Effect of an Educational Intervention on the Accuracy of Data Submitted to a National Quality Registry.

Background: We hypothesize that data-entry errors within the National Cardiovascular Data Registry® (NCDR) ICD Registry™ may be an important reason behind labeling many cases as nonevidence-based.

Objective: To describethe frequency of data-entry errors in implantable cardioverter-defibrillator (lCD) implant data from our institution and develop a plan for quality improvement using the Deming cycle.

Methods And Results: We assessed data of patient report forms from2007to 2010 and compared these data with forms submitted from 2011 to 2012 after implementation of a continuous multicomponent staff education and training program. Of 211 ICD implants between 2007 and 2010, 36 (17%) were labeled nonevidence-based. Twenty-four (11.4%) resulted from misclassification due to data entry errors and 12 (5.7%) were actual nonevidence-based. Postintervention, review of 97 submitted patients' data revealed one (1%) data-entry error and three (3.1%) actual nonevidence-based implants.

Conclusions: Multicomponent educational intervention was effective in reducing errors in data sub- mitted to the NCDR ICD Registry.