Investigating the conversion rate of routine referrals for patients presenting with symptoms of possible cancer.

Date of Approval
Application Number
21_000696
Technical Summary

National Institute of Health and Care Excellence (NICE) guidelines recommend routine referral for suspected cancer for patients presenting with low-risk features, and urgent referral for high-risk features. Our previous work found that some patients receive a routine referral instead of an urgent referral. Routine referral is more likely to result in longer diagnostic timelines, linked to worse outcomes, so it is important to highlight features where routine referral results in a high number of patients being diagnosed with cancer. We will examine patients who presented with one or more of the following high and low cancer risk symptoms in 2016-2017:

High risk features
Dysphagia
Jaundice
Lumps/masses
Haematuria
Breast lump
Post-menopausal bleeding

Low risk features
Weight loss
Anaemia
Abdominal pain
Back pain
Recurrent urinary tract infection

Hospital Episode Statistics Outpatient data will be used to derive whether a referral was made, and if it was routine or urgent. Information on cancer diagnoses will be obtained via linkage to the cancer registry.

Mixed effects models (logistic regression for binary outcomes, parametric survival models for intervals) will be used to assess the variability in routine referrals between patient groups, presenting features and between practices/GPs. Cancer diagnoses will be explored across a number of different time windows (30 days, 60 days, 90 days, 180 days and 1-year). Timeliness of cancer diagnoses will be investigated in terms of stage at diagnosis (early vs advanced), survival, and secondary care interval (days from referral to diagnosis).

Benefit to the public health of patients
A greater understanding of which patients receive routine referrals following presentation with symptoms of possible cancer, how many, the proportion of people referred diagnosed with cancer, and the timeliness of such diagnoses will facilitate identification of patient groups who are disadvantaged in the cancer diagnostic pathway. Findings may refine cancer guidance for GPs.

Health Outcomes to be Measured

1. A routine referral recorded in Hospital Episode Statistics Outpatient (HES OP) data after presenting with one or more symptoms of interest during the 2016-2017 time period. We will initially consider referral requests recorded in secondary care data within two weeks of presentation in primary care, but will explore the appropriateness of other timings.
2. Cancer diagnosis as recorded in cancer registry within 30 days, 60 days, 90 days, 180 days and 1 year of presentation with one or more symptoms of interest.
3. Cancer diagnosis as recorded in cancer registry within 30 days, 60 days, 90 days, 180 days and 1 year of the date the patient was seen in secondary care.
4. Stage at diagnosis of cancers diagnosed within 1 year of presentation; early (TNM stages I and II) vs advanced (TNM stages III and IV).
5. Secondary care interval; days from referral to cancer diagnosis.
6. Survival; days from presentation/referral to death.

Collaborators

Gary Abel - Chief Investigator - University of Exeter
Luke Mounce - Corresponding Applicant - University of Exeter
Bianca Wiering - Collaborator - University of Exeter
William Hamilton - Collaborator - University of Exeter

Linkages

HES Outpatient;NCRAS Cancer Registration Data;No additional NCRAS data required;Patient Level Index of Multiple Deprivation