- Design
- Observational case-control and case-only analyses of linked primary care and registry data
- Population
- 223,422 patients in English primary care, cancer cases and matched controls
- Primary outcome
- Non-cancer diagnoses before cancer linked to cancer and its outcomes
- Effect
- 37 interim diagnoses across 7 cancer sites; 18 linked to worse outcomes
This BJGP study linked English primary care records to the cancer registry for 223,422 patients, comparing people diagnosed with cancer with matched controls. It looked for non-cancer diagnoses made in the year before cancer that share symptoms with it, and tested whether they were linked to stage, emergency presentation and survival.
It identified 37 such 'interim' diagnoses across seven cancer sites. Eighteen were consistently associated with worse cancer outcomes and eight with better ones. Reflux disease, gastritis, oesophagitis or peptic ulcer, for example, were linked to worse outcomes in pancreatic, colorectal and ovarian cancer.
The practical message is about how a plausible benign label can close a diagnosis too early. A label such as reflux is often right; the risk lies in treating it as settled when symptoms persist. The design is observational and cannot show that any individual case was missed.
- Safety-net every working diagnosis that could be explaining a cancer symptom.
- Set a review date when symptoms are attributed to reflux, gastritis or irritable bowel.
- Reconsider the label if symptoms persist despite treatment.
- Record what the patient should return with and when.
Why it matters
A reasonable first diagnosis is also the commonest way a cancer goes unrecognised for months.
Don't overread it
Observational; it identifies risky labels, not individual missed diagnoses.
The statistics, in plain English
A case-control design compares people who got cancer with people who did not, so it shows which earlier diagnoses were more common before cancer. Bonferroni correction was used to limit chance findings across many tests. It cannot say how many cancers were actually delayed by a wrong label.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for family medicine, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free