[Cost variation in care groups?]

Area of Science:

• Health economics
• Healthcare management
• Diabetes care research

Context:

• Benchmarking healthcare provider performance is crucial for quality improvement.
• Traditional methods of cost analysis may not accurately reflect care group performance.
• Diabetes patient data from Dutch care groups provide a large-scale dataset for analysis.

Purpose:

• To evaluate the suitability of simple mean costs for comparing diabetes care groups.
• To determine if total patient costs accurately indicate care group performance.
• To compare the outcomes of simple mean cost analysis versus a generalized linear mixed model.

Summary:

• A study analyzed 2009 insurance claims for 104,544 diabetes patients in the Netherlands.
• Simple mean costs showed no significant differences between care groups, while a generalized linear mixed model revealed some significant variations.
• The ranking of care groups varied depending on the analysis method used.

Impact:

• The choice of benchmarking methodology significantly impacts the perceived differences between healthcare providers.
• Simple mean cost calculations can overestimate inter-group cost disparities.
• Generalized linear mixed models provide a more accurate comparison of care group performance, with diabetes-specific hospital costs being a more sensitive indicator than total curative care costs.