Algorithm development for identifying breast cancer incident cases and epidemiological updates: A cohort study based on multiple secondary sources

Purpose: This study aimed to develop and validate an algorithm for identifying incident breast cancer (BC) cases using Healthcare Utilization Databases (HUDs) and to assess BC incidence trends in the Marche Region, Italy, from 2010 to 2021. Methods: This population-based longitudinal study included women aged ≥ 18 years residing in Marche. The HUDs Algorithm was developed to identify new BC cases using hospital discharge, outpatient, and beneficiary databases, and it was validated against the Cancer Registry by evaluating agreement, sensitivity, and positive predictive value (PPV). Age-standardized BC incidence rates were estimated. A Poisson regression model was used to assess trends, including comparisons between pre/post COVID-19 pandemic periods. Results: Validation results showed a sensitivity of 81.2 % and PPV of 85.0 %. A total of 18,158 incident BC cases were identified, with a mean incidence rate of 224.7 per 100,000 person-years (95 % CI: 221.5–228.0). No significant increase in BC incidence was observed over time, but a marked decline occurred in 2020–2021, likely due to COVID-19-related disruptions. Conclusions: HUDs can be a valuable complementary data source, providing additional information useful for timely epidemiological surveillance and supporting rapid public health responses in cases where Cancer Registry data are delayed. Further refinements and integration with other data could enhance the accuracy of the HUDs Algorithm.