Primary care referral practice, variability and socio-economic deprivation in colorectal cancer.

Abstract

The reasons for pre-hospital delay of the diagnosis of cancer are multifactorial, but include a physician-related component. Urgent cancer pathways and direct-to-test approaches have been implemented, but the emergency presentation of colorectal cancer (CRC) remains little changed over recent years. We examined the variability between primary care providers in referral patterns and its effect on outcome. A retrospective analysis was performed of a prospectively maintained database for 2009-2014 in a UK district hospital providing bowel cancer screening and tertiary rectal cancer services. Of 1145 CRC patients, 937 (81.8%) were diagnosed with a symptomatic cancer; 229/937 (24.4%) initially presented as an emergency. Between 44 primary care providers, emergency presentation varied between 8.3% and 57.1%. Patients of providers with high levels of emergency presentations (HV) had more advanced cancers than those of providers with medium (MV) or low levels (LV) [103/253 (40.7%), 154/461 (33.4%), 65/223 (29.1%); P=0.025] and a lower 3-year overall survival (50.2%, 57.8%, 65.6%; P=0.013), but with no difference in case-mix or deprivation levels. In adjusted analysis, this difference remained significant (advanced disease, OR 1.663, P=0.011; 3-year hazard ratio 1.479, P=0.010; comparing HV with LV). Conversely, elective suspected cancer referrals were less often used amongst diagnosed cancers [LV 136/223 (61.0%), MV 228/461 (49.5%), HV 114/253 (45.1%), P<0.001] with limited evidence for a more selective approach in the use of the 2-week rule amongst all 2-week rule referrals [LV 136/2508 (5.4%); MV 228/4239 (5.4%); HV 115/1526 (7.8%); positive cancer diagnosis, P=0.005]. Significant variability in emergency presentation of CRC requires local audit and examination of the reasons for delay in diagnosis and targeted measures to improve performance in non-emergency referral pathways.