Online decision support for persons having a genetic predisposition to cancer and their partners during reproductive decision‐making

Abstract A nationwide pretest–posttest study was conducted in all clinical genetic centres in the Netherlands, to evaluate the effects of an online decision aid to support persons who have a genetic predisposition to cancer and their partners in making an informed decision regarding reproductive options. Main outcomes (decisional conflict, knowledge, realistic expectations, level of deliberation, and decision self‐efficacy) were measured before use (T0), immediately after use (T1), and at 2 weeks (T2) after use of the decision aid. Paired sample t tests were used to compute differences between the first and subsequent measurements. T0–T1 and T0–T2 comparisons indicate a significant reduction in mean decisional conflict scores with stronger effects for participants with high baseline decisional conflict. Furthermore, use of the decision aid resulted in increased knowledge levels and improved realistic expectations. Level of deliberation only increased for participants with lower baseline levels of deliberation. Decision self‐efficacy increased for those with low baseline scores, whereas those with high baseline scores showed a reduction at T2. It can be concluded that use of the decision aid resulted in several positive outcomes indicative of informed decision‐making. The decision aid is an appropriate and highly appreciated tool to be used in addition to reproductive counseling.


| INTRODUC TI ON
Most hereditary cancer syndromes follow an autosomal-dominant inheritance pattern, implying that there is a 50% risk of transmitting a pathogenic variant to offspring, with a high risk of a future malignancy as a consequence. For the relatively frequent breast cancer gene mutations in BRCA1 or BRCA2, this implies risks of 27%-57% and 6%-40% of developing breast respectively ovarian cancer by the age of 70 (Brohet et al., 2014;Chen & Parmigiani, 2007). Persons having a genetic predisposition to cancer and their partners have to make fundamental decisions about future reproduction and face difficult challenges (Dekeuwer & Bateman, 2013;Derks-Smeets et al., 2014;Donnelly et al., 2013;van Asperen et al., 2002). Couples have three options to fulfill their wish for a child that is genetically related to both parents.
The first option is natural conception without genetic testing, implying acceptance or taking the risk of passing on the pathogenic variant. Furthermore, there are two options for having a genetically related child to both parents without a pathogenic variant. The first option is natural conception with prenatal diagnosis (PND), offering the choice to terminate the pregnancy if the fetus has the pathogenic variant (de Die-Smulders, de Wert, Liebaers, Tibben, & Evers-Kiebooms, 2013). The second option is preimplantation genetic diagnosis (PGD). PGD offers the option to obtain embryos by in vitro fertilization (IVF) and screen them for the familial pathogenic variant. Only embryos without the pathogenic variant are transferred into the uterus (de Die- Smulders et al., 2013). Levels of awareness for PND (61%) and PGD (66%) are similar, and couples consider PGD (80%) to be more acceptable for hereditary cancer compared to PND (26%) (Gietel-Habets et al., 2017).
Couples may experience difficulties with reproductive decision-making (Dekeuwer & Bateman, 2013;Dommering et al., 2010;Ormondroyd et al., 2012), and it was reported that for some, even years later, the impact of reproductive decision-making still had an influence on their lives at a daily basis (Derks-Smeets et al., 2014). In deliberating the options, couples consider personal values and (dis)advantages of the options, such as physical (e.g., burden of PGD treatment), psychological (e.g., loss of sense of romance), social (e.g., elimination of the pathogenic variant in family line), ethical (e.g., moral duty to protect the child), and practical considerations (e.g., reimbursement of treatment) (Derks-Smeets et al., 2014). Which reproductive option suits them best, should ideally be decided in an informed decision-making process by an educated and empowered couple, supported by a dedicated health care provider. In order to promote informed reproductive decision-making, the use of decision aids can be effective (Derks-Smeets et al., 2014;Juraskova et al., 2014;O'Connor & Jacobsen, 2003;Quinn et al., 2010;Stacey et al., 2017). The present study is part of a larger study on the development and implementation of an online decision aid, developed in accordance with the International Patient Decision Aids Standards (Reumkens, Oudheusden, et al., 2018;Volk, Llewellyn-Thomas, Stacey, & Elwyn, 2013). In this study, we report on the effects of the decision aid evaluated in a nationwide pretest-posttest study in all clinical genetic centres in the Netherlands. Hippel Lindau disease, Li-Fraumeni syndrome, familial atypical multiple mole/melanoma syndrome (FAMMM). Furthermore, couples levels and improved realistic expectations. Level of deliberation only increased for participants with lower baseline levels of deliberation. Decision self-efficacy increased for those with low baseline scores, whereas those with high baseline scores showed a reduction at T2. It can be concluded that use of the decision aid resulted in several positive outcomes indicative of informed decision-making. The decision aid is an appropriate and highly appreciated tool to be used in addition to reproductive counseling.

K E Y W O R D S
counseling, decision aid, genetics, hereditary cancer, informed decision-making, oncology, patient participation, reproductive decision-making needed to have the intention to have children within the next 5 years, and had not yet made a definitive decision regarding their preferred reproductive option. Both partners had to be 18 years or older and both partners needed to have sufficient knowledge of the Dutch language.

| Procedures
Eligible couples were provided with an information brochure including a link to an online registration page. After registration, both partners received an informed consent form by e-mail. After providing online informed consent, both partners were individually directed to an online (baseline) questionnaire (T0). Questionnaires were completed separately by both partners. Subsequently, they received a personal login code for the decision aid. It was allowed to use the decision aid together. Duration of use and page visits were monitored. Immediately after use of the decision aid, participants were directed to the second questionnaire (T1). Two weeks after baseline, participants were asked by e-mail to complete a third questionnaire (T2). A reminder was sent to participants who did not complete the T1 questionnaire within 1 day, or the T2 questionnaire within 7 days. An incentive of 15 euros in vouchers was provided after completion of all questionnaires. This study was approved by the medical ethics committee of Maastricht UMC+ (METC 14-5-089).

| Content of the decision aid
An extensive explanation of the developmental process and the specific content of the decision aid are provided elsewhere . Overall, the decision aid contained: 1. Information about the risk of transmitting the pathogenic variant to offspring and couples' options to have genetically related children.
2. Treatment burden of reproductive options and the chances of different outcomes (e.g., risk of miscarriage after PND) presented in multiple suitable formats using text and videos (e.g., verbal, and population diagrams) (Reumkens, Oudheusden, et al., 2018;Trevena et al., 2013). 7. Information regarding the scientific resources used to underpin the decision aids content, the development team, funding resources, and contact information.

| Instrumentation
Gender, age, educational level, carrier status, disease type, number of children and couples' planning for having children were assessed at T0. Less than primary education, primary and lower secondary education were considered as low education levels. Upper secondary and post-secondary non-tertiary education were considered as middle education levels. Tertiary education was considered as a high education level. At T0 and T2, couples were also asked if they already had a consultation with a healthcare provider. The main subject of this consultation (1 = solely focusing on the reproductive options, 2 = focusing on the consequences of having the pathogenic variant, the reproductive options concerned only a small part) and the profession of the healthcare provider were assessed.
The primary outcome measure, that is, participants' level of having made an effective decision. Each item was scored on a 5point Likert scale ranging from 0 (strongly agree) to 4 (strongly disagree). Total scores ranged from 0 (no decisional conflict) to 100 (extremely high decisional conflict) (O'Connor, 1995a). As the items in the effective decision subscale could not be completed by couples who did not have a preferred reproductive option in mind, a combined score was also calculated for the four other subscales. These 12 items were summed, divided by 12, and multiplied by 25. Total scores ranged from 0 (no decisional conflict) to 100 (extremely high decisional conflict).

Participants' current knowledge of the three reproductive options
(at T0, T1, T2) was assessed by 15 items (Gietel-Habets et al., 2017). Three questions measured participants' knowledge of natural conception without genetic testing (e.g., "When opting for natural conception, there is a 50% risk of transmitting the pathogenic variant to offspring"; 1 = correct, 2 = incorrect, 3 = not sure), five questions measured knowledge of PND (e.g., "PND takes place during pregnancy") and seven questions measured knowledge of PGD (e.g., "IVF is necessary to perform PGD"). One point was provided to each correctly answered question, with a maximum score of 15.
Realistic expectations regarding the reproductive options (T0, T1, T2) were assessed by three questions (i.e., "What is the extra risk of miscarriage due to PND?", "What is the chance of pregnancy after one IVF treatment with PGD?", "What is the risk of complications with PGD?"). These questions contained 8 to 11 answer options.
One point was provided to each correctly answered question, with a maximum score of 3 (O'Connor, 1995b).
Evaluative items (T1). Participants were asked to give an overall appreciation score for the decision aid at a scale from 1-10 and to indicate in open-ended questions positive and negative features and possibilities for improvements. Furthermore, 10 items (e.g., perceived efficiency and active trust) were used to assess user perceptions (Crutzen et al., 2014) including two items of the system usability scale (SUS) (Brooke, 1996). Each item was scored on a 5-point Likert scale ranging from 0 (totally disagree) to 4 (totally agree).

| Data analysis
Data from the baseline characteristics were analyzed by means of descriptive statistics. Cohen's d was used to report effect sizes; Cronbach's alpha was computed to assess reliability. Furthermore, an intra-couple correlation test was performed before evaluating effects. We compared two models to test for intra-couple correlation regarding the main outcome (decisional conflict); one linear mixedeffects model in which clustering within participants over time and within couples was corrected for, and one model without correction for clustering within couples. Both models yielded similar results, and a likelihood-ratio test showed that correction for the clustering of observations within couples did not lead to a better model fit (likelihood ratio = 0.00, p = 1.000). Therefore, all participants were analyzed as independent from each other and therefore we chose to report the simpler model without correction for clustering and used the paired sample t test to compute differences between the first and subsequent measurements. For in-depth analyses, a median split was performed for all main outcome measures. Analyses were performed using IBM spss version 23 and r version 3.3.3. p-values of <0.05 were considered to indicate statistical significance.  the participants, the reproductive options had been the main topic.

| Baseline characteristics
The majority of the participants had heard of PND (73.0%) and PGD (89.6%) before participation in this study and most of them had also received information: on PND: 59.1%, on PGD: 69.6%. A little over half of the participants (51.4%) had a preferred reproductive option in mind at baseline.

| Effects of the decision aid
As shown in Table 2 As shown in Table 2, the mean level of knowledge (range 0-15) significantly increased from 9.28 at baseline, to 13.16 at T1 (ES = −1.37) and 12.63 at T2 (ES = −1.11). In-depth analyses (Table 3) indicated that knowledge scores significantly increased for both participants with high (>10) and low (≤10) baseline knowledge levels.
As shown in Table 2, with a mean score of 23.23 (range 6-30), the level of deliberation was relatively high at baseline and did not show an overall increase over time. However, in-depth analyses ( No effect was found for participants with higher baseline levels of deliberation (>24). Table 2

| Depth of use of the decision aid
As shown in Table 4, both users with low engagement (≤15 pages) and users with high engagement (>15 pages) showed decreased decisional conflict scores, increased knowledge levels, and increased TA B L E 3 In-depth analyses for main outcome measures based on median split baseline scores

| Evaluation of the acceptability
The mean score on the Preparation for Decision Making Scale  Post-hoc analyses indicated that baseline knowledge levels were identical for participants with low and high baseline scores of deci-

| Study limitations
The use of a pretest-posttest design restricts the internal validity, as maturation and history effects as well as effects due to repeated testing cannot be controlled for. Although the execution of meas-

| Research recommendations
The majority of the participants in this study were highly educated (57%). Although this is in line with general characteristics of oncogenetic counselees (Giessen van der, 2017), this number is notably high compared to the numbers in the general Dutch population (30%) (CBS, 2016). This further exposes the need for research on measures to improve referral of patients with a lower educational background. Furthermore, as the reproductive decision is often not implemented within several months after reproductive counseling or after reviewing the decision aid, a long-term follow-up to measure decision adherence (e.g., 18 months after reviewing the decision aid) would be useful.

| Practice implications
Use of the decision aid resulted in several positive outcomes indicative of informed decision-making which may lessen the negative psychological impact of decision-making on couples' daily life and well-being. The decision aid is an appropriate and highly appreciated tool to be used in addition to reproductive counseling.
In-depth analyses showed that the couples who are in the highest need of reproductive decision support are those who are the most supported by the decision aid which increases the overall impact of the decision aid. Currently, we are conducting an explorative implementation study to clarify optimal timing of providing the decision aid and how to incorporate the decision aid in daily practice.
To further increase the impact of the decision aid, the content of the tool will be adapted to other hereditary conditions. Supporting Information Data S1-S3.

| CON CLUS ION
The current findings indicate that the decision aid can be effective in supporting persons having a genetic predisposition to cancer and their partners in making an informed decision regarding reproductive options. Further research is needed to indicate prolonged effects on informed decision-making and informed choice.

ACK N OWLED G EM ENTS
I confirm that the work was conducted to fulfill a degree requirement or as part of training.

| COMPLIAN CE WITH E THIC AL S TANDARDS
All procedures performed in this study were in accordance with the