5th October 2018, Volume 131 Number 1483

Zoe Windner, Sue Crengle, Brandon de Graaf, Ari Samaranayaka, Sarah Derrett

Colorectal cancer (CRC) is a significant contributor to morbidity and mortality in Aotearoa/New Zealand, comprising the second-greatest subset of cancer registrations and deaths in 2013.1 Diagnosis can be complex and many patients present acutely.2 Existing research found 34% of New Zealanders with colon cancer first present to an emergency department (ED).3 Late-stage diagnoses are overrepresented in New Zealand; 24% of colon cancers in New Zealand are metastatic,3 compared with 19% and 17% in Australia and the UK respectively, despite the countries having comparable healthcare systems.4 New Zealand’s later staging at CRC diagnosis predicts poor outcomes, particularly for Māori.5

Diagnostic delay is implicated in this later staging, but research evidence is conflicting.6,7 Aggressive disease often muddles the relationship by producing more concerning symptoms, encouraging earlier presentation.8 However, it is known that those diagnosed while asymptomatic tend to have better prognoses.9 Delay is also known to cause psychological distress, further influencing outcomes.10,11 To assess delay, the Model of Pathways to Treatment provides an analytic framework, dividing the process into relevant intervals: symptom appraisal (internal to the patient), healthcare-seeking (external to the patient) and diagnosis.12 Holistic evidence is limited, but patient, care and disease factors appear consistently influential.13 Delays appears multifactorial and patient-centred, with attempts to rationalise symptoms and reluctance to seek care among diagnostic barriers identified in other countries.19,20

With the introduction of a nationwide bowel screening programme targeting those aged 60–74 years, New Zealand’s diagnostic process is in a state of change; currently 7 of 20 district health board regions are participating in the programme.14,15 Understanding the present nature of diagnostic facilitators and barriers is potentially useful in understanding the future impact of this screening programme. Additionally, the screening programme excludes those under 60 years of age; 21% of those diagnosed with CRC each year.1 Understanding diagnostic facilitators and barriers, from the perspective of the patient, may help ensure diagnostic pathways are as effective and efficient as possible.

In New Zealand, existing research begins at referral, usually by a general practitioner (GP). This identifies the most significant period in the diagnostic pathway as the time from referral from the GP to the first assessment by a specialist (FSA).16 However, specific events preceding the FSA have not yet been explored in New Zealand. Australian studies have focused on the perspectives of healthcare professionals (HCP).17,18 While research has broadened knowledge about diagnostic facilitators, like age and pain,17 research examining the patient perspective and experience of the diagnostic pathway is currently absent in New Zealand.

The overall aim of this research is to describe and begin to understand the experiences of patients diagnosed with CRC in Aotearoa/New Zealand. Specifically, our objectives are to:

  1. Describe the characteristics of participants diagnosed with CRC in our sample.
  2. Describe their pathways to diagnosis, their points of contact with the healthcare system and the patient experience at each of these points.
  3. Understand the experience of specific symptoms and the role of symptoms in the diagnostic pathway.
  4. Begin to understand factors which may influence the diagnostic pathway and any delays in diagnosis.

Methods

A comprehensive cross-sectional questionnaire was designed for online administration using LimeSurvey.21 Invitations were placed on two Facebook groups run by national patient and family charity, Bowel Cancer New Zealand (BCNZ). Invitations were also posted on the BCNZ website and newsletter. These were followed by a newspaper article.22 All interested potential participants were screened. Eligible participants were patients diagnosed in New Zealand, or their immediate support people, and aged over 18 years. The paper analyses responses from patients only.

The questionnaire included questions on demographics, CRC characteristics, help-seeking behaviours and the diagnostic pathway. Demographic information collected included age, gender, ethnicity, region of residence, education, family history of CRC, usual healthcare contacts (prior to diagnosis) and health insurance status. Ethnicity options were derived from Ministry of Health guidelines and level one prioritisation was used for analysis.23 Patient residential regions were grouped into Ministry of Health Cancer Networks.3

All patients were asked the year in which they were diagnosed, and this was categorised for analysis (grouped as 0–5 years, 6–10 years, or 11 or more years ago). All were also asked to report the site of their primary tumour as in the colon or rectum, their CRC stage at diagnosis, and the healthcare sector (public or private) in which the diagnosis was made.

Symptoms were assessed using multiple-choice checklists with free-format fields. Facilitators of, and barriers to, seeking help were recorded in the same way. Symptom concern was assessed on a visual analogue scale from 1 (not at all worrying) to 5 (extremely worrying). Symptom-to-diagnosis interval (SDI) was recorded for symptomatic patients. This was free-format and checked against the separate pathway timeline. SDI was then classified as less than six months, or six months or more.

Pathway questions incorporated patient experience factors from the New Zealand Health Survey and known obstacles to early cancer diagnosis.20,24–26 All respondents were asked a cycling set of questions on the patient pathway: the type of HCP, the patient experience and the HCP response; the cycle was repeated until diagnosis was reached. Patient experience questions were derived from the New Zealand Health Survey.24 Two of the six question themes were covered as follows: “At this time, how good was this person at listening to you?” and “At this time, how good was this person at taking your problems seriously?”. These employed a five-point Likert-like scale for responses: ‘very good’, ‘good’, ‘neither good nor bad’, ‘poor’ or ‘very poor’. A ‘good experience’ required responses to both questions to be ‘very good’ or ‘good’.

Data was checked for inconsistencies and duplicates, and analysed with Stata, version 15.1.27 Characteristics associated with delay in diagnosis were investigated using the Chi-squared test, Fisher’s exact test, and univariate and multivariate logistic regression.

Results

Sociodemographic information for the 98 patients is shown in Table 1. Our sample was younger, with more females than the population diagnosed with CRC in New Zealand.1

Table 1: Sociodemographic characteristics of study participants.

Characteristics

n (%)

Age at diagnosis

39 years or younger

40–49 years

50–59 years

60–69 years

70 years or older

Missing

17 (17)

24 (24)

31 (32)

18 (18)

6 (6)

2 (2)

Gender

Female

Male

Missing

76 (78)

21 (21)

1 (1)

Ethnicity

New Zealand European/Pākehā

Māori

Other

83 (85)

8 (8)

7 (7)

Region by Ministry of Health Cancer Network3

Northern (Northland, Auckland)

Midland (Waikato, Lakes, Bay of Plenty, Gisborne)

Central (Taranaki, Manawatu, Hawke’s Bay, Wellington)

Southern (South Island)

23 (23)

13 (13)

23 (23)

39 (40)

Table 2 presents factors relating to our participants’ diagnoses, symptoms and care. Most were diagnosed recently (within the last five years) after experiencing symptoms, and first sought care for these symptoms from a GP.

Table 2: Cancer and care characteristics reported by study participants.

Characteristics

n (%)

Time of diagnosis

0–5 years ago (2013–2018)

6–10 years ago (2008–2012)

11 or more years ago (2007 or earlier)

Tumour site:

Colon

Rectum

Unsure

Stage at diagnosis:

I

II

III

IV

Unsure

76 (78)

10 (10)

12 (12)

 

57 (58)

37 (38)

4 (4)

 

17 (17)

26 (27)

45 (46)

8 (8)

2 (2)

Symptomatic

 93 (95)

Symptoms

(multiple symptoms possible; does not add to 100%)

Change in bowel habit

Rectal bleeding

Abdominal pain

Unexplained weight loss

Low energy

Anaemia or iron deficiency

 

52 (53)

55 (56)

39 (40)

10 (10)

40 (41)

23 (24)

Factors encouraging healthcare-seeking

(multiple responses possible; does not add to 100%)

Worried about symptoms; unsure what they could represent

Worried about symptoms; suspecting bowel cancer

Worried about symptoms; suspecting another disorder

Acute exacerbation of symptoms

Routine medical appointment

Family or personal medical history

 

59 (60)

17 (17)

27 (28)

12 (12)

16 (16)

13 (13)

Barriers to healthcare-seeking

(multiple responses possible; does not add to 100%)

Acceptable explanation for symptoms

Waiting for symptoms to resolve without intervention

Found symptoms embarrassing or private

Lack of access

Sought care immediately; no barriers reported

 

41 (42)

27 (28)

12 (12)

3 (3)

25 (26)

First HCP contacted

General practitioner

Emergency department

Other

81 (83)

10 (10)

7 (7)

HCP visits before first specialist appointment (FSA)

0–1

2–3

4 or more

Missing

53 (54)

27 (28)

6 (6)

12 (12)

Most patients (n=72, 73%) reported more than one symptom and several symptom combinations appeared important. Of those experiencing altered bowel habit, 35 (67%) also reported rectal bleeding; 17 (74%) of those reporting anaemia or iron deficiency also reported low energy; and 8 (80%) of those experiencing unexplained weight loss also reported abdominal pain. The most common trigger symptom (immediately preceding the first HCP contact) was rectal bleeding (n=27, 28%).

Symptom concern was greatest for a palpable mass, where the mean level of concern was 4.5 on the five-point scale. Symptom concern was least for anaemia or iron deficiency and low energy, where the mean level was 3.1 on the five-point scale.

The most common facilitator of healthcare-seeking behaviours was worry about symptoms, although most participants reported not knowing what their symptoms could represent. The most common barrier to healthcare-seeking behaviours was providing an alternative explanation for symptoms experienced. Encouragingly, few reported embarrassment about symptoms, but most who did were experiencing an altered bowel habit (n=10, 26%) or rectal bleeding (n=11, 27%).

Most patients first approached someone who was not a HCP (n=62, 77%) and the first HCP approached was the GP in most cases. For 17 (17%) patients, the reported pathways to diagnosis were incomplete or unclear. The remaining 81 (83%) patients reported a clear and complete pathway to diagnosis and these are shown in Figure 1, where the area of each circle is proportional to the number of patients at each step. This figure has been simplified so that only pathways followed by two or more patients (total n=55, 56%) have been shown. Where additional steps to diagnosis have been omitted, as in the remaining 26 cases, this is shown with a dotted line.

Figure 1: Pathway map showing the patient-reported diagnostic process (circle area proportional to n). 

c 

The most common pathway shown is linear—patients first approached a non-HCP, then a GP, then a specialist, where they received their diagnosis. However, less direct pathways are also present, with repeated visits to one provider.

Twenty-three (25%) patients reported an SDI of less than three months, 41 (44%) less than six months and 66 (71%) less than 12 months. Table 3 reports associations between patient-reported characteristics and SDI.20,25,26 It appears that younger age, less formal education, poor/neutral first healthcare experience, higher number of visits and diagnosis in the public healthcare sector are associated with longer SDI.

Table 3: Patient characteristics according to patient-reported symptom-to-diagnosis interval (SDI).

Patient characteristics

SDI less than six months

n (%)*

SDI of six months or longer

n (%)*

OR (95% CI)**

P value**

Age

<50 years

50–59 years

60 years or older

13 (32)

11 (27)

15 (37)

28 (49)

20 (35)

9 (16)

1.00

0.84 (0.31–2.26)

0.28 (0.10–0.80)

0.05

 

Ethnicity

NZ European / Pākehā

Māori

Other

36 (88)

1 (2)

4 (10)

47 (82)

7 (12)

3 (5)

1.00

5.36 (0.63–45.56)

0.57 (0.12–2.73)

0.23

Any tertiary qualification

No tertiary qualification

34 (83)

6 (15)

37 (65)

20 (35)

1.00

3.06 (1.10–8.53)

0.03

 

No family history of CRC

Any family history of CRC

26 (63)

15 (37)

36 (63)

21 (37)

1.00

1.01 (0.44–2.33)

0.98

Tumour site

Colon

Rectum

Unsure

26 (67)

13 (33)

2 (5)

31 (56)

24 (44)

2 (4)

1.00

0.84 (0.11–6.37)

 

0.57

 

Stage at diagnosis

I or II

III or IV

Unsure

20 (49)

21 (51)

0 (0)

23 (42)

32 (58)

2 (4)

1.00

1.33 (0.59–2.99)

0.50

 

Asymptomatic

1 symptom

>1 symptom

1 (2)

20 (49)

20 (49)

4 (7)

20 (35)

33 (58)

4.80 (0.48–47.68)

1.00

2.07 (0.87–4.91)

0.16

No intermittent symptoms

Any intermittent symptom

17 (41)

23 (56)

14 (25)

39 (68)

1.00

1.70 (0.74–3.90)

0.21

Good experience at first HCP appointment

Poor or neutral experience at first HCP appointment

36 (88)

5 (12)

33 (58)

24 (42)

1.00

3.70 (1.25–10.96)

0.02

<3 HCP visits before diagnosis

3 or more HCP visits before diagnosis

25 (61)

12 (29)

21 (37)

28 (49)

1.00

2.78 (1.14–6.77)

0.02

 

No healthcare insurance

Healthcare insurance

19 (46)

19 (46)

30 (53)

22 (39)

1.00

0.73 (0.30–1.70)

0.47

Diagnosed in the private system

Diagnosed in the public system

18 (44)

19 (46)

12 (21)

40 (70)

1.00

3.16 (1.27–7.86)

0.01

*Do not always sum to 100% because of missing values.
**Excludes missing responses. P-values refer to the association between the outcome and whole variable and were obtained using univariate logistic regression with SDI of six months or more as the outcome. 

With one exception, there were no significant changes in any of the associations presented in Table 3 after adjusting for age using multivariable logistic regression. The exception is the significance of the patient experience at the first HCP appointment, which moved to P=0.06 after adjustment.

Age was also classified as in the new screening programme, where those eligible are aged between 60 and 74 years.14 Those inside this age group (aged 60–74 years) were less likely to experience delay (n=8, 38% with SDI six months or longer) than those outside this age group (aged <60 or >74 years), where 48 (67%) reported delay (P=0.02).

Discussion

This study appears to be the first to present patient perspectives of the pathway to a diagnosis of CRC in New Zealand. The majority of patients sought care for general symptom worry. While few suspected bowel cancer, those who did were most likely to be experiencing a change in bowel habit, rectal bleeding or both. Paradoxically, patients with these symptoms were also most likely to feel embarrassed about their symptoms. While this did not appear to influence delay (overall or in healthcare-seeking), these are important perspectives to focus on in public and HCP awareness programmes.

Most patients first sought the opinion of a non-HCP (usually a partner or friend) before turning to their GP. Non-specific symptom concern appears to be the most important facilitator of this step. Most visits in each patient’s pathway were with this GP. This puts a resourcing and accessibility target on primary care. Very few patients in our study were unable to access an FSA after consulting a GP. However, many reported considerable delay between referral and this scheduled FSA. In this time between referral and the scheduled FSA, many decided to seek other care (eg, through a GP or ED) because of their symptoms.

Delay was significant in our sample. We found those ineligible for the new screening programme to be more likely to report delay; but in our sample, these people were nearly all younger than the age of entry to the bowel screening programme. Age is already built into procedural criteria.28 The new screening programme may further constrain the diagnostic resources available to this group, which may exacerbate this delay.29 Care is needed to ensure timely diagnosis for this screening-ineligible younger group, a significant proportion of those diagnosed with CRC each year.1

Our study found those reporting a good experience at their first appointment were less likely to report delay than those reporting a poor or neutral first experience. However, the cause-and-effect nature of this relationship is unclear. It is possible that experiencing delay makes a patient more likely to report poor or neutral experiences and more research is needed to clarify this.

Those diagnosed in the publicly-funded healthcare system were more likely to report delay. However, delay does not appear to be associated with insurance status. Those in public care do not seem to seek care later, indicating the difference may lie in the care received or resources available in the public system. Regardless of insurance status, it is likely that socio-economic status is related to this finding—whereby wealthier New Zealanders are able to afford to pay for private healthcare services out-of-pocket to hasten diagnosis (eg, private colonoscopy or CT colonography) and poorer New Zealanders do not have such access, leading to inequities. While it is beyond the scope of this descriptive study to explore this further, this warrants further investigation in New Zealand.30 More research is also required into the relationship between ethnicity and delay, particularly in Māori. Our study indicates a tendency towards diagnostic delay for Māori, but our sample size is underpowered to determine the strength of this association. Similarly, and importantly, there is also a tendency towards delay for patients experiencing intermittent symptoms. Future, larger studies should investigate this, because any such relationships are identified this would be important for public awareness programmes and planning.

The primary limitation in this descriptive study is our sampling. Our sample was younger, with more females and fewer metastatic cases than the population diagnosed with CRC in New Zealand.1 This introduces potential confounders, the effects of which are unknown. Given that our participants are not representative of the New Zealand population diagnosed with CRC, it is important for further studies with larger, more representative samples to be undertaken. We also relied on self-reported data, which we could not verify against routinely collected data sources. Several relationships, such as those between SDI and ethnicity or symptom intermittency, did not reach statistical significance in our sample but appear worthy of investigation in future studies.

Despite these limitations, we have contributed first insights into the patient perspective of the pathway to diagnosis for a major cancer in Aotearoa/New Zealand. Our findings are potentially relevant for three main groups: the general population, HCPs and policymakers. Firstly, for individuals in the general population, although symptoms were the primary driver of healthcare-seeking, attribution of non-specific symptoms to a serious cause was uncommon in our sample. This suggests that greater awareness of common symptoms may be required, which can direct future education initiatives. It is also possible that awareness about the importance of reporting intermittent symptoms to HCPs is also a useful focus for public awareness campaigns. Secondly, for HCPs, we have confirmed general practice as an important source of care in the diagnostic pathway. Delays in diagnosis appear prevalent, and several groups appear to be at particular risk. Careful attention must be paid to Māori, those aged <60 years and those with less formal education. Finally, for policymakers, we have provided first insights into the patient pathway and have suggested areas worthy of investigation in future studies. Associations between delayed diagnosis and publicly funded care in particular ought to be investigated further.

Summary

Bowel cancer is a significant health issue in New Zealand, with over 3,000 people diagnosed each year, but little was known about the process of diagnosis. We surveyed a small group, of mostly younger patients, to gain some insight into this process for the first time. Most of our sample first approached a layperson, then a general practitioner. The majority of our sample took over six months to reach diagnosis; this was more likely if aged under 60 years, without a tertiary qualification, or were diagnosed in the public sector. Our research is not representative of the whole population diagnosed with bowel cancer, but provides some early indicators and highlights the potential for improvement in the healthcare system.

Abstract

Aim

To understand colorectal cancer (CRC) symptoms experienced by Aotearoa/New Zealand patients and to describe patient-experienced pathways and factors which may be associated with delayed diagnosis.

Method

Ninety-eight patients diagnosed with CRC, recruited via a national charity, completed a questionnaire. Questions included demographics, symptoms, help-seeking and diagnostic pathways followed.

Results

Of 98 participants, 72 (73%) were aged under 60 years; most were symptomatic (n=93, 95%) and first discussed symptoms with someone who was not a healthcare professional (HCP) (n=71, 79%). The first HCP approached was usually a general practitioner (n=81, 83%). Symptom-to-diagnosis interval (SDI) was often six months or more (n=52, 56%) among our younger cohort. Delay was more likely if patients were younger (P=0.05), without a tertiary qualification (P=0.03), reported a poor/neutral experience at their first related HCP appointment (P=0.02), or were diagnosed in the public sector (P=0.01).

Conclusion

Few patients initially suspected bowel cancer or reported embarrassment seeking care; those who did were most likely to experience changes in bowel habit or bleeding. Our study is small, and not representative of all those diagnosed with CRC in New Zealand; yet it provides important first insights into patients’ diagnostic experiences.

Author Information

Zoe Windner, Student, Dunedin School of Medicine, University of Otago, Dunedin;
Sue Crengle, Associate Professor, Department of Preventive and Social Medicine, University of Otago, Dunedin; Brandon de Graaf, Data Manager/Programmer, Department of Preventive and Social Medicine, University of Otago, Dunedin; Ari Samaranayaka, Biostatistician, Biostatistics Unit, Dunedin School of Medicine, University of Otago, Dunedin; Sarah Derrett, Professor, Department of Preventive and Social Medicine, University of Otago, Dunedin.

Acknowledgements

Zoe Windner was supported by a Bowel Cancer New Zealand Summer Student Scholarship.

Correspondence

Sarah Derrett, Department of Preventive and Social Medicine, Dunedin School of Medicine, PO Box 56, Dunedin 9054.

Correspondence Email

sarah.derrett@otago.ac.nz

Competing Interests

Professor Sarah Derrett is a member of the Executive of Bowel Cancer New Zealand and Associate Professor Sue Crengle is a medical advisor to Bowel Cancer New Zealand, both in unpaid and voluntary capacities. Dr de Graaf reports affiliation with Bowel Cancer New Zealand during the conduct of the study. Ms Windner reports grants from Bowel Cancer New Zealand during the conduct of the study.

References

  1. Ministry of Health. New Cancer registrations 2015 [Internet]. Wellington, New Zealand: Ministry of Health. 2017 Dec 14 [cited 2018 Feb 16]. Available from http://www.health.govt.nz/publication/new-cancer-registrations-2015
  2. Cubiella J, Valentín F, Vega P. Colorectal cancer diagnosis: Pitfalls and opportunities. World J Gastrointest Oncol. 2015 Dec 15; 7(12):422–433.
  3. Jackson C, Sharples K, Firth M, et al. The PIPER Project: An Internal Examination of Colorectal Cancer Management in New Zealand. 2015 Aug 7 [cited 2018 Feb 16]. Available from http://www.fmhs.auckland.ac.nz/assets/fmhs/sms/ctnz/docs/THE%20PIPER%20PROJECT%20Final%20deliverable%20report%207%20August%202015%20(HRC%2011_764%20FINDLAY).pdf
  4. Maringe C, Walters S, Rachet B, et al. Stage at diagnosis and colorectal cancer survival in six high-income countries: A population-based study of patients diagnosed during 2000-2007. Acta Oncologica. 2013 Apr 15; 52(5):919–932.
  5. Hill S, Sarfati D, Blakely T, et al. Survival disparities in Indigenous and non-Indigenous New Zealanders with colon cancer: the role of patient comorbidity, treatment and health service factors. J Epidemiol Community Health. 2010 Jan 7; 64:117–123.
  6. Tørring ML, Murchie P, Hamilton W, et al. Evidence of advanced stage colorectal cancer with longer diagnostic intervals: a pooled analysis of seven primary care cohort comprising 11,720 patients in five countries. Br J Cancer. 2017 Sep 5; 117(6):888–897.
  7. Ramos M, Esteva M, Cabeza E, et al. Relationship of diagnostic and therapeutic delay with survival in colorectal cancer: A review. Eur J Cancer. 2007 Nov; 43(17):2467–2478. 
  8. Crawford SC, Davis JA, Siddiqui NA, et al. The waiting time paradox: population based retrospective study of treatment delay and survival of women with endometrial cancer in Scotland. Br Med J. 2002 Jul 27; 325:196.
  9. Neerincx M, Buffart TE, Mulder CJJ, et al. The future of colorectal cancer: implications of screening. Gut (Br Med J). 2013 Sep 14; 62(10):1387–1389.
  10. Miles A, Mcclements PL, Steele RJ, et al. Perceived diagnostic delay and cancer-related distress: a cross-sectional study of patients with colorectal cancer. Psycho-Oncology. 2016 Feb 11; 26(1):29–36.
  11. Engel GL. The Need for a New Medical Model: A Challenge for Biomedicine. Sci. 1977 Apr 8; 196(4286):129–136.
  12. Scott SE, Walter FM, Webster A, et al. The Model of Pathways to Treatment: Conceptualisation and integration with existing theory. Br J Health Psychol. 2012 Apr 27; 18(1):45–65.
  13. Olesen F, Hansen RP, Vedsted P. Delay in diagnosis: the experience in Denmark. Br J Cancer. 2009 Dec 3; 101 Suppl 2:S5–S8.
  14. National Screening Unit, Ministry of Health. About the National Bowel Screening Programme. [Internet] Wellington, New Zealand: Ministry of Health. Available from: http://www.timetoscreen.nz/bowel-screening/about-the-national-bowel-screening-programme/ [cited 2018 Sep 28].
  15. Love T, Poynton M, Swansson J. The cost effectiveness of bowel cancer screening in New Zealand: a cost-utility analysis based on pilot results. Wellington, New Zealand: Ministry of Health. 2016 Jul [cited 2018 Feb 16]. Available from http://www.health.govt.nz/system/files/documents/publications/appendix4-cost-utility-analysis-based-on-findings-of-the-pilot-results.pdf
  16. Murray M, Brown J, Hinder V, et al. The colorectal cancer patients’ journey: the Auckland region. N Z Med J. 2011 Mar 25; 124:1331.
  17. Lynch BM, Baade P, Fritschi L, Leggett B. Modes of presentation and pathways to diagnosis of colorectal cancer in Queensland. Med J Australia; Pyrmont. 2007; 186(6):288–291.
  18. Harris MF, Pascoe S, Crossland L, et al. Patients with colorectal cancer: A qualitative study of referral pathways and continuing care. Australian Family Physician. 2012 Nov; 41(11):899–902.
  19. Hall N, Birt L, Banks J, et al. Symptom appraisal and healthcare-seeking for symptoms suggestive of colorectal cancer: a qualitative study. Br Med J. 2015 Oct 9; 5(10).
  20. Langenbach MR, Schmidt J, Neumann J, Zirngibl H. Delay in Treatment of Colorectal Cancer: A Multifactorial Problem. World J Surg. 2003 Mar; 27:304–308.
  21. Weller D, Vedsted P, Rubin G, et al. The Aarhus statement: improving design and reporting of studies on early cancer diagnosis. Br J Cancer London. 2012 Mar 27; 106(7):1262–1267.
  22. Goodwin E. Bowel symptoms often ignored [Internet]. Dunedin, New Zealand: Otago Daily Times. 2018 Jan 11 [cited 2018 Feb 16]. Available from http://www.odt.co.nz/news/dunedin/bowel-symptoms-often-ignored#comment-7008
  23. Health Information Standards Organisation. Ethnicity Data Protocols [Internet]. Wellington, New Zealand: Ministry of Health. 2017 Oct 6 [cited 2018 Feb 26]. Available from http://www.health.govt.nz/system/files/documents/publications/hiso-10001-2017-ethnicity-data-protocols.pdf
  24. Ministry of Health. Patient Experience 2011/12: Key findings of the New Zealand Health Survey [Internet]. Wellington, New Zealand: Ministry of Health. 2013 Sep 12 [cited 2018 Feb 16]. Available from http://www.health.govt.nz/publication/patient-experience-2011-12
  25. Siminoff L, Thomson M, Dumenci L. Factors associated with delayed patient appraisal of colorectal cancer symptoms. Psycho-Oncology. 2014 Feb 26; 23(9):981–988.
  26. Salander P, Bergenheim AT, Hamberg K, Henriksson R. Pathways from symptoms to medical care: a descriptive study of symptom development and obstacles to early diagnosis in brain tumour patients. Family Practice (Oxford University Press). 1999 Apr 1; 16(2):143–148. 
  27. StataCorp. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC. 2017.
  28. Ministry of Health. National Referral Criteria for Direct Access Outpatient Colonoscopy – updated November 2015 [Internet]. Wellington, New Zealand: Ministry of Health. 2015 Nov [cited 2018 Feb 16]. Available from http://www.health.govt.nz/our-work/diseases-and-conditions/cancer-programme/bowel-cancer-programme/background-reports
  29. Robertson L, Roberts G, Cooney T, et al. Bowel cancer screening. The New Zealand Medical Journal [Internet]. 2017 Sep 1[cited 2018 Feb 16];130(1461). Available from http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2017/vol-130-no-1461-1-september-2017/7356
  30. Sharples K, Firth M, Hinder V, Hill A, Jeffrey M, Sarfati D, Brown C, Atmore C, Lawrenson R, Reid P, Derrett S, Macapagel J, Keating J, Secker A, De Groot C, Jackson C, Findlay M. (2018) The New Zealand PIPER Project: Colorectal cancer survival according to rurality, ethnicity and socioeconomic deprivation – results from a retrospective cohort study. N Z Med J;131(1476):24-39: ISSN 1175-8716.