Pre-meeting surveys to scope the origin and current use of PIL and other resources received low response rates: n=16/104 (15%) for the first meeting (Lynch) and n=23/147 (16%) for the second (haematology).

Results from the Lynch pre-meeting survey showed that 11/16 responders provided a PIL. Nine out of 11 were locally written and curated PIL and 2/9 were created with patient involvement. Fifteen out of 16 responders signposted patients to charities or support organisations, the vast majority to Lynch syndrome UK. In response to a question about what additional resources would be helpful, comments were made about gene-specific risks/management, as well as PIL for different stages of the genetic testing pathway.

Four out of 23 responders to the haematology pre-meeting survey indicated they provided a PIL, and these were locally written/curated. Nine out of 23 responders signposted patients to charities or support organisations. Named charities included MDS UK Patient Support Group, Leukaemia Care UK, Macmillan Cancer Support and Blood Cancer UK. Comments showed a demand for PIL to address somatic versus germline genetic variants, familial implications, predictive testing and gene-specific risks/management.

PIL in current use were added to a shared Google drive by 7/23 regional genetics services in the UK and three specialist genetics service or patient charities. These varied in length, content and format. There was a lack of patient co-design or at least notation of this on the PIL. Outreach to services that were non-responders will be undertaken by working groups overseen by the AGNC, in preparation for future work to develop condition specific PIL.

Over 100 invitations were sent inviting patients and professionals to attend one or both meetings, share with their team and/or suggest relevant stakeholders. Interest in the meetings was universal, but availability to attend and complete the pre-meeting surveys was limited due to time pressures, clinics and other commitments. Three patients and 17 professionals attended both meetings, but only voted once (on day 2). All other participants attended one meeting and voted once in the polls. There were 48/61 engaged with polls in the first meeting and 43/57 in the second.

Recommendations for clinical practice are presented in table 2. Detailed poll results are presented in online supplemental table. Questions were grouped into seven sections/subheadings to address the following topics: diagnostic genetic/genomic testing, patients with a GPV in a cancer susceptibility gene, predictive genetic testing, PIL format, PIL content, risk communication and communicating uncertainty.

Recommendations for clinical practice from the UK Cancer Genetics Group (UKCGG), Cancer Research UK (CRUK) funded CanGene-CanVar programme and the Association of Genetic Nurse Counsellors (AGNC) on co-design of patient information leaflets (PILs) for germline predisposition to cancer

Consensus was reached on all statements when voting across both days was considered. The same statements were presented at both meetings. There were two statements where consensus was not reached on day 2 only, one regarding including links to peer support groups (agree/strongly agree: day 1=87%; day 2=79%+16% neutral/no opinion, online supplemental table, Section 5) and one regarding phrasing subheadings in the form of questions (agree/strongly agree: day 1=84%; day 2=65%+26% neutral/no opinion, online supplemental table, Section 4).

There was some minor revision of the statements agreed in real time on day 2 shown with tracked changes (online supplemental table). There was little opportunity to explain complicated concepts due to the character limit for Slido. Rewording was based on in-meeting feedback and aimed at increasing statement clarity.

A descriptive summary is presented below, under poll topic heading.

Diagnostic genetic/genomic testing

Most genetic testing discussions occurred within clinical genetics services. This may not be representative of the proportion of tests undertaken within clinical genetics versus a mainstream setting but rather could reflect the fact that most meeting participants were from clinical genetics. High-level consensus was reached regarding the offer of a PIL at the time of diagnostic genetic testing. Chat analysis showed that participants did not feel this needed to be extensively detailed, especially since some genetic tests are broad and most patients do not have a GPV identified. A shorter PIL was suggested, which could be replaced by a longer, more specific and detailed PIL if a GPV was identified.

Patients with a GPV

Most genetic test results were delivered by specialist clinical genetics services, with a minority by oncology. Again, this may be representative of participant specialty rather than an overall practice in the UK. High-level consensus was reached regarding the offer of a gene-specific PIL at this stage in the pathway of care. Chat comments suggested it was acceptable for the PIL to be comprised of mostly generic information if it accompanied a personalised clinical letter.

Predictive genetic testing

Most discussions took place within specialist clinical genetics services. High-level consensus was reached regarding the offer of genetic counselling and a PIL at the time of predictive testing.

PIL format

Most people felt that up to two sides of A4 paper should be the maximum length. Chat comments showed that longer PIL, such as The Royal Marsden Beginner’s Guide to Lynch syndrome, could also be useful, but this is rarely printed due to length. High-level consensus was reached on the need to include sections with subheadings. There was verbal and chat discussion about whether PIL subheadings should be presented in the form of questions. Participants felt this could make the PIL appear more personal but could also reduce relevance for some patients, dependent on the topic.

PIL content

Many consensus statements on day 2 were revised live, based on participant feedback. Several referred to inclusion of certain information, such as reproductive risks. Discussion suggested some sections would not be relevant to many patients. Changes are shown in online supplemental table, mostly adding ‘where relevant’ to reflect that it would only be appropriate in specific situations, for example, involving a patient of reproductive age. For GPV in many cancer susceptibility genes, there is insufficient evidence to provide personalised risk estimates. It was felt that healthcare professionals should not overemphasise the possibility of this where data is scarce and there are no management guidelines. Preferences for terminology to describe results from cancer susceptibility gene testing ranked ‘mutation’ below gene alteration, gene change and pathogenic variant, which fits with a general trend away from using mutation in clinical practice due to its potential negative connotations.

Risk communication

Polling questions revealed the importance of showing visual presentations of the chance of getting cancer in the future rather than only describing risk in words. This can be achieved with numbers, pictures and graphics. Discussion highlighted the icon arrays in the NICE patient decision aid for Lynch syndrome: Should I take aspirin to reduce my chance of getting bowel cancer?34 as particularly helpful.

Communicating uncertainty

Consensus statements showed the importance of conveying the origin of uncertainty and ranked showing the range of known risks above other options. In situations where this is not possible, the chat suggested it would be acceptable to convey the amount of uncertainty in words, for example, ‘some uncertainty’ or ‘a lot of uncertainty’.