Diagnostic timeliness in adolescents and young adults with cancer: a cross-sectional analysis of the BRIGHTLIGHT cohort

Background Adolescents and young adults (AYAs) are thought to experience prolonged intervals to cancer diagnosis, but evidence quantifying this hypothesis and identifying high-risk patient subgroups is insufficient. We aimed to investigate diagnostic timeliness in a cohort of AYAs with incident cancers and to identify factors associated with variation in timeliness. Methods We did a cross-sectional analysis of the BRIGHTLIGHT cohort, which included AYAs aged 12-24 years recruited within an average of 6 months from new primary cancer diagnosis from 96 National Health Service hospitals across England between July 1, 2012, and April 30, 2015. Participants completed structured, face-to-face interviews to provide information on their diagnostic experience (eg, month and year of symptom onset, number of consultations before referral to specialist care); demographic information was extracted from case report forms and date of diagnosis and cancer type from the national cancer registry. We analysed these data to assess patient interval (time from symptom onset to first presentation to a general practitioner [GP] or emergency department), the number of prereferral GP consultations, and the symptom onset-to-diagnosis interval (time from symptom onset to diagnosis) by patient characteristic and cancer site, and examined associations using multivariable regression models. Findings Of 1114 participants recruited to the BRIGHTLIGHT cohort, 830 completed a face-to-face interview. Among participants with available information, 204 (27%) of 748 had a patient interval of more than a month and 242 (35%) of 701 consulting a general practitioner had three or more prereferral consultations. The median symptom onset-to-diagnosis interval was 62 days (IQR 29-153). Compared with male AYAs, female AYAs were more likely to have three or more consultations (adjusted odds ratio [OR] 1.6 [95% CI 1.1-2.3], p=0.0093) and longer median symptom onset-to-diagnosis intervals (adjusted median interval longer by 24 days [95% CI 11-37], p= 0.0005). Patients with lymphoma or bone tumours (adjusted OR 1.2 [95% CI 0.6-2.1] compared with lymphoma) were most likely to have three or more consultations and those with melanoma least likely (0.2 [0.1-0.7] compared with lymphoma). The adjusted median symptom onset-to-diagnosis intervals were longest in AYAs with bone tumours (51 days [95% CI 29-73] longer than for lymphoma) and shortest in those with leukaemia (33 days [17-49] shorter than for lymphoma). Interpretation The findings provide a benchmark for diagnostic timeliness in young people with cancer and help to identify subgroups at higher risk of a prolonged diagnostic journey. Further research is needed to understand reasons for these findings and to prioritise and stratify early diagnosis initiatives for AYAs.