary supplement use, exercise habits, smoking, and alcohol consumption. They were also asked to explicitly state whether they wanted to be informed about their APOE genotype. Following this first (initial measurement) meeting, the participants completed a secure online survey (Webropol). Of the 70 TERVAS participants who volunteered to participate, 61 completed the close-ended questionnaire on ethics, and of the 211 newly recruited participants, 189 did the same (n = 250). The newly recruited participants were then randomized into the intervention group (who received their APOE allele status at the beginning of the study) and the control group (who had the option to know their APOE allele status at the end of the study or opted not to know). The TERVAS participants knew their APOE allele status already. The total duration of the ApoE4mot project was a year and a half. At the conclusion, the participants in the control group were given the opportunity to receive their APOE genotype.

The purpose of the close-ended questionnaire was to understand the perceptions of the participants on the ethical issues discussed in the literature related to genotype disclosure (Tables 1, 2). Another purpose of the questionnaire was to probe the participants’ knowledge on basic and study-related genetics, which also reflects to some extent the effectivity of the compulsory genetics lecture they were given before they decided to sign their informed consent forms (Table 1).

The questionnaire had altogether 4 major categories of questions (Table 2). The questions were mainly based on the literature with modifications and additions made by the research group members. Thorough discussions among the research group members, who are also experts in the related fields, led to the finalizing of the questionnaire following several steps of modifications. Except for the demographic questions, the answering options provided for each question were “agree,” “do not know,” and “disagree.” As this questionnaire was part of a large set of questionnaires related to the nutrigenetic study, a short response scale was preferred, as it offers more clarity to the responders compared to a 5-point or a longer scale [33] and also enables a quicker response [34]. A neutral option of “do not know” was provided to avoid forcing the participants to choose either a positive or a negative option.

This questionnaire formed a part of a set of questionnaires related to the ApoE4mot nutrigenetic study. The entire set of questionnaires was provided as a Webropol online survey, the web link of which was sent to each of the participants by email. The survey was provided after the informed consent form was signed. The exact time and place to fill the questionnaires were at the participants’ discretion. The filled questionnaires bore only the participants’ research number but no personal information to protect their privacy. In case of the few who did not have access to email, the questionnaires were sent and received by regular mail and coded afterwards.

The demographic data were analysed by using means and frequencies. Frequencies were calculated for the responses of all the questions. The responses from the knowledge assessment and the concern categories were transformed into knowledge and concern scales, respectively. For the knowledge scale, the responses were first transformed to “correct,” “incorrect,” and “do not know” and scored accordingly (Table 3). The questions (2 of the 15 questions) from the knowledge assessment category with >99% correct responses were removed from score calculations. For concern scale, higher scores indicated more concern (Table 3). The questions in the category “opinions on the processes of disclosure of information” were not used for rating scale, as this category comprised few questions with a wide range of constructs related to the different aspects of the topic (Table 1). The responses from these questions have been treated as individual categorical items and have been analysed and presented as such.

The relationships between the scores and the responses from the demographic and other categorical variables were assessed by Mann-Whitney test for the dichotomous variables and by Kruskal-Wallis test for variables with >2 categories. The statistical associations between the categorical variables were determined using χ² test of independence. The relationship between the continuous variables (age, knowledge, and concern scores) was assessed using Pearson product-moment correlation coefficient. For all the statistical calculations, IBM SPSS version 25 software was used.

The internal consistency of a scale reflects the homogeneity of the underlying constructs between the questions of the scale [35, 36]. Cronbach’s alpha was used for measuring the internal consistency of the knowledge assessment and concern scales. Cronbach’s alpha for the knowledge assessment scale was calculated to be 0.59 and for the concern scale was found to be 0.58. For calculating Cronbach’s alpha of the concern scale, the negatively worded items were reversed before the calculation. Although an alpha value of ≥0.7 has been mentioned in the literature as desirable, strong arguments against such threshold has also been put forward by many authors (see [36] and the included references). Apart from the similarity of the underlying constructs, the value of Cronbach’s alpha depends on factors such as number of questions and range of response scale (i.e., number of response options) as well as on the sample of participants on whom the questionnaire has been tested [35, 36]. We considered the alpha values obtained for our knowledge assessment and concern scales acceptable for this study, as the number of items in the scales was limited, and we probed different concepts in the knowledge scale and also different aspects of concerns related to participation in the concern scale (Table 1).

Ethical approval for the ApoE4mot project was gained from the Regional Ethics Committee of Tampere University Hospital on December 14, 2016 (decision number 12/2016).

Of the 281 participants of the ApoE4mot project, 250 (89%) answered the close-ended questionnaire. Of these 250 participants, 61 were from the TERVAS study (referred to as “old” participants) and 189 were recruited newly for this study (referred to as “new” participants). All the participants (n = 250) were native Finnish citizens with ages ranging between 27 and 72 years (mean 51.7 years). A small, negative correlation was observed between the age and the knowledge score (r = −0.122, p = 0.054). No statistically significant association was observed between age and the concern score. Majority of the participants (80%) were female. No significant association was observed between the gender and the knowledge or concern scores. All the participants had received at least basic education equivalent to primary and lower secondary education (Table 4). The primary and lower secondary groups recorded a statistically significantly higher knowledge score (Md = 13.0; Kruskal-Wallis H = 19.581, p = 0.001) compared to the other groups (with median scores ranging between 11 and 12). However, no correlation was observed between the level of education and knowledge scores when the 5 groups were considered together. Participants living with children younger than 15 years had higher concern scores although not statistically significant (p = 0.07). The “old” participants had higher knowledge scores (not statistically significant, p = 0.085) than the “new” participants. No other significant differences were observed between the response patterns of the “old” and the “new” participants.

The knowledge of genetics and the genetic basis of AD and CVD development were assessed through a set of 13 questions (Fig. 2). The knowledge scores obtained ranged between 2 and 13, with mean ± SD = 11 ± 1.7 (obtainable range: 0–13). All the questions were correctly answered by 20.8%; 12 out of 13 questions were correctly answered by 26.8%. The lowest score of 2 was obtained by 0.4%. The question with the most incorrect and “do not know” responses was “It is possible to have an AD gene but no symptoms of AD,” where 41.6% answered correctly, 38.8% answered incorrectly, and 19.6% did not know. To the rest of the questions, at least 70% gave correct responses (Fig. 2). We observed statistically significant differences in the responses between the 4 pairs of the same questions, one related to AD and the other related to CVD (Table 5). The CVD-related questions had higher proportions of correct answers compared to the AD-related questions. It was also observed that those who gave the correct answer to “It is possible to have an AD gene but no symptoms of AD” and those who gave correct answer to “It is possible to have a CVD gene but no symptoms of CVD” had statistically significantly higher knowledge scores (p = 0.00 for both).

The participants’ concerns regarding participation in research using genetic information were assessed through 7 questions. The concern scores of the participants ranged between 7 and 17, with mean ± SD = 8.2 ± 1.7, n = 250 (obtainable range: 7–21, with 21 as the maximum concern). Almost half of the participants (50.8%) scored 7, and 17 was scored by 0.8% of the participants.

The responses to 4 of the questions suggested a lack of concern among >90% of participants (Table 6). When asked about concerns related to the hypothetical use of their genetic information in projects without their knowledge, 6% felt concerned. The information of being genetically predisposed to AD and CVD would be stressful, thought 10% of the participants. Almost an equal proportion (10.4%) thought that disclosing information on their genetic susceptibility would be stressful to their family members and relatives. These 2 were found to be the major concerns of the participants. Similarly, high proportions (16 and 24%, respectively) of participants were discovered to respond “do not know” for these 2 questions.

As mentioned earlier, one of the major concerns of the participants was related to the disclosure of their APOE ε4 carrier status to their family members and relatives (Table 6). According to the responses to the questions on their opinion about disclosure (Table 7), majority of the participants (89%) mentioned that they would prefer to share their susceptibility information with their family members themselves, and 87% disagreed with the notion that a doctor should disclose the information. However, less agreement between the responders could be observed with the last 3 questions (Table 7).

A negligible negative correlation was observed between the knowledge and concern scores (r = −0.065, n = 250, p = 0.304). A higher concern score (p = 0.061) was observed for those who responded incorrectly or “do not know” to the question “It is possible to have an AD gene but no symptoms of AD,” which was also the question with the least correct responses in the knowledge category (Fig. 1). Statistically significantly higher concern scores were observed for those who had responded “do not know” compared to those who had responded “disagree” to the following questions from the disclosure category: “I feel the medical doctor/researchers should inform my family members” (p = 0.017), “I feel they (family members) should be asked whether they want to know before they are given the information” (p = 0.042), and “I feel discussing about my genetic susceptibility with a medical professional will be helpful for me” (p = 0.00). No such correlations were observed with knowledge scores for these responses.

Statistically significantly higher concern score was observed in those who had responded “do not know” compared to those who had responded “agree” to the following question from disclosure category: “I feel disclosing health information about my family members and/or relatives for the study is acceptable” (p = 0.003). In addition, a statistically significantly (p = 0.00) higher knowledge score was observed for those who had responded “agree” compared to those who had responded “do not know” to the same question. Among those who had responded “disagree” to the question “I feel discussing about my genetic susceptibility with a medical professional will be helpful for me” in the disclosure category, a statistically significantly higher proportion had responded “disagree” compared to “do not know” also for the following questions in the concern category: “I feel knowing about my genetic susceptibility to AD and CVD will be stressful for me” (p = 0.005) and “I am concerned that disclosing my genetic susceptibility to AD and CVD will be stressful for my family members and relatives” (p = 0.00). These 2 questions had the most “agree” and “do not know” responses in the concern category (Table 6).

The most important finding of this study was the high knowledge on study-related genetics and basics of genetics accompanied with low concerns related to the issues of disclosing genotype information. The high knowledge scores may reflect the effectivity of the study-related information and on basic genetics provided to the participants (both through written information and lectures). Similarly, adequate support in the form of genetic counselling and opportunity to discuss matters with professionals may have reduced concerns in the study participants. This is in accordance with the literature [1, 20], where adequate provision of study-related information and provision of genetic counselling have been strongly promoted as important ethical requirements for disclosing genotype information. Another important finding of this study was the preference of most of the participants to communicate their genetic risk information to their family members themselves over a professional. Disclosing genetic risk information to family members can be a highly sensitive issue and can bring forth a complex mix of emotions affecting personal relationships [37, 38]. However, by preferring to communicate their APOE genetic risk information to the family members themselves, the participants may not only gain emotional support from their family members but considering the purpose of this nutrigenetic study, they can also get support for the desired changes in their lifestyle. This active attitude of the participants makes them more like collaborators in the ApoE4mot project than mere research participants.

Although the overall concern score was quite low for our participants, the main concerns were related to the development of stress in them and their family members on knowing the APOE allele status. This awareness of a possible development of stress may help in preparing them for the disclosure of genotype information and may also aid in communicating their genetic risk information to their family members. As majority of the participants disagreed with feeling concerned on their family members obtaining knowledge on their health information, the results display their positive attitude towards sharing their genetic risk information with their family members. Similar to our finding, Hamilton and co-workers [37] observed in their study that one of the processes undertaken by the participants when disclosing their genetic test results for Huntington’s disease or hereditary breast and ovarian cancer was to consider its effects on themselves and their family members. The ApoE4mot participants also showed little concern about their health information being discovered by their employers and their health insurer, that is, they did not anticipate negative consequences under such circumstances. This attitude may be particular to the Finnish population and their trust in their healthcare system. Finland has a well operating state-provided healthcare system; hence, one does not necessarily need to have an additional health insurance. In addition, Finnish population is more aware and accepting towards the hereditary aspect of diseases and health research in general [39]. A survey conducted in Finland >2 decades ago on the general public and the family members of patients with a hereditary disease, aspartylglucosaminuria, showed a positive attitude towards genetic testing [40]. As discussed by Klitzman [41], genetic identity does not only concern the individual but has social implications as well. The openness of the Finnish society, and the general acceptance of this notion, may also assist in understanding the social context of an individual’s genetic identity. Such an attitude, instead of further complicating the already difficult process of revealing genetic risk information, may help in understanding the situation as part of population health issue. As the equality and non-discrimination policy have been strongly maintained in healthcare availability in Finland, this can also affect the attitude of the Finnish people.

Regarding the knowledge assessment, the most common misconception among the participants in our study was that having an APOE risk allele would invariably result in AD and/or CVD. More participants thought this to be true for AD than for CVD, suggesting a difference in perception between these 2 diseases. This may also reflect the difference in awareness between these 2 diseases, as Finnish population has been aware of the high mortality rate due to cardiac diseases for decades [42]. In the literature, AD has been perceived to be strongly associated with genetic predisposition among lay people and relatives of AD patients [43, 44]. Roberts and Connell [43] have discovered that most of the studied first-degree relatives of AD patients wrongly assumed that most cases of AD are hereditary in nature. According to Leventhal’s “common sense model of illness representation” (discussed, e.g., in [45]), health risk information activates the cognitive representation of threat, which helps in identifying and selecting coping mechanisms to the disease threat. Simultaneously, an emotional response to the threat is also activated, both of which influence the reaction to the risk information [45]. Similarly, Koenig and Silverberg [46] have discussed the “disease prototype,” which is a pre-existing idea about a disease in the individual’s mind and is strongly based on his/her demographic and psychosocial background. It has been suggested that understanding and retaining the disease risk information is dependent on how similar the provided information is with the existing disease concepts in an individual’s mind [45, 46]. As discussed by Marteau and Weinman [45], genetic factors are often considered by general public to have a more deterministic role in disease development than negative health behaviour.

Our study involved a specific population, the native Finnish population, which may not reflect the perceptions more generally. Finland has a high literacy rate, which was also reflected in the demographics of our participants, all of whom had basic education. Due to the existence of thorough health records and the openness towards health research of the Finnish population, it is expected that research using genotype data will only increase in Finland in the coming years [39]. Hence, it is important to know the perceptions of this population regarding the commonly discussed ethical aspects of genetic research in the literature. The global generalizability of the study results is also limited due to the high number of female participants (80%), although we did not observe any significant gender-based differences in the response patterns, as observed by Ashida and co-workers [38] in their study on communication of genetic test results. Our questionnaire was developed based on the literature in combination with extensive discussions among experts in the relevant fields focussing on the study objectives. Keeping in mind the practicality and easiness of implementation, we preferred the close-ended and more structured form of the questionnaire. Also, the close-ended form helped us to know how the views of the participants compared to those discussed in the literature. The questionnaire was implemented for the first time in this study. Based on these findings, we aim to further develop this close-ended questionnaire, which will be eventually validated. We hope that this kind of short close-ended questionnaire can be developed to aid in studies about ethical aspects such as whether the participants’ understanding of the study is adequate and whether the participants’ interests are protected in research disclosing genotype information.

In summary, majority of the ApoE4mot participants answered all the questions in our close-ended questionnaire assessing the perceptions of the participants regarding their knowledge on genetics and the concerns discussed in the literature. The participants were found to have a good study-related and basic genetic knowledge. They did not consider privacy or confidentiality issues related to their health information to be a major concern. Their main concern was the possibility of stress for themselves and their family members due to discovering their APOE allele status. They preferred to communicate their APOE allele status themselves to their family rather than having a professional to disseminate the information. Their active participation in this regard reflected their willingness to act as collaborators in the study. The high level of basic education in Finland and the availability of information and professional support from the beginning of the ApoE4mot project have probably contributed to these outcomes.

We are obliged to Dr. Matti Estola for statistical help. We are thankful to Ms. Heli Karjalainen for her contribution to the study. We also warmly thank all those who participated in the study.

Participants have given their written informed consent. The study protocol has been approved by Regional Ethics Committee of Tampere University Hospital on December 14, 2016 (decision number 12/2016).

The authors have no conflicts of interest to declare.

The University of Eastern Finland, Finnish Cultural Foundation, and Academy of Finland (decision number 292653). They had no role in the design, analysis, or writing of this article.

Formulation of the research and research questions: A.H., R.T., K.V., and A.-M.P. Designing the study: K.V., S.D., A.-M.P., A.H., and R.T. Carrying out the study: S.D., K.V., and M.T. Analysing the data: S.D., K.V., and A.-M.P. Writing the first draft: S.D., K.V., and A.-M.P. Correcting and accepting the final form of the manuscript: S.D., K.V., A.-M.P., M.T., A.H., and R.T.