Using English National Cancer Registration and Linked Health Datasets to Assess Variation in Diagnostic Pathway Length for Colorectal and Lung Cancer Patients by Stage and Route to Diagnosis

Abstract

Background: Understanding factors that contribute to longer diagnostic pathways is important to improve efficiency of these pathways and can provide evidence for the implementation of the forthcoming 28-day Faster Diagnostic Standard (FDS) in England. This analysis uses linked national cancer registrations and other health datasets to define diagnostic pathway length and examine variation by route to diagnosis (RtD), stage and patient characteristics for colorectal and lung cancer patients. Aim: To achieve a more in-depth understanding of the diagnostic pathway for colorectal and lung cancer patients and identify particular factors associated with longer diagnostic pathways. Methods: English cancer registrations (2014 & 2015) diagnosed with colorectal and lung cancers (C18-20, C33-34) were linked to the hospital episode statistics, diagnostic imaging dataset, cancer waiting times and RtD data. Patients with multiple diagnoses or unknown RtD were excluded. To construct the pathway length, a start date was derived by defining the earliest relevant event (referral into/appointment in secondary care or diagnostic procedure) from available datasets in the 6 months prediagnosis. The pathway length was determined for each cancer site separately, by stage, RtD and patient characteristic. Regression analysis produced odds ratios (OR) of having a longer diagnostic pathway while controlling for other factors, including age, sex, comorbidities and deprivation. The longer pathway was defined as longer than the median days per cancer site. Results: Of 64,320 colorectal and 71,526 lung patients included, 99.5% and 99.8% respectively had at least one relevant first event recorded. The median pathway length (days) was 26 (IQR 11-56) for colorectal and 35 for lung (15-83). Pathway length decreased significantly with later stage (stage 1-4 - colorectal: 35 to 20, lung: 75 to 25) with significant variation also by presentation route and comorbidity score. Regression analysis showed that, after adjustment for other factors (including stage), patients on a GP referral route had an increased odds of a long pathway compared with the two week wait route (an urgent GP referral with a suspicion of cancer) (colorectal aOR: 4.5, lung aOR: 2.5). Patients diagnosed via emergency presentation route, which are predominantly late stage, had the shortest pathway length and reduced ORs of having a longer diagnostic pathway (colorectal aOR: 0.2, lung aOR: 0.4). Certain patient characteristics are also associated with longer diagnostic pathway length. Conclusion: There is substantial variation in diagnostic pathway length by stage and route for both sites and in many cases these pathways exceeded 28-days (colorectal: 45.3%, lung: 56.4%). Vague symptoms, comorbidities and other patient characteristics may make cancer more difficult to diagnose. Factors associated with longer waits could support the creation of targeted initiatives to reduce the diagnostic pathway length.