The relationship between the diagnostic interval length and survival in colon cancer: The role of diagnoses in the emergency department.

Abstract

47 Background: The diagnostic interval, defined as the time from first cancer-related healthcare encounter to cancer diagnosis, is used to quantify delays in diagnosis and as a quality metric in understanding barriers to optimal cancer outcomes. Unlike with treatment delays, however, the relationship between the diagnostic interval length and survival is unclear. Our study aims to investigate the relationship between the diagnostic interval and all-cause mortality in individuals with colon cancer, with focus on the setting in which the cancer diagnosis originated. Methods: We conducted a retrospective cohort study of individuals diagnosed with colon cancer in Ontario, Canada from 2007-2019 using population-based administrative data. We calculated the diagnostic interval as the number of days from first colon-cancer-related healthcare visit to the cancer diagnosis date. The diagnosis was classified as emergent if the first colon-cancer-related visit occurred in the emergency department. We modelled the relationship between the diagnostic interval and overall survival using multivariable Cox proportional-hazards models with a restricted cubic spline (RCS) for the percentile length of the diagnostic interval, stratifying models by emergent diagnosis status. Results: We identified 52,916 individuals with colon-cancer; 33,475 (63%) had died at the end of follow-up. Median overall survival was 5.4 years (95% CI 5.2 – 5.5 years). Among the 17,201 individuals diagnosed emergently, both median survival (3.1 years [2.0 – 3.2]) and diagnostic intervals (49 days [IQR 3 – 149 days]) were shorter than individuals diagnosed non-emergently (6.9 years [6.7 – 7.0], 164 days [IQR 57 – 310]). The relationship between the diagnostic interval and survival depended on whether a diagnosis was initiated emergently (Table 1). Emergent diagnoses had consistently increased hazards of death with shorter diagnostic intervals and decreased hazard at longer intervals, whereas non-emergent diagnoses followed a more U-shaped distribution, with increased hazards at relatively shorter and longer diagnostic intervals. Conclusions: The relationship between the colon cancer diagnostic interval and overall survival is moderated by whether a patient’s cancer journey begins in the emergency room. This information is critical to understanding how the diagnostic interval may be integrated as a quality metric into system level performance assessment. Hazard ratio of death at varying diagnostic interval percentiles (reference=50 th percentile). Diagnostic interval (percentile) Diagnosed emergently (HR [95% CI]) Diagnosed non-emergently (HR [95% CI]) 10 th 1.29 (1.23 - 1.36) 1.16 (1.12 - 1.21) 25 th 1.28 (1.22 - 1.34) 1.01 (0.98 - 1.06) 40 th 1.13 (1.09 - 1.18) 0.97 (0.94 - 0.99) 60 th 0.92 (0.89 - 0.95) 1.02 (1.01 - 1.04) 75 th 0.91 (0.87 - 0.95) 1.04 (1.00 - 1.08) 90 th 0.92 (0.87 - 0.98) 1.09 (1.04 - 1.14)