A test for blood in the poo, the Faecal Immunochemical Test (FIT), is accurate for finding patients at high risk of bowel cancer in screening. Recently, FIT has been used for patients with symptoms that may suggest bowel cancer, to improve diagnosis. FIT is recommended for patients with symptoms in England, but how it’s used, the effect on bowel cancer diagnosis and impact medications have on the result is unknown.
Objective:
• Assess how accurate FIT is for bowel cancer and other conditions in patients with symptoms.
• Investigate what differences in FIT use and accuracy exist by region, patient age, sex, ethnicity and deprivation, and how these impact route to diagnosis.
• Explore the impact that medications (such as blood-thinners) have on the accuracy of FIT.
• Evaluate links between FIT result and death, from bowel cancer or other causes.
• Describe what non-bowel cancers are found after FIT.
Design: We will look at data routinely collected for patients who have undertaken a FIT for symptoms (2015-present). We estimate this will include approximately 260,000 patients, with no requirement for further patient input.
Outcomes and patient benefit: Exploring FIT usage is important to understand variations in the country. Calculating its accuracy will highlight for the first time its effectiveness in diagnosis of bowel cancer and other bowel conditions across the entire country in patients suspected of having cancer. Highlighting clinical benefits, as well as inequalities in access, will help adoption of this test in areas that would benefit the most.
The Faecal Immunochemical Test (FIT) is used for bowel cancer screening and increasingly for patients with symptoms of possible colorectal cancer (CRC). There are no national-level studies on inequalities in use and accuracy for CRC and other bowel conditions in symptomatic patients.
Objectives:
•Evaluate the performance characteristics of FIT for CRC, Inflammatory Bowel Disease and polyps.
•Describe inequalities in use of FIT for symptomatic patients across the UK, by sociodemographics (age, sex, ethnicity, deprivation), region, and how these impact test performance for CRC and time from test to CRC diagnosis.
•Quantify the impact medications (anticoagulants/antiplatelets/proton-pump inhibitors) have on FIT’s performance for diagnosing CRC.
•Describe the relationship between FIT result and all-cause mortality.
•Quantify extra-colonic cancers diagnosed after FIT.
Design:
Observational study using CPRD Aurum combined with HES for patients who have completed a FIT (excluding screening) 2015-present (estimated >260,000 patients). CRC diagnosis will be defined using a combination of CPRD, HES and ONS, with FIT result and sociodemographics from CPRD.
Performance characteristics at various thresholds will be calculated and Receiver-Operating Characteristic, Precision-recall and decision curves plotted to compare performance for CRC between regions, sociodemographic groups and medication usage. FIT usage will be compared to a baseline population of patients in each region, with differences in usage by sociodemographics assessed using Poisson regression. One/two-year mortality and extra-colonic malignancies will be compared between different FIT thresholds with logistic regression and time to death with Cox regression.
Benefit:
This work is novel, will justify increasing usage of FIT and will highlight disadvantaged groups to target to increase healthcare equity.
Variations in performance by demographics/medication usage will inform risk-prediction in groups at higher risk from procedures, where different thresholds for investigation may improve risk-benefit balance, and association with all-cause mortality and extra-colonic cancers will inform decision-making for patients with raised FIT and no CRC.
Primary outcomes:
• Calculate the performance characteristics of FIT for diagnosing CRC and other significant bowel pathologies (Inflammatory Bowel Disease and colonic polyps);
Secondary outcomes:
• Describe use of FIT in the UK, how this differs by region and patient demographics (such as age, sex, ethnicity and deprivation) and what impact these factors have on test performance for CRC and route to CRC diagnosis (defined as time from FIT test to definitive diagnosis, assessed across different FIT result categories);
• Calculate the effect medications (eg anticoagulants, antiplatelets and proton-pump inhibitors) have on the performance characteristics FIT for CRC;
• Explore the relationship between FIT result and CRC-related, cause-specific and all-cause mortality;
• Quantify the rates of extra-colonic cancers (those elsewhere in the gastrointestinal tract or elsewhere in the body) in patients having FIT, and whether FIT result impacts the probability of having an extra-colonic cancer.
David Humes - Chief Investigator - University of Nottingham
Alastair Morton - Corresponding Applicant - University of Nottingham
Colin Crooks - Collaborator - University of Nottingham
Joe West - Collaborator - University of Nottingham
HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation;CPRD Aurum Ethnicity Record