Guideline discordant diagnostic care: when do primary care referrals not reflect guidelines for suspected cancer?

Talk Code: 
Gary Abel
Bianca Wiering, Georgios Lyratzopoulos, John Campbell, William Hamilton
Author institutions: 
University of Exeter Medical School, University College London


A key service aimed at improving UK cancer survival rates (through improved diagnostic timeliness) is the fast-track referral system known as the two-week wait pathway. Guidelines advise general practitioners which patients warrant a two-week wait referral. However, guidelines are only effective to the degree that they are implemented. This study aims to investigate how often GPs follow the referral guidelines and whether certain types of patients are more or less likely to be referred.


We used linked primary care (Clinical Practice Research Datalink) and secondary care (Hospital Episode Statistics) data. Patients presenting with any one of 6 symptoms (haematuria, rectal bleeding, breast lump, post-menopausal bleeding, dysphagia or anaemia) for the first time during 2014-2015 were included (for ages where NICE guidelines recommend two-week wait referral). For patients presenting with haematuria, either the second or third GP visit was defined as the index consultation where patients had received treatment for urinary tract infection during earlier visits. A composite outcome of a two-week wait referral or urgent referral was used since both reflect the need for an urgent clinical assessment. Multilevel logistic regression was used to investigate whether the tendency to refer varies symptoms and patient characteristics (age, gender, ethnicity, comorbidities and deprivation level). Further analyses investigated whether patients not receiving a referral were diagnosed with cancer within one year of the index consultation.


Preliminary results for referrals made within 14 days of first presentation with a recorded symptom indicate a high percentage of patients do not receive a two-week wait or urgent referral, varying from 81.8% (rectal bleeding) to 33.3% (breast lump). There is evidence that referrals are associated with age and comorbidities (p<0.001), with young patients and patients with comorbidities less likely to receive a referral. Associations between patient characteristics and referrals differ by symptom. More deprived patients with breast lumps, female patients with anaemia or haematuria and non-white patients with rectal bleeding were less likely to receive a referral. Furthermore, interim findings suggest that around 2.3% of patients not receiving a referral were later diagnosed with cancer.


Guideline recommended actions often do not occur for patients presenting with common possible cancer symptoms. Given that a significant number of such patients go on to be diagnosed with cancer, better adherence to referral guidelines could potentially improve the cancer diagnostic process for many people. Appreciating which patient groups are at greater risk of non-referral against guideline recommendations can help target improvement efforts. For example, these findings may be used by schemes such as CRUK cancer facilitators and Macmillan GPs to directly influence practice. Alternatively, they could inform educational materials and novel interventions targeting early diagnosis.

Submitted by: 
Gary Abel
Funding acknowledgement: 
This work was funded by Cancer Research UK (EDAGI) Grant No. C50627/A25622. WH is Director, GL Associate Director, and GAA Senior Investigator of the multi-institutional CanTest Collaborative, which is funded by Cancer Research UK (grant number C8640/A23385). GL is supported by a Cancer Research UK Clinician Advanced Scientist Fellowship (grant number C18081/A18180)