Authors' objectives: Faecal immunochemical tests may be better than symptoms alone at identifying which patients who present to primary care with symptoms are at high risk of colorectal cancer and should have a colonoscopy. This could reduce waiting lists and patient anxiety/discomfort and enable earlier treatment of colorectal cancer. The threshold used will affect how well faecal immunochemical tests work, with a higher threshold resulting in fewer referrals but a greater chance of missing disease. What is the most clinically effective and cost-effective way to use faecal immunochemical tests to reduce the number of people without significant bowel pathology who are referred to the suspected cancer pathway for colorectal cancer, taking into consideration potential colonoscopy capacity constraints for urgent and non-urgent referrals? Tests were HM-JACKarc, OC-Sensor, FOB Gold, NS-Prime, QuikRead go, IDK TurbiFIT, IDK Hb, IDK Hb/Hp complex and IDKHb+Hb/Hp ELISAs. Early diagnosis and treatment of colorectal cancer (CRC) in people presenting to primary care with symptoms can improve survival and cure rates. The introduction of the National Institute for Health and Care Excellence (NICE) National Guideline 12 (NG12) in 2015 to expand symptoms-based criteria for referral to secondary care led to an increase in the number of urgent 2-week wait (2WW) suspected cancer pathway referrals, but no corresponding increase in the proportion of patients investigated through 2WW who had cancer. This has led to pressure on colonoscopy capacity and to long waiting times in some areas of the UK, especially in the non-urgent [18-week wait (18WW)] referral pathway. Quantitative faecal immunochemical tests (FITs) are designed to detect occult (small amounts) of blood in stool samples (faecal haemoglobin) using antibodies specific to human haemoglobin. They are currently used in patients with low-risk symptoms in primary care [as described in NICE Diagnostics Guidance 30 (DG30)], but not in patients with high-/medium-risk symptoms as defined in NG12, who are instead referred directly to secondary care. There is evidence that FITs are a better predictor of CRC risk in patients than symptoms alone and could result in fewer referrals of people without CRC to the 2WW pathway. Therefore, triage with FITs for all patients could avoid unnecessary referrals, patient anxiety, time off work and loss of economic productivity, and rare adverse events associated with colonoscopy such as bleeding, perforation and death. Those who are likely to have CRC could be prioritised more effectively, potentially reducing time to diagnosis. The released colonoscopy capacity could allow non-urgent referrals to be seen more quickly. The extent to which colonoscopy capacity is released and time to diagnosis is affected will depend in part on the threshold used to define a positive FIT result, with a higher threshold resulting in fewer referrals but a greater chance of missing disease. The decision problem in the NICE scope was ‘What is the most clinically and cost-effective way to use quantitative FITs to reduce the number of people without significant bowel pathology who are referred to the suspected cancer pathway for CRC, taking into consideration potential colonoscopy capacity constraints for urgent and non-urgent referrals?’ Eight FITs were within the scope of the assessment, namely HM-JACKarc, FOB Gold, OC-Sensor, NS-Prime, IDK TurbiFIT, IDK Hemoglobin ELISA (IDK Hb), IDK Hb/Hp complex ELISA (IDK Hb/Hp) and QuikRead go. The decision problem was addressed through a systematic review of evidence relating to the accuracy of the tests, a statistical synthesis to pool data across studies, and an economic model that aimed to estimate the cost-effectiveness of FIT strategies based on diagnostic accuracy, the number of colonoscopies undertaken and the impact on time to diagnosis.

Authors' results and conclusions: Syntheses of sensitivity and specificity were conducted for HM-JACKarc (n = 16 studies), OC-Sensor (n = 11 studies) and FOB Gold (n = 3 studies). No synthesis was conducted for QuikRead go, NS-Prime IDK Hb or IDK Hb/Hp as there was only one study for each. No eligible studies were found for IDK Hb+Hb/Hp or for IDK TurbiFIT. Other outcomes (e.g. patient acceptability) were also synthesised. Model results suggest that faecal immunochemical tests generate a positive incremental net monetary benefit compared with current care, typically in the range of £200–350 per patient, regardless of the threshold used, for the majority of faecal immunochemical tests strategies assessed. These conclusions were robust to the sensitivity analyses undertaken. For all faecal immunochemical test brands, there are strategies at which the incremental net monetary benefit is positive compared with current care. The exact brand and threshold(s) that generate the greatest incremental net monetary benefit could not be robustly determined due to the similarity of incremental net monetary benefit values, parameter uncertainty and the possibility of omissions from the model structure. Clinical evidence results Forty-nine studies were included in the review, across all tests and all subgroups and outcomes. There were risk of bias and/or applicability concerns with all the studies. Studies recruiting in secondary care generally scored as being at high risk of bias for patient selection, as some primary care patients were not recruited, and studies recruiting in primary care generally scored as being at high risk of bias for the reference standard, as not all patients received colonoscopy or CTC. Various other sources of bias were also noted. As there was only a small number of head-to-head comparative studies, comparative test accuracy was not formally quantified. Considering a threshold of 10 µg/g, the results for sensitivity and specificity, respectively, were as follows: HM-JACKarc (n = 16 studies), 89.5% [95% credible interval (CrI) 84.6% to 93.4%] and 82.8% (95% CrI 75.2% to 89.6%); OC-Sensor (n = 11 studies), 89.8% (95% CrI 85.9% to 93.3%) and 77.6% (95% CrI 64.3% to 88.6%); and FOB Gold (n = 3 studies), 87% (95% CrI 67.3% to 98.3%) and 88.4% (95% CrI 81.7% to 94.2%). No synthesis was conducted for QuikRead go, NS-Prime or IDK tests, as there was only one study for each. The estimates of sensitivity and specificity at 10 µg/g, respectively, were as follows: QuikRead go, 92.90% [95% confidence interval (CI) 68.5% to 98.7%] and 70.10% (95% CI 66.1% to 73.8%); and NS-Prime, 71.40% (95% CI 35.9% to 91.8%) and 83.60% (95% CI 78.2% to 87.9%). The study of IDK Hb and IDK Hb/Hp only reported data at 2 µg/g, and the sensitivity and specificity were calculated as follows: IDK Hb 87% (95% CI 84.4% to 89.6%) and 88.1% (95% CI 85.6% to 90.6%); and IDK Hb/Hp, 82.6% (95% CI 79.6% to 85.6%) and 80.8% (95% CI 77.7% to 83.9%). No diagnostic test accuracy data that met the inclusion criteria for the review were found for the combined use of IDK Hb + Hb/Hp or for IDK TurbiFIT tests. The sensitivity analyses showed that the results were similar when studies were subgrouped according to population type (all patients presenting to primary care, high-risk patients, low-risk patients) and when subgrouped according to the reference standard used, as the credible intervals overlapped. Four studies reported data using a dual FIT strategy. Where both were reported, the sensitivity was higher but specificity lower using dual FIT (either FIT positive) than when using only the first FIT result. The three included comparative diagnostic test accuracy studies concluded that there were some differences between tests, but none concluded whether (and what) different FIT cut-off values would be required for each test. Across patient characteristic subgroups (anaemia, n = 11 studies; age, n = 3 studies; sex, n = 3 studies; and people taking medications that may affect FIT results, n = 3 studies), evidence was limited and sometimes inconsistent. It was not possible to conclude what or whether different FIT thresholds may be required. No studies were identified according to ethnicity or for people with other blood disorders that may affect FIT results. Eight studies reported data on the accuracy of FITs for AAs and IBD. Uncertainty was high, with a large amount of heterogeneity between studies. Eleven studies reported test failure rates largely between 2% and 5%. The non-return rate in the study most closely matching the decision problem was 9.4%. For dual FIT, non-return rates appeared generally higher. Two studies reported patient perspectives. The authors concluded that most patients found FIT acceptable, but strategies are needed to engage patients with more negative views of FIT, and shared decision-making should be considered for patients dissatisfied with relying on a negative FIT result. Generalisability may have been affected by the fact that all patients included had been referred to secondary care. One study reported on the impact of sociodemographic factors on FIT return rates and found higher return rates for female patients than for male patients, for patients aged ≥ 65 years than for those aged

Authors' methods: Systematic review, meta-analysis and cost-effectiveness analyses were conducted. Searches across four databases and six registries were conducted (December 2022). Diagnostic accuracy studies conducted in patients presenting to or referred from primary care with symptoms suggestive of colorectal cancer using any reference standard were included. Risk of bias was assessed with quality assessment of diagnostic test accuracy studies version 2. For each test, sensitivity and specificity were pooled at all reported thresholds and summary estimates were provided at all possible thresholds within the observed range. Comparative accuracy between tests was considered. Other outcomes, for example test uptake, failure and patient acceptability, were also extracted. Clinical evidence review methods Searches were conducted across four databases and six registries in December 2022. The titles and abstracts of records retrieved were screened by one reviewer, with the first 20% checked by a second reviewer before the remainder were screened. Records for which the full text was obtained were checked for inclusion by two reviewers. Data extraction and quality assessment were conducted by one reviewer and checked by a second. Study quality was assessed using quality assessment of diagnostic test accuracy studies version 2. As no randomised controlled trials were identified, studies were included if they reported the diagnostic test accuracy of FIT in patients presenting to primary care, or referred from primary care, with signs or symptoms of CRC. Studies reporting data for ‘dual FIT’, whereby patients are asked to provide two samples from different bowel movements, were also included. All thresholds for defining a FIT were eligible for inclusion. The reference standard was not restricted but expected to comprise colonoscopy, computed tomography colonography (CTC), other imaging tests or records follow-up. Studies were also subgrouped according to several patient characteristics (anaemia, age, sex, ethnicity and medication or other blood disorders that might affect FIT). Test failure rates, uptake of FITs, time to colonoscopy, time to diagnosis and patient-reported outcomes such as health-related quality of life, preference and anxiety were also sought. The statistical synthesis pooled estimates of sensitivity and specificity at all reported diagnostic thresholds and provided summary estimates at all possible thresholds within the observed range. Studies were synthesised for each test separately. Sensitivity analyses investigated the effects of population type and reference standard, where data allowed.

Project Status: Completed

URL for project: https://www.journalslibrary.nihr.ac.uk/programmes/hta/NIHR135637

Year Published: 2025

URL for published report: https://www.journalslibrary.nihr.ac.uk/hta/AHPE4211

URL for additional information: English

English language abstract: An English language summary is available

Publication Type: Full HTA

Country: England, United Kingdom

DOI: 10.3310/AHPE4211

Organisation Name: NIHR Health Technology Assessment programme

Contact Address: NIHR Journals Library, National Institute for Health and Care Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK

Contact Name: journals.library@nihr.ac.uk

Contact Email: journals.library@nihr.ac.uk