Use of the faecal immunochemical test for patients with lower risk symptoms of colorectal cancer in primary care: a pragmatic prospective study in the East of England

UEG Week 2021 Poster Presentations

Introduction
In England, patients meet criteria for urgent/fast-track referral when presenting with high-risk symptoms of colorectal cancer (CRC) that have a positive predictive value ≥3%1. Patients presenting with lower risk symptoms do not meet criteria for fast-track referral but may still have cancer. Recognising this, the faecal immunochemical test (FIT) was introduced to triage patients with lower risk symptoms in English primary care from 20182. Evidence is scarce on symptom presentation which prompts a general practitioner (GP) to request a FIT, and on factors associated with a positive FIT result.

Aims & Methods
We aimed to assess GP use of FIT for patients presenting in primary care with lower risk symptoms of colorectal cancer. This was a pragmatic, prospective, multi-method study of patients offered a FIT in two regions of the Eastern Cancer Alliance in England between January and June 2020. Data were collected from consenting primary care patients aged ≥40 years (questionnaire), their GPs (via lab requests) and labs (FIT and routine blood test results). We used descriptive statistics and bivariate analyses to characterise patient socio-demographic and clinical features, and binary logistic regression to identify factors independently associated (p<0.05) with a positive FIT result (defined as fHb ≥10 µg Hb/g faeces). We also assessed concordance between patient- and GP-reported symptoms (% concordance and kappa statistics).

Results
Full datasets were available for 310 patients: median age 70 (IQR 61-77) and 53.0% female. 23.2% had a positive FIT result. The commonest patient-reported symptoms were change in bowel habit (68.7%) and fatigue (57.4%), while the commonest GP-reported symptoms were abdominal pain (24.8%) and change in bowel habit (24.2%). GPs reported symptoms less often and for fewer symptoms (median 1 symptom, IQR 1-1), compared to patients (median 5 symptoms, IQR 3-6). Symptom concordance ranged from 44.2% for fatigue to 80.3% for unexplained weight loss. Kappa agreement was universally low across symptoms. Five factors were found to be independently associated with a positive FIT result. Patients reporting symptoms of diarrhoea (OR 2.05 (95% 1.08-3.86)) or rectal bleeding (OR 3.53 (95% CI 1.75-7.10), taking anticoagulants (OR 4.39 (95% CI 1.83-10.5) or aspirin (OR 3.43 (95% CI 1.38-8.54), or having a low blood haemoglobin result (OR 2.72 (97%CI 1.29-5.75)) had increased odds of having a positive FIT result.

Conclusion
This study demonstrates that certain patient clinical characteristics (symptoms on presentation, medications and blood test results) are associated with a positive FIT result. Furthermore, patients report on more symptoms than their GPs record about them. This discrepancy may have implications for patient management and raises questions about the use of GP-reported data when investigating patient pathways towards timely detection of colorectal cancer.