Introduction: The impetus toward robust linkage between patient-centered, high quality care and reimbursement continues to gain traction with the advent of alternative payment models, quality registries and national healthcare policy. Our objective was to construct a broad composite score across all disease states, determine drivers of quality-variation and benchmark treatment centers.
Methods: We used a claims-based algorithm to describe quality in kidney cancer care across the disease spectrum. SEER-Medicare data were used to identify patients with renal cell carcinoma diagnosed 2004-2013 including demographic, socioeconomic, oncology and treatment-center data. We identified 11 quality metrics from a previously published Delphi process and expanded on a prior 5-metric renal cancer quality score.
Results: 33,434 patients with kidney cancer treated at 1,021 hospitals were identified. 20% of patients were non-white, and 30% rural. 54% of hospitals were teaching hospitals with median annual Medicare nephrectomy volume of 8 (IQR 2-16, note that Medicare represents a fraction of the patient-mix, actual volume is higher). The sample size allowed statistical power to study “learning curve” procedures such as caval thrombectomy and partial nephrectomy, with 1,930 (5.6%) having caval thrombi and 13,314 (42%) having cT1a masses. Additionally, there were 5,274 patients with metastatic disease. Rate of hospital compliance with metrics are shown in Table. Variation in metric-compliance by hospital nephrectomy volume was dramatic in some measures and nearly flat in others (Figure).
Conclusions: We identify and describe variation in consensus-based quality metrics in renal cancer allowing for future construction of an expanded claims-based kidney cancer score that can be operationalized in quality registries and reimbursement models. Source of
