Framework for Measurement and Prevention of Human Error in Service Delivery

• Activity Number: 358
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Quality and Productivity
• Author(s): Larisa Shwartz*+ and Genady Grabarnik
• Companies: IBM Thomas J. Watson Research Center and St. John's University
• Keywords: human error ; service delivery ; automation and process improvement

Abstract

The theory of human error addresses interactions between human performance variability and situational constraints. The occurrence and frequency of human error more greatly depend on interactions with environment than stable and inherent characteristics of the operator or task. In our study, we applied two approaches to understanding the pervasiveness of human error. We analyzed incident tickets from a large service provider and surveyed service delivery experts (SMEs). We requested that the SMEs classify incidents into three categories: 'human error,' 'not a human error,' or 'maybe caused by human'. We discovered a significant discrepancy in the estimation we got from these approaches. We examined and evaluated the reasons for the discrepancy and concluded that quantitative measures, such as error rates, could be inadequate and misleading. Thus, meticulous and consistent descriptions of the error-prone conditions must be considered. Furthermore, we considered an impact of automation on human error prevention and process improvement, and provided recommendations for the usage of these methods, illustrating them on specific use-cases.