Thursday, July 10, 2003

This presentation is part of : Health Informatics

Analysis of Medical Device Errors Using Grounded Theory

Juliana J. Brixey, MSN, MPH, RN, Doctoral Student, Todd R. Johnson, PhD, Associate Professor, Jiajie Zhang, PhD, Associate Professor & Associate Dean for Research, and James P. Turley, PhD, RN, Associate Dean for Academic Affairs. School of Health Information Sciences, University of Texas Health Science Center - Houston, Houston, TX, USA
Learning Objective #1: Describe how grounded theory was used to analyze medical errors
Learning Objective #2: Explain the need for an improved medical errors reporting form

Objective: The objective of this study was to use grounded theory to code and analyze MedWatch reports received by the US Food and Drug Administration for volumetric infusion pump errors. The infusion pump was chosen as the medical device to study because of continued reported user error and patient harm.

Design: The study was a retrospective secondary analysis of reported volumetric infusion pump errors that are publicly available on the World Wide Web.

Sample: MedWatch reports received by the FDA from July 31, 1996 through June 27, 2002 for specific infusion pumps were selected for analysis. The sample consisted of 312 MedWatch reports.

Concepts: Volumetric infusion pump errors.

Methods: The Manufacturer User Facility Device Experience Database was queried using search terms for specific brand names of commercially available infusion pumps. The reports were stored in a Microsoft Access Database. The free text section for each report was selected and imported into NVivo. Constant comparative coding was used for analysis of the free text.

Findings: A limited framework began to emerge from the reports. A lack of detail in the free text of the MedWatch reports hindered development of a comprehensive, cognitive framework to explain user error involving volumetric infusion pumps.

Conclusion: Although grounded theory has the potential to be a useful methodology to code and analyze medical device errors, the lack of detail in the MedWatch hinders theory development.

Implications: It is suggested that the MedWatch form be modified to support reporting of user error. Improved reporting would lead a better understanding of medical device errors.

This project is supported by a training grant from the Keck Center for Computational and Structural Biology of the Gulf Coast Consortia (NLM Grant No. 5 15LM07093), Agency for Healthcare Research and Quality (Grant No. PO1 HS11544), and National Medical Technology Testbed (Grant No.DAMD17-97-27016).

Back to Health Informatics
Back to 14th International Nursing Research Congress
Sigma Theta Tau International
10-12 July 2003