Contemporary volume-outcome relationships for percutaneous nephrolithotomy: results from the Nationwide Inpatient Sample

Abstract

Introduction/objective: We sought to examine the contemporary relationship between case volume and outcome for percutaneous nephrolithotomy (PCNL) using a publically available administrative database.

Methods: A weighted sample of 7785 patients was obtained from the 2010 Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample (NIS). ICD-9-CM diagnostic codes were used to identify patients with urolithiasis (592.0, 592.1, and 592.9) who underwent percutaneous nephrostomy with fragmentation (5504). Charlson Comorbidity Indices (CCI) were calculated based on diagnostic codes for all patients. Hospital case volume was quartile classified and we then compared key outcomes (the complication rate, transfusion rate, length of stay [LOS], and in-hospital mortality rate) by volume quartile. We then performed multivariate analysis to examine the effect of CCI, annual volume, and age on key outcomes.

Results: The overall complication rate was 17% in the weighted sample. In univariate analysis, statistically significant variation in the complication rate, CCI, transfusion rate, and in-hospital mortality was noted with regard to the hospital volume. The complication rate and transfusion rates varied by case volume, but in a nonlinear fashion, wherein rates were highest at the lowest and highest volume centers. CCI was strongly predictive (p<0.001) of complications and LOS in the multivariate analysis. Case volume was only predictive (p=0.042) of LOS in the multivariate analysis.

Conclusion: Annual case volume of the treating center was associated with shorter LOS after PCNL, but case volume was not independently predictive of complication or transfusion in multivariate analysis. CCI was a strong independent predictor of complication and LOS.