Algorithm development for identifying breast cancer incident cases and epidemiological updates: A cohort study based...

Area of Science:

• Epidemiology
• Public Health
• Health Informatics

Background:

• Breast cancer (BC) surveillance is crucial for public health.
• Timely identification of BC cases is essential for effective epidemiological monitoring.
• Cancer Registries are valuable but can experience data delays.

Purpose of the Study:

• To develop and validate an algorithm using Healthcare Utilization Databases (HUDs) for identifying incident breast cancer cases.
• To assess breast cancer incidence trends in the Marche Region, Italy (2010-2021).
• To evaluate the impact of the COVID-19 pandemic on breast cancer detection rates.

Main Methods:

• A population-based longitudinal study of women aged 18+ in Marche, Italy.
• Development of the HUDs Algorithm using hospital discharge, outpatient, and beneficiary databases.
• Validation against the Cancer Registry, assessing sensitivity (81.2%) and positive predictive value (PPV) (85.0%).
• Estimation of age-standardized BC incidence rates and trend analysis using Poisson regression.

Main Results:

• The HUDs Algorithm demonstrated good performance with 81.2% sensitivity and 85.0% PPV.
• A total of 18,158 incident BC cases were identified, with a mean incidence rate of 224.7 per 100,000 person-years.
• No significant long-term increase in BC incidence was observed, but a decline occurred in 2020-2021, attributed to COVID-19 disruptions.

Conclusions:

• Healthcare Utilization Databases (HUDs) offer a valuable supplementary data source for BC surveillance.
• The HUDs Algorithm can support timely epidemiological surveillance and public health responses.
• Further integration and refinement of the HUDs Algorithm can improve accuracy and utility.