Why does New Zealand have such poor outcomes from colorectal cancer?: the importance of the pre-diagnostic period

KEYwords: Bowel cancer; equity; primary health care

Colorectal cancer (CRC; cancer of the colon, rectosigmoid and rectum, or bowel cancer) is common in New Zealand (NZ), with over 3000 new cases diagnosed annually.¹ Survival post-diagnosis depends on the extent of disease (stage) at diagnosis, ranging from a 90% 5-year relative survival rate for early stage disease (localised disease, cancerous cells confined to the colon or rectum, American Joint Committee on Cancer (AJCC) stage I, IIA and IIB), to 14% for late or advanced-stage disease (distant disease, cancerous cells found in other organs or distant lymph nodes, AJCC stage IV).² Diagnosis at an early stage and subsequent intervention are critical to ensuring positive outcomes for patients.

The distribution of stage at diagnosis for NZ patients diagnosed with CRC has been published by the PIPER (Presentations, Investigations, Pathways, Evaluation and Rx) project, the largest study of CRC in NZ to date.³ For colon cancer, the distribution of disease stage at diagnosis was 12% stage I, 27% stage II, 25% stage III and 24% stage IV. For rectal cancer (reported as non-metastatic vs. metastatic only), 19% had metastatic disease at diagnosis.⁴ These distributions are comparable to that of an unscreened United Kingdom population.⁵ Stage at diagnosis for NZ patients also varies by ethnicity, with Māori and Pacific patients having higher proportions of late-stage disease than non-Māori/non-Pacific (35%, 31% and 23% respectively).

Disease stage at diagnosis, survival outcomes and the poor distribution of stage at diagnosis, may explain why CRC survival in NZ is poor among international comparisons, particularly when compared to Australia.^6–8 The CONCORD-3 study (an international comparison of 18 cancers across 71 countries) reports the 5-year net survival from colon and rectal cancer in 2010–14 as 64% and 66% respectively in NZ versus 71% for both in Australia.⁶ Similar 5-year net survival rates were also reported from the International Cancer Benchmarking Partnership SURVMARK-2 study, ranking NZ the third worst for survival from both colon and rectal cancer out of seven countries (Australia, Canada, Denmark, Ireland, NZ, Norway and the United Kingdom).⁷

These large-scale international studies support previous findings of NZ-based researchers who found that the 5-year relative survival for 2006–10 was 5% less than in Australia.⁸ Survival post-diagnosis also varies depending on patient ethnicity and socioeconomic status. The PIPER project identified significant survival disparities for Māori and Pacific patients, and for people living in areas of high deprivation.⁴ When investigating survival inequities between ethnic groups, controlling for disease stage significantly reduced the disparity for Māori patients, confirming the importance of early diagnosis in this population.⁴

Indicators of deficiencies in the pre-diagnostic pathway for CRC include diagnosis being made via emergency department (ED) and obstructive disease at initial diagnosis. In the PIPER study, 31% of patients were diagnosed following ED presentation, and 19% with obstruction.⁴ These indicators were worse for Māori patients living in areas with the greatest deprivation (socioeconomic status) and for rural patients.⁴ A NZ report on national performance indicators for bowel cancer between 2013 and 2016 found that 26% of patients were diagnosed following ED presentation and that this varied by District Health Board (DHB),⁹ and was higher for people aged <50 or >75 years, Māori, Pacific, and people living in areas of high social deprivation, confirming that inequalities in access to primary care and diagnostic services exist.

A staged roll out of a national bowel screening programme has been under way in NZ since July 2017, following a 6-year pilot in Waitemata DHB.¹⁰ At the time of writing, 10 of the 20 DHBs are participating in the programme (Hutt Valley, Wairarapa, Waitemata, Southern, Counties Manukau, Nelson Marlborough, Hawkes Bay, MidCentral, Whanganui and Lakes).¹⁰ However, even with a screening programme, most bowel cancers are still diagnosed symptomatically and limitations to access remain, including the age band covered by the programme (60–74 years) and disparities in participation.¹⁰ CRC occurs across all age groups, and there are patient groups who are more likely to be diagnosed at a younger age, particularly Māori and Pacific peoples.³ A position statement from Te Ohu Rata O Aotearoa, Māori Medical Practitioners Association, highlights that more than half of all cases of CRC occurring in Māori patients are diagnosed before age 60 years.¹¹ As they emphasise, ignoring the different distributions in age at diagnosis between populations will result in increased inequities for Māori patients diagnosed with CRC.¹¹ The incidence of CRC in patients aged <50 years is increasing in both the NZ population¹² and internationally.² Thus, the known limitations of screening coverage combined with increasing incidence in younger patients is reflected in our high rates of late stage at diagnosis and poor survival outcomes. This makes it important to deepen our understanding of the pathway to diagnosis for NZ patients diagnosed with CRC, and what we can do to intervene.

The aim of this paper is to identify and summarise research undertaken in NZ to investigate factors affecting the pre-diagnostic period for patients with CRC, which may contribute to late stage at diagnosis and poor survival.

This scoping review searched for published studies of factors contributing to late stage at diagnosis that included NZ patients’ data. Both qualitative and quantitative studies were included. Original research articles examining the pre-diagnostic period for patients diagnosed with CRC in NZ published between 2009 and 2019 were searched using the PubMed Central database. The pre-diagnostic period was defined as the time between the discovery of symptoms (or receipt of the invitation letter for bowel screening) and diagnosis. A Medical Subject Heading (MeSH) term search of the PubMed database for ‘colorectal neoplasms’ was combined with additional headings or subheadings including ‘diagnosis’ and ‘primary care’ and ‘New Zealand’ (see Appendix 1). Abstracts were reviewed for all articles, where available.

Editorials, letters to editors, and review articles were excluded. Full-text articles for all relevant studies were obtained, reviewed and data abstracted by one author (MJF). Reference lists of the full-text articles were also reviewed to identify any additional studies to be assessed for inclusion.

Data were abstracted into a pre-populated proforma for each study. Results were considered within the Model of Pathways to Treatment framework (Fig. 1),¹³ an internationally recognised theoretical framework for examining pathways to diagnosis. The framework considers four key intervals in the pre-treatment period: appraisal, help-seeking, diagnostic and pre-treatment. The framework allows for the consideration of contributing factors (patient, health-care provider and systems, and disease) and their impact on the intervals as facilitating or impeding progress through the pathway.¹³ For this study, only the first three intervals are relevant.

Figure 1. Model of Patient Pathways to Treatment. Reproduced from Walter et al. 2011 with permission.²⁸

Database searching yielded 83 results. Following removal of duplicate records (n = 22) and 53 exclusions, eight relevant studies were reviewed (see Fig. 2). Reasons for exclusion were (in frequency order): study examining post-diagnostic pathway, diagnostic test parameters or secondary care; editorial; study examining a related diagnosis; letter to editor; review article; clinical guideline document; health economics study; pharmacy-based study; summary paper post conference. One search result could not be accessed for review and one article was identified through reference searching. Included studies were mainly qualitative (five of eight) and conducted >5 years ago (pre-implementation of the screening pilot). Five studies addressed research questions specific to CRC screening. However, the topics explored in these studies included factors that are relevant to the appraisal, help-seeking, and diagnostic intervals, hence their inclusion in this study. Four studies included Māori in their design (see Table 1). Collated findings are grouped into subheadings based on the intervals of the Models of Pathways to Treatment framework.

Figure 2. PRISMA 2009 diagram.

Table 1. Summary of included studies

Studies examining perceptions to CRC screening identified the need to raise awareness of CRC in the public profile.^14–17 They suggested that a multiple media campaign to raise awareness of CRC was necessary and could address many of the perceived inhibitory factors to screening, including patient factors surrounding reticence and concern regarding ability to collect faecal specimens, and health-system factors including perceived poor test reliability. Disease factors relating to lack of specific symptoms and perceived slow development of CRC were seen by patients as positive reasons to undergo screening.

In a qualitative study by Windner et al., 95% of participants reported being symptomatic, with 73% reporting more than one symptom. The most common ‘trigger’ symptom was rectal bleeding.¹⁸ In considering the pathways within this interval, this research found that most patients consulted a non-health-care professional before consulting a health-care professional, usually a general practitioner (GP).

The critical role of GPs in CRC diagnosis and screening was re-emphasised multiple times by earlier studies examining screening perceptions. No studies aimed to quantify the specific time frame of this interval, but one study examined time frames that included this interval.¹⁸ Windner et al. captured self-reported symptom-to-diagnosis interval for all symptomatic patients in their cohort. This time frame includes the appraisal period, the help-seeking interval; and the diagnostic interval: 25% reported <3 months, 44% <6 months and 71% <12 months. Patients aged <50 years were statistically significantly more likely to report a symptom-to-diagnosis interval of ≥6 months than patients in the screening programme range of ≥60 years.¹⁸ However, although this study is the most recent and one of the most in-depth, its design means it is probably not representative of the NZ population.

Windner et al. asked direct questions about the help-seeking interval.¹⁸ Disease factors identified as facilitating help-seeking behaviour were non-specific symptom concern. Conversely, an acceptable alternative benign explanation for symptoms was the most commonly identified inhibitory factor. Raising public awareness of CRC in the appraisal interval would likely also have an impact on the help-seeking interval, as would the role and relationship with patients’ GPs. No studies attempted to quantify the time frame of this interval, while recognising the challenges of measuring this interval specifically.

Windner et al. reported 54% of participants had zero to one and 6% had four or more visits with a health-care professional before diagnosis. The two quantitative studies largely focused on this interval. Tiong et al. compared their cohort to national and international targets for wait-times between referral to colonoscopy and referral to first treatment, and found that 44% and 21% met the 42 day and 62 day targets respectively. They also identified an increased pre-hospital delay (symptom onset to first specialist appointment) for patients with systemic symptoms and altered bowel habit.¹⁹ Murray et al. also report on the period between referral to first treatment, with 68% meeting the comparative UK target of 62 days (median length 35 days).²⁰ There were no significant differences between 2001 and 2005 cohorts or by ethnicity. The greatest delays in this study were seen in the interval from initial referral to first specialist appointment.²⁰

This study found that little research has been undertaken about the pre-diagnostic period and its relationship to late diagnosis of CRC in NZ patients. Most studies we found are nearly 10 years old and repeatedly highlight the need for increased public awareness of CRC in NZ to assist self-appraisal, help-seeking and screening participation. They also emphasised the fundamental role GPs and primary health play in a CRC diagnosis and facilitating screening. Qualitative studies demonstrated a failure to meet national and international targets for timeliness, particularly when looking at the period from referral to first specialist appointment, diagnosis or treatment, although delays were not shown to be associated with late-stage diagnosis. A comprehensive mixed-methods approach, including analysis of time frames and qualitative assessment of factors influencing these time frames, has not been undertaken by any one study. Many gaps exist in our understanding of patient, health-care provider, system and disease factors that facilitate or inhibit the pathway to diagnosis for patients diagnosed with CRC in NZ. This study also highlights the lack of information on Māori and Pacific populations, who have poorer CRC outcomes.

Relevant research is being conducted internationally. The International Cancer Benchmarking Partnership is a collaboration to explore population- and health-care-related factors affecting cancer survival outcomes in Australia, Canada, Denmark, Norway, Sweden and the UK.²² Published work to date has studied primary care physician-reported access to investigations, timeliness of test results and wait times for secondary care specialist assessment, and the readiness of primary care physicians to investigate or refer to secondary care following symptoms indicative of cancer.²³ They have also reported research on diagnostic routes and time intervals from first symptom to initiation of treatment.²⁴ Both topics were assessed for differences between the six study countries and impact on reported survival figures. These large-scale studies identified a correlation between readiness to refer or investigate suspected cancer symptoms for CRC and survival,²³ and showed that wide variations in time intervals exist between the countries, suggesting that improvements could be made in expediting diagnoses.²⁴ They were unable to establish any correlation between greater time intervals and survival (countries with poorer survival did not consistently have longer time intervals).²⁴ The authors of both studies acknowledged the need for more detailed examination and understanding of factors affecting readiness to refer (including changing access to investigations, quality and utility of clinical guidelines, and relationships between primary and secondary care)²³ and length of time to diagnosis (noting that, in many cases, longer periods included more investigations).²⁴

Many factors influencing the pre-diagnostic pathway are likely to be population- and health-system-specific. A 2014 NZ study used the International Cancer Benchmarking Partnership survey instrument to survey 192 GPs about a range of cancer types and found that NZ GPs have poor access to colonoscopy compared to other jurisdictions with similar primary care-led health services.²⁵ This work also suggested that NZ GPs are less likely to refer patients at risk of CRC, although it could not address why this may be. Perhaps poorer access to colonoscopy means that GPs are more reluctant to refer and apply a higher threshold before referring for colonoscopy. The critical role of GPs and primary care was highlighted by several studies included in this review and is identified by the Ministry of Health as being ‘key’ in the success of the bowel screening programme.¹⁰ Accordingly, we urge support to facilitate GPs in the CRC pre-diagnostic pathway more effectively, through improving our knowledge and understanding of the current inhibitory factors and implementing evidence-based changes to mitigate these factors and improve timely diagnosis for patients.

Perceived delay in CRC diagnosis is important to NZ patients. The 2015 Health and Disability Commissioner report on delayed diagnosis of cancer in primary care between 2004 and 2013 indicated that delays in diagnosing CRC were a large source of complaint and were over-represented relative to its incidence in the population.²⁶ Of 197 complaints, 54 (27%) pertained to a diagnosis of CRC.²⁶ The report suggests that the most common issue for CRC complaints related to non-specific or atypical symptoms presentation.²⁶ However, absence of appropriate examination where symptoms were present was significantly associated with delayed CRC diagnosis.²⁶ For 72% of the CRC cases reviewed, the outcome was death or terminal illness, further emphasising impact of the late stage of diagnosis of CRC in NZ.²⁶ The report found that the total number of cancer complaints made to the Health and Disability Commissioner over the 10-year period significantly increased from 2004 to 2013.²⁶ Although this report is now 6 years old, it is likely that similar issues still exist, as evidenced by a 2019 article from the Associate Commissioner, Jane King, in NZ Doctor, describing a case seen four times over a 9-month period, initially for perianal itch and irritation, progressing to rectal bleeding and change in bowel habit.²⁷ Failure to conduct a rectal examination and insufficient clinical records were found by the Health and Disability Commissioners’ clinical advisor to be a breach of the NZ code of Health and Disability Consumers’ Rights.²⁷ Clear pathways and interventions based on knowledge of facilitatory and inhibitory factors to diagnosis, along with adequate support and prompt and appropriate follow-through from the secondary care sector, are needed to support the primary sector in this crucial role.

The authors are currently undertaking a project using both quantitative and qualitative research methods to examine the pre-diagnostic period for patients diagnosed with CRC in the Midland region of NZ. We hope that this research will identify where the greatest barriers are in the pre-diagnostic period, to drive targeted interventions to reduce late stage diagnosis of CRC. Given the critical role of GPs in the CRC pre-diagnostic pathway, this research, along with any subsequent potential interventions, will include input from colleagues working in primary care.

There is a paucity of recent research examining the pre-diagnostic period for patients in NZ diagnosed with CRC. Given the poor distribution in NZ of stage at diagnosis and survival outcomes known by international comparisons, inequities in outcomes by ethnicity, limitations of the current screening programme, differing age distributions for Māori and Pacific populations, and increasing rates of CRC diagnosis at younger ages, greater understanding is needed of how we can improve stage at diagnosis, via thorough examination of the pre-diagnostic pathway and implementation of facilitatory factors. Work to date highlights the critical role of GPs in this pathway, and the need for carefully designed and evaluated public awareness campaigns for CRC.

The authors declare no competing interests.

This study was supported by the Health Research Council of New Zealand (HRC grant number 17/147).