Can incorporating lower risk symptoms into urgent cancer referral guidance detect cancers earlier?

Talk Code: 
4A.5
Presenter: 
Sarah F. Moore
Twitter: 
Co-authors: 
Sarah F. Moore, Sarah Price, Richard D. Neal, Willie Hamilton
Author institutions: 
University of Exeter

Problem

Reducing the threshold for urgent cancer referrals in England is a potential path to identifying cancers earlier. The current threshold is set at a positive predictive value (PPV) of ≥3% for presenting features resulting in a diagnosis of cancer. Our previous study showed that reducing this to ≥2% or ≥1% would increase the numbers of patients eligible for referral by 8% and 136%, respectively, across 11 cancer sites.What we now need to explore is whether patients presenting to their GP with a ≥3% risk feature also present with a lower risk feature in the preceding year. If this were the case, lowering the referral threshold might allow them to be identified earlier, resulting in improved outcomes if diagnosed with cancer.

Approach

This observational cohort study used data from the Clinical Practice Research Datalink. Information on PPVs of potential features of 11 cancer sites (bladder, breast, colorectal, endometrium, kidney, larynx, lung, oesophago-gastric, ovary, pancreas, prostate) was collated and stratified into bands of PPV≥3%, 2-2.99% and 1-1.99%. For each cancer site studied, we identified patients presenting in 2016 with a feature meeting at least one criterion for PPV≥3%. The earliest occurrence was nominated the index date. We then identified patients who presented with any features meeting criteria in the 2-2.99% or 1-1.99% bands within a year before the index date. For multiple instances of meeting a criterion, the furthest instance from the index within a year was used. Analysis was conducted using Stata and results presented as detailed descriptive statistics.

Findings

In 2016, of 150,921 eligible patients in our sample, 8576 presented with a feature with a PPV of ≥3% for one of the 11 cancers. Of those, 365 (4.2%) and 1147 (13.3%), respectively, presented with a feature meeting 2-2.99% PPV and 1-1.99% PPV criteria in the preceding year. The percentages were heterogeneous across cancer sites, ranging from 0 (bladder, breast, endometrium, kidney, pancreas) to 9% (larynx) for those meeting 2-2.99% PPV criteria and from 0 (endometrium, larynx, ovary) to 30.7% (oesophago-gastric) for those meeting 1-1.99% PPV criteria in the preceding year.

Consequences

This information could underpin changes to NICE guidance on thresholds for urgent cancer referral in England. Although limited by lack of linked cancer referral and diagnosis data, the detail of progression of specific features from low to high risk across multiple cancer sites will allow for a tailored approach to any future reduction in referral thresholds. Further work to estimate the potential impact on referrals and diagnoses will strengthen this evidence and increase its potential for influence on policy.

Submitted by: 
Sarah Moore
Funding acknowledgement: 
This project is funded by the National Institute for Health and Care Research (NIHR) School for Primary Care Research as part of a GP Career Progression Fellowship awarded to SFM. SP is funded by the National Institute for Health Research Policy Research Program, conducted through the Policy Research Unit in Cancer Awareness, Screening and Early Diagnosis, PR-PRU-1217-21601. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.