Cancer diagnoses after emergency GP referral or A&E attendance in England: determinants and time trends in Routes to Diagnosis data, 2006-2015

Abstract

Background: Diagnosis of cancer as an emergency is associated with poor outcomes but has a complex aetiology. Examining determinants and time trends in diagnostic routes can help to appreciate the critical role of general practice over time in diagnostic pathways for patients with cancer.

Aim: To examine sociodemographic, cancer site, and temporal associations with type of presentation among patients with cancer diagnosed as emergencies.

Design and setting: Analysis of Routes to Diagnosis data, 2006-2015, for patients with cancer in England.

Method: The authors estimated adjusted proportions of emergency presentation after emergency GP referral (GP-EP) or presentation to accident and emergency (AE-EP), by patient sex, age, deprivation group, and year of diagnosis using multivariable regression.

Results: Among 554 621 patients presenting as emergencies, 24% (n = 130 372) presented as GP-EP, 62% as AE-EP (n = 346 192), and 14% (n = 78 057) through Other-EP sub-routes. Patients presenting as emergencies were more likely to have been GP-referred if they lived in less deprived areas or were subsequently diagnosed with pancreatic, gallbladder, or ovarian cancer, or acute leukaemia. During the study period the proportion and number of GP-EPs nearly halved (31%, n = 17 364, in 2006; 17%, n = 9155 in 2015), while that of AE-EP increased (55%, n = 31 049 to 68%, n = 36 868).

Conclusion: Patients presenting as emergencies with cancers characterised by symptoms/signs tolerable by patients but appropriately alarming to doctors (for example, pancreatic cancer manifesting as painless jaundice) are over-represented among cases whose emergency presentation involved GP referral. Reductions in diagnoses of cancer through an emergency presentation likely reflect both the continually increasing use of 2-week-wait GP referrals during the study period and reductions in emergency GP referrals.

Keywords: early diagnosis; emergencies; patients; population groups; referral and consultation.

Figure 1.

Adjusted numbers of patients with cancer diagnosed via emergency presentation sub-route or any other diagnostic route, estimated using a multivariable multinomial logistic regression, for AE-EP, GP-EP (reference outcome), Other-EP, 2WW, non-emergency GP referral, screening, and ‘Other’, fitted to all patients diagnosed with cancer (including non-EP), 2006–2015 (N = 2 619 067), where independent variables were sex, age group, deprivation group, cancer, and year of diagnosis. Predicted numbers of AE-EPs, GP-EPs, Other-EPs, and non-EPs were derived by multiplying predicted proportions of these outcomes by the number of observed cancer cases (including non-EP), per year. AE-EP = emergency presentation through presentation to accident and emergency. EP = emergency presentation. GP-EP = emergency presentation through a GP referral. GP ref = non-emergency GP referral. Other-EP = emergency presentation through routes other than through the GP or accident and emergency, for example, through referral during an inpatient admission. 2WW = 2-week-wait referral.

Figure 2.

Adjusted proportions of emergency presentation type among emergency presenters, over time, estimated using a multivariable multinomial logistic regression, for AE-EP, GP-EP (reference outcome), and Other-EP, fitted to all patients diagnosed through EP in 2006–2015 (n = 554 621), where independent variables were sex, age group, deprivation group, cancer, and year of diagnosis. AE-EP = emergency presentation through presentation to accident and emergency. EP = emergency presentation. GP-EP = emergency presentation through a GP referral. Other-EP = emergency presentation through routes other than through the GP or accident and emergency, for example, through referral during an inpatient admission.