Abstract
Background/Aim: To investigate factors that affect colorectal polyp or colorectal cancer (CRC) detection amongst patients referred urgently to colorectal services for suspected bowel cancer. Patients and Methods: This was a prospective observational study at a UK colorectal centre (2017-2018). Logistic regression determined odds ratios for colorectal polyp or CRC according to age, gender, previous polyp or cancer, and the 6 NICE referral (NG12) categories. Results: A total of 605 patients were included in the study; median age 66 (IQR=54-76); 47.9% male. Nineteen (3.1%) patients had CRC and 64 (10.6%) had polyps. No individual variable increased the likelihood of CRC detection, but male patients had a higher likelihood of having either polyp or CRC (OR=1.72; 95%CI=1.07-2.80; p<0.05). Conclusion: At the point of an urgent referral to a colorectal clinic, the likelihood of CRC detection appears to be unaffected by age, gender, or any individual referral criterion. However, overall disease detection may be more likely amongst male patients.
In the UK, best practice guidelines from the National Institute for Health and Care Excellence (NICE) state that patients should be referred by their General Practitioner to a colorectal clinic urgently [i.e. within the 2-week wait (2WW) pathway] if they have ‘red flag’ symptoms, signs, or blood test results that might indicate risk of colorectal or anal cancer (1). The aim of the pathway is to identify and treat patients with colorectal or anal cancer as soon as possible, without any of the delays that might be expected during a more routine referral. Patients referred by this pathway are usually assessed by a colorectal surgeon, who decides whether further endoscopic or radiological investigation is warranted. Alternatively, patients may bypass this clinic and go straight-to-test (2-4). Cancer detection rates amongst patients referred by the 2WW pathway range from 4% (5) to 14% (6), with an average of 7.7% (7). The average polyp detection rate has been reported to be 10.8% (7).
In 2015, NICE reduced the threshold for 2WW referral criteria in order to improve the detection of colorectal cancers in patients that might otherwise not fulfil the previous criteria. Iron-deficiency anaemia, occult faecal blood, and suspected masses palpated in the rectum, anus or abdomen were added to the list of criteria alongside changes in bowel habit and weight loss (1). Many NHS Trusts anticipated higher numbers of referrals, and instigated mechanisms to filter referrals and ensure appropriate adherence to guidelines by GPs. Our NHS Trust implemented an electronic referral proforma for GPs that included mandatory indicators of which NICE criteria were relevant for every referral. Although a recent systematic review and meta-analysis of 2WW referrals has analysed the colorectal cancer detection rate before 2015 (7), to our knowledge there have been no studies that have reported the detection rates for polyps or colorectal cancer after the lower threshold guidance was published.
The current study aimed to investigate the detection rates for polyps and colorectal cancer using the most recent referral criteria, and to determine whether any referral criteria or patient characteristics were associated with higher detection rates.
Patients and Methods
Study design. A prospective observational study was undertaken at a single UK NHS trust in the West Midlands from 1st May 2017 until 14th August 2018 following institutional approval. Due to the non-interventional, observational nature of the data, no formal research ethics submission was required according to the HRA tool (8).
Data collection. Data recorded included gender, age, type of referral (i.e. which NICE referral categories had been ticked by the GP referrer), type of investigations, and outcome (polyp or colorectal cancer detection). These prospectively collected data were then cross-referenced with data obtained from the Somerset Cancer Register by the Performance and Informatics team, including the patients who had a diagnosis of colorectal cancer during the same time period at the NHS Trust.
Referral details. GP electronic proformas and referral letters were examined for all patients who were referred on the urgent 2WW pathway. The category of referral was recorded as one of the 6 categories listed in NICE Guideline NG12 (1), which were: (a) aged 60 and over with iron-deficiency anaemia or changes in bowel habit; (b) aged 50 and over with unexplained rectal bleeding; (c) aged under 50 with rectal bleeding and any of the following unexplained symptoms of abdominal pain, change in bowel habit, weight loss, or iron-deficiency anaemia; (d) aged 40 and over with unexplained weight loss and abdominal pain; (e): rectal or abdominal mass; and (f) unexplained anal mass or unexplained anal ulceration. All electronic referrals required at least one (but allowed more than one) of these referral categories to be ticked.
Outcomes. The main outcome measures were: (a) the detection of colorectal cancer; (b) the detection of polyps; and (c) detection of either polyp or cancer (as a composite outcome measure).
Data analysis. Summary data are presented as N and % for categorical data, or median and interquartile range (IQR) for continuous data. Binomial multivariate logistic regression was performed to analyse the odds ratios [with 95% confidence interval (CI)] of the outcome measures according to each of the independent variables (age, gender, referral category, and previous polyp or colorectal cancer). A p-value of <0.05 was considered statistically significant.
Results
Patient characteristics. A total of 1323 patients were seen in the outpatient clinic during the study period, of which 605 (45.7%) were referred as part of the 2WW pathway and eligible for inclusion in this study. Patient characteristics and referral categories are shown in Table I. Of all 2WW patients, 290 (47.9%) were male. The median age was 66 (IQR=54-76) years. Figure 1 illustrates the number of referrals and polyp/CRC detection according to age category.
Method of referral. Methods of referral and the detection of polyps or cancer are illustrated in Table I. The most common NICE referral category was “aged 60 and over with iron-deficiency anaemia or changes in their bowel habit”, of which 5/252 (2.0%) had a diagnosis of colorectal cancer and 18/252 (7.1%) had a diagnosis of polyp. This NICE referral criterion was the least likely to yield a positive diagnosis of either polyp or cancer (OR=0.38; 95%CI=0.15-0.92; p=0.037) (Table II). The remainder of referral categories did not affect the likelihood of a polyp or cancer detection. The least common referral criteria were for rectal or anal masses or ulcers.
Investigations. There were 537/605 (88.8%) patients who had investigations arranged, of whom 346/537 (64.4%) had a colonoscopy, 116/537 (21.6%) had an oesophagogastroduodenoscopy, 54/537 (10.1%) had a flexible sigmoidoscopy, 110/537 (20.5%) had an abdominal CT scan, 91/537 (16.9%) had a CT virtual colonoscopy, and 12/537 (2.2%) had an abdominal ultrasound. Of the 68/605 (11.2) patients who were not investigated further, the reasons included: no indication for investigation or inappropriate referral (n=27); unfit for investigation or surgery (n=17); already had been investigated recently (n=10); benign anorectal pathology (n=7); and patients declined further investigation (n=7).
Cancer detection. Out of all 2WW referrals, there were 19 (3.1%) patients diagnosed with colorectal cancer. They included 12 male and 7 female patients, with a median age of 68 (IQR=54-74) years. Table II shows the likelihood of colorectal cancer detection according to age, gender, and category of referral. Neither age (OR=1.02; 95%CI=0.98-1.05; p=0.350) nor gender (OR=1.97; 95%CI=0.77-5.44; p=0.169) were statistically significant predictors of colorectal cancer. Similarly, previous diagnosis of colorectal polyp or cancer was not a significant factor in the prediction of colorectal cancer (OR=1.42; 95%CI=0.31-4.70; p=0.597). None of the individual NICE referral criteria were more likely to result in cancer detection.
Polyp detection. Out of all 2WW referrals, there were 64 (10.6%) patients diagnosed with polyps. They included 38 male and 26 female patients, with a median age of 62 (IQR=54-70) years. Table II shows the likelihood of polyp detection according to age, gender, and category of referral. Although male patients (OR=1.59; 95%CI=0.93-2.74; p=0.093) and those with previous polyps or colorectal cancer (OR=2.01; 95%CI=0.89-4.24; p=0.077) were more likely to have polyps, these factors did not reach statistical significance. None of the NICE referral criteria made polyp detection more likely. Conversely, one of the categories (aged 40 and over with unexplained weight loss and abdominal pain) appeared to have a statistically significant reduction in the likelihood of polyp detection (OR=0.30; 95%CI=0.08-0.92; p=0.049).
Combined polyp or cancer detection. Using the composite outcome measure of both colorectal polyp and cancer detection, male gender increased the likelihood of this outcome (OR=1.72; 95%CI=1.07-2.80; p=0.028) (Table II). Patients referred as “aged 60 and over with iron-deficiency anaemia or changes in their bowel habit” were less likely to have this outcome (OR=0.38; 95%CI=0.15-0.92; p=0.037), but no other category had an effect (Table II).
Discussion
The main finding from the current study is that amongst our UK cohort of 605 patients referred urgently for suspected bowel cancer, there were no particular NICE referral criteria that increased the likelihood of colorectal cancer detection. However, male gender increased the likelihood of disease detection (either polyp or cancer). Cancer detection was lower than expected, although polyp detection was similar to that which was expected. These findings may suggest that at the point of referral, patients cannot be triaged into those who are more likely to have disease based on the basic information provided by the GP referrer.
The overall detection rate for colorectal cancer in this study cohort was 3.1%, which is lower than that reported in the literature before the 2015 change in NICE guidance. It is possible that although referrals and investigations are increasing, the overall detection is decreasing. Other investigators have reported a similar trend, with increasing numbers of referrals but a reduction in cancer detection from 7.9% to 4.7% (9). Since polyp detection in the study cohort was very similar to the detection rate reported in the literature (approximately 11%), it is possible that this may represent the estimated baseline prevalence of polyps amongst symptomatic patients in this population. Further data would be required to investigate this hypothesis in more detail.
There is some evidence that GP referral straight-to-test may be justified and of benefit to patients in terms of reducing the delay to investigation and treatment (10). The current study showed that 88.8% of patients required endoscopic or radiological investigation. The majority of patients who did not require investigation were unfit for surgery, had recently been investigated, or declined investigation. If referrals for such patients were avoided, then the percentage of patients requiring investigation would have approached 100%, supporting a strategy of straight-to-test referrals for appropriately selected patients who are fit for (and would agree to) intervention.
With the reduced threshold for referral following the update in NICE Guidance, it is all the more important for successful communication and collaboration between primary care and colorectal services. There have been previous reports of patients being referred urgently on the 2WW pathway without necessarily fulfilling the NICE criteria (11), and the data from the current study are in keeping with such reports. There have been calls to review or modify the 2WW referral pathway due to lack of adherence to guidance and poor effectiveness in cancer pick up rate (12-14). Some have proposed improving awareness of best practice in primary care (15) or by developing specific electronic referral protocols (16). Furthermore, there may be a place for routine faecal immunochemical testing for risk stratification for symptomatic patients (17, 18). Such developments and initiatives will be even more important if straight-to-test referrals are to increase in practice.
With an increasing number of referrals via the 2WW pathway, some have advocated a nurse-led service with the addition of telephone triage services for those that might not require examination before investigation (19). There is some evidence that patients may even prefer nurse-led to doctor-led 2WW colorectal clinic appointments (20). Based on the data from the current study, important triage questions to determine whether further investigations are warranted would include those that determined: (a) the patient's willingness to be investigated further; (b) whether the patient has already been investigated recently; and (c) the patients' fitness for invasive intervention. Whether patients have an initial telephone consultation, or are reviewed in clinic, or sent straight for test, it is vital to ensure that patients are involved in the decision-making together with their primary care and colorectal clinicians.
Limitations. The current study is an observational study from a single UK NHS Trust, with all of the risk of bias and limitations of translatability that come with this study design. Furthermore, the current study compared cancer and polyp detection rates to those reported in the medical literature rather than from the same institution. This was considered to be a suitable alternative to conducting an uncontrolled before-after study, due to the flaws and criticism for that type of design (21).
Conclusion
In a cohort of UK patients referred urgently to colorectal services for suspected bowel cancer by their General Practitioners according to updated NICE guidelines, there were no patient characteristics or individual referral criteria that made colorectal cancer more likely. Cancer detection was lower than expected, which may be due to the lower threshold for referral in the new NICE guidelines. Close collaboration and communication between primary care and colorectal services is paramount for the effective detection of colorectal cancer in symptomatic patients.
Footnotes
Authors' Contributions
AB, CE, and DNN designed the study. AB, DNN, and QA collected data. DNN analysed the data. The first draft of the manuscript was written by DNN, and was critically appraised and revised by QA, AB, and CE. All Authors approved the final version.
Conflicts of Interest
All Authors declare that they have no conflicts of interest related to this study.
- Received January 21, 2020.
- Revision received January 28, 2020.
- Accepted January 31, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved