What would I do? Perspectives on the factors underlying Lynch syndrome genetic testing and results sharing decisions for high‐risk colorectal cancer patients

Abstract Objective Universal tumour testing for Lynch syndrome (LS) in all incident colorectal cancers (CRCs) and sequential diagnostic genetic testing is cost‐effective in Australia. Because of this, our study aimed to understand factors underlying possible decisions faced by tumour test‐positive CRC patients and their at‐risk relatives throughout the LS diagnosis pathway. Methods Semi‐structured telephone interviews were conducted with 23 participants, using four hypothetical scenarios. Vignette‐guided closed‐ and open‐ended questions asked about LS genetic testing uptake, discussing diagnosis with at‐risk relatives, and risk‐reducing interventions. Personal perspectives on genetic testing were collected pre‐post vignette discussion. Inductive thematic analysis was performed on open‐ended questions. Decisional pathway diagrams were developed to convey factors influencing complex decision‐making processes. Results Participant responses incorporated unfolding scenario information, resulting in three decision themes: (1) wanting to know one's LS status; (2) informing family about LS; (3) navigating risk‐reducing interventions. Across all themes, ‘knowledge’ emerged as a facilitator, and ‘negative emotional experience’ as a barrier. Personal supportive views toward genetic testing increased post‐interview. Conclusions When communicating with tumour test‐positive CRC patients or their relatives about LS genetic testing, providing guidance/resources to inform decisions around risk‐reducing interventions and informing family members is critical. Scenario‐driven interviews provide insight into what individuals might do when facing complex healthcare decisions and could aid informed decision‐making. This approach may be applicable in other conditions, particularly with mainstreaming being increasingly introduced into the genetic context.


| INTRODUCTION
Lynch syndrome (LS) is a hereditary cancer-susceptibility disorder associated with an increased risk of developing a range of cancers, particularly colorectal cancer (CRC). 1 Universal tumour testing in all incident CRC cases to identify tumours with DNA mismatch deficiencies (dMMR) followed by sequential diagnostic germline genetic testing to confirm LS is cost-effective in Australia and other developed countries. 2,3 It's cost-effectiveness improves if more mutation carriers (i.e. probands) undertake diagnostic genetic testing and communicate results with at-risk family members, so family cascade testing rates increase. 2 This approach is likely to be more systematically implemented in Australia. 4 Universal LS genetic testing in all CRC cases is not currently cost-effective, but may eventually become cost-effective should costs decrease.
CRC patients with dMMR tumours have a high chance (up to 67%) of carrying an underlying germline LS mutation. 5,6 If patients are found to be probands, their at-risk relatives can undergo predictive genetic testing to clarify their own cancer risks 7,8 (further details in Supporting Information). Even if a germline LS mutation is not detected despite dMMR ("Lynch-like syndrome"), patients may still have elevated cancer risks and may benefit from personalised surveillance plans. 9 However, Australian evidence has found that only around half of tumour-test positive cases choose to have genetic testing. 10 Therefore, it is important to understand factors affecting decision-making processes regarding LS genetic testing for individuals diagnosed with tumour-test positive CRC, and if LS confirmed, their at-risk relatives.
These individuals must consider possible consequences, including increased cancer risks to themselves and at-risk relatives (e.g., children, siblings), risk-reducing interventions (e.g., colonoscopic surveillance, surgery), and life insurance implications. [11][12][13] To provide effective decision-making support to affected individuals, healthcare providers must understand facilitators and barriers associated with these decision-making processes. To date, Australian studies have restricted decision-making research to those already commencing LS testing and/or are confirmed carriers. 10,[14][15][16][17][18] CRC is the second most common cancer in Australia, 19 and universal genetic testing for LS on CRC may eventually become cost-effective. 2 Therefore understanding perspectives around genetic testing and subsequent decisionmaking processes is important.
This study aimed to understand perspectives towards LS genetic testing for newly diagnosed tumour-positive CRC patients and relevant events following LS diagnosis. Specific aims were to identify and explore: (i) key considerations and decision-making processes when facing hypothetical scenarios about LS testing; and (ii) facilitators and barriers associated with key considerations identified.  (Table S1). May 2018 were invited to participate in a telephone interview. Those interested were given a study verbal outline and consent form.

| Participant recruitment
Consenting participants were given an LS factsheet and interview vignettes (further details in Supporting Information). Inclusion criteria were: no current or previous cancer diagnoses, aged 18 years or over, and ability to communicate in spoken English.

| Interview schedule development and data collection
Vignettes about a family affected by LS were developed through consulting representatives from consumer organisation Lynch Syndrome Australia (LSA). Vignettes follow a proband 'James' and sister 'Sarah' throughout stages of the LS diagnostic pathway. Scenarios included LS genetic testing uptake following tumour test-positive CRC diagnosis, discussing subsequent LS diagnosis with at-risk relatives, considering risk-reducing interventions, and involving healthcare professional support. The interview schedule was piloted and refined.
In one-on-one telephone calls (conducted by Yoon-Jung Kang, April Morrow, and Victoria Freeman), participants were first asked for personal opinions about genetic testing and given a verbal summary of the LS factsheet. They were presented with the four-part evolving scenario, each followed by closed-and open-ended questions, and asked to answer from the protagonist's perspective. At conclusion, participants were asked if personal views about genetic testing had changed. Interviews were audio recorded, transcribed 588 -TIERNAN ET AL. verbatim, and checked for accuracy (Victoria Freeman, Gabriella Tiernan).

| Quantitative analysis
Responses to five close-ended questions were independently categorised (Victoria Freeman, Gabriella Tiernan) then discussed to achieve consistency.

| Thematic analysis
All transcripts were de-identified and imported into NVivo 12 Plus software (QSR International Pty Ltd 2018). An inductive thematic analytic approach was taken by two researchers (Gabriella Tiernan, Victoria Freeman). 21 Transcripts were initially reviewed independently without coding and preliminary notes taken. A blinded iterative complete coding approach was taken. 22 Coding structures were shared, discrepancies discussed, and a common structure was formed.
This was applied to re-code all transcripts. Similar codes were grouped to develop and refine key themes and subthemes based on quotation salience and volume. A third independent reviewer resolved discordances (Natalie Taylor). Decision pathway diagrams were developed to convey factors influencing decision-making processes around key themes. During analysis write-up, key themes or subthemes were further refined (Gabriella Tiernan, Victoria Freeman) with input from the research team (Natalie Taylor, Yoon-Jung Kang, April Morrow).
Those initially conditionally in favour [that is, provided specified conditions were met] (n = 4), and neutral, open-minded or unsure (n = 4) towards genetic testing, became in favour by conclusion.
One participant remained unsure and two participants initially favouring genetic testing changed their views to conditionally in favour of, or unsure about testing. Of the 23 participants asked at interview conclusion, 91% (n = 21) were in favour of genetic testing.

| Theme
Working through the vignettes, participants felt wanting to know of one's LS status was a key factor in this process.

| Initial considerations
Personal circumstance, personal beliefs, and knowledge levels about LS and genetic testing were initially raised and may immediately influence the decision to know of one's LS status ( Figure 1A).

| Positive reinforcers and enablers
Knowledge was perceived as a positive reinforcer in deciding to uptake testing as well as enabler which may overcome barriers to testing (e.g., fear and mistrust of procedure) (A.Fr.4). -591 James and Sarah undertaking genetic testing and being informed of their LS status (Table S3; Figure 1A).

| Theme
Participants felt the decision to inform at-risk family members about LS was a key consideration in the LS diagnosis scenario.

| Initial considerations
Family reactions, degree of closeness to family members, age of family members, and timing of genetic information disclosure were immediately considered, "I don't think it would be necessary to tell her kids before getting the test…best to get the test and if it comes back positive then tell her kids and any other immediate family…" CP48 ( Figure 1B). If a proband decided against informing at-risk family, some participants felt it was the healthcare professional's responsibility to disclose their results to family (enabler).

| Negative reinforcers
Patient privacy was a negative reinforcer where some participants felt there was no justification to disclosing LS positive results to family members without the proband's consent, "I don't see why automatically the doctor should be able to say to the family member 'Hey, James has got Lynch syndrome' as opposed to Ebola or something that's highly infectious" P42.  (Table S3). Strong family history of cancer, potential to prevent cancer in family members and believing family members need to know of the proband's LS diagnosis informed this view, "it affects more than me" P1. If a proband decided not to share genetic results with atrisk family members, approximately half (n = 11/23) (Table S3)

| Decisional outcomes
Generally, participants believed at-risk family members should be informed of the proband's LS diagnosis.

| Theme
Participants felt navigating decisions related to risk-reducing interventions were important throughout the siblings' LS diagnosis scenario.

| Positive reinforcers and enablers
The knowledge that these risk-reducing interventions were the best available options to approaching a LS diagnosis was viewed as an enabler and positive reinforcer: "if they're at high risk and they've been told that that's the best…way to do it… [that] would be, a pretty good motivating factor." CP45 (C.Fr.5.Cs), (C.Fr.6. RRS).

| Negative reinforcers
One participant did not answer in favour of annual surveillance and expressed this may be too frequent given the invasiveness of the procedure (C.Br.7.Cs). Another participant expressed they may not uptake these risk-reducing interventions if other options were available.
Only one participant decided against prophylactic surgery due to negative physical implications (C.Br.8.Cs).

| Subthemes
Cancer prevention was a key factor influencing the decision to-  (Table S3).

| DISCUSSION
We found that individuals perceive wanting to know one's LS status, Recent public funding changes for LS genetic testing in Australia 4 may progressively shift clinical responsibilities for diagnosing probands from genetic specialist services to oncologists or treating clinicians (known as 'mainstreaming' of genetic counselling and testing). 26 Given challenges reported by non-genetic counsellor healthcare professionals in delivering genetic counselling 27 and encouraging probands to share information with relatives, 28 these scenarios have potential to be further developed as a tool to support clinicians in enabling informed decisionmaking.
Key considerations around LS genetic testing identified in this study align with findings from prior studies, including with familial cancer clinic attendees 29 or those recently diagnosed with CRC. 30 In our study, knowledge emerged as a fundamental facilitator in decision-making processes across all three themes. Knowledge acted as a facilitator to aid decision-making and reduce perceived negative enforcers or barriers around genetic testing. TIERNAN ET AL.
We also found negative emotional experience (e.g., fear and anxiety) was a key barrier across all themes. Participants expressed concerns about financial implications associated with genetic testing and subsequent risk-reducing interventions and insurance implications. These concerns are well documented. 17,31,32 Our study occurred before protective genetic discrimination regulations were introduced. Public funding was also not available for LS testing at time of interview. Consistent with our findings, a previous study reported misconceptions among both patients and providers about potential high out-of-pocket costs associated with genetic testing as barrier to LS genetic referral. 33 Whilst insurance reforms and new public funding may alleviate financial concerns, further efforts may be needed to ensure providers are aware of changes and that patients receive appropriate information to make informed choices about genetic referral and testing.
Most participants supported recommended risk-reducing interventions, but were also concerned about potential associated sideeffects and discomforts. A recent Australian nation-wide audit of risk management in LS carriers across 12 familial cancer clinics reported main reasons for clinical recommendation non-adherence, which included feeling uncomfortable and/or disruption associated with colonoscopy, and, for female LS carriers, younger age and family incompleteness. 34 Vignettes described a male proband and female sibling. It would be interesting to explore whether switching genders would elicit alternative perspectives, given gender has been found to influence communication around LS. 35 We recruited participants without cancer history. Given universal tumour testing for LS in all incident CRC cases is likely to occur in Australia, it is important to understand these individual's perspectives on LS genetic testing. However, Australian research has focused on high-risk populations (individuals with dMMR tumours or confirmed LS) and is unlikely to capture testing pathways relating to systematic LS testing in all CRC cases. 10,[14][15][16][17][18] Therefore, this study provides insight into potentially effective and feasible approaches to facilitating decision-making regarding LS genetic testing and related events, when systematic LS testing is introduced. Decision-making processes are likely not limited to LS, therefore it is possible for these findings to be expanded across other hereditary conditions with similar potential actionability. We have developed a microsimulation modelling platform 'Policy1-Lynch' to evaluate the predicted impact, benefits, harms and cost-effectiveness of various testing and risk management strategies for LS. 2 Results from this study can be used as a range of parameter inputs to 'Policy1-Lynch' to investigate impacts of likely testing uptake and test result sharing rates on the cost-effectiveness of systematic LS testing.

| Study limitations
This study did not specifically investigate decisions of patients with an unconfirmed MSI status CRC diagnosis. Should Australia adopt universal genetic testing, 4 our findings may still be relevant to these patients.
Lack of participant demographic information limited sub-group analysis; however, our preliminary findings can be used as a foundation for future investigation towards specific differences in views between demographic groups within a larger sample. Most participants (18/23) were recruited from a cancer fundraising event and a suboptimal response rate led to additional convenience sampling. Therefore, views may not represent the broader population due to small sample size as well as potential favourable attitudes towards cancer related issues. Despite this, a variety of views towards genetic testing and vignettes were expressed. To account for potential positivity bias to influence convenience sample responses, we also checked the thematic integrity without convenience sample data and found no compromise.
Interviews examined anticipated behaviours and intentions, but evidence for the extent to which vignette-driven interviews can predict future real-life healthcare decisions varies. For example, in breast cancer genetic testing, differences have emerged between hypothetical scenario and real-life decision-making. 36 By contrast, in-depth consideration involved with hypothetical scenarios 37 and use of effective methodologies within vignette-driven interviews can illustrate predictive relationships between hypothetical responses and future decisions. 38 Valuable insights can also be gained by understanding motivations and thought-processes underlying LS testing decisions when framed hypothetically. 39 This study did not control for social-desirability bias, that is, participants modifying responses according to perceived interviewer expectations. 40 Future investigations may benefit from specific measures to minimize social-desirability bias (e.g. intentional rapport building). 41

| Clinical implications
This study highlights challenges faced by individuals moving through the LS diagnostic pathway, and key factors impacting decisionmaking about genetic testing, uptake of risk-reducing measures, and information dissemination to at-risk relatives. Results emphasise the healthcare professional's role in communicating with patients about LS and facilitating information dissemination to at-risk relatives. Given attitude changes pre-post interviews, these vignettes may have potential as interventions (targeting both healthcare professionals and patients) to support informed decision-making towards LS testing, and to aid implementation of mainstreaming genetic testing.

DATA AVAILABILITY STATEMENT
The data supporting the findings of this study are available in the supplementary material of this article.