2.1 Participants and Recruitment

This analysis used data collected in the Travelling for Treatment study [18], a longitudinal cohort study of rural cancer survivors and their caregivers, tracking patient-reported outcome measures over time (up to 5 years post-recruitment). The Travelling for Treatment study team comprised early-career to senior researchers with backgrounds in psycho-oncology, public health and cancer survivorship. Methods for participant recruitment and data collection have previously been described in detail [7], with further information outlined elsewhere [18-21]. Briefly, between September 2017 and June 2020, consecutive adults who had a cancer diagnosis or were undergoing tests for cancer, staying at one of Cancer Council Queensland's (CCQ) subsidised accommodation lodges, were invited to participate in the study. To be eligible to stay at one of CCQ's lodges, guests must be required to travel > 50 km (i.e., > 31 miles) to receive cancer treatment or follow-up care in Queensland, Australia. Where possible, study invitation packs were distributed by lodge staff to guests during their stay. For guests who checked in after hours, an invitation pack was sent via mail to their home address. One week post-distribution, a research volunteer contacted guests to follow up on the study invitation and respond to any questions. Ethical approval was obtained from a recognised institutional Human Research Ethics Committee (H17REA152). This analysis is reported according to the Consolidated Criteria for Reporting Qualitative Research (COREQ) (Table S1) [22].

2.2 Data Collection

Over the 4-year study recruitment period, participants completed a one-on-one structured interview (approx. 45–60 min) either during their stay at the lodge or via phone, which was neither audio nor visually recorded. Participants and interviewers had no prior relationship. The interviews were conducted by a team of 44 trained interviewers. To ensure consistency and reliability in the data collection across the large team, all interviewers completed mandatory, study-specific training that covered core interviewing skills, ethical considerations and protocols for managing participant distress. Standardised interview guides and ongoing supervision were used to support uniformity in the conduct of interviews. Participants also completed a written, self-administered questionnaire that collected information on sociodemographic and clinical characteristics. This analysis reports on qualitative information collected in the interview about participants' self-reported reasons for perceived delays in their initial pathway to initial cancer detection and treatment.

3.1 Pathway to Initial Cancer Detection and Treatment

In three open-ended interview questions, participants were asked to recall whether they perceived any delays at three distinct steps: (i) seeking medical attention, (ii) receiving their diagnosis and (iii) commencing treatment in their pathway to initial cancer detection and treatment. If a participant reported perceiving a delay at any of these steps, they were prompted to provide further information on reasons for the perceived delay. Participants' responses were recorded during the interview on a paper interview guide by the interviewer and were subsequently entered into an online data file by a researcher post-interview.

3.2 Demographic and Clinical Characteristics

Participants were asked to self-report their age, sex, country of birth, highest education level, annual household income, access to private health insurance, marital status and cancer type. Residential street addresses were geocoded and mapped to the 2011 Statistical Area Level 2 (SA2) boundaries using MapMarker Australia Version 15.16.0.21 and MapInfo Pro Version 15.0 SA2. Geographical remoteness was determined using the Australian Bureau of Statistics Accessibility/Remoteness Index of Australia (ARIA), and relative area-level disadvantage was identified using the Socio-Economic Indexes for Areas (SEIFA) [23, 24]. Cancer type was self-reported in the questionnaire and, where possible, verified against the population-based Queensland Cancer Register (QCR). Self-report data were relied upon where the diagnosis could not be verified by the QCR (n = 71, 10%).

3.3 Data Analysis

Study participants were included in this analysis if they had a confirmed cancer diagnosis and provided at least one response to one of the interview questions about perceived delays in the pathway to initial cancer detection and treatment. Descriptive statistics were used to summarise the sociodemographic characteristics of the study participants. Chi-square tests and Fisher's exact T-tests were used to compare the characteristics of participants included in this analysis with those excluded.

The number and proportion of participants who did and did not perceive a delay were calculated for each of the three steps: (i) seeking medical attention, (ii) receiving their diagnosis and (iii) commencing treatment. Additionally, the number and proportion of participants who perceived a delay at one, two or all steps were calculated. Descriptive statistics were used to summarise the sociodemographic characteristics of the participants who did and did not perceive a delay at each step. Chi-square tests and Fisher's exact T-tests were used to compare the characteristics of these groups of participants. If the difference between the two groups was statistically significant for variables with more than two categories (e.g., age and cancer type), post-hoc chi-square tests were conducted to compare each category between the two groups of participants. When determining statistically significant differences in post hoc analyses, a Bonferroni correction for multiple comparisons was applied.

Participants' descriptions of self-reported reasons for perceived delays at each step were analysed using conventional content analysis, as described by Hsieh and Shannon, whereby codes and categories were derived inductively from the interview data rather than based on literature [25]. First, two authors (A.T.-S. and I.B.) familiarised themselves with the data by reading through the interview responses several times. Second, for each of the three steps separately, responses were coded by one researcher (IB) using Microsoft Excel and documented in a coding framework using Microsoft Word, which included representative quotes alongside codes to illustrate the analytic findings. Then, codes were categorised based on the perceived source of delay: (i) personal (i.e., related to the person with cancer), (ii) healthcare professional or (iii) healthcare system (i.e., organisational processes, policies). Codes that did not fit into one of these categories (e.g., public holidays) were categorised as (iv) other. The coding framework was reviewed by a second researcher (A.T.-S.) to ensure all codes were mutually exclusive before independently applying it to the full dataset. Any discrepancies in coding between the two researchers were resolved by a third researcher (E.J.). The final coding frameworks for each step are shown in Tables S2–S4.

Concept maps were developed to visually summarise participants' perceived sources of delay and self-reported reasons for these perceived delays at each of the three steps. For each of the three steps, descriptive statistics were used to summarise the number and proportion of participants reporting each category (i.e., personal, healthcare professional, healthcare system and other) as the perceived source of delay. The number and proportion of participants reporting each code (i.e., self-reported reason for the perceived delay) were also summarised. Findings are presented below, alongside representative participant quotes to illustrate the analytic findings.

Finally, the number and proportion of participants whose perceived delay in commencing treatment (i.e., Step 3) was an actual delay were determined based on previous analyses conducted in this cohort [26]. Briefly, for participants with a relevant Optimal Cancer Pathway (OCP) for their cancer type and self-reported dates for both diagnosis and treatment commencement, actual delays in commencing treatment were determined by comparing the number of weeks between diagnosis and starting treatment with recommended timeframes provided in the relevant OCP [27]. Participants whose timeframe between diagnosis and treatment commencement exceeded the number of weeks recommended by the relevant OCP (based on cancer type and modality of initial treatment) were classified as experiencing an actual delay in commencing treatment.

4.1 Participant Characteristics

Of the 811 participants who consented to participate in the Travelling for Treatment study over the 4-year study period, 686 (85%) were eligible for inclusion in this analysis (Figure S1). Participants who did not complete an interview at study recruitment (n = 99) or did not respond to any questions relating to delays in their pathway to initial cancer detection and treatment (n = 15), along with participants who were undergoing investigations for cancer but were not eventually diagnosed with cancer (n = 11) were excluded from this analysis. Compared to those excluded from this analysis, participants included in this analysis were more likely to be born overseas (χ²(1) = 27.079, p = < 0.001, phi [φ] = 0.183) and have completed tertiary education (χ²(1) = 5.827, p = 0.016, φ = 0.088). Participants included in this analysis were also more likely to be diagnosed with breast cancer (χ²(1) = 10.176, p = < 0.001, φ = 0.113) and less likely to be diagnosed with cancers of unknown primary origin (χ²(1) = 170.328, p = < 0.001, φ = 0.462) (Table S5).

Participant characteristics are shown in Table 1. Half of the participants (50%) in this analysis completed their interview within 6 months of receiving their cancer diagnosis. Participants' ages ranged from 26 to 93 years (median = 66 years, interquartile range = 58–72 years), 53% identified as male and 47% as female and 80% were born in Australia. The most common primary cancer diagnoses were breast (18%), head and neck (15%), prostate (12%) and skin (12%) cancer. Most participants lived in inner (45%) and outer (43%) regional areas. A small proportion of participants (4%) were classified as living in a major city; however, they were included in the study as they had travelled > 50 km to access cancer treatment or follow-up care. Over two-thirds of participants (68%) were in a relationship, de facto or married. Less than half (42%) had completed university or vocational level education. Participants' annual household income varied, with almost half (47%) earning under AU$30,000 per year. Most participants (79%) did not have private health insurance that partially or fully covered their cancer treatment.

4.2 Perceived Delays in the Pathway to Initial Cancer Detection and Treatment

Almost half of the participants (n = 320, 47%) perceived a delay at one or more of the three steps: (i) seeking medical attention, (ii) receiving their diagnosis and (iii) commencing treatment. Of these, 206 (64%) participants perceived a delay at only one step, 99 (31%) perceived a delay at two steps and 15 (5%) perceived a delay at all three steps.

4.3 Perceived Delays in Seeking Medical Attention

Almost a fifth of participants (n = 132, 19%) perceived a delay in initially seeking medical attention related to their subsequent cancer diagnosis. Characteristics of participants who did and did not perceive a delay are shown in Table 1. Characteristics only significantly differed with respect to age (χ²(5) = 11.499, p = 0.042, φ_c = 0.131) and cancer type (χ²(8) = 27.867, p = < 0.001, φ_c = 0.202) (Table S6). Participants aged 40 to 49 years (χ²(1) = 9.939, p = 0.002, φ_c = 0.122) were more likely to perceive a delay in seeking medical attention. Participants with gynaecological cancer (χ²(1) = 12.195, p = < 0.001, φ = 0.133) and those with cancers of unknown primary origin (χ²(1) = 7.739, p = < 0.005, φ = 0.106) were also more likely to perceive a delay in seeking medical attention (Table S6).

Participants' self-reported reasons for perceived delays in seeking medical attention are shown in Figure 1, with participant quotes presented in Table 2. Of the participants who perceived a delay in seeking medical attention, half (n = 67, 51%) perceived this was due to personal factors, such as misinterpreting (n = 19, 28%) or ignoring (n = 19, 28%) their signs and symptoms, travelling for leisure (n = 4, 6%) and competing caregiving responsibilities (n = 4, 6%) or employment commitments (n = 3, 4%). Almost one-third (n = 42, 32%) perceived that their delay was related to factors associated with healthcare professionals. For example, participants perceived inattention from the healthcare professional (n = 26, 62%) with multiple consultations required before their concern (i.e., their cancer-related symptom/s) was addressed (i.e., via testing or referrals to specialists). In other cases, participants reported a misdiagnosis (n = 8, 19%) by the healthcare professional, which resulted in action and/or treatment undertaken for the misdiagnosed condition. Participants perceived these to delay their efforts in seeking medical attention for their current cancer, including delays in obtaining referrals for cancer-related testing or specialist consultation. Some participants (n = 26, 20%) perceived that their delay was related to factors associated with the healthcare system, particularly difficulties in obtaining appointments with general practitioners (n = 7, 27%) or specialists (n = 4, 15%) and at cancer screening or testing facilities, such as mammograms (n = 4, 15%). A small number of participants perceived that their delay was due to other reasons (n = 5, 4%), such as initially receiving negative test results (n = 4, 80%) or the impact of public holidays such as the Christmas and New Year's period (n = 1, 20%).

4.4 Perceived Delays in Receiving Their Cancer Diagnosis

Almost a quarter of participants (n = 161, 23%) perceived a delay in receiving their diagnosis. Characteristics of participants who did and did not perceive a delay in diagnosis are shown in Table 1. Characteristics only significantly differed with respect to age (χ²(5) = 13.528, p = 0.019, φ_c = 0.142), geographical remoteness (χ²(3) = 7.907, p = 0.048, φ_c = 0.108) and highest level of education completed (χ²(1) = 4.145, p = 0.042, φ = 0.080) (Table S6). Participants who perceived a delay in receiving their diagnosis were more likely to have completed university or vocational education compared to those who did not perceive a delay. While the overall chi-square tests identified a significant association between delay status and both age and geographical remoteness, no specific age category or geographical area showed a statistically significant difference after controlling for multiple comparisons (Table S6). Thus, the difference observed between those who did and those who did not perceive a delay may not be attributable to any one category within each variable, but rather to the overall distribution of ages and geographical remoteness.

Participants' self-reported reasons for perceived delays in receiving their cancer diagnosis are shown in Figure 2, with participant quotes presented in Table 3. Over half (n = 86, 53%) perceived that the delay was related to their interactions with, or the actions of, healthcare professionals. For example, the healthcare professional required multiple opinions or further tests or scans prior to confirming the diagnosis (n = 30, 35%) or being initially misdiagnosed (n = 24, 28%). Around one third of participants (n = 57, 35%) perceived delays due to challenges accessing the healthcare system, including waiting lists for specialists (n = 16, 28%) and tests or scans (n = 11, 19%). In some cases (n = 15, 9%), participants perceived that the delay in receiving their diagnosis was due to personal reasons, such as their comorbidities (n = 6, 40%), competing caregiving responsibilities (n = 3, 20%) and travelling for leisure (n = 3, 20%). Other delays (n = 19, 12%) included the perceived impact of natural disasters (n = 2, 11%), public holidays (n = 4, 21%) and receiving negative test results (n = 10, 53%). A small number of participants (n = 11, 7%) perceived that they were delayed in receiving their diagnosis, but did not provide further information on reasons why.

4.5 Perceived Delays in Commencing Treatment

Delays in commencing treatment were perceived by almost a quarter (n = 157, 23%) of participants. Characteristics of participants who did and did not perceive a delay in commencing treatment are shown in Table 1, with no statistically significant differences observed between the two groups (Table S6).

Participants' self-reported reasons for perceived delays in commencing cancer treatment are shown in Figure 3, with participant quotes presented in Table 4. More than one third (n = 57, 37%) perceived the delay to be related to challenges accessing the healthcare system, particularly waitlists for treatment (n = 39, 68%), waiting for the appropriate healthcare professional or healthcare facility (n = 10, 18%) and travel distances to treatment facilities (n = 3, 5%). Almost one third (n = 49, 31%) perceived the delay was related to personal factors, such as their comorbidities delaying commencement (n = 33, 67%) or their personal choices regarding treatment, including seeking further information (n = 3, 6%) or initially declining treatment (n = 2, 4%). Over a quarter of participants (n = 44, 28%) perceived the delay in commencing treatment was related to their interactions with, or actions of, healthcare professionals, for example, prolonged treatment planning (n = 21, 48%), having treatment rescheduled or cancelled due to the healthcare professionals' personal commitments (n = 8, 18%), or miscommunications within the care team (n = 5, 11%). Other delays (n = 13, 8%) included the perceived impact of natural disasters (n = 2, 15%) and public holidays (n = 7, 54%). A small number of participants (n = 16, 10%) perceived that they were delayed in commencing treatment, but did not provide further information on reasons why.

4.6 Perceived Versus Actual Delays in Commencing Treatment

Of the 157 participants in this analysis who perceived a delay in commencing treatment, 108 (69%) had data available on actual treatment delay status, with over half (n = 62, 57%) experiencing a treatment delay according to the relevant Optimal Care Pathways for their cancer type based on their treatment modality.

5.1 Implications for Policy and Practice

In Australia, Optimal Care Pathways (OCPs) are available for 24 cancer types, with the implementation of the OCPs a key objective of the newly launched Australian Cancer Plan [45]. Population data show a positive correlation between the use of OCP-aligned care and improved cancer outcomes [46]. In line with findings from our analysis, implementation of the OCPs may contribute to improved care for cancer survivors living in rural and remote areas by reducing delays in cancer care, especially for diagnosis and treatment [46]. Further, recent funding awarded for the development of an OCP specifically tailored for cancer survivors in rural and remote areas is a key step in reducing diagnosis and treatment delays among this cohort [47]. Findings from the current analysis provide insights into factors that may underlie or contribute to suboptimal care for rural cancer patients that could be addressed in the rural-specific OCP. Additionally, the findings also highlight the importance of improved clinician-patient communication regarding recommended timeframes, particularly for diagnosis and treatment, and survivors' knowledge of and access to the OCPs. With an estimated 169,500 cases of cancer diagnosed in Australia in 2024, and almost one-third of Australians living in rural and remote areas, improving rural cancer care has the potential to benefit more than 55,000 people each year [5, 48].

5.2 Strengths and Limitations

This analysis included a large sample of rural cancer patients with a diverse range of cancer diagnoses. To our knowledge, this is one of the first studies to investigate rural cancer survivors' self-reported reasons for perceived delays in their pathway to initial cancer detection and treatment. A key limitation of this study is that self-reported reasons for perceived delays were not systematically assessed, rather were reported by participants in response to an open-ended question (i.e., ‘If yes, why?’). Therefore, not every reason or their true prevalence may have been captured. Furthermore, due to the wording of the question and how participants' responses were recorded, perceived delays relating to seeking medical attention could not be categorised as delays in the patient interval (i.e., from first symptom to first visit to a healthcare professional) or the primary care interval (i.e., from first visit to a healthcare professional to referral).

Additionally, all participants in this study had accessed subsidised accommodation in a major city during their treatment or follow-up care; therefore, compared to the broader population of rural cancer survivors, participants in this study may have been less likely to perceive delays in receiving their diagnosis and/or commencing treatment. The sample also only included a very small number of cancer survivors from remote or very remote areas (4%), and it is likely that those living in these areas would face additional challenges in seeking medical attention and accessing healthcare that may not have been fully captured in this analysis. Finally, it is not known whether the self-reported reasons for perceived delays in this study are rural-specific or applicable to cancer survivors more generally.