A microsimulation model based on healthcare pathways to estimate the impact of COVID-19 pandemic–induced delays on breast cancer mortality

Abstract

Objectives

This study aimed to develop a microsimulation model to estimate the impact of a care disruption during the COVID-19 pandemic on breast cancer mortality using an administrative database.

Study Design and Setting

Patient flows and pathways were assessed from the French hospital discharge database for four French cancer centers for patients with breast cancer from 2018 to 2021. Patients' return dynamics were compared to time series predictions to determine flow differences. Forecasted and observed patients were matched through healthcare pathways to account for possible delay heterogeneity among patients with breast cancer. Healthcare pathways were modeled and analyzed as sequences of states defined based on hospital treatments. Unmatched patients were reconsidered for matching the next month, with incrementing delays. We derived the number of expected additional cancer deaths at 5 years and the associated relative mortality rate using hazard ratios (HRs) associated with delays extracted from the literature. A deterministic sensitivity analysis was performed on HRs. Confidence intervals were computed for each outcome based on 1000 bootstrap replications.

Results

A forecasted population of 8125 patients with incident breast cancer was analyzed. An overall decrease of 20.8% in flows was estimated during the first lockdown. For the year following the beginning of the lockdown, 24.8% of patients were expected to receive delayed care, resulting in a 4.6% excess cancer mortality rate at 5 years among the 8125 forecasted patients.

Conclusion

Using an innovative approach based on patient-level data from an administrative database, our study further strengthens previous estimates of excess breast cancer mortality following the COVID-19 pandemic.