Melanoma Diagnostic Processes and the Occurrence of Advanced Disease in Ontario

Abstract

Background: Reasons for the occurrence of advanced melanoma and an understanding of melanoma diagnostic processes and their relationship with advanced disease is understudied. Methods: This thesis examined the melanoma diagnostic process in Ontario using abstracted pathology reports and administrative data held at ICES. Objective 1 described the occurrence of advanced melanoma using abstracted pathology reports on a 65% random sample of people diagnosed with melanoma in Ontario from 2007–2012. We used modified Poisson regression to identify factors associated with advanced melanoma, stratifying by ulceration, which is a marker of clinical appearance and prognosis. Objectives 2–4 used ICES data. Objective 2 described the melanoma diagnostic interval (DI) for all Ontario melanoma patients diagnosed from 2007–2019. We used quantile regression to evaluate the association between patient-, disease-, and system-level factors and the length of the DI and primary care (PCI) and specialist care (SCI) subintervals. Objective 3 used latent class cluster analysis to group patients who experienced similar diagnostic processes. Objective 4 used quantile regression with restricted cubic splines to examine the relationship between melanoma thickness and DI length, stratified by ulceration. Results: Advanced melanoma was diagnosed in 26% of patients, and age, sex, socioeconomic status, and health region were associated with an increased risk of advanced disease. Ulceration attenuated these effects. Median DI was 36 days (Interquartile range [IQR]: 8–85 days), PCI was 22 days (IQR: 6–54 days) and SCI was 6 days (IQR: 1–42 days). Different factors were associated with the lengths of the DI, PCI, and SCI. We identified four diagnostic pathways: “primary care only”, “referred to specialist for immediate action”, “multiple visits and procedures in SCI”, and “specialist care only”. The distribution of patient-, disease-, and system-level factors, along with the lengths of the DI, PCI, and SCI varied across pathways. Finally, we found a complex relationship between thickness and DI duration, which differed by ulceration. Conclusions: We found independent risk factors for advanced melanoma in Ontario. This research identified variation in the melanoma diagnostic process within, and across, Ontario. We did not find a significant association between a longer DI and thicker melanomas.