2.1 Design

The constructivist approach, which asserts that reality perception is constructed from individual and socio-historical contexts,³⁵ informed this qualitative study. A thematic analysis was conducted, guided by selected grounded theory techniques including constant comparative, iterative, concurrent data collection and analysis, and predominantly inductive analysis.³⁶ While grounded theory is a contested approach,³⁵ the “Strauss” led approach supports using grounded theory techniques to develop themes.³⁷

2.2 Participants

Purposive sampling was used whereby only SCORE trial participants who expressed willingness (via consent-form tick-box) to participate in an interview about their experiences on study completion (12 months) were invited. Eligible participants were invited via email or letter, provided with an information statement, and followed up via phone call. Participant consent was verbal and audio recorded on interview commencement. Participants were selected to ensure even distribution between those allocated to the shared and usual care arms of the SCORE trial, and to ensure representation across all enrolling hospital sites.

2.3 Data collection

SCORE trial investigators²⁹ developed two sets of semi-structured interview guides, with a shared care set specific for participants in the shared care arm, and a usual care set specific for participants in the usual care arm (Appendices 1 and 2). Development of these guides was informed by implementation outcomes relevant to the individual consumer level, as defined by Proctor et al.²⁶ and aimed to understand the acceptability and appropriateness of the SCORE shared care intervention from the patients' perspectives, and their preferences for follow-up care. Audio-recorded interviews were conducted by the lead author (CG) between August 2021 and August 2022 either via video or telephone. CG is a registered psychologist and was trained in research interviewing techniques by PS, an experienced qualitative research interviewer. The interview for shared care participants explored their experiences of shared care and follow-up care preferences. The interview for usual care participants explored their experiences of usual care and attitudes towards greater GP involvement. All interviews explored participant experiences of receiving telehealth-delivered care during the COVID-19 pandemic.

2.4 Data analysis

Audio-recorded interviews were transcribed verbatim. Initial analysis (CG) involved open coding, with labels created to describe segments of transcript content. Codes with similar meaning were grouped into categories.³⁸ Similarities, differences, and relationships across categories were considered to inform the theme names, aligned with the study goals.³⁸ NVivo software (QSR International Pty Ltd) supported data management.

CG coded all transcripts and a subset of five transcripts were reviewed and co-coded by an experienced qualitative researcher (CW) to improve analytic trustworthiness. CG and CW first met after each coding the initial transcript and compared perspectives on the codes. CW supported CG's approach to coding. Throughout the analysis, CG and CW met to review and define emerging categories and themes within the data. Minor differences were reconciled upon discussion. This method of inter-rater analysis was employed to increase interpretative rigour.³⁹

Demographic information was extracted from the interviews and summarised using descriptive statistics.

Consolidated criteria for reporting qualitative data (COREQ)⁴⁰ were adhered to, although ‘participant checking’ was conducted through inviting participant feedback on the interviewer's summary of their statements at the end of interviews. This ‘participant checking’ conversation was included in the verbatim transcription. All participants confirmed their experiences had been captured accurately, and were able to elaborate, or clarify, as necessary. Overall findings were not returned to participants due to potential burden⁴¹; and no available evidence indicates that the strategy improves the quality of findings.⁴²

2.5 Ethics

Ethics approvals were received from Peter MacCallum Cancer Centre (Ref: HREC/72311/PMCC-2020) and Swinburne University (Ref: 20215904-8162) ethics committees.

3.1 Facilitators and barriers to implementing shared care

Five themes were developed on the facilitators and barriers to implementing shared care as perceived by those who had experienced either 12 months of shared or usual survivorship care in the SCORE trial: communication with and between health care providers; relationships with GPs; accessibility of GPs; satisfaction with cancer treatment; and participant beliefs, attitudes and values. In data illustrations, participants P1S, P3S, P4S etc., received shared care, and P2U, P5U, P7U etc., received usual care.

3.2 Communication with and between health care providers

Alternatively, when they perceived good specialist-GP communication, participants were “more comfortable” (P10U) seeing GPs for follow-up care.

Participants also identified the need for clarity of roles between specialists and GPs, sometimes querying who was “the first port of call?” (P12S). They expressed confusion around appointment schedules and attributed this to poor hospital communication.

While some participants stated that informative resources helped clarify shared care expectations, few recalled receiving any information, or if they did, did not use them.

3.3 Relationships with GPs

Many described the advantages of having “good,” longstanding relationships with GPs, including feeling comfortable to “talk more freely,” (P10U) believing that the GP “really cares” (P17S) and that GPs can be trusted. Participant P10U detailed the benefits of being able to “explain in my language how I am feeling” when their GP is from the same cultural and linguistic background.

3.4 Accessibility of GPs

Participants found their GPs to be more geographically accessible and convenient, particularly those living rurally who had GPs in the same town, or “only 10 min down the road” (P11S). Some mentioned they could contact/see their GP more frequently and more urgently than oncologists if needed.

However, some participants had difficulty making GP appointments due to limited availability, or GP leave-taking and inability to find another. This increased distress about unmet needs. GP affordability was also cited as an access barrier when their GP visit was not fully covered by the Medicare rebate.

3.5 Satisfaction with cancer treatment

Participants from all hospital sites spoke positively about experiences of hospital-based care during cancer treatment, describing it as “excellent,” (P19U) and “terrific” (P11S). Some preferred continuity of care with the same hospital staff rather than GP involvement. Participants also asserted that specialists are experts and felt more confident in hospital-based follow-up care. P15S stated “I probably had every confidence in the hospital because I think they did a marvellous job [during treatment].”

3.6 Beliefs, attitudes, and values

Many participants believed GPs could not expertly manage their survivorship care (especially if they had delays in diagnosis). P10U said “It makes you feel better … Because he [oncologist] knows the treatment that I had. Yes, he is a specialist in cancer, not like a GP.” Some also believed that GPs were redundant in survivorship care, with participant P7U stating “I don't see why I would need the GP for this.”

Others referenced the importance of reducing hospital system pressures and believed GPs were well-positioned in the shared care model to provide test results and follow-up. P1S stated “I think you're probably not wanting to take up hospital resources with getting notification of results … If it's just for results like blood tests and stuff like that, the GP could do that.”

3.7 Perspectives on using telehealth-delivered health care

Four themes were developed on perspectives about using telehealth for survivorship care accessed through GPs and hospital-based specialists: telehealth was convenient and more comfortable; satisfaction with quality of care; associated challenges; and future preferences (Appendix 4).

3.8 Telehealth appointments were comfortable and convenient

Many participants felt more relaxed during telehealth appointments due to being more comfortable at home, and less aware of the doctor being busy. Some participants found telehealth appointments less confronting as they avoided hospital waiting rooms, which were described by one participant as “very tense” (P2U). Another stated, “not that I was worried, but you see other people and you think ‘what are they going through?’” (P5U). Most felt telehealth saved travel and hospital waiting times.

3.9 Satisfaction with quality of care

For many participants, unless they believed they required a physical examination, for example, they felt unwell, there was no perceptible difference in quality of care. P4S said, “I feel it's exactly the same really. At that point they are only following up the results of bloods and scans so, it's no different.” Participants spoke of being “able to communicate” (P5U) effectively via telehealth. They described their interactions as “personable” (P12S) and “exactly the same” [as face-to-face] (P11S).

3.10 Challenges of telehealth appointments

A few participants cited difficulty setting up video appointments, due to not having the technology or required knowledge. Some participants found telephone appointments “impersonal” (P8S) and rushed. Incidental social interactions in the hospital were also missed with telehealth appointments for example, visits to the wellness centre.

3.11 Future preferences

However, some still preferred to see their doctor face-to-face for ease of communication. P18U said, “If I want to see the doctor, I will go and see [them] in person… First of all, I haven't got a computer, that's one thing. Second, I've got poor hearing.”

4.1 Communication

The shared care intervention followed a study protocol²⁹ with defined roles for the GP and specialist. However, participants still expressed confusion about the follow-up schedule, and roles and responsibilities of each provider when the model of care was not specialist-led. Similar experiences have been reported in GP-led models of follow-up for CRC survivors.⁴³ While participants in the shared care arm of the SCORE trial received SCPs to facilitate communication, many could not recall using these and still found that their GPs were unsure of the study protocol, leading to distress that their care needs were not being met. Better communication tools are needed between hospital-based care providers, GPs, and patients as shared care models have been deemed less acceptable when survivors perceived poor communication between care providers.^{18, 23} A survivorship care coordinator role may assist in the implementation of shared care by scheduling tests, following up results, sending appointment reminders, and serving as a point of contact and support for both patients and providers.²⁴ Research suggests that GPs would prefer clear follow-up protocols and easy access to hospital-based specialist teams, rather than additional training to implement shared care.²⁴ Electronic medical records that facilitate sharing of health information between GPs and specialists in real time might enhance the efficiency of communication.¹⁰

4.2 Survivor beliefs and attitudes

Consistent with existing literature, survivors in our study were often concerned about their GP's knowledge of cancer survivorship, and doubted their ability to provide effective survivorship care.^{16, 23} Some participants also believed that, because specialists are experts, GPs are not needed in survivorship care. This belief is incongruent with growing evidence which shows that GPs can play a vital role in survivorship care^{12, 15} and specialist-led models leave survivors with unmet needs.^9-13 This inconsistency suggests there is a need to better: explain to patients the purpose of survivorship care and the roles of the GP and specialist (and survivorship care coordinator); prepare patients for alternative models of survivorship care; and reassure patients that shared care is a safe and appropriate alternative to specialist-led follow-up.^{9, 17}

4.3 Relationships with GPs

Generally, participants who experienced shared care were satisfied with having their GP involved while remaining connected to their specialist. Emerging data from the SCORE trial show that CRC survivors exposed to shared care prefer this model.⁴⁴ Studies suggest there is not a “one size fits all” model of survivorship care and recommend personalising care pathways based on individuals' risk of recurrence and long-term effects, existing physical and psychosocial issues, circumstances, and capacity to self-manage.^{9, 10, 24} We suggest including quality of the patient-GP relationship in this criteria, as this was found to be both a facilitator and barrier to shared care, depending on whether it was perceived to be “good” or “poor”. Similarly, Lisy et al²⁴ found that patients having a regular, trusted GP was a facilitator to shared care. Personalised follow-up care pathways may lead to better patient outcomes with more efficient use of oncology time, fewer visits for survivors, and reduced overall costs.⁹

4.4 Telehealth-delivered care

Importantly, the results also suggest that when only receiving test results, survivors are satisfied with accessing survivorship care via telehealth (both telephone and video), and generally perceived no difference in quality of care received when attending these appointments in person or via telehealth. This finding is significant given the uptake of telehealth during and since the global COVID-19 pandemic, and the limited data on telehealth use in adult survivors of CRC.³² Telehealth-based follow up can create care environments at home, work or wherever is convenient for the survivor of cancer.³⁴ It is recommended that patient preferences for telephone/video appointments versus in person be considered when offering a particular model of follow-up and, in line with Chan et al,³² that telehealth be offered to supplement in-person appointments and not as a stand-alone model of follow-up.

4.5 Study limitations

The results of this study should be interpreted in light of several limitations. Firstly, we explored perspectives of CRC survivors treated in the public health system in Melbourne, Australia. These results may not be generalisable to other cancer types, to people treated in the private system, or to other countries with different health care systems. Second, the study participants were interviewed at various time points, with some completing their participation in the SCORE trial more than 12 months before they were interviewed. This may have led to recall bias in interviews and accounted for difficulty recalling survivorship care resources and other aspects of their survivorship care. We also acknowledge the potential risk of selection bias with interviews of participants in the SCORE trial who had already agreed to being randomised to shared care, and may have more favourable attitudes towards GP involvement. Additionally, we only explored patient perspectives on their first 12 months of survivorship care, and not the longer-term acceptability of the model for example, at two, three, or five years-post treatment. It is possible that patient perspectives (including confidence in their GP) may change over time as GPs become more experienced in the provision of shared care. Cancer-specific issues and concerns might also be less prominent years later.

4.6 Clinical implications

It is also suggested that researchers and healthcare practitioners evaluate future implementation of shared care against implementation outcomes,²⁶ as well as clinical and service outcomes.