Individual and ecological factors associated with early detection of nodular melanoma in the United States

Abstract

The aim of the study was to identify individual and ecological factors associated with early diagnosis of nodular melanoma (NM). Using cross-sectional, prospective data from 18 Surveillance, Epidemiology, and End Results registries, we performed multiple logistic regression to generate odds of thick (>4 mm) versus thin (≤2 mm) NM using patient-level demographics and tumor characteristics as well as county-level socioeconomic status, healthcare access, and preventive service use as predictors. We identified 10 475 patients with NM. Divorced, separated, and widowed individuals had increased odds of thick versus thin NM compared with married individuals [odds ratio (OR): 1.47, P<0.001], as did never married individuals (OR: 1.63, P<0.001). Individuals who lived in areas with high (≥4/100 000) dermatologist density had decreased odds of thick versus thin NM (OR: 0.75, P=0.02). Women aged 18-65 years who lived in areas with high Papanicolaou test use and women aged 40 years and older who lived in areas with high mammography use had decreased odds of thick versus thin NM (OR: 0.60, P<0.001 and OR: 0.68, P<0.001, respectively). High school completion, median income, insurance status, family medicine physician density, and internist density in one's area of residence were not associated with odds of thick versus thin NM. In contrast to other types of melanoma, NM thickness at diagnosis is not associated with the socioeconomic environment in one's area of residence. Instead, it is related to factors associated with incidental or formal detection. This information should be used to better target melanoma education and early detection programs.