Disparities in timeliness of cancer diagnosis across a multi-site academic health system

Background: Early cancer diagnosis is crucial for improving survival and quality of life. Understanding which patient populations are at higher risk for late-stage diagnosis within a healthcare system is a crucial step toward achieving health equity. This study examines demographic, clinical, and socioeconomic factors associated with late-stage diagnosis, defined as stage III/IV diagnoses using the American Joint Committee on Cancer (AJCC) criteria. Methods: A retrospective cohort study of 12,721 patients diagnosed with breast, colorectal, or lung cancer within the University of California health system database. Multivariable logistic regression was used to identify various factors associated with late-stage diagnosis. Results: 14.1% of patients were diagnosed at a late-stage (stage III/IV). In adjusted analyses, patients with colorectal (OR: 4.76; 95% CI: 4.12-5.50) and lung cancer (OR: 12.82; 95% CI: 11.01-14.92) had significantly higher odds of late-stage diagnosis compared to breast cancer (reference group). A higher Area Deprivation Index (ADI) was significantly associated with increased odds of late-stage diagnosis (OR: 1.03 per 1-unit increase; 95% CI: 1.01-1.05). Neither race nor ethnicity was an independent predictor of late-stage diagnosis in the final adjusted model. Conclusion: Cancer type is the strongest predictor of a late-stage diagnosis. Additionally, greater neighborhood-level socioeconomic disadvantage (measured by ADI) is a significant, independent risk factor. These findings highlight the need for targeted screening and diagnostic interventions that incorporate geospatial analyses, particularly for lung and colorectal cancer patients in socioeconomically disadvantaged communities.