Healthcare professionals’ perspectives on implementation of universal tumor DNA testing in ovarian cancer patients: multidisciplinary focus groups

In a qualitative study, semi-structured online focus group interviews were held with professionals involved in the Tumor-First workflow—including gynecologic oncologists, pathologists, clinical geneticists, and clinical laboratory geneticists and clinical scientists in molecular pathology. The latter two groups are non-medical doctors specialized in genetic testing and data interpretation and will be referred to as clinical laboratory specialists. A group setting was preferred to stimulate multidisciplinary discussions among professionals. A focus group setting allowed participants to talk and discuss—freely but with some guidance—the factors critical for the implementation of the Tumor-First workflow. This study was assessed by the Medical Ethical Committee (CMO) of region Arnhem-Nijmegen, who declared there was no need for ethical approval (number 2021-7290). We report this study in accordance with the COREQ checklist [17].

In the Netherlands (population of 17 million), around 1300 patients are diagnosed yearly with OC [18]. Here, surgical care is centralized in gynecologic oncology centers and surgical procedures are performed by gynecologic oncologists (gynecologists who performed an additional 2 year subspecialty ‘gynecologic oncology’) [19]. There are nine gynecologic oncology centers (seven university hospitals, one non-academic hospital and one cancer-center), each having connections with regional hospitals. In previous national multidisciplinary meetings regarding genetic testing for OC, the Tumor-First workflow emerged as the preferred workflow to replace the ‘old’ one (all OC patients referred to the clinical geneticist for germline testing), as illustrated in Fig. 1. The level of experience with the Tumor-First workflow varies across the gynecologic oncology centers, varying from more than two years’ experience (three centers), one to two years’ experience (three centers), to less than a year of experience (three centers). The level of experience with the workflow and some general information were collected by telephoning gynecologic oncologists, pathologists and clinical geneticists from the expert centers.

We organized three focus groups between February 8th and March 22nd 2021 for all nine gynecologic oncology centers in the Netherlands. As the familiarity with the Tumor-First workflow varied somewhat between these centers, we strived for a mix of skills in each focus group. We invited a gynecologic oncologist, a clinical geneticist, a pathologist, and a clinical laboratory specialist (clinical laboratory geneticist or clinical scientist in molecular pathology) from every gynecologic oncology center. Representatives were selected in consultation with the centers. Additionally, all gynecologic oncologists and pathologists were asked to invite a colleague from a hospital in their region to participate. Invitations were sent per email. Professionals who were unable to attend were asked to delegate a colleague.

A semi-structured interview guide was developed based upon the various phases of the Tumor-First workflow and the theoretical framework by Flottorp [16]. The interview guide distinguished the three phases of the workflow; (1) patients are informed about the workflow and the tumor DNA test is initiated, (2) tumor DNA test is performed and test result is communicated to professionals (3) patients are notified about the test result and advised on (a referral for) germline testing in case of an aberrant tumor test. We use the term ‘aberrant test’ in case a pathogenic variant in an OC risk gene is detected, and ‘normal test’ in case no pathogenic variant is detected. The phases included in the interview guide are illustrated in Fig. 2. For each phase we started to retrieve information about the current the workflow, followed by a discussion on possible barriers while exchanging solutions already considered. Possible barriers and enablers have been addressed in a free and open discussion by an experienced moderator (author: JB) who knew the seven domains from the framework of Flottorp by heart. In each focus group, two key persons in this field were present to support the moderator in answering specific questions from participants. Focus groups were conducted online using the Zoom videoconferencing platform, and lasted 90 to 120 min. The zoom meetings were recorded (video and audio) and transcribed non-verbatim. Participants were informed about the recording and were given the opportunity to object. All gave permission to be recorded. To ensure the privacy of participants, transcripts were deidentified before analysis. Original transcripts were in Dutch and quotes have been translated for this paper. Contextual additions to the quotes are displayed using […].

The main results of the focus groups were summarized and sent to the participants (including those who were unable to attend) to allow them to provide comments and additions. Subsequently, the complete transcripts were uploaded to the program ATLAS.ti GmbH Version 8.4.20 to aid the qualitative analysis. Transcripts were analyzed deductively using a framework approach [20]. We prepared a framework combining the theory from Flottorp with the phases of the Tumor-First workflow [16]. After familiarization, reading and rereading of the first transcript, a first draft of codes was prepared by two project members independently (authors: VW, YS). Thereafter, the transcript was coded by both project members and, if necessary, codes were added, merged or altered. This process was repeated for the following two transcripts. Ultimately, a final codebook was prepared and transcripts were coded again by both project members to test reliability. Another author (JB) was consulted in case of disagreement. This codebook including relevant text passages was discussed to identify core themes regarding Tumor-First testing. These themes were connected to the seven implementation domains [16] for presentation in this report.

Three focus groups were organized, and a total of 53 invitations to participate in one of those focus group discussions were sent out. Forty professionals accepted the invitation and participated, six declined, three did not reply and four dropped out after signing up. Table 1 provides an overview of the professional disciplines and number of years of experience with Tumor-First workflow of the participants in each focus group. In comparison, the participation of gynecologic oncologists was somewhat higher in the first and third focus group, whereas in the second focus group more pathologists participated. The level of experience was well distributed across the three focus groups.

Healthcare professionals from all four disciplines (gynecologic oncologists, pathologists, clinical geneticists, and clinical laboratory specialists) have an optimistic view on the Tumor-First workflow for OC patients and they endorse national implementation. They were all familiar with the concept of the Tumor-First workflow and no questions were raised on the usefulness or feasibility of this diagnostic workflow. Healthcare professionals indicated that the main advantage is that a tumor DNA test provides information for both heredity and treatment. At the outpatient clinics, OC patients often have questions with regards to cancer risk for family members and with the Tumor-First workflow it can be reassuring that there is a standard pretest.

Testing of tumors is very useful and provides a lot of information Clinical geneticist_6

It is interesting that it brings together consequences for treatment and for family Gynecologic oncologist_9

The level of experience with the Tumor-First workflow varied among centers and professionals. The willingness and need to help those with less experience emerged from the focus group discussions. Furthermore, factors critical for implementation of the Tumor-First workflow were discussed, with a focus on how to promote implementation (what do healthcare professionals need?). Table 2 provides a summary of these themes, and Fig. 3 presents example quotes.

Besides the overall endorsement for implementation, healthcare professionals emphasized points for improvement related to the Tumor-First workflow. First of all, all professionals agreed that the quality of the tumor DNA test is of critical importance for the reliability of the workflow’s outcome. As the tumor DNA test is much more laborious and complex compared to the germline test, minimal criteria for the quality of the tumor DNA test are essential for all professionals to trust the results. Thus, it has to be clarified which laboratories are able to perform a tumor DNA test of sufficient quality. Completely aligning test quality among laboratories is complicated, however, due to variation in techniques. There is need for general parameters on the minimal quality of the test.

Additionally, with regards to the tumor DNA test, some technical questions were raised. Firstly, it was questioned which genes the tumor DNA test should cover in addition to BRCA1 and BRCA2, since the germline test panel includes more OC risk genes. Secondly, a question arose regarding histological triage, specifically whether it is possible to select OC patients for tumor testing based on their histology. Thirdly, the tumor DNA test is performed on formalin-fixed paraffin-embedded tissue and fixed cytological material, but there was some ambiguity as to which tissue types are most suitable to be tested; e.g., resections versus biopsies and cytological material. Advice has to be prepared by experts in the field to answer these questions and provide clarity for all professionals involved.

Another theme that emerged entails referral criteria in case of a normal tumor DNA test result. Healthcare professionals indicated there could be additional reasons to refer a patient for genetic counseling, for example a family history of certain types of cancer or incomplete test results due to technical limitations. A concern was expressed that gynecologic oncologists might be less aware of a family history suggestive of familial cancer due to the new Tumor-First workflow. There is a need for an establishment of national referral criteria to the clinical geneticist in case of a normal tumor DNA test result, and healthcare professionals have to be aware of these criteria.

Lastly, there appeared to be variation in the professional discipline initiating the tumor DNA test in the different centers. Some centers are used to the pathologist routinely initiating the test after diagnosis of OC, whereas in other centers an active request by the gynecologic oncologist is necessary. It remains questionable whether this should be harmonized, as initiation in itself is more important than by whom the test is initiated. It would be beneficial to establish which professional discipline initiating the test best facilitates adherence to the workflow.

At the level of professional interactions, healthcare professionals indicated it to be essential to keep everyone sufficiently informed on the different phases of the workflow and the results. This includes exchange of information among healthcare professionals from different disciplines, but also between healthcare professionals and patients.

Firstly, there appeared to be variation amongst gynecologic oncologists in what information they provide to patients regarding initiation of the tumor DNA test. Also, opinions on the amount of information a patient should receive varied, similar to opinions on whether or not explicit (written) consent is necessary for performing a tumor DNA test. In general, professionals agreed that patients must receive sufficient information on the tumor DNA test before they can consent to this test. To ensure this, there is a need for a minimal set of information that is standardly explained to a patient. Additionally, an opt-out option has to be incorporated into the workflow to provide the opportunity for patients to decide to not be included in a tumor DNA test procedure.

Another theme focuses on reporting the test result in electronic medical records, being another task of the gynecologic oncologist. Healthcare professionals agreed on the importance of clearly reporting the tumor DNA test result, and it has to be distinguishable from a germline test result. They also emphasized the importance of the findability and accessibility of the tumor DNA test result in the long term, indicating that current electronic medical records generally fail in this regard. There appeared to be a need for agreement on where and how to report the tumor DNA test result in commonly used electronic health record systems (e.g., HiX and Epic).

Furthermore, the responsibility of informing patients about the tumor DNA test result has to receive special attention, similar to the responsibility of communication to medical oncologists and general practitioners. There appeared to be variation among centers in who takes this responsibility (mostly the gynecologic oncologist). Center-specific and region-specific agreements are essential in this regard. Moreover, clinical geneticists have to be able to see patients with an aberrant tumor DNA test on short notice, to avoid any unnecessarily long period of uncertainty for the patient.

Monitoring patient inclusion in the Tumor-First workflow was an important topic discussed related to the level of professional interaction. Healthcare professionals emphasized the importance of including all patients in the Tumor-First workflow. Patients should not be missed, neither during the initiation of the tumor DNA test nor the referral for germline testing (based on an aberrant tumor DNA test result). Designing checks during several points of the workflow was proposed as a safety guard. Monitoring patient inclusion in the Tumor-First workflow may be essential both at a local and national level.

Additionally, the level of awareness in non-expert centers has to receive specific attention. In comparison to gynecologic oncology centers, there seems to be less awareness in non-expert centers about initiating a tumor DNA test. Also, professional collaboration among disciplines is more complicated in non-expert hospitals, as they sometimes collaborate with multiple expert centers, i.e. they collaborate with one center for a gynecologic oncology expert and with another for a clinical genetics expert.

At the level of patients, healthcare professionals perceive variation in the needs and preferences of patients regarding the Tumor-First diagnostic workflow. Considering that patients are ill, vulnerable and in the middle of the diagnostic process, one must avoid information overload and not provide them with too much information on the tumor DNA test. On the other hand, some patients come up with questions regarding heredity and cancer risk for family members, in which case healthcare professionals should always consider the needs and preferences of patients on an individual basis when providing them information.

At the level of professionals, it stood out that a small number of healthcare professionals have insufficient knowledge on the technical aspects of the tumor DNA test; particularly the terms ‘somatic’ and ‘germline’ caused confusion. Educational training would be helpful in this regard. Moreover, some healthcare professionals were unaware of the specific activities of other professionals within the current workflow for OC patients. Interdisciplinary communication within centers can be encouraged.

When focusing on incentives and resources, healthcare professionals mentioned that hospitals (specifically non-expert centers) have less incentive to initiate a tumor DNA test if they need to pay for the test themselves. Therefore, it is of critical importance that a national solution for financing of the tumor DNA be established. This national solution is complicated by healthcare insurance arrangements (social/legal level) as well as by the fact that every hospital has its own internal procedures of division of finances (organizational level).

Additionally, at the level of incentives and resources, clinicians (predominantly gynecologic oncologists) would appreciate assistance during the implementation of the workflow. Information material for patients, letters explaining the tumor DNA test results and a tool for assessing family history were examples of assisting material that would be useful. Also, educational training for professionals and assistance when communicating to non-expert centers would be of added value.