Persistent inconsistencies in patient cost variability within the French DRG classification system over the 2012–2019 period

This paper evaluates the effectiveness of the 2009 French Diagnosis-Related Group (DRG) classification reform, which introduced four severity levels within each DRG, ranging from low to very high, with corresponding increases in fixed-price reimbursements. Notably, the reform incorporates the Medicare Severity Diagnosis-Related Group (MS-DRG) system, first implemented in the United States in 2007, giving the French system international relevance. The French Public Health Insurance system (NHI) reimburses both public and private healthcare establishments through a DRG-based payment system. This study focuses on variations in hospital resource costs for four different health conditions. The paper begins by discussing the theoretical challenges of constructing DRG categories, particularly the trade-off between greater clinical detail (granularity) and the risk of distorting incentives for hospital efficiency. It then presents an empirical analysis of hospital resource cost variations both within and between DRGs for the same pathology or clinically meaningful group (DRG-root), using data from 2012 to 2019. Our findings suggest that a one-size-fits-all approach to severity classification is inadequate. In some cases, broader categories improve statistical validity, while in others, more granular distinctions are necessary. We conclude that a tailored, case-by-case approach is the most effective solution. Specifically, the analysis reveals significant overlap in confidence intervals for hospital resource costs across DRG severity levels, suggesting that the current classification system fails to effectively capture cost differences related to severity. Additionally, a large portion of cost variation within DRGs is driven by factors unrelated to severity, such as hospital-specific characteristics. Overall, the results underscore the need to revise the current DRG system in France in order to reduce financial discrepancies and to prevent incentives for patient selection, especially before implementing bundled payment models that include both inpatient and outpatient care.