Our experience of classifying incident data

Each year Northern Rail produces an Operational Risk Analysis report, which identifies and documents patterns in their operational incidents. Each report is based on incidents which occurred in the period from January through... [ view full abstract ]

Each year Northern Rail produces an Operational Risk Analysis report, which identifies and documents patterns in their operational incidents. Each report is based on incidents which occurred in the period from January through December and is published by the end of March the following year. The contents of this analysis have grown organically over the years as the company’s understanding of human factors and incident causation has developed and influenced organisational processes.
The Operational Risk Analysis is largely based on factors highlighted by the incident investigators. At the beginning of this project incident investigators identified, among other things, the type of human error that caused the incident (ie slip, lapse, mistake, violation) and the non-technical skills involved. Investigators were also tasked with identifying any underlying factors relating to the individual, job or organisation, although these were not categorised in further detail and therefore not analysed within the Operational Risk Analysis.
In 2012, Northern Rail approached RSSB for a review of their Operational Risk Analysis, with a focus on whether it is accurate, and whether it reaches its full potential in terms of the information that can be extracted from the data, its presentation and the conclusions that can be drawn. As part of this project, RSSB and Northern Rail worked together to create a classification framework for incident data. This framework was applied retrospectively to the 2012 and 2013 incident data to understand its potential benefits. Northern Rail have since integrated the framework into their incident investigation processes and are producing an analysis of their incident data from 2014.
This paper will describe the approach taken to producing the incident data classification framework and supporting materials, and discuss Northern Rail’s experiences of using it.