April 22, 2021
A chronic disease’s estimated annual cost varies depending on the method used to arrive at the calculation, with inconsistent differences across diseases. Researchers presented the finding in a poster at the AMCP 2021 virtual meeting.
“Studies on health care cost often rely on administrative claims data,” they wrote. “While a variety of methods exist to estimate annual costs, the limitations that arise due to different assumptions are rarely discussed.”
For the study, researchers compared three methods to estimate annual costs for two chronic disease populations: 72,840 women with endometriosis, and 14,117 patients with type 2 diabetes and chronic kidney disease. Patients in each cohort were identified using IBM MarketScan claims data and were followed until disenrollment, end of data, or a maximum follow-up of 365 days.
Method A calculated annual cost only using patients with 365 days of follow up, according to the study. Method B calculated annual mean population costs by dividing the sum of costs for all patients by the sum of follow-up days for all patients, then multiplying by 365. For Method C, annual patient-level mean costs were calculated by dividing each patient’s cost by their number of follow-up days, then multiplying by 365.
Method C produced the highest estimates, researchers found. Total costs were $32,379 for endometriosis, and $31,743 for type 2 diabetes and chronic kidney disease. Meanwhile, Method A produced the lowest estimates, with mean total costs of $20,950 for endometriosis and $27,816 for type 2 diabetes and chronic kidney disease. Using Method B, annual costs were $22,868 for endometriosis and $29,952 for type 2 diabetes and chronic kidney disease, according to the study.
Researchers surmised that cost estimates using Method A may be less generalizable because of a selection bias of healthier, lower cost patients. On the other hand, Methods B and C may overestimate costs due to extrapolation of uneven costs throughout the year. Considering skewed distributions of costs within claims data, Methods A and C may be preferable to Method B because they assess cost variability, researchers wrote.
Nonetheless, “transparency of methodology and sensitivity analyses of multiple approaches are necessary to understand how health care costs may be over- or underestimated,” they concluded.
Kelly A, Kuang K, Gibbs L, Luo K, Petruski-Ivleva N. Comparing methods for estimating annualized costs using administrative claims. Poster presented at AMCP 2021 Virtual Meeting; April 12-16, 2021.