A PROMISE for improvement: the ProRail Management Information for Safety and Environment database Linda Wright ProRail
Overview Background to the PROMISE database and role in the SMS Purpose of PROMISE The generic framework for a reporting system (decisions at each stage) Implementation of the database Reliability of the data Evaluation of the database Working on Safety 2006 / 2
Background PROMISE and role in the SMS SMS introduced in 2004 – reporting and investigation of incidents is a central function in the SMS Old database did not fulfill the requirements of the SMS Minimum requirement was the continuation of a recently introduced monthly report Working on Safety 2006 / 3
Working on Safety 2006 / 4
Purpose of PROMISE Monitoring of incidents: gain statistically reliable quantitative insight into factors or combination of factors that give rise to incidents Register and analyse data on cause and context Help the regions to learn from each other and identify countermeasures Evaluate the functioning of the SMS via the monthly reports and subsequent analyses Working on Safety 2006 / 5
Algemeen: in vergelijking tot afgelopen maand is er weer lichte verbetering te zien. Er zijn meer positieve dan negatieve ontwikkelingen te zien. Verbetering (van rood naar geel) op Ontsporing (1.1.4), Bijna-aanrijdingen derden (3.2.1) en (geel naar groen) Botsing obstakel (1.1.3), TOBS (1.2.4, vorige maand nog in kritisch gebied) en Aanrijding personeel/gereedschap (4.1.1). Positieve trendwijzigingen bij: Botsing stootjuk (1.1.2), Afvallen sein (1.2.3), Spoorstaafbreuk (1.2.8) en Gevaarlijke situaties werkzaamheden (4.2.2). Nieuwe rode vlakken zijn Gescheurde tong/puntstuk (1.2.6 – de negatieve trend van de afgelopen maanden heeft doorgezet), Vandalisme (3.2.4) en Lekkage ballast (5.1.1 – echter zeer laag aantal in absolute zin). De grafieken 1.2.7, 2.1.4, 3.2.4 en 4.2.2 zijn in vrijwel alle regio’s “rood”. Negatieve trends gesignaleerd bij Wissel open gereden (1.2.7) en Vee op baan (1.2.12). Opvallend is dat “Niveau I” bij overwegen wat slechter uitziet, vooral door de gevolgen van incidenten (doden/gewonden), terwijl “Niveau II” een overwegend positievere trend laat zien t.o.v. afgelopen maand. Working on Safety 2006 / 6
The generic framework for reporting Comprises 7 modules which should be prepared and planned prior to launching the system. Each will be discussed with reference to how ProRail implemented PROMISE Working on Safety 2006 / 7
1. Detection and reporting Type of reports Publicity & training Design reporting route Accidents, near misses, dangerous situations None required Use of current reporting route Working on Safety 2006 / 8
2. Selection of events for investigation Rules required Based on potential rather than actual consequenses Certain incidents always investigated Events per month – Severity/ Frequency matrix; Training of safety advisors in S/F matrix Working on Safety 2006 / 9
3.Detailed description & deeper study Select a method (e.g. Fault tree analysis) Detailed investigation necessary – training of investigators Causal trees Training in two levels of investigation – facts gathering & analysis Working on Safety 2006 / 10
4. Classification of root causes Select a taxonomy (the minimal requirements) Training Inter-rater reliability Set of root causes & not simply “the main cause” PRISMA-Rail (trial before implementation) Training for analysts Demonstrated on a small scale Database designed for multiple causes Working on Safety 2006 / 11
5. Computation: recognise patterns Design for statistical analyses Perform analysis Able to extract what is entered in different forms. Graphs, tables, word & excel files Monthly & quarterly reports, causal analysis Working on Safety 2006 / 12
6. Interpretation & implementation recommendations Recommendations are data driven Feasibility group Feedback channel for organisational learning Based in monthly reports and supplemental analysis by the Regions LCV & RVC Safety advisors monthly meeting, Management team, LCV & RVC Working on Safety 2006 / 13
7. Evaluation How to evaluate the recommendations Specific action fields in PROMISE Over time evaluate relationship between countermeasures, causes and events. Working on Safety 2006 / 14
Implementation of PROMISE Introduced January 2005 Incorporated in the SMS Training sessions in input & output 2 administrators Users group once a month PRISMA-Rail support group Quality control Working on Safety 2006 / 15
Reliability of the database 3 steps to data entry: Selection, assigning case type and incident category Inter-rater reliability of the incident types: showed good reliability for all steps Plans to perform same test 2 x year Small scale PRISMA-Rail reliability trial showed good reliability between 2 raters Larger scale inter-rater reliability for PRISMA-Rail planned for next year Working on Safety 2006 / 16
Evaluation of PROMISE Audit currently being performed New data source being implemented from November New training in Severity/ Frequency Matrix New inspection module for proactive information Better and faster management reports Working on Safety 2006 / 17