Objective: Evidence suggests that patients requiring high-risk procedures benefit from care at institutions providing a large volume of these procedures. Our objective was to determine whether there is a volume-outcome relationship among intensive care unit patients receiving renal support therapy in two different healthcare systems (France and the United States).
Design: Retrospective cohort study.
Setting: Two multicenter intensive care unit databases: CUB-Réa (France) and Project IMPACT (United States).
Patients: All nonsurgical adults requiring renal support therapy from 1997 to 2007 were included.
Measurements and Main Results: We assessed association of annual renal support therapy volume with intensive care unit and hospital mortality using multivariable modeling, accounting for clustering and adjusting for age, comorbidities, admitting diagnosis, illness severity, pre-intensive care unit length of stay, admission source, and hospital and intensive care unit characteristics. Our final cohorts were 9,449 patients treated in 32 intensive care units in CUB-Réa and 3,498 patients treated in 76 intensive care units in Project IMPACT. Patient demographics did not differ between cohorts. Renal support therapy delivery varied widely across intensive care units (3–129 patients per year in CUB-Réa, 1–66 in Project IMPACT). Overall intensive care unit and hospital mortality rates were 45% and 49% in CUB-Réa and 34% and 47% in Project IMPACT. After adjustment for patient, intensive care unit, and hospital characteristics, there was no association between renal support therapy volume and intensive care unit or hospital mortality whether we treated volume as a continuous measure or quartiles. Higher renal support therapy volume was associated with shorter length of stay only in CUB-Réa.
Conclusions: There is a large variation in annual renal support therapy volume across intensive care units in France and the United States but no association of higher volumes with improved outcomes.
From the Clinical Research, Investigation, and Systems Modeling of Acute Illness Laboratory (YLN, EBM, LW, JMK, GC, JAK, DCA), Department of Critical Care Medicine, University of Pittsburgh Medical Center, Pittsburgh, PA; Centre d'Epidémiologie Clinique, Hôpital Hôtel-Dieu (YLN), Service de Réanimation Médicale, Hôpital Cochin (JDC), and Service de Réanimation Médicale, Hôpital Saint-Antoine (BG), Assistance Publique des Hôpitaux de Paris, Paris, France; Department of Biostatistics (LW) and Department of Health Policy and Management (JMK, DCA), Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA; Département de Biostatistiques (PA), Hôpital Ambroise Paré, Assistance Publique des Hôpitaux de Paris, Boulogne, France; Unité 707 (BG), Institut National de la Santé et de la Recherche Médicale, Paris, France; and Université Pierre et Marie Curie Paris VI (BG), Paris, France.
This work was performed at the Clinical Research, Investigation, and Systems Modeling of Acute Illness Laboratory, Department of Critical Care Medicine, University of Pittsburgh Medical Center, Pittsburgh, PA.
Dr. Nguyen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Supported, in part, by a grant provided by the 2008 European Critical Care Research Network Levi-Montalcini Biomedical Science Award of the European Society of Intensive Care Medicine (Brussels, Belgium) and by grant NHLBI K23 HL078760 from the National Heart, Lung, and Blood Institute (Bethesda, MD).
Dr. Kellum consulted for and received grants from Gambro (Stockholm, Sweden) and Baxter (Deerfield, IL); however, these disclosures are not associated with this study. The remaining authors have not disclosed any potential conflicts of interest.
Statistical code: Available from Dr. Weissfeld. E-mail: firstname.lastname@example.org
Data set Project IMPACT: available for purchase from Cerner (Kansas City, MO; http://www.cerner.com/piccm).
Data set CUB-Réa: available through the Collège des Utilisateurs de Bases de données en Réanimation network (Paris, France; http://www.pifo.uvsq.fr/hebergement/cubrea/cr_index.htm).
Address requests for reprints to: Yên-Lan Nguyen, MD, Clinical Research, Investigation, and Systems Modeling of Acute Illness Laboratory, Department of Critical Care Medicine, University of Pittsburgh Medical Center, 3550 Terrace Street, Pittsburgh, PA 15261. E-mail: email@example.com