Abstract 19419: Algorithm-Based Vendor Selection for Implantable Cardiac Devices Reduces Costs and Avoids Appearance of Conflict of Interest

• Published in: Circulation (New York, N.Y.) N.Y.), 2012-11, Vol.126 (suppl_21)
• Main Authors: Henrikson, Charles A, Stecker, Eric C, Stajduhar, Karl, Clark, Thomas, Kron, Jack
• Format: Article
• Language: English
• ISSN: 0009-7322; 1524-4539
• DOI: 10.1161/circ.126.suppl_21.A19419

Abstract only Introduction: In most hospitals, device vendor selection for an individual case is at the discretion of the implanting physician. This exposes the physician to the appearance of conflict of interest and may increase device costs. Methods: At a single academic center, implanting physicians and electrophysiology lab staff constructed and implemented a device selection algorithm based on referring provider request, device features, and equal distribution of cases among the four vendors in our geographic area. Costs for comparable devices were also incorporated into the algorithm, and lower cost devices were selected if all other factors were equal. For each individual case, the algorithm was followed by technical and administrative staff who selected and notified the vendor. Information was solicited from the implanting physicians to inform algorithmic selection. The selection algorithm could be overruled at discretion of implanting physician on clinical grounds. When manufacturer-specific features were selected or non-algorithmic clinical factors incorporated into the decision, the rationale was documented. Results: Over a three-year period of implementation, this algorithm resulted in a distribution of cases among vendors (15.4% Biotronik, 19.0% Boston Scientific, 35.4% Medtronic, and 30.0% St Jude Medical). Imbalances were due to referring provider preferences and the incorporation of cost considerations into the algorithm. In no instance did the implanting physician overrule the algorithm. There were no complaints from referring providers. As compared to prior to the implementation of the algorithm, approximately $750,000 was saved over three years, an 8.3% reduction in device costs. This was largely on the basis of algorithmic selection of less expensive (older generation) devices for appropriate patients. After two years, when vendor contracts were renegotiated, substantial further cost savings were realized. Conclusions: A device selection algorithm removes the daily task of vendor selection from the physician and removes the appearance of conflict of interest. The algorithm also results in a substantial cost savings for the hospital. We hope this model will find broad application in the electrophysiology community.