To our knowledge, this is the largest study reported to date of germline genetic testing using a digital pathway in cancer patients. We have demonstrated logistical feasibility and acceptability of a patient-centred, digital pathway for mainstreaming of BRCA-testing for unselected female BC patients. Via randomised comparison, we also showed non-inferiority against a range of outcomes of digital delivery of pre-test information compared with pre-test genetic counselling consultation.

We observed attrition upstream of participants (i) consenting to the research study and (ii) receiving the pre-test information (19.3% of those initially expressing interest, Figure 2). At both stages, attrition may have partly been a consequence of the demands of the research study through which the BRCA-testing was being delivered. For example, there was need for both research consent as well as standard genetic testing consent, which in some cases proved confusing to patients. Furthermore, being a research study placed additional time requirements on participants for completion of questionnaires. However, this attrition may also reflect barriers to inclusion within the study or progression through the pathway; there was limited opportunity for collection of data within this research study regarding non-participation.

The demographics of participating patients are broadly reflective of BC patients nationally, albeit slightly younger in average age of diagnosis (59 compared with 62 years old). However, there plausibly may be socio-demographic differences in characteristics between all patients presenting in clinic for BC management during the study window and those taking up BRCA-testing. These may reflect biases relating to those offered the study, those expressing interest following offer, and those progressing digitally to the point of test uptake (genetic test consent). Such factors are relevant to the generalisability of the findings; similar factors may well pertain to uptake of BRCA-testing through existing standard pathways and warrant further exploration to improve equity of access to BRCA-testing more generally. However, these biases should not be differential between the two arms of the study, and as such, for those who do opt for testing via a digital pathway, the observation of superior test uptake in the fully-digital arm is of note.

Indeed across all outcomes, our randomised comparison demonstrated non-inferiority of digital delivery of pre-test information compared with a pre-test genetic counselling consultation. Though it is important to consider that our comparator arm was a “partially-digital” pathway, limited to comparison of just the pre-test information, rather than a true standard-of-care arm for BRCA-testing. Nonetheless, our findings relating to equivalence of patient-reported anxiety and satisfaction were promising. Additionally, the relatively low demand for clinical support via the genetic counselling hotline similarly indicate low proportions with substantial anxiety or need for extra support throughout the process.

The genetic counselling hotline was intended to provide optional access to a GN/GC to supplement digital pre-test information. Thus, patient contact with a GN/GC is not necessitated within the proposed pathway but rather allows patient-centred access to support if required. This differs from previous studies in cancer patients, which have largely focussed on implementing digital information or tools in addition (or prior) to genetic counselling (18, 19).

Two exceptions to this, with similar findings, are a US study, reported by Swisher et al., that reported noninferiority of patient distress at three months and no statistically significant differences in anxiety, depression, or decisional regret between participants who did not have mandatory pre-test counselling (received information in the format of a video) compared to those who had pre-test genetic counselling (20). This study was outside of mainstream settings, for participants recruited from the public with a personal or family history of breast or ovarian cancer.

Another study reported by Sie et al. (2013), gave 161 selected BC patients referred to clinical genetics services within the Netherlands the option to proceed with testing via a fully digital route, as alternative to conventional consultation (there was no genetic counselling hotline in this study). Sai et al. also reported high uptake in the digital pathway, that more patients preferred testing without prior face-to-face counselling, and similar outcomes in relation to distress (anxiety) and satisfaction between the digital arm and conventional clinical genetics consultation arm (21).

Within our study, we also compared knowledge between the two arms and our findings reveal that, whilst lower, mean patient knowledge scores in the fully-digital arm lay within the pre-set margin of non-inferiority. Therefore, the standardised material we produced with our patient involvement group can be deemed suitable and sufficient for providing pre-test information generating acceptable knowledge levels, as compared to a GC/GN consult. However, a priority area for future exploration is improving inclusivity and accessibility of digital materials to meet diverse patient needs and to ensure that we are not exacerbating health inequalities. For example, for different learning styles, reading levels, or languages, and accessible content for those with hearing or sight loss, which could be achieved via more interactive tools, use of audio-visual content, or physical paper-based information as an adjunct.

Previous studies of digital pathways have largely focused on community-based ascertainment for genetic testing, for example BRCA founder mutation testing for those with Jewish ancestry. Our study is the first focussed on using a digital pathway for patients under active oncology management within UK NHS. At both sites involved in the study, HCP satisfaction with the pathway was overall high, with expression of readiness for broader rollout with number of benefits identified in previous survey. Our clinician satisfaction data may reflect bias regarding those clinicians electing to participate in the survey; furthermore, the two trusts involved in the study had already established mainstreamed pathways for BC GT and thus their clinician population may not be reflective nationally. Improvements in HCP feedback relating to turnaround times were observed compared to the pilot study, likely through improvements in laboratory test turnaround times (which improved over the study period following COVID19-related laboratory delays and optimisation of copy number variant analysis for saliva-derived DNA).

In terms of the patient population, we offered GT to all women attending for BC diagnosis, management, or follow-up. Under current UK guidelines, fewer than 20% of women with BC are currently eligible for BRCA-testing (22). UK health economic analysis undertaken in 2018 suggest universal testing of BC patients to be economically impactful within NICE willingness to pay thresholds of £30,000/QALY up to a per-patient cost of £1,626 (payer perspective) and £1,868 (societal perspective) (23). Of note, recent guidance from the American Society of Clinical Oncology (ASCO) has recommended testing for BRCA1 and BRCA2 in all BC cases up to the age of 65, including those diagnosed historically(24). However, detection of PVs will be higher in cases with younger onset, higher-grade, bilateral and/or hormone-receptor negative disease, and where a relevant family history is present (22). There is an inherent tension in balancing enhancement of detection rate against the complexity and imperfect sensitivity incurred from test eligibility criteria.

Nevertheless, any substantial expansion in BRCA-testing of BC cases, for example testing of all cases arising age ≤65 years, would require higher-throughput clinical and laboratory systems. We propose that such systems should (i) incorporate more generic patient information materials and clinical workflows to enable scale, (ii) retain availability where required of expert clinical input individualised to patient requirements and (iii) attune flexibly to local clinical pathways and informatic workflows, especially as GT timed at point of diagnosis requires efficient timely delivery. Additional work is required to understand necessary facilitators and adaptations to optimise equity of access; pathway-specific health economic analyses will also be important.

Overall, this study demonstrates that, supported by a genetic counselling hotline, a fully-digital pathway may be a suitable alternative to conventional models of pre-test information-giving, sampling and consenting, and could support end-to-end management of genetic testing for a large proportion of BC patients. Where expansion of germline genetic testing is limited by clinical capacity, pathways such as BRCA-DIRECT, implemented within mainstream oncology clinics may offer a viable and acceptable approach for the majority of patients to minimise the patient-facing and administrative burdens to clinicians of BRCA-testing, whilst providing flexible patient access to clinical genetics expertise. Thus enabling both clinical genetics and oncology professional time to be utilised more effectively for management of those with a positive result or supporting a minority of patients for whom a standardised, digital approach is unsuitable.