TRANSFUSION PRACTICE Errors in patient specimen collection: application of statistical process control Walter Sunny Dzik, Neil Beckman, Kathleen Selleng, Nancy Heddle, Zbigniew Szczepiorkowski, Silvano Wendel,and Michael Murphy for the Biomedical Excellence for Safer Transfusion (BEST) Collaborative BACKGROUND: Errors in the collection and labeling of blood samples for pretransfusion testing increase the risk of transfusion-associated patient morbidity and mor- tality. Statistical process control (SPC) is a recognized method to monitor the performance of a critical process. An easy-to-use SPC method was tested to determine its feasibility as a tool for monitoring quality in transfu- sion medicine. STUDY DESIGN AND METHODS: SPC control charts were adapted to a spreadsheet presentation. Data tabu- lating the frequency of mislabeled and miscollected blood samples from 10 hospitals in five countries from 2004 to 2006 were used to demonstrate the method. Control charts were produced to monitor process stability. RESULTS: The participating hospitals found the SPC spreadsheet very suitable to monitor the performance of the sample labeling and collection and applied SPC charts to suit their specific needs. One hospital moni- tored subcategories of sample error in detail. A large hospital monitored the number of wrong-blood-in-tube (WBIT) events. Four smaller-sized facilities, each fol- lowing the same policy for sample collection, combined their data on WBIT samples into a single control chart. One hospital used the control chart to monitor the effect of an educational intervention. CONCLUSION: A simple SPC method is described that can monitor the process of sample collection and label- ing in any hospital. SPC could be applied to other criti- cal steps in the transfusion processes as a tool for biovigilance and could be used to develop regional or national performance standards for pretransfusion sample collection. A link is provided to download the spreadsheet for free. E rrors in the collection and labeling of patient samples used for pretransfusion testing are known to be an important source of transfusion- related patient morbidity and mortality. In England, the Serious Hazards of Transfusion program has identified errors made during sample collection and label- ing as important components of transfusion error. 1 In the United States, Linden and colleagues 2 documented that 14 percent of ABO-incompatible transfusions results from sample collection errors. Many more “near-miss” events occur, but adverse consequences are often prevented by stringent criteria for sample acceptance and by careful comparison of the patient’s results to historical results (Fig. 1). 3 Thus, process failure in sample collection is many times more frequent than would be estimated by counts of ABO-incompatible transfusions. Process failure rates can be estimated by tracking two general types of errors, mislabeled samples and miscol- lected samples. 3 A miscollected sample represents a ABBREVIATIONS: BEST = Biomedical Excellence for Safer Transfusion Collaborative; SPC = statistical process control; UCL = upper control limit; WBIT = wrong blood in tube. From the Massachusetts General Hospital, Boston, Massachusetts; the Laboratory Service, National Blood Service, Birmingham, UK; the Department of Transfusion, Ernst-Moritz- Arndt-Universität, Greifswald, Germany; the Department of Medicine, McMaster University, Hamilton, Ontario, Canada; the Department of Pathology and Medicine, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire; the Blood Bank, Hospital Sirio Libanes, São Paulo, Brazil; and the Department of Transfusion, John Radcliffe Hospital, Oxford, UK. Address reprint requests to: Walter Sunny Dzik, MD, Blood Transfusion Service, J-224, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114; e-mail: sdzik@partners.org. Received for publication January 14, 2008; revision received April 18, 2008; and accepted April 23, 2008. doi: 10.1111/j.1537-2995.2008.01824.x TRANSFUSION 2008;48:2143-2151. Volume 48, October 2008 TRANSFUSION 2143