Chapter 6: Information Collection Analysis and Dissemination

In order to advance safety, it is crucial for railway companies and regulators alike to have the right data at the right time. The importance of sound data for critical analysis and interpretation cannot be overstated. Similarly, providing clear information to the public on the state of railway safety is equally significant and plays a vital role in the development of public policy. Railway safety data collection, analysis and dissemination was a recurring theme brought to our attention throughout the consultations. Generally speaking, there is dissatisfaction from all quarters on this issue. As a Panel, even after using publicly available data from government sources and commissioning a statistical study,1 we still experienced some difficulty in determining the true state of railway safety in Canada, due to deficiencies in the data.

In looking back at recommendations from previous reviews of railway safety,2 we note that many of the same themes were raised, such as insufficient data, an absence of thorough analysis and a lack of performance indicators. Similar deficiencies still exist today. As noted earlier in our report, measuring railway safety using the data currently collected does not provide a comprehensive or unambiguous portrait of how safe the system is or should be. We fully recognize that measuring railway safety is a complex topic involving a number of various entities. Despite efforts over the years to improve upon this aspect of railway safety, we believe there is still much room for improvement and that a high priority needs to be placed on achieving results.

6.1 Responsibilities for Information Collection, Analysis and Dissemination

A number of parties collect, analyze and disseminate railway safety statistics and information, including the Transportation Safety Board of Canada, Transport Canada, provincial governments and the railway industry itself.

As previously noted, the Transportation Safety Board (TSB) of Canada is an independent agency reporting to Parliament through the President of the Queen's Privy Council for Canada. Its main purpose is to advance transportation safety by conducting independent investigations of railway, marine, aviation and pipeline accidents. Simply put, the mandate of the TSB is to answer three questions: What happened? Why did it happen? What can be done to reduce the risk of it happening again? Inherent in this process are the functions of identifying safety deficiencies, making recommendations to correct them and reporting to the public the results of investigations and findings. The TSB does not assign fault or determine liability, and its findings and recommendations are not binding on the parties implicated. There is, however, an obligation on federal ministers to provide formal responses to TSB recommendations in terms of action taken or planned.

The Board follows a rigorous investigation process including validation of information and facts to ensure fairness and accuracy in reporting before it publishes its reports on individual accident investigations. The TSB does not have the resources to undertake an in-depth investigation of all accidents. In rail for example, out of approximately 1,200 rail accidents per year, only a dozen or so are actually fully investigated and reported annually. Deciding to proceed to a full investigation largely depends on whether or not the TSB concludes the effort will result in advancing safety. It can take up to one or two years for an investigation report to be completed and made public.

Another important function fulfilled by the TSB is that of publishing, in aggregate form, statistics on transportation occurrences, consisting of both accidents and incidents. The TSB compiles monthly statistics reported to it and produces a year-to-date update and an annual report at the conclusion of each year. The data is disseminated to the general public through the TSB website. The Board is also required to report annually to Parliament on its activities. TSB data is the principal source of occurrence information used by other organizations such as Transport Canada.

Accident reporting requirements imposed on the transportation industry are spelled out in regulations to the Canadian Transportation Accident Investigation and Safety Board Act. The regulations define what constitutes a “reportable railway accident” and a “reportable railway incident.” The regulations also stipulate what information must be reported to the TSB as soon as possible in the case of a “reportable” accident or incident. Such information includes the date and time of the accident, location, a description of the accident or incident, the extent of damage to rolling stock, and anticipated time of arrival of wreck-clearing equipment.

6.1.1 Transport Canada

Two organizations within Transport Canada are involved in collecting data that relates to the safety of railway operations - the Rail Safety Directorate and the Transport Dangerous Goods Directorate. As part of its mandate for regulatory oversight of Canada's railways, the Rail Safety Directorate requires industry to report information to the department on matters specific to safety management systems. In addition, railway safety inspectors have the power to request documents and/or information for the purposes of ensuring compliance with the Railway Safety Act (RSA). The Directorate also collects various information from industry and elsewhere in order to track the activities of the industry and its own railway safety inspectors for internal planning purposes.

A data system being developed in the Rail Safety Directorate, known as the Rail Safety Integrated Gateway (RSIG), will draw upon a variety of information sources, both internal and external, and help the Directorate with trend analysis, risk management and decision-making activities related to such things as targeting safety inspections and deployment of its resources. For accident data, rather than create a duplicate database, the Directorate mainly relies on the TSB database. The Directorate does not undertake any formal accident investigations, as this is the domain of the TSB, but it may investigate for compliance with rules, regulations and standards as established under the RSA.

The second organization within Transport Canada which actively collects railway safety information is the Transport Dangerous Goods Directorate. This directorate is the federal regulator for the handling and transport of dangerous goods and the principal source of information on occurrences involving dangerous goods, not only for rail but also for the other modes of transportation. Under the Transportation of Dangerous Goods Act and its regulations, an accident-reporting regime has been established principally so that a response to any actual or potential spill can be immediately initiated. The Canadian Transport Emergency Centre (CANUTEC) is operated by Transport Canada to assist emergency response personnel in handling dangerous goods emergencies and also acts as an information and research centre. The Transport Dangerous Goods Directorate is therefore in a position to collect, maintain and disseminate information on dangerous goods occurrences and make this information available to users.

In addition, Transport Canada annually produces Transportation in Canada, in which it reports in summary format on its yearly activities, and in which there is an array of statistical information covering all transportation modes. However, when it comes to railway safety information, there is little in the report outside the TSB rail occurrence data.

6.1.2 Provinces and Industry

Provincial authorities, to varying degrees, record information related to railway activity, including accident data for railways that fall under their jurisdiction. Due to the limited authority of the federal government over all railway operations in Canada, as discussed in Chapter 4, there is no requirement for provincially regulated railways to report accident information to a federal agency such as the TSB or Transport Canada. Nevertheless, the TSB collects data from some provincially regulated railways, but does not mix or publish it with data collected from federally regulated railway companies.

Individual railway companies record safety data for their own purposes, usually as part of their internal processes for effective management. Safety data related to individual railway operations is vital for the company to self-assess, maintain a safe operation, gauge performance, plan maintenance activities and reduce risk. On behalf of its members, the Railway Association of Canada collects and disseminates annual industry statistics. For example, its annual report, Railway Trends, contains an industry snapshot of railway safety, along with many other indicators of railway operations in Canada.

6.2 Data Deficiencies

There are three distinct yet intertwined phases associated with the statistical assessment of railway safety - data collection, data analysis and information dissemination. Advancing railway safety must involve all three phases and be a common objective, contributing to public safety, safety of railway employees, protection of property and the environment, and also benefiting railway companies in terms of efficiency and profitability.

During our consultations with stakeholders, and through additional research studies, we received a consistent message that the existing state of railway safety data does not adequately reflect or help advance railway safety to the extent necessary, due to a number of shortcomings. Deficiencies in data collection, analysis and dissemination were frequently pointed out.

There is a widely held view that the TSB's published data on railway occurrences does not provide a comprehensive or fully accurate picture of railway safety in Canada. For example, there can be problems for railway companies in interpreting the reporting requirements set out in the TSB regulations. The railway has to determine if, as a result of an accident, rolling stock sustains damage that affects its safe operation. There can be ambiguity in how this is determined and how consistently it is determined within a company and among many companies.

Recently, it came to light that due to a difference in interpretation of the TSB reporting requirements, the CN data going back to 2002 had been under-reported. This required a revision to the statistics covering the five-year period to 2007. Any ambiguities from this period would most often be related to more minor types of incidents. As a result of this situation, we understand that the TSB is in the process of revising its accident/incident reporting regulations and that they should come into effect in 2008.

The TSB largely relies on the railway to report occurrences to it, without any structured or consistent process for validation or challenge of what is being reported to the Board. Most frequently, a railway company accident is not formally investigated by the TSB and assigning a cause of the accident is left to the individual railway company. Often the cause is not immediately known and is left blank when data is initially reported to the TSB. All too frequently, the cause is never reported to the TSB; it does not follow up to determine the cause and, therefore, the occurrence database is left incomplete. For example, the number of main track derailments in the TSB database that were not coded with an attributable cause grew from less than 10 per cent in 1999 to close to 50 per cent by 2006 - hence the database limits insights and conclusions that can be drawn.3

Also brought to our attention were problems associated with the lack of severity indicators that would help gauge the gravity of an accident. The only apparent indicator reported is the number of rail cars derailed per accident, but this says little in terms of the accident's severity or actual consequences. A rail car, whose wheels leave the track at slow speed, but which remains upright, is counted the same as a car travelling at high speed that falls on its side and spills its contents across the track. In the U.S., under the Federal Railroad Administration (FRA) reporting requirements, the main criterion for reporting a railway accident is that damage to railway equipment exceeds a specific dollar value (currently standing at US$8,200).4 This gives some limited insight into severity.

Canada does not use dollar value as a reporting criterion or as a severity indicator. Unless ranges for the value of damage were introduced (e.g., $8,200 - $15,000; $15,000 - $50,000; $50,000 - $200,000; etc.), a simple single threshold value adds little to understanding degrees of severity. On the other hand, one of the benefits of using a dollar value as the basis for reporting an accident is that it is clear, assuming the damage is easily assessed, and can be readily translated into monetary terms. The Panel favours the exploration of new ideas for illustrating the severity of railway accidents in Canada.

The criteria for reporting accidents and incidents in Canada are less precise than those in the U.S., where a threshold dollar value is used. Canadian criteria are more open to interpretation by the railways as mentioned above. To be reported to the TSB, the regulations stipulate that rolling stock must sustain damage that affects its safe operation.

The phrases, “safe operation,” or “poses a risk or threat,” introduce a subjective element into the determination of whether an accident or incident is reportable. Clear guidelines issued by the TSB and understood by the railways would be needed to accurately capture the data. Unfortunately, this is not the case as we have seen in the TSB's recent experience with CN data.

Generally speaking, comparing Canadian and U.S. railway safety data as collected by national authorities is like trying to compare apples and oranges given the different reporting requirements of each national authority. Canadian railways are required to report to U.S. authorities on their U.S. operations according to U.S. reporting rules. For comparison purposes, the companies also record whether an accident occurring in their Canadian operations would meet the U.S. reporting threshold.

TSB definitions:

reportable railway accident” means an accident resulting directly from the operation of rolling stock, where

(b) the rolling stock

(iv) sustains damage that affects its safe operation, or

(v) causes or sustains a fire or explosion, or causes damage to the railway, that poses a threat to the safety of any person, property or the environment;

reportable railway incident” means an incident resulting directly from the operation of rolling stock, where

(a) a risk of collision occurs,

(g) any crew member whose duties are directly related to the safe operation of the rolling stock is unable to perform the crew member's duties as a result of a physical incapacitation that poses a threat to any person, property or the environment

Transportation Safety Board Regulations (SOR/92-446)

Using accidents per million train miles, where accidents are defined in accordance with U.S. criteria, CN and CP compare favourably to U.S. railways, based on their overall North American operations.5 It is interesting to note that if the U.S. definition of accidents were adopted in Canada instead of the TSB criteria, the number of TSB reportable accidents would drop by 90 per cent or more.6 We believe that analysis of accidents is enhanced by having more data, not less, and would not be in favour of revising accident reporting criteria or severity indicators that would result in fewer reported accidents and incidents than is the case under current TSB requirements.

The data normalization measures employed are inadequate to reflect workload changes in the railway industry over time. Tonnage carried by the railways has steadily risen7 and outside third party influences, such as urban growth, have resulted in greater opportunities for human/vehicular interfaces with railway rolling stock. At present, the most common measure used to normalize the number of accidents into a comparable rate of accidents, taking into account changes in the amount of rail activity, is the accident rate per train mile, usually expressed as accidents per million train miles.

Number of train miles is not the only activity measure that can be used to normalize the data. When using accidents per million train miles, it becomes evident that the data provides almost no additional knowledge concerning railway trends over time, as the figures simply mirror the absolute number of accidents per year.8 It would be useful for decision makers to have a more informative basis for normalization of the number of accidents in the railway industry. Using billion gross ton miles, for instance, has been suggested as a promising alternative for freight movements and one that can capture changes in railway workload and productivity, such as heavier train loads.9

Another limiting factor is that TSB data reflects only information obtained from federally regulated railways and does not capture the entire population of railway operations in Canada, which also includes many provincially regulated railways.10 Furthermore, the size of the industry is not well accounted for in the data because of changes in the number of railway companies under federal jurisdiction, which can vary as federal lines are abandoned, provincial short lines are created, or, as in the case of BC Rail being acquired by CN, large provincial railway operations are absorbed into the federal domain. To our knowledge, the accident data is not adjusted to accommodate such circumstances.

While the TSB produces excellent detailed accident investigation reports and is recognized as a world leader in accident investigation methods and procedures, it can only manage to undertake a small fraction of accident investigations in any given year. Recognizing that the Board can issue interim recommendations or safety advisories, the Panel is concerned that it often takes one to two years for a final accident report to be produced and published, thus slowing the flow of information to the public - a public that can become anxious when it witnesses a high-profile accident or recurring accidents in the same geographic area.

With its inherent problems, railway safety data, therefore, presents numerous challenges when trying to analyze its true meaning. Further, the TSB is not responsible for safety oversight and cannot be expected to conduct detailed analysis of the data and trends that should feed into safety policy, regulations, oversight and corrective measures - this is largely the domain of the regulator.

The authors of two of the research studies that we commissioned criticized Transport Canada, Rail Safety Directorate for doing too little in the way of systematic data-gathering and analysis and for being reactive rather than proactive.11 For railway occurrence data (both accidents and incidents), the Rail Safety Directorate relies largely on information obtained from the TSB database - a source of information that has limitations, as described above. On the other hand, railway companies complain that Transport Canada makes too many ad hoc requests for information without adequately communicating why the information is being requested - whether it is part of an audit, an inspection, or an investigation, or is related to safety management systems monitoring. Railway companies also complain of a lack of feedback from the audit and data collection activities conducted by the regulator. There does not seem to be a coherent approach to data collection in the Directorate.

More open communication between Transport Canada and companies is needed in this regard. More fundamental, though, is the absence of a steady flow of the right information and data for policy development and regulatory oversight purposes. As pointed out by the researchers, there seems to be limited analysis of the data by Transport Canada and an absence of effort to assess the overall safety performance of the railway industry.12

Beyond re-quoting TSB occurrence data, no other indicators are publicly available to give insight into whether the railway system in Canada is safe or is getting safer. While the department does use data collected from various sources to help set internal priorities, target inspections and regularly monitor railway company compliance, more needs to be accomplished at the macro level to report these results to a broader audience. For example, setting performance targets with the industry and having it demonstrate continuous safety improvement to the regulator, both as an industry and by individual company, can help transform a philosophy of “accidents happen” into a more proactive attitude.

Direction 2006, the 10-year program to reduce rail crossing and trespassing accidents by 50 per cent, may not have fully achieved its goal. Nevertheless, it resulted in significant reductions through education, engineering, research and enforcement. It is the view of the Panel that the railway industry, with the support of regulators, needs to demonstrate the same continuous advancement in safety as it does in cost efficiencies.

It is also necessary for the regulator to acquire information from individual railway companies in order to fulfill its regulatory oversight role, whether through inspection, investigation or audit activities. Difficulties in obtaining safety-related information from companies have been experienced, whether due to a lack of clarity in the regulations or in the rationale for requesting the information. The railway industry has been criticized for being uncooperative in providing data or making it more difficult than necessary for inspectors to obtain data.

Railway companies have been moving towards highly technical applications in generating data from inspection activities related to monitoring track and equipment, and processing this data for purposes of planning a proactive maintenance program. The industry also conducts internal investigations into accidents and incidents. Such processes and information should assist industry in streamlining the reporting of necessary information to regulatory authorities in a timely fashion. We recognize that many small railways may not be as capable as larger railways of capturing and reporting data electronically to the regulator, and therefore, reasonable accommodation should be provided in such cases.

One element not considered within our mandate is railway security. There may also be security data requirements that need to be considered and possibly incorporated into any revised railway safety data collection and analysis activity of Transport Canada. The issue of limited departmental resources was also brought to our attention as a factor in how much can be achieved within the existing Rail Safety Directorate's suite of responsibilities. We address this issue more fully in Chapter 11.

6.3 Data Collection

As we heard during our consultations and through numerous written submissions, the data collected by government entities, principally the TSB and Transport Canada, does not provide an accurate or comprehensive view of railway safety in Canada and focuses mainly on accidents and incidents. Transport Canada accesses the TSB database for railway occurrence information, and supplements it by collecting data on activities related to its regulatory oversight role. The department also maintains a separate database for dangerous goods accidents. This information is not collected and coordinated with the goal of providing Transport Canada, Rail Safety Directorate with the capacity to effectively monitor the overall safety of the industry or assess the effectiveness of its programs, regulations, rules and standards.

The current data does not allow the regulator to assume its full responsibility, and immediate effort is needed to determine what information is required to satisfy government entities and to assist the industry in pursuing safer railway operations.

At present, there is an absence of meaningful performance measures and proactive/ predictive indicators of safety. To rely on limited TSB accident data does not provide a sufficiently thorough understanding of railway safety from a broad public safety perspective. It is necessary for Transport Canada to consider the range of railway safety information needs on behalf of the government. Coordinating the effort and working closely with all involved to determine these needs will be important in achieving a data collection and reporting protocol that is predictable, regular and electronically supported. This should also reduce the need for the regulator to make ad hoc requests of the industry for information. When required to do so, such requests should be fully explained and justified.

Ultimately, Transport Canada is responsible to the Government of Canada and the public for a safe national railway system and, therefore, the department must be at the centre of the effort to understand and interpret the data, and to translate it into policies, programs and regulations that advance railway safety.

The TSB should continue with its mandate of investigating and reporting on accidents, but it is not necessary for it to collect railway accident data - the regulator should assume this role. Other countries separate these two roles, one of the reasons being the maintenance of neutrality of the investigative body. It is not necessary for such a body to be involved in the collection and analysis of the data that feeds back into the regulatory process.

Examples in other jurisdictions where data collection and analysis is much more advanced include the U.K. Rail Safety and Standards Board (RSSB) and the U.S. Federal Railroad Administration (FRA) Office of Safety Analysis.

The FRA maintains a database on its website for the purpose of making railway safety information readily available to a broad constituency that includes FRA personnel, railway companies, research and planning organizations and the public. Nearly 700 railways report information to the FRA and data queries can reach back 10 years or more for trends. A portion of the database is secure and accessible only to FRA staff. Visitors to the public website have access to a wide array of railway safety information covering accidents and incidents. For example, data for accidents by railway company and by state, for number of inspections and for highway-rail crossing accidents are readily available. Users can download a variety of safety database files, order publications, and view current statistical information on railroad safety as they require. The FRA and stakeholders work together to make changes to the database and proposed changes go through the normal Notice of Proposed Rulemaking process for comment before changes are finalized.

The U.S. National Transportation Safety Board (NTSB), which is similar to the Canadian Transportation Safety Board, conducts investigations into transportation accidents, including those involving railways. NTSB final accident reports used to take approximately two years to be published, but the Board has since reduced this to 15 months and is targeting 12 months. Factual reports written by NTSB investigators are usually available to the public within six months after an accident. The NTSB does not investigate close calls or near misses; nor does it collect railway accident data.

The RSSB is a not-for-profit company, owned by the major railway industry stakeholders in the U.K., and independent of any individual railway company. Its mission is to lead and facilitate the industry's work to achieve continuous improvement in the health and safety performance of railways, and to facilitate the reduction of risk to passengers, employees and the public. It produces an annual safety performance report that analyzes, measures and communicates the industry's safety performance and contains a comprehensive statistical analysis for a wide range of safety performance indicators. The Rail Accident Investigation Branch is a separate independent organization in the U.K. that investigates railway accidents and incidents. The main regulatory body is the Office of Rail Regulation, whose principal function is to regulate the national rail network.

Canada may benefit from examining these and other jurisdictions with respect to accident investigation, data collection processes and activities, as well as ways of measuring railway safety performance.

Recommendation 25

Transport Canada should be responsible for railway safety data collection and ensure that the needs of government agencies are met and that there is no duplication or confusion for reporting entities and stakeholders. There should be a regular timetable for reporting, and ad hoc demands for information or requests must be accompanied by valid reasons and should be kept to a minimum.

The foundation of any database consists of collecting appropriate information that will be subjected to analysis and will generate meaningful interpretations designed to address underlying questions. In this case, the objective is railway safety and the desire to monitor and continually improve upon safety performance. To assess the safety performance of an organization or an industry requires an appropriately selected “basket of measures.” Data can be drawn from a variety of sources, so it is important to understand what the objective is from the start, what performance indicators are important and useful and how the information can be collected and analyzed. In a railway safety setting that involves both private companies and public safety entities, joint cooperation and action is necessary to achieve this common goal.

The Report of the 1994 review of the Railway Safety Act, On Track: The Future of Railway Safety in Canada, pointed to the need for the regulator and industry to work together towards assessing data requirements and collecting and analyzing data in order to produce meaningful performance standards so that the regulator could perform its role and industry could be held accountable for safety plans.

The 1998 Review of Railway Safety, found that, “current data and analysis do not provide an adequate basis for determining past influences on safety nor predicting future safety performance,” and “railway safety management system information, important for assessing system health, is not reported to or collected by Transport Canada.” Further, “limited analysis is performed with existing data largely due to resource pressures,” and “research on safety issues is also limited.”

Despite earlier observations and recommendations, to date, it appears that little action has been taken to establish this essential foundation. Our view of railway safety is that it cannot be advanced if you cannot rely on accurate, robust and timely data. As pointed out earlier, except for a few cases, railway safety is not generally improving to the degree necessary, particularly when it comes to main track derailments and non-main track accidents. A collaborative approach between regulators and the industry in determining data requirements is strongly encouraged. As lead regulatory authority, Transport Canada should address data collection and analysis issues, including the range and type of data required, reliability of data, normalization factors, the capture and use of untapped railway safety data, as well as establishing new measures of industry performance.

Recommendation 26

Transport Canada should give the highest priority to putting in place a robust program of data collection and analysis in order to measure railway safety performance, and should be provided with the necessary resources to do so.

An important consideration for government agencies is to ensure that industry is not overburdened with multiple or duplicate reporting requirements. The filing of information should be made as simple as possible. Any paper-based reporting is inefficient, can lead to transcription errors and should be eliminated to the extent possible. One comprehensive railway safety database would help achieve a number of objectives and eliminate some of the problems discussed earlier. Establishing a secure database, available to “licensed” stakeholders on a shared basis, and enabling automatic reporting and updating with online access, could meet the needs of both government entities and industry. Creating a public access component to extract data is also envisioned.

Recommendation 27

A secure electronic database should be established to enable electronic filing of railway safety data by railway companies.

When a railway accident occurs, many parties need to be notified. Railway companies maintain up-to-date call-out and notification lists involving federal, provincial, local and private entities, as do the emergency response plans of local and provincial organizations. Immediate reporting of accidents is important in order to mount an adequate response to the emergency as soon as possible. At present, there is no central reporting requirement for all railway accidents, as is the case for accidents involving the transportation of dangerous goods.13

In an effort to capture the entire picture with respect to railway accidents, it is preferable for all such accidents to be reported to a central point from which other levels of government and agencies can be notified. As described in greater detail in Chapter 8, there is a need for improved coordination of response to railway accidents that can involve many entities, from first responders such as local fire, ambulance and police departments, to national and provincial government organizations.

Recommendation 28

Transport Canada, in consultation with other departments and agencies, should create a one-stop reporting system for immediate reporting of accidents and for disseminating that information throughout all levels of government and agencies.

While some provincially regulated railway safety data is collected by the TSB, it is not reflected in the TSB published statistics. The TSB data cannot therefore reflect the performance of the railway industry as a whole. This is not sufficient for the regulator to gain a full picture of railway safety in Canada. As discussed above, it would be useful for the Rail Safety Directorate to monitor overall safety - something that it cannot currently do by relying on partial data collected either by itself or by the TSB. The Rail Safety Directorate must become a comprehensive data centre in which provincial railway safety data is also collected and assessed.

The addition of provincial data should improve the identification of safety issues for the rail industry in Canada and more specifically for the shortline railway industry.

Government of Nova Scotia Submission, page 6.

The regulator needs to gain a view of the entire railway system, particularly when there are running rights and agreements between federally regulated and provincially regulated railways and when provincial railways operate over federal track. Recognizing that the federal authority does not extend over all railways in Canada, Transport Canada should work with provincial and territorial authorities responsible for railway activity, ideally through the Federal-Provincial Working Group on Railway Safety referred to earlier. Together, they should consider a program of capturing and reporting on provincial railway safety data in order to monitor and understand national railway safety in a more comprehensive manner.

Recommendation 29

Transport Canada should work with the provinces to develop a comprehensive database, including both provincial and federal railway safety data.

The information-gathering powers of a railway safety inspector are found in s. 28 of the RSA. For the purpose of ensuring compliance with the Act, regulations, rules, and orders, inspectors can “carry out any inspection” that may require the railway “to produce any document for inspection” and may copy or seize any property with respect to administration and enforcement of the Act. This section of the Act, as currently written, limits the inspector's ability to fully and efficiently carry out the duties necessary for the department to monitor and assess industry compliance. Rather than having to go on site, the railway safety inspector should have the authority to request information under this section from any location and have it made available through electronic means, if it exists in that format.

The intent is not to create more work for the railways or insist that hard copies of documents need to be converted into electronic form, but rather to facilitate reasonable requests from railway safety inspectors in the course of their duties. Commercially sensitive information would not normally be required, and if so, could still be kept confidential by Transport Canada under Access to Information Act provisions. Timely response from the railway company is also an important requirement to enable an inspector to conduct his duties. The purpose of these proposed amendments is to streamline their work. They are not intended to enable inspectors to pursue more ad hoc requests for data, as these should be reduced through the creation of a regularized data-reporting requirement referred to in earlier recommendations.

Recommendation 30

Section 28 of the Railway Safety Act should be amended to clearly state that:

  • a railway safety inspector, for the purposes of exercising an audit or inspection function, may require any person to provide information or copies of any existing documents in any format (electronic or hard copy) specified by the railway safety inspector;
  • the request may be made from any location for documents stored at any location; and
  • the regulated party must provide the requested information or document in a timely manner.

6.4 Data Analysis

Understanding how railway safety data will be analyzed, what indicators are important and how performance will be measured are issues that need to be considered as part of the process of determining basic information and data collection needs. These issues must be considered in a collaborative manner between the industry and government entities interested in understanding and advancing railway safety - thus ensuring that data collection and analysis are not separate activities when planning a robust information system. It is key to establish, early in the process, how the raw data will be utilized to produce meaningful measures of performance, trends and benchmarks. Rather than relying on data that has been traditionally collected, new measures of performance, benchmarking and leading indicators need to be creatively considered and established through a collaborative effort.

By establishing performance standards, the Rail Safety Directorate could use them to gauge overall compliance with railway safety regulations, rules and standards and thereby be able to target front-line safety inspections and SMS audits. Such standards could also be useful in considering where changes to existing regulations and rules may be needed and determining if safety of the railway industry is improving. The move to performance standards should eventually lead to fewer prescriptive rules and regulations and a greater reliance on SMS. Setting targets based on sound performance measures benefits both the regulator and the industry in that such targets represent goals to be achieved. Measurement in any management process is key to continuous improvement.

Existing measures have traditionally focussed on outcomes commonly expressed in terms of accidents and injuries - reactive to past events rather than forward-looking measures (i.e., leading indicators), such as those used in business (sales projections, return on investment or expected profits). Measures, both reactive and proactive, that reflect the breadth of the railway industry are required to effectively understand safety performance. Well-designed measures should be accepted by, and meaningful to, those involved in the activity being measured and those who need to use the measures. They should be simple, unambiguous, understandable, repeatable and objective. They should also be capable of showing trends, be cost-effective in terms of data collection, and provide timely information for decision-makers.

In economic terms, a leading indicator is a measurable factor that changes before the economy starts to follow a particular pattern, and is used to help predict changes in the economy.14 In railway safety, the identification of leading indicators is made more difficult because so many variables affect railway safety, including capital spending, employee work schedules and overtime, quality of employee training, track and equipment employed, and weather conditions. Nevertheless, it should be possible to identify some leading indicators worth tracking in order to be more predictive and proactive regarding railway safety.

For example, capital expenditures on track infrastructure and track inspection information may be correlated with track-related accidents and incidents. Train accidents due to human error may benefit from analysis of the amount and quality of the training provided to operators, their knowledge of the rules and observations of unsafe behaviours. Examining near misses and incident data also has potential. Determining whether a strong linkage exists between the data collected and the leading indicator would require commitment of resources. Developing some leading indicators should be considered in conjunction with other performance measurement activities that we have suggested. Benchmarking and trend analysis are vital in maintaining and enhancing railway safety and should be addressed by both the regulator and the industry as a “value-added” task.

Relevant and up-to-date statistics are critical for all stakeholders in order to make appropriate safety risk assessments and regulatory enforcement decisions. ... TSB statistics have limitations for properly assessing safety performance and safety risks.

Railway Association of Canada, RAC Submission to Panel (February 2007), page 17.

It is our belief that the Rail Safety Directorate needs to establish a strong, centrally located office of railway safety data analysis that would integrate the data from such sources as regional and national audits, SMS, inspections and evaluations, rail accident investigations, coroner reports, railway companies and public complaints. The Rail Safety Integrated Gateway (RSIG) database is a natural starting point, but to fulfill this enhanced role the Directorate requires sufficient and trained personnel. The data then could be shared widely outside the Directorate with the TSB, other parts of Transport Canada, provincial authorities, industry and the public.

As mentioned earlier, in 1994, the authors of On Track and the 1998 Review of Railway Safety recommended that the regulator needed to fulfill its role as a monitoring and auditing organization by collecting and analyzing data and producing performance indicators to measure success. At present, we detect an impasse between the regulator and the industry with respect to determining what information is required and how it will be captured and shared. Industry feels that there are too many ad hoc requests for information from Transport Canada and requests for information that are questionable in terms of the regulator's role. On the other hand, Transport Canada believes that it is entitled to any information it feels is required under the RSA.

The parties need to move beyond these viewpoints and look at the need for information from a more holistic perspective, incorporating rigour and predictability. A collaborative effort is needed to develop a national, systematic approach to data collection and analysis, based on transparency, trust and a sense of common purpose. To succeed, the task should not become bogged down due to ingrained positions or lack of a specified completion date.

The RSA provides the Governor in Council authority, under section 37 (Maintenance and Production of Safety Records), to make regulations pertaining to filing with the Minister information suitable for monitoring safety performance or predicting potential changes in levels of safety. This provision of the Act has not been used by the regulator, as there are no regulations in place. As discussed above, a well thought-out, comprehensive set of data requirements capable of producing relevant information and performance indicators of the state of railway safety in Canada would be beneficial.

Railway companies would know what is required of them and could accommodate regular information submissions rather than reacting to ad hoc information requests. Companies would establish more definite targets to improve safety and the regulator could gauge progress, assess the impact of its actions more scientifically through trend analysis, and make necessary corrections at a systems level. In addition, the Rail Safety Directorate would be better positioned to target inspections, conduct audits and justify rule-making and regulatory decisions more effectively. Using section 37 may be the appropriate route to enshrine data collection and analysis requirements.

As was recommended above, Transport Canada should give the highest priority to putting in place a robust program to measure railway safety performance leading to an enhanced focus on trend analysis, risk reduction, and strategies to address problem areas. As demonstrated through targeted programs, such as Direction 2006 100 Chapter 6: Information Collection, Analysis and Dissemination and a vigorous transportation of dangerous goods regime, reductions in crossing accidents and rail occurrences involving dangerous goods have been accomplished. There is reason to believe that such a focus can be extended to railway safety with similar results. As with data collection, the establishment of performance measures and data analysis needs a collaborative approach between the regulator and the industry so that meaningful output can be agreed to, accepted and acted upon.

Recommendation 31

Transport Canada should take a more active role in trend analysis and benchmarking of railway performance. This should involve a collaborative approach with government and industry stakeholders to develop appropriate and meaningful measures of risk and safety performance. To this end, Transport Canada must work with stakeholders to:

  • define data requirements;
  • develop reporting and data sharing mechanisms;
  • develop regulations requiring the industry to report data and performance measures; and
  • publish safety performance results.

6.5 Information Dissemination

Communicating safety information to the public is an important function of government, both from the standpoint of accountability and transparency and to advance safety. Demonstrating to all stakeholders that railway transportation is safe, particularly when so many large and small communities are bisected by railway lines carrying all types of cargo, helps to dispel any suggestions by the media that the system may be unsafe - something that can occur after a dramatic railway accident or series of accidents. On the other hand, if members of the public consider a system to be unsafe, they may become more engaged and participate in the development of public policy.

In a railway safety context, publicly reporting information involves both the regular dissemination of statistical and performance data and reporting on the causes of specific accidents. At present, the TSB is the prime disseminator of accident data and accident investigation reports. If our recommendations for an enhanced role for the regulator were to be implemented, Transport Canada would be responsible for providing regular statistical information on railway safety to Canadians. The TSB would continue with its present role of producing and publishing accident investigation reports.

It is also necessary to provide the public with reliable, unbiased, factual information on individual railway accidents as soon as possible, to allay any fears or misconceptions about what happened. This is better than creating uncertainty, which encourages media conjecture and the potential for exaggeration of the relevance of individual occurrences. Waiting months or even years for an official investigation report to be issued does not adequately serve the public interest. While the detailed investigation reports are very useful and should continue, there is a need for the TSB to issue statements of fact as soon as practical after all significant accidents (e.g., those involving loss of life/serious injuries, environmental damage, having a high public interest, or exhibiting abnormal circumstances) in a similar fashion to the accident briefs issued by the U.S. National Transportation Safety Board.

Providing meaningful information on railway safety data and performance, including reporting on the performance of individual railways, aids in transparency and accountability on the part of the railway and the regulator and can help them to achieve greater safety. It should not be problematic to publicly identify individual company performance if the data is treated with rigour, is reliable and portrays meaning, and if the regulator and the industry have agreed on reporting parameters and protocols. Public opinion can often lead to positive action with respect to safety matters when the principals involved strive to make needed improvements.

Reporting on enforcement activities and specific infractions that are serious enough to put the public or environment at risk should be made public. Transport Canada already issues information for some of the other transportation modes on enforcement and compliance actions. This information includes fines that are levied against marine polluters and aviation operators, and indicates when certain Air Operator Certificates have been revoked. It is the view of the Panel that public safety is enhanced when such information is communicated to a wide audience and that the effect of public opinion on influencing railway safety should not be ignored.

Recommendation 32

To ensure that the public is informed on rail safety issues, the Government should make public:

  • purely factual information on a significant rail accident as soon as possible after the occurrence;
  • railway safety performance data (including information by company); and
  • information on enforcement actions.

One of industry's concerns is that if it provides information to the government, that information becomes subject to the Access to Information Act and is accessible by request from the public. While it is true that the information collected by Transport Canada would fall under the Act, the Act also provides certain protections to third party15 information. Basic tenets of the Access to Information Act include transparency and accountability. The Panel strongly opposes any effort that would use the Act to prevent safety problems from coming under public scrutiny, since railway safety is improved, not by keeping information secret, but through accountability and transparency.

At the same time, we understand that some information collected by the TSB and the aviation mode of Transport Canada may be excluded from disclosure by statute or regulation for the purposes of conducting a thorough investigation. We have some sympathy with industry's position that only information necessary for the administration and enforcement of the RSA should be collected by the regulator, and that there may be circumstances that warrant its protection once collected and in the government's possession. Both Transport Canada and the railway industry should review this issue and clarify the rights and obligations of each party.

Recommendation 33

Transport Canada, in consultation with industry, should determine whether, and to what extent, information provided by a railway company under the Railway Safety Act should be privileged information.


1 Joseph F. Schulman, CPCS Transcom Limited, The State of Rail Safety in Canada (August 2007).

2 Railway Safety Act Review Committee, On Track: The Future of Railway Safety in Canada, Report of the Railway Safety Act Review Committee (December 1994); and Transport Canada, Review of Railway Safety Act Amendments and Safety Oversight and Regulatory Compliance Mechanisms: Report of the Transport Canada Project Team (January 1998).

3 G.W. English and T.W. Moynihan, TranSys Research Ltd., Causes of Accidents and Mitigation Strategies (July 2007), Figure 2; section 5.1.

4 A FRA reportable accident includes any collision, derailment, fire, explosion … that results in total damages to all railroads involved greater than the current reporting threshold. The threshold value is periodically updated by the FRA to reflect cost increases. In 2007, the threshold was raised from US$7,700 to US$8,200 and includes damages to rolling stock as well as signals, track, track structures or roadbed including the cost of labour.

5 Schulman, State of Rail Safety, op. cit., section 8.

6 Ibid., section 7.3.

7 From 1995 to 2005, the volume of tonnage carried by the railway sector in Canada increased approximately 25 per cent. Traffic increased from 292 million to 368 million tonnes of rail freight from 1995 to 2005. See Transport Canada, Transportation in Canada 2006, Annual Report Addendum (May 2007), Table A6-8.

8 Schulman, State of Rail Safety, op. cit., section 3.1.

9 Ibid., section 9.2.1.

10 However, it should be pointed out that if a provincially regulated railway has an accident while operating on federally regulated track, such an accident is required to be reported to the TSB, in accordance with its accident reporting regulations.

11 English and Moynihan, Causes of Accidents, op. cit., section 5.1; Milt Poirier, QGI Consulting Ltd., Performance Measurement in Railway Safety (July 2007) section 6, “Transport Canada.”

12 English and Moynihan, Causes of Accidents, op. cit.; Poirier, Performance Measurement, op. cit.

13 As mentioned previously in this chapter, CANUTEC acts as a reporting centre for accidents involving dangerous goods.

14 Schulman, State of Rail Safety, op. cit., page 33.

15 “ ‘Third party', in respect of a request for access to a record under [the] Act, means any person, group of persons or organization other than the person that made the request or a government institution,” Access to Information Act (R.S., 1985, c. A-1), s. 3. Provisions related to how government institutions must handle third-party information are set out in section 20 of the Act.

Date modified: