evaluation survey nsa - final

27
Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector ERA Questionnaire on the Use of Accident Precursor Data by NSAs Survey evaluation report Version 2 (Final, including additional comments from NSAs) Date 06 February 2013 Written by Vojtech EKSLER Controled by Iskra STOJANOVA Approved by Christopher CARR

Upload: vuongque

Post on 14-Feb-2017

223 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

ERA Questionnaire on the Use of Accident Precursor Data by NSAs

Survey evaluation report

Version 2 (Final, including additional comments from NSAs)

Date 06 February 2013

Written by Vojtech EKSLER

Controled by Iskra STOJANOVA

Approved by Christopher CARR

Page 2: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Contents

Introduction……………………………………………………………………………………………………………………………………….1

Objectives………………………………………………………………………………………………………………………………….………2

Methods……………….………………………………………………………..…………………………………………………………………3

Results.………………………………………………………………………………………………………………………………………..…….4

Conclusions……………………………………………………………………………………………………………………………………...13

ANNEXES

Annex A: Replies to the survey……………………………………………………………………………………………………….…14

Page 3: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Summary

The Agency has been reviewing the framework for railway safety performance monitoring as part of

the revision of the CSTs/CSIs frameworks. Six accident precursors are nowadays collected at EU level

under CSIs reporting framework, but there are no further provisions in place for the precursor data

monitoring and management. However, the common EU regulatory framework for railway safety

may imply a common approach to accident precursor monitoring and management at different levels

of safety management.

This questionnaire was aimed at understanding the accident and incident data reporting regimes that

are in place in all Member States, with the view to provide evidence for eventually reviewing the

framework for accident/incident data monitoring in the EU.

The answers provided to ERA by all NSAs show that monitoring of non-significant accidents is

widespread at country level in the EU. Most typically, the countries collect data on the six types of

accidents defined under CSIs for which the outcome threshold criteria were not met (accidents with

no fatality, or serious injury, or significant damage). Non-significant collisions, derailments and fires

are the most commonly reported non-significant accidents in Member States. The purpose of

collecting non-significant accidents (numbers) is the general monitoring of railway safety at country

level. The data is seldom used for planning supervision activities by the NSA.

The majority of countries collect accident precursors in addition to those six defined under CSIs. On

average, an additional 11 accident precursor indicators are monitored at a national level. The

indicators are usually country-specific and only a few national indicators find equivalence in other

countries. The accident precursor indicators collected by more than one country are highlighted in

this report.

The management of precursors data shows several common points (e.g. NSA is typically the keeper

of the database and its key user), but there are great differences between NSAs when it comes to the

purpose and use of this data. Notably, it is rare to see a systematic use of precursor indicators by

NSAs for determining supervision needs at country level. The differences in reporting regimes and in

the analytical practice by the NSA further highlights the absence of a common approach of NSAs to

precursor data. It became obvious that the tools for managing accident precursor data are country-

specific: each MS has developed its own database and applies different reporting provisions. Last,

but not least, two reporting regimes have been identified in respect to the type of information

collected: about ten countries systematically and instantly collect information about particular

occurrences (including underlying data), while up to ten countries only collect the aggregated data at

regular time intervals.

There is certain support among NSAs for the extension of precursors monitoring at EU level, yet they

do not see a need for an urgent action. Some NSAs take a view that the precursor indicator analysis

fall under the competence of NSAs, rather than of the Agency or the EC. It was suggested that the

Agency may however provide guidance on the use of precursor data by the NSAs.

Page 4: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 1

1 Introduction

‘Precursors to accidents’ are indicators of incidents that under other circumstances could have led to

an accident. Precursors include for example near-misses, incidents, technical, human and

organisational failures. As accidents on railways are rare, the monitoring of less serious events

occurring on railways is an essential tool in a proactive Safety Management System (SMS). In Europe,

the monitoring and assessment of precursors take place at three levels of railway safety

management: EU level by ERA, national level by NSAs and operational level by RUs and IMs.

Precursors are measures based on events occurring in the accident causal chain (typically when one

or more safety barriers have been breached). They may include system or component failures,

human errors and operational conditions. They are sometimes classified as either leading or lagging.

The Railway Safety Directive (RSD) introduced a framework for a common monitoring of railway

accident data, with the focus given to outcome indicators allowing for the safety monitoring and

management at the EU level. Under this framework, a limited number of precursors are subject to

reporting by NSAs to the Agency under the Common Safety Indicators (CSIs). The indicators reported

nowadays to the Agency are:

broken rails

track buckles

signals passed at danger (SPAD)

wrong-side signalling failures

broken wheels

broken axles

The figures on certain types of precursors could be used by the Agency as complementary

information that can be used for the monitoring of the safety and of the quality of the underlying

SMS. They could ultimately be used as an input for the revised assessment method for the

assessment of CSTs/NRVs.

Accident precursors can be subject of investigation by the National Investigation Bodies (NIBs), who

may investigate “incidents which under slightly different conditions might have led to serious

accidents”. The number of accident precursors investigated by NIB is however limited (about 15 per

year for all EU countries). This makes it impossible to perform aggregated data analysis on accident

investigation data.

Accident precursor indicators are used by National Safety Authorities (NSAs) as measures of safety

performance, complementary to outcome indicators (accidents, casualties, ….). They enable the NSAs

to plan (prioritize) their supervision activities, through the identification of risk areas and of the

underperforming Railway Undertakings (RUs) and Infrastructure Managers (IMs).

The harmonised EU railway safety regulatory may imply a harmonised approach to the monitoring of

precursors by NSAs. Currently, the Agency does not have knowledge about the extent of a systematic

monitoring of precursors by NSAs. This type of information is not systematically assessed by any of

the current Agency’s activities, including the cross-audit programme.

Page 5: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 2

In order to determine needs for accident precursor analysis and management at different levels of

safety management, the Agency has commissioned in 2012 the Transport Research Laboratory (TRL)

to conduct on its behalf the “Prospective study into harmonized train accident precursors analysis

and management”. The study, whose results are due in March 2013, has three main objectives:

Establish a theoretical basis for the monitoring and management of precursors by defining a common set of causal trees and by identifying those precursor indicators with highest relevance to the safety management at EU/NSA/RUs, IMs levels.

Provide an overview on the precursors monitoring practice at the level of NSA and RU/IM in a sample of EU countries in order to understand the current practice and the motivations behind it.

Determine a potential harmonised set of precursor indicators for different levels of safety management/governance (RU/IM, NSA, ERA).

It was agreed between the TRL and the Agency that the survey on the use of accident precursors data

by NSAs should be carried out by the Agency, as it has the direct access to the NSAs through its NSA

network and it further allows it to obtain additional information that are out of scope of the study

carried out by the TRL.

The survey was thus prepared and run by the ERA in the two last months of year 2012. Its results are

presented in this evaluation report and will also be incorporated into the study by TRL.

2 Objectives

Given the need to better understand the NSA approaches to the use of accident precursor data at

country level in particular and the NSA approaches to railway safety monitoring in general, the

objectives of this survey was twofold:

Assess which accidents other than significant accidents the NSAs monitor and how they manage the precursor data at the national level

Assess which accident precursors / indicators the NSAs systematically monitor and how they manage these data at the national level

The survey also sought to better understand the distribution of responsibilities in respect of

precursor data collection and management at country level, the extent of the use and of the

dissemination of these data at national level. Last, but not least, the questionnaire sought to collect

views of the NSAs on the need for more harmonization in the use of accident precursor data in

Europe.

Page 6: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 3

3 Method

In order to gather the information of interest, a survey was run among all NSAs. The standard MS

Excel-based questionnaire was prepared and distributed to the NSA network members via email on

16 November 2012. The questionnaire uses the forms that allow the respondent to provide

standardized answers, but without being limited on the extent and nature of information provided.

Free text boxes were largely used in the form. The background for the questionnaire together with

contact information was provided on a dedicated sheet of the same Excel file. The answered

questionnaires were sent by NSAs to ERA via email. In some instances, the questionnaire file was

accompanied with an extract of national legislation or glossary of terms.

The questionnaire was split into three parts. The first part (Part A) aimed at understanding which

accidents, other than significant accidents as defined in 149/2009/EC, are collected at national level.

The second part (part B) aimed at getting an overview of all accident precursor indicators monitored

at the national level. The last part of the questionnaire addressed the issues relevant to data

management (database, publication, analysis). Last, but not least, the questionnaire sought opinions

of NSAs on a need to harmonize more accident precursor indicators at EU level.

The replies were received from all 26 NSAs by the end of the year 2012. Some of them had to be

followed due to the lack of clarity or incompleteness of the reply. All replies were eventually

compiled in one file and analysed in MS Excel. The summary of the replies together with the analysis

of findings provide the core of this report. While the key conclusions are presented at the end of the

report, the discussion about their practical implication is limited. This reflects the objectives of the

survey to provide an overview and understanding, rather than providing advice, or decisions.

The preliminary survey evaluation report is being sent to the NSAs for their verification, and for

additional comments. The results shall be presented at the workshop on accident precursors

organized by ERA and TRL in early 2013, where its conclusions shall be subject to discussion by

participants. The final version of the report shall be published on the Agency’s internet webpage.

Page 7: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 4

4 Results

Other than significant accidents monitored at EU countries

Altogether 20 out of 27 EU member states monitor (collect number of events) other than significant

accidents. These countries typically collect data of at least two specific indicators.

Bulgaria, Germany, Spain, Luxembourg, Norway, Romania, and Slovakia monitor only significant

accidents as introduced by the RSD in its Annex I.

There is a clear pattern for the 20 countries in the list of accident indicators: most of them collect the

data for five (six) types of accidents that would have been classified as significant (in RSD) had they

resulted in the casualties/material damage required for the classification as significant accident.

Often, these accidents that do not meet criteria of “significant accident” are depicted as “minor

accidents” in the national regulations/classifications.

In Austria, Czech Republic, Ireland, Slovenia and the UK, the number of various types of collisions is

separately monitored. The examples include collisions during shunting operations (AT, CZ, SI, UK),

collisions on sidings (IE), collisions with buffer stop (UK), collisions with animals (IE), collisions with

vehicle outside of level crossing (CZ, IE, UK), roll back collision (UK), open door collision (UK).

Besides significant and non-significant accidents for five (six) types and the disaggregated collision

type, there are only a few accident types other than those defined in the RSD reported to be

collected at MS level: These include electrocution (NL), Accidents by shunting (SI, IE), or accidents at

sidings (IE). It seems that only in some Member States, the historical classification of accidents was

replaced by the classification introduced by the RSD, so that two reporting regimes exist at country

level in several countries. One country (IT) seems to use UIC classification for non-significant

accidents.

Accidents other than significant accidents are typically reported by IMs (19 countries out of 20). The

RUs seldom report these events less often: the practice exists in 11 countries). In Sweden, anyone

can report the event, while in Netherlands a public internet platform has been set up to allow all

users, including members of the public, to enquire about an occurrence. In all countries, these events

are reported routinely, but the frequency of reporting differs greatly. The frequency of reporting of

accidents other than significant accidents is shown in Figure 1. For each level of reporting frequency,

the number of countries is shown in the figure, followed by the relative proportion of the 18

countries that provided this information. In 10 out of 18 countries in which non-significant accidents

are reported and for which the frequency of reporting was provided, the data are reported daily.

The monitoring of non-significant accident data is typically part of the general railway safety

monitoring, complementing the CSI reporting. Up to 15 countries reported that the monitoring is

part of a general railway safety monitoring, while 10 countries stated that one of the end purposes is

the supervision of RUs/IMs. It is not clear why 5 of the countries that collect non-significant accident

data do not use these data for the supervision of RUs/IMs/ECMs and what information they use for

planning supervision activities at the national level. Belgium, Estonia and Sweden listed NIB

investigation among the objectives of the reporting. In Belgium, the NIB is the holder of the national

Page 8: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 5

database of accidents and incidents. In Estonia and Sweden, the data are reported to the NIB, which

then provides data to the NSA.

Figure 1: Frequency of reporting for non-significant accidents

The complete list of non-significant accidents monitored at the national level is summarized below,

showing the name of the indicator and the countries in which this indicator is collected.

Non-significant accident Countries

Train collision (CSIs) AT,BG,CZ,DK,FI,HU,IT,LT,LV,NO,RO,UK

Collision of trains CZ,NO,FR

Collision with an obstacle CZ,EE,FR,LV,NO

Collision with a vehicle on level crossing

UK

Collision with an animal IE,NO,UK

Collision with vehicle outside level crossing

CZ,UK

Open door collision UK

Roll back collision UK

Train derailment (CSIs) AT,BG,CZ,DK,FI,HU,IT,LT,LV,NO,RO,UK

Shunting movements AT

Outside mainline FR

Siding, line under works IT

Accidents to persons caused by rolling stock in motion (CSIs) AT,BG,CZ,DK,FI,HU,IT,LT,NO,RO,UK

Level crossing accident (CSIs) AT,BG,CZ,DK,FI,HU,IT,LT,NO,RO

Other LC accidents FR

Fires in rolling stock (CSIs) AT,BG,CT,CZ,DK,FI,HU,IT,LT,NO,RO,UK

Accidents involving dangerous goods (CSIs) AT,BG,RO

Other (CSIs)

Electrocution NL

Shunting accidents NO

Striking buffer stop SE

Struck by large falling object SE

Struck by missiles (airborne objects such as stones)

SI

Struck by large falling object UK,

Infrastructure failure (bridge, UK

10; 56%

1; 5%

1; 6%

2; 11%

2; 11%

2; 11% day

3 days

week

month

3 months

year

Page 9: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 6

tunnel collapse)

Fire/smoke on track/station

Other accidents with system error

Action of sabotage

Table 1: List of non-significant accidents reported at national level

Accident precursors monitored at EU countries

Altogether 21 out of 27 NSAs reported monitoring precursors other than the six introduced under

the CSI framework. The countries that do not monitor additional precursors are mostly different from

those which do not monitor any non-significant accidents. Only in Spain, Norway and Romania is the

accident and incident monitoring limited to CSIs.

For 22 countries in which accident precursors are monitored (except those reported under CSIs), the

average number of those precursor indicators is 12 (with median 9). In six EU countries

(BE,EE,FR,IT,LV,UK) the number of additional precursors at the national level is greater than 15. The

UK is the country with by far the highest number of precursors monitored at the national level.

Additional precursors are typically reported by IMs (in 16 countries) and often also by RUs (12

countries). In AT, BE, BG, CZ, DK, LV, SI the precursors are reported by both RUs and IMs. In all 20

countries, except Germany, the precursors are reported routinely. Only in Austria, Hungary,

Lithuania, Latvia, Portugal and the UK is the reporting done daily. Monthly reporting is almost as

common as daily reporting as shown in Figure 2.

Events / Frequency

day 3 days week month 2 months 3 months year

Non-significant accidents

AT,CZ,FI,HU, LT,LV,NO,PT, PL,SE,UK

BE BG CT,NL

FR,IT DK,SI

Accident precursors

AT,HU,LT,LV, NO,PT,UK

BE

BG,CT, EE,FI,NL

CZ FR,IT DK,SI

Table 2: Frequency of reporting for non-significant accidents and accident precursors (other than CSIs)

Table 2 provides an overview of reporting regimes for non-significant accidents and for accident

precursors. It appears that in 13 countries, the reporting regime in terms of data reporting frequency

is the same for the two groups of events. Six countries allow more time for precursors reporting

compared to non-significant accidents.

Page 10: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 7

Figure 2: Frequency of reporting for precursors to accidents

The motivation for monitoring of precursors is similar to the motivation for reporting non-significant

accidents. In 14 countries, general safety monitoring is the main objective. Twelve responding NSAs

stated that the supervision (and/or the planning of supervision activities) is the motivation for the

accident precursors monitoring. This is more than in the case of non-significant accidents in both

absolute and relative terms.

The complete list of precursors other than the six CSI precursors is shown in Table 3. Only the name

and the code for country are shown in this simplified overview.

Precursor Countries

Runaway train BG CT CZ FR PL PT SI UK

Train accepted on busy track without collision (because of wrong routing): (unauthorized admitting into occupied track sections)

BG HU LV PL PT AT UK

Train sent on busy intersection or in wrong direction (because of wrong routing) (wrong routing) (acceptance or routing of a train along an unprepared route) (misrouting)

BG HU LV PL PT UK

SPAD other movements (e.g. shunting movement; due to untimely close position of the signal, or untimely shut down of the signal; after the changing of the signal indicator when the train is too close to the signal to stop)

AT BE CZ FI

Irregular operation of a level crossing installation compromising the traffic safety BE DK EE UK

Breakdowns or technical problems with rolling stock, compromising the security (e.g. hot axe box, coupling breakage at a passenger train, poor brake equipment)

BE SE FR UK

Driving without authorization (departure without authorization) AT FR

Affecting the safe operation of railway crime, e.g. signaling cables and other equipment stolen AT SE

Attempted suicide BE DK UK

Other fires or explosion in the cargo and explosion in the rolling stock ( (fires in the rolling stock) IT (fire extinguisher efficiency) PL (in train) UK (due to rolling stock failures; vandalism; other train fires))

BE CZ

Malicious acts that endanger safety (e.g. placement of various materials or objects on the tracks, throwing objects or shooting at or from a train, sabotage of safety devices; killed animals; objects on line)

BE BG UK

Technical error in brakes; (failed or seriously impaired braking ability); serious brake failure

DK SE UK

6; 35%

1; 6%

0; 0%

5; 29%

1; 6%

2; 12%

2; 12% day

3 days

week

month

2 months

3 months

year

Page 11: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 8

Errors in signaling (technical errors on signal equipment that have resulted in a less restrictive signal message + signals with broken lamps ) (other signaler error))

DK SE UK

Broken barriers in level crossings ((collisions with gates or barriers at level crossings) DK (Irregularity in level crossing))

SE UK DK

Train disconnection CZ FI

Incident involving the transport of dangerous goods CZ EE

Train passes a level crossing without any form of warming system and/or protection activated

CZ DE

Risk of injury to person caused by rolling stock in motion] (a nearby miss); (near misses with pedestrians and road vehicles)

DK UK

Table 3: List of accident precursors reported at national level by more than one country

Accident precursors database

The first question on the precursor data management concerned the availability of a database on

precursors at country level. (Q: Is there an electronic database at the national level set up for the

purpose of collecting information about indicators of precursors to accidents?)

The answers of 27 NSA shows that in just half of 26 countries, there is a database in which the

precursor data are collected (see Figure 3). In 11 out of 13 countries with a database containing

precursors, the database is owned by the NSA. In five countries (CZ, DK, LV, NL, SE) the database is

strictly internal, with no access provided to third parties. In two countries (AT, BE) the database is

maintained by the NIB, with access provided to the NSA. In one country (PT), the database is

maintained by the IMs with no access provided to the NSA. Similarly to Portugal, the IM in the

Netherlands maintains its own database, which exists in parallel with the NSA database. Finland

reported to have scheduled a launch of a database in 2013.

The restriction of the access (internal to NSA) is sometimes justified by the inclusion of private

information, such as names of persons (LV). In one country the access is provided not only to the NIB,

but also to the RUs and IMs (BG). No other NSA reported to provide access to the database to a RU.

Figure 3: Availability of a database on accident precursors at country level

Yes

(AT,BE,BG,CZ,DK,EE,FR,LV,NL,NO,PT,SE,

UK); 13

No

(CT,EL,ES,FI, HU,IE,IT,LT, LU,PL,RO,SI

,SK); 14

Page 12: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 9

Precursors database management

The second question on the precursor data management concerned the ownership of the database

on precursors at the national level and the reporting entities. (Q: If yes, who maintains the database

and who submits the data?)

In nine countries the database is maintained by the NSA and fed by IMs and RUs. One country (LV)

reported that also ECMs and keepers provide data to the database. In Austria, Belgium and Hungary

the database is kept by the NIB with access provided to the NSA. In Portugal and Slovakia the

database is owned by the IM who does not provide access to the NSA. This could be a situation in

other countries as well, as seen from the example of the Netherlands where two databases

containing precursor data exist in parallel (see Table 4).

Regime Database keeper Reporting entities Country

1 NSA RU+IMs CZ,DK,EE,FI,FR,NL1,NO,SE

2 NSA RUs+IMs+ECMs+keepers LV

3 NIB RU+IMs AT,BE,HU

4 IM IM PT,SK, (NL)

5 Other RU+IMs UK

Table 4: Database and reporting management regimes per country

France reported that a team of five persons works with the database at the NSA . They classify,

analyse and use data, including accident/incident reports. In the UK, the database regime is different

than in other countries, with RSSB maintaining the database at the national level.

1 Two reporting flows (databases) exist in the Netherlands

Page 13: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 10

Publication of precursors data

The third question on the precursor data management concerned the dissemination of precursors at

the national level. (Q: Do you publish this information? (e.g. in annual reports))

In most countries, the precursor data are not published on a regular basis at the national level. Only

in a few countries, a specific report(s) is published containing information on precursors. Austria is

the only country in which the publisher is the NIB and not the NSA (see Table 5).

Regime Who publishes Country

1 NSA EE,CT,IE,NL,RO,UK

2 NIB AT

Table 5: Countries in which the precursor data are regularly published

Several other countries indicated that their precursor data are available in their Annual (safety)

report to the Agency, but in reality, the annual reports typically only contain the data on the six

precursors introduced by the RSD (under CSIs). The dissemination of accident precursor data seems

to be most developed in the UK, through the publications of RSSB.

Page 14: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 11

Analysis of precursors data

The forth question on the precursor data management concerned the analysis of precursors at the

national level. (Q: Do you systematically analyse the data collected?)

In 19 countries, the precursor data are analysed - typically by the NSAs. However, only in 12 of these

countries, the data analysed are collected into a database and in only 4 countries the precursor data

are stored in the database and published on a regular basis. This is presented in Table 6, in which the

answers to three questions relevant to precursor data management were standardized.

The regime of analysing the precursors is shown in the table as well. In six countries, the analysis of

precursor data is done annually, in four countries it is carried out monthly and in two countries it is

weekly. Only Norway and the UK reported the continuous, systematic analysis of precursor data at

national level.

Country Database Published Analysed How

AT Annually

BE

Annually

BG

Annually

CT

Monthly

CZ

DE

DK

EE

EL

ES

FI

Monthly

FR

Quarterly / daily

HU

Occasionally

IE

Annually

IT

Annually

LT

Annually

LU

Monthly

LV

Monthly

NL On request

NO Continuously

PL

PT

Annually

RO

Weekly

SE

Weekly

SI

SK

UK Continuously

Table 6: Countries in which the precursor data are regularly published

Page 15: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 12

Need for harmonization of more precursor indicators

The last question of the survey aimed at understanding the view of the NSAs on the need to extend

the monitoring of precursors at the national level (extension of precursors reported under CSIs). (Q:

Do you think that there is a need to harmonise at EU level additional indicators related to precursors

of accidents?)

The NSAs have no common view on the question of whether more precursor indicators should be

collected and monitored at the EU level. Eight countries consider that there is a need to harmonize

additional precursor indicators at EU level, with some of them listing the indicators currently missing

under CSIs. Three countries see a space for more precursor indicators harmonized at EU level, but

they put forward conditions, or further considerations that should be taken into account. Eleven

countries do not see the need for more precursor indicators at EU level, with some of them being

categorical in their reply. One NSA recommended that ERA develop guidance for NSAs enabling them

to work with precursor data more effectively (see Table 7).

Table 7: Extended monitoring of precursors under CSIs?

Arguments put forward by those NSAs being in favour of extended precursors monitoring at EU level

(through CSIs) include the need to better understand the functioning of the railway system, an

extension of possibilities in benchmarking countries, or background data for rules preparation, or

even investment programmes. The same NSAs however warn against quick actions and plead for a

deep discussion about the appropriate indicators, including their common definitions.

Arguments put forward by those NSAs not being in favour of extended precursors monitoring at EU

level include their belief that the compatibility/validity/reliability of the current six precursor

indicators must first be assured, the belief that analysis of precursors falls better under the NSA

competence, or that the necessary knowledge is not yet available (e.g. in respect of human factors

related precursor indicators).

Yes; 8

Possibly; 3

No; 11

Guidance for NSA; 1

N/A; 4

Yes

Possibly

No

Guidance for NSA

N/A

Page 16: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 13

5 Conclusions

Monitoring of non-significant accidents at country level is widespread in EU countries. In some

countries, the RSD reporting regime (CSIs) has not replaced the original reporting system (and the

relevant classification) and two reporting regimes exist in parallel. Various types of collisions are

monitored at country level, in contrast to one single type of collision introduced under CSIs. Non-

significant collisions, derailments and fires are the most commonly reported non-significant

accidents. The purpose of collecting non-significant accident numbers is the general monitoring of

railway safety at country level. It is usually not used by the NSA for planning its supervision activities.

A majority of countries monitor precursors other than those defined in the RSD, Annex I as CSIs. On

average, an additional 11 precursor indicators exist at national level. The indicators are very much

country specific and it is rather difficult to find common patterns among countries. Yet a short list of

commonly used precursor indicators could have been drafted. The management of precursor data

manifests some common points (NSA is typically the keeper of the database and its key user), but

there are great differences between NSAs when it comes to the purpose and use of precursor data.

Notably, it is rare to see a systematic use of precursor indicators for determining supervision needs at

country level. The differences in reporting regimes and in the analytical practice by the NSA further

underline the lack of a common approach among NSAs on the use of precursor data. This may imply

a need for defining a common approach to precursor data analysis by the NSAs at the EU level.

There is certain support among NSAs for the extension of precursors monitoring at EU level (through

additional precursors defined under CSIs), yet the NSAs do not see a need for any urgent action.

There is also certain feeling that the precursor indicator analysis might fall under the competence of

NSAs, rather than ERA. The Agency may however provide guidance on the use of precursor data by

the NSAs.

Page 17: Evaluation Survey NSA - Final

Evaluation of ERA Questionnaire on the Use of Accident Precursor Data by NSAs Safety Unit, Regulation and Safety Monitoring Sector

Page | 14

Annex: Detailed evaluation (overview of replies)

Page 18: Evaluation Survey NSA - Final

Page | i

A You collect at least the statistics on significant accidents that must be reported to ERA (Annex I of the RSD as amended by Directive 2009/149/EC).

In the table below, please provide details about the accidents (other than significant accidents) that you monitor (i.e. where you collect the number of events).

NSA Yes How many

Who reports this? Way of reporting

Part of which NSA activity / objective? RU IM Other/Who?

Routinely

On request Yes/No If yes, how often?

AT yes 8 x x 0 Yes immediately 0 CSIs have priority

BE yes 2 X X 0 Yes within 3 days of the

occurrence Legal obligation NIB database+Annual report RU/IM

BG yes 5 - X - Yes Weekly - Investigation+ Safety monitoring+supervision

of RUs/IMs/

CT yes 1 0 x 0 Yes Monthly 0 General safety monitoring + supervision of IM

CZ yes 1 X X NIB if

investigates Yes

200 accidents + 800 incidents a year

incidents on request

General safety monitoring

DE no 0 0 0 0 0 0 0 0

DK yes 6 x x heritage railways

yes once a year 0 The data is used by Trafikstyrelsen when

visiting the companies and as a way to assess their general safety level.

EE yes 4 0 X 0 0 0 ASAP by phone or

e-mail, after by official note.

NIB investigation / NSA general safety monitoring

EL yes 5 0 0 0 0 0 0 0

ES no 0 0 0 0 0 0 0 0

FI yes 6 X X 0 Yes

Short information continuously,

aggregated numbers monthly

and annually

Also more information on

request, if needed on

individual cases

All national railway safety indicators for monitoring the safety of the Finnish railway

system + supervision of Rus/IMs

FR yes 7 0 X 0 Yes Quarterly 0 General safety monitoring + supervision of

RUs/IMs

Page 19: Evaluation Survey NSA - Final

Page | ii

HU yes 2 0 X 0 Yes Daily report 0 General safety monitoring

IE yes 12 0 0 0 0 0 0 0

IT yes 2 x x 0 y every 3 months 0 monitoring

LT yes 6 0 x 0 Yes Daily 0 General safety monitoring + supervision of

RUs/IMs

LU no 0 0 0 0 0 0 0 0

LV yes 3 x x Yes Every day 0 Monitoring of safety and supervision planning

of IM, RU or ECM

NL yes 1 x x internet yes monthly yes no

NO no 0 0 0 0 0 0 0 0

PL yes 1 0 X 0 Yes After accident

occurrence 0

Within safety monitoring process and within supervision process over IMs and RUs.

PT yes 1 0 X 0 Yes daily 0 General safety monitoring + supervision of

RUs/IMs

RO no 0 0 x 0 Yes 0 0 General safety monitoring

SE yes 3 x x Anyone Yes As soon as possible after it has occurred

0 Reporting to the Swedish NIB. Part of the

Swedish NSA supervision of RUs/IMs.

SI yes 1 x x / Yes Annual / supervision

SK no 0 0 0 0 0 0 0 0

UK yes 26 X X 0 Yes Daily 0 Supervision of IM

Page 20: Evaluation Survey NSA - Final

Page | iii

B You already collect six accident precursors (or indicators) that must be reported to ERA (Annex I of RSD as amended by Directive 2009/149/EC).

In the table below, please provide details of all other accident precursors / indicators that you systematically monitor (including incidents and near-misses but excluding broken wheels and axles, broken rails, track buckles, wrong side signalling failures and SPADs).

NSA Any

(other than 6)

Nr

Who reports this? Way of reporting

Part of which NSA activity / objective? RU IM Other/Who?

Routinely On request

Yes/No If yes, how often?

AT Yes 8 x x 0 Yes 2) 0 as required

BE Yes 17 x x 0 Yes within 3 days of the occurrence Legal

obligation NIB database Annual report RU/IM

BG Yes 8 х х - Yes Monthly - Safety monitoring + supervision of RUs/

IMs

CT Yes 3 0 X 0 Yes Monthly 0 General safety monitoring + supervision

of IM

CZ Yes 9 x x 0 Yes within 60 days of the occurrence of

the accident 0 General safety monitoring

DE Yes 3 x 0 ECM/Keeper No 0 yes supervision of RU/ECM/keeper

DK Yes 10 x x heritage railways

yes once a year 0 same as above

EE Yes 18 0 X 0 Monthly

Incidents which take place within a calendar month by a summary which shall be submitted not later than on the fifth day of the following month.

0 NSA general safety monitoring

EL No 0 0 0 0 0 0 0

ES No 0 0 0 0 0 0 0

FI Yes 12 X 0 0 Yes Monthly If needed See table A

FR Yes 18 0 X 0 Yes Quarterly 0 Supervision of RUs/IMs

HU Yes 4 0 X 0 Yes Daily report 0 0

IE Yes 15 0 0 0 0 0 0 0

IT Yes 18 x 0 0 yes every 3 months 0 monitoring

LT No 3 0 x 0 Yes Daily 0 Yes

LU No n/a 0 0 0 0 0 0

LV Yes 19 x x NIB Yes Every day 0 Monitoring of safety and supervision

planning of IM, RU or ECM

NL Yes 3 no x contractor yes monthly yes yes

NO No 0 0 0 0 0 0 0

Page 21: Evaluation Survey NSA - Final

Page | iv

PL Yes 6 - X - Yes after accident occurrence - Within safety monitoring process and

within supervision process over IMs and RUs.

PT Yes 13 0 X 0 yes daily 0 General safety monitoring + supervision

of RUs/IMs

RO No 0 0 x 0 Yes 0 0 General safety monitoring

SE Yes 7 0 0 0 0 0 0 0

SI Yes 5 x x / Yes Annual / supervision

SK No 0 0 0 0 0 0 0

UK Yes 73 0 x 0 Yes Daily 0 Supervision of IM

Page 22: Evaluation Survey NSA - Final

Page | v

i Is there an electronic database at the national level set up for the purpose of collecting information about indicators of precursors to accidents?

AT Yes, the RUs/IMs send the data (according to the national law called "Meldeverordnung 2006") to the NIB. The NIB runs a database with all accidents and incidents (see also NIB Safety Report 2011, chapter IX - Statistic, see link below)

BE Yes, there is a national database of which the access is restricted to the NIB and NSA

BG Yes, but database is internal only for NSA, and access is restricted to NIB, IM and RUs.

CT No

CZ Yes, there is an internal database of NSA.

DE No

DK There is an electronic database which is confidential. The collected data is used when assessing which companies have a good safety performance and which does not. When assessing this it is both evaluated how the company performs compared to other companies but also whether the company is improving or regressing in performance. Companies with problems receive more audit visits from Trafikstyrelsen than the other companies.

EE All railway accidents, incidents and collisions occurring in Estonia are entered in the module. The module enables significantly better processing and analysis of accident-related data in the future, thus also the adoption of necessary safety improvement measures more operatively than before as well as the preparation of statistical reports in a given format both for the European Railway Agency and for Statistics Estonia.

EL No

ES No

FI

No. At the present we have an excel table and get information in paper. Next year we will have a database for accidents and incidents (collecting data on individual cases). That will not be public. All the RUs and IMs will be able to report cases via internet form and the big RU will send data digitally from their database to the database of the NSA.

FR There is an electronic database called VIGIE used to collect all data about incident/accident information. It is used to monitor safety level and to animate experience feedback at national level.

HU There's no public database available on the NSA's website. List of incidents/accidents available on NIB's website. (http://www.kbsz.hu/hu/vasuti-koezlekedes/hirek)

IE No

IT no

LT How do you manage these data (reporting, analysis, etc.)?

LU n/a

LV Yes. This database is for internal usage within NSA only. It is due to the information is keeping sensitive data (names of train drivers, victims, etc.).

NL Yes, no public access. Public can ask for data. One database is at de IM, and one at the NSA.

NO Yes, access restricted to NSA.

PL No.

PT Yes; IM has a database to record all incidents that occur in the network. Access restricted to IM

RO There is not an national electronic database.

SE Yes, but the database has recently been launched (in June 2012). The data is only available for the Swedish NSA.

SI No. IM and RU sent the report once a year.

SK We do not gather these data. The only purpose for getting them from the IM is the annual safety report.

UK Yes, the Safety Management Information System (SMIS)

Page 23: Evaluation Survey NSA - Final

Page | vi

ii If yes, who maintains the database and who submits the data?

AT NIB maintains the database, RUs and IMs submit information on each occurrence immediately if persons significant injured or killed and precursors at endangering, otherwise until next day

BE The NIB maintains the national database, and the RU's and the IM have to report to the NIB on accidents / incidents within 3 days of the occurrence

BG The database is maintains only by NSA.

CT N/A

CZ NSA maintains the database. RU+IM provide information on precursors within 60 days of the occurrence of the accident.

DE 0

DK The NSA maintains the database. The RU, IM and heritage railways submit data.

EE NSA maintains and NSA submits the data. Due to the fact that the data will be provided by RU/IM's by paper to NIB and NSA. We are planning to make this information providing electronically - directly to the module (database) by RU /IM.

EL -

ES 0

FI Next year the database will be maintained by the NSA and all the RUs and IMs will submit the data.

FR A team of five persons analyse, classify and use data (e.g. for safety alert or control strategy) including incident/accident reports or indicators from RU/IM. Data come from IM for essential but also RU for specific incident where additional information are required.

HU NIB maintains the database, RU+IM have to submit information.

IE Not applicable

IT N.A.

LT 0

LU n/a

LV

NSA maintains the database and put information daily according to the received information from IM, RU, ECM, keeper, etc. The information to NSA has to be sent every day up to 09:00 morning or call to inspectorate. Nevertheless when accident happens it shall be investigated by the commission (which consists of members of RU and IM). The report shall be sent to the State Police, Investigation Bureau, the SRTI, the IM and RU involved in accident. The information from the report has to be put in database according to prepared classification. For example, derailment, collision, broken rails, etc.

NL The IM respectively the NSA.

NO NSA maintains the database. RU and IM have to submit precursors within 8 days of the occurrence.

PL -

PT IM maintains the database, and it is its own staff that submit the data.

RO 0

SE The Swedish NSA maintains the database and submits the data.

SI N/A

SK Database is maintained by the IM, who also submits the data.

UK The Railway Safety and Standards Board (RSSB) maintain the database. Data is submitted Network Rail (principle IM) and all RUs.

Page 24: Evaluation Survey NSA - Final

Page | vii

iii Do you publish this information? (e.g. in annual reports)

AT the NIB publishes a public Safety Report (see link below)

BE No, only in the NSA annually report the indicators imposed by the annex I of the safety Directive are reported

BG No.

CT The IGC includes an Annex C1 (safety related incidents previously included in reports published by the Channel Tunnel Safety Authority) in its annual safety report which includes these incidents.

CZ No, they are not published.

DE No, the information is not made publicly available. It might be communicated to a wider circle of concerned parties, but not to the public.

DK we publish the data in the annual safety report which we also submitted to ERA. It is mostly the ERA-defined indicators that is reflected upon in the report.

EE Yes, in our annual book, also we propose information about accidents after each quarter on our website. www.tja.ee. Also we make two press notice every year (half year conclusion)

EL NO

ES No

FI We do not publish the detailed data on individual cases. We publish our Annual safety report with the trends and the aggregated data on our web page.

FR No.

HU No.

IE Statistics are published by the NSA in its annual report

IT Only the most significant issues are reported in the annual safety report.

LT 0

LU n/a

LV

Partly. Every quarter we do analysis for railway traffic accidents and precursors. We send (electronically) this document to the Ministry of Transport for information (restricted access). The report which consists of CSIs data we publish in our home page www.vdzti.gov.lv in our annual activity and safety reports (publically available). We just started to think on exchange of information among involved parties.

NL Yes, the information is published in several reports, like the NSA safety report of the railway

NO Yes. For instance in an annually report published on our website. We publish statistic on our website. We can also use morning seminars to give such information.

PL The information concerning the accidents is published in annual safety report (sent also to ERA). Everyone who is interested in subject has an access to report because report is published in Official Journal of Ministry of Transport.

PT We only publish the common safety indicators in our annual report, which access is public via our website.

RO Yes, the data regarding accidents are published on the Romanian Railway Authority site. There is also a full report published annually.

SE Some data are published in the NSAs annual report, for example accidents and number of persons killed and seriously injured.

SI No.

SK Annually in a safety report published on the website of the NSA.

UK Yes, RSSB publishes data in a range of general and topic based safety performance reports; ORR publishes summarised data on the National Rail Trends (NRT) portal.

Page 25: Evaluation Survey NSA - Final

Page | viii

iv Do you systematically analyse the data collected?

AT yes concerning to the common safety indicators (supervision, preparation of the annual report)

BE Annually, in connection with the preparation of the annual report of the NSA to ERA, the data of the NIB database are checked by the NSA on the base of the data in the annual reports of the RUs and IM. Occasionally in connection with supervisory activities and in case of major accidents or incidents (to fined trends).

BG Yes, annually in connection with preparation of the annual report.

CT Yes, the Channel Tunnel Safety Authority, on behalf of the IGC, reviews the data on a monthly basis in order to inform its supervisory activities.

CZ Ne, we have it only for general safety monitoring. We use this information for testing drivers. Extraordinary exams of drivers are based on this information.

DE Data is evaluated, if necessary further supervision activity follows.

DK Data is evaluated, if necessary further supervision activity follows.

EE The Technical Surveillance Authority constantly analyses the circumstances of railway accidents, in order to find suitable solutions in collaboration with the railway infrastructure managers and organisations dealing with railway safety. The Technical Surveillance Authority also values strong cooperation with local governments in railway planning activities, to ensure that the railway safety activities would be, first and foremost, based on the needs of the region.

EL NO

ES No

FI Sure. We follow the safety indicators monthly and analyse the reasons for the changes if the safety indicators show increasing risk level.

FR There is a systematic analyse of all data collected and stored in the database. There is a quarterly meeting with IM to confirm data and to comment trends.

HU Occasionally in connection with supervisory activities.

IE Statistics are analysed annually in the NSA's statistical report

IT yes. At least annually in connection with the preparation of the annual safety report. Occasionally in connection with audits and other supervision activities

LT Example: Yes, in an annually published report with restricted access (for RU/IM)

LU n/a

LV Yes. Every month, every quarter and annually. We look tendencies and compare with previous years – increase of derailments or increase of unauthorised movements, etc. The tendencies are also compared with train km. Analyses are as assistant for planning of annual Inspectorate's supervision activities

NL The data is only analysed on request for several purposes, like the annual NSA safety report, to inform the inspectors and to inform the politicians.

NO Yes, continuously

PL The collected data are analysed occasionally. The data collected are analysed in connection with preparation of the annual safety report. The analysis are also performed monthly in order to show the tendency in accidents and to compare with a corresponding period in previous year.

PT Yes, annually in connection with the preparation of the annual report. Occasionally in connection with supervisory activities (to confirm trends ...).

RO Yes, quarterly.

SE Yes, at least every week. The Swedish NSA reports seriously accidents and near accidents to the Swedish NIB. We also use the analysed information in risk-based supervision.

SI No.

SK Analysis is included in the annual safety report that is also submitted to ERA.

UK Yes, in the Annual Report to ERA, in the ORR annual and quarterly health & safety reports and RSSB analyses data via models like the Train accident precursor indicator model, the safety risk model, in risk assessments, for safety performance models, in organisation specific data profiles and in response to data requests.

Page 26: Evaluation Survey NSA - Final

Page | ix

iv Do you think that there is a need to harmonise at EU level additional indicators related to precursors of accidents?

AT No, not at the moment. If there is a need, it is necessary to determine clear definitions

BE No, the current list of indicators related to precursors of accidents in Annex I was already extensive. Due to occasional national hot items, more (less important) precursors are used by the NSAs and NIB with a reporting limited to the national level.

BG We think that indicators of accident precursors should be used more frequently by the NSAs in their processes. ERA could however draft a guidance for NSAs on accidents as follows: runaway train, train uninsured with required brake mass and wrong routing

CT No, we think that there is a need to ensure that there has been harmonisation of existing indicators before adding more.

CZ It might be useful up to some level. Firstly we have to define purpose of this statistics. The purpose might be information for new rules preparation, investment into the signalling systems and so on.

DE Yes, if they prove to be relevant (as regards statistical significance) on EU-level. For example, wheel distortion / wheel shift are in our view an increasing problem in several ways and therefore could be relevant.

DK Yes, this would be a good idea. But first of all the effort should concentrate on how the member states and the companies understand and define the different indicators to secure a common understanding and definition. As our last meeting in the working party showed "track buckles" is not understood in the same way across Europe. This could also be the case for other indicators.

EE The current list of indicators related to precursors of accidents in Annex I is sufficient.

EL 0

ES 0

FI There is a need to harmonise additional indicators related to precursors of accidents, but there is also need not to make too many changes too quickly. The development of safety indicators is continuous work both on European and on national level.

FR I don't think that there is a need to harmonise at European level more indicators. The fact is that the current list of precursors is not sufficient to monitoring safety levels. But I think that additional indicators can be collected at national level relevant to specific need.

HU No. The current list of indicators related to precursors of accidents in Annex I is sufficient.

IE I believe that this possibility should be explored. The focus should be on precursors and non-significant accidents with high potential for harm, damage or loss. These could also include events of misbehaviour (e.g., near miss or non-significant accident at a level crossing) but not misbehaviour with criminal intent.

IT yes. We think that the collection and analysis of precursors of accidents can help to prevent major accidents

LT 0

LU As far as now the RU's and IM only report to the Luxembourgish NSA the indicators related to precursors of accidents as defined in Annex I of the RSD. If the analysis of this questionnaire brings out that there is a need to report additional indicators related to precursors of accidents, the Luxembourgish NSA would support this initiative.

LV Yes, but there is necessary deep research. Human factor precursors have to be developed, for example violation of technological processes or unauthorised movements. It shows management or monitoring gaps in railway companies. In our practice 40-50 % (it depends on year) the cause (or precursor) of all railway traffic accidents is a misdeal of railway staff or management. At this moment it is only SPAD at the list of CSIs.

NL Maybe one, an indicator based on a definition of the quality of data. A train drives can miss information about the track in front of him and hits an obstacle. Or the data analyst gets information from RU and IM and cannot use 7% of the information due to the lack of essential components in the short messages. Guidance could help to calculate the quality of raw information.

NO No. We think that the current list of indicators related to precursors of accidents in Annex I is sufficient.

Page 27: Evaluation Survey NSA - Final

Page | x

PL According to NSA PL, there should be monitored the railway events caused by failures of bogie (e.g. hot axle box, metallic element from the rails stuck to the surface of the wheels (PL: nalepy na zestawach kołowych), loose wheel tyre (PL: luźne obręcze), failures of spring etc.)) and wrong routing, failure of pantograph of electric locomotives, dropping of catenary.

PT Yes. Because in that way they could be standardised with the same definition all over the Europe, and also there will be the possibility to benchmark performances.

RO No, if all the member states will use the same definitions.

SE I think that the current list of indicators in Annex I is sufficient. The challenge now is to use them for national analysis and to improve the quality of the collected data.

SI No. The current list of indicators related to precursors of accidents in Annex I is sufficient.

SK 0

UK No. The current system needs to work properly with the existing CSIs before adding new categories.