Safety
Opinion
A better way is needed by the Chemical Safety and Hazard Investigation Board to assess the progress made in improving process safety. Two key modifications are examined here.
The U.S. Chemical Safety and Hazard Investigation Board's primary mission is to prevent major accidental chemical releases. The nature of the prevention effort facing the Board can only be understood in relation to the responsibilities and authorities it is given under law and the type of major accidents falling within its purview. Successful programs require suitable metrics for measuring if an organization is making progress in accomplishing its mission, as well as identifying where the organization should focus its efforts and resources.
At present, suitable data systems do not exist for generating the metrics needed by the Board, or for that matter, for the industry to guide its prevention efforts and measure its success. We shall briefly examine the deficiencies of existing databases in generating metrics on the progress in preventing major accidents and then propose steps to generate a best practical approach for doing so. Board's primary mission and responsibilities
The purpose of Section 112(r) of the Clean Air Act Amendments of 1990 (112(r)) is the prevention of major accidental chemical releases. Therefore, the Board's mission is the same as those of EPA's Risk Management Program (RMP) and OSHA's Process Safety Management (PSM) regulation required under 112(r). The Board's responsibilities and authority that are most pertinent to the topic of metrics, can be stated by extracting language from (6)C and F of 112 (r) as follows:
"The Board shall ...
* "... investigate (or cause to be investigated), determine and report to the public in writing the facts, conditions, and circumstances and the cause or probable cause of any accidental release resulting in a fatality, serious injury or substantial property damages.
* "... issue periodic reports to the Congress, federal, state and local agencies, including the EPA and OSHA, concerned with the safety of chemical production, processing, handling and storage, and other interested persons recommending measures to reduce the likelihood or consequences of accidental releases and proposing corrective steps, etc.
* "... establish by regulation [the] requirements binding on persons for reporting accidental releases into the ambient air subject to the Board's investigatory jurisdiction.
* "... [be] authorized to conduct research and studies with respect to the potential for accidental releases, whether or not an accidental release has occurred, where there is evidence which indicates the presence of a potential hazard or hazards."
The Board's activities
There are four major categories of activities that the Board employs to discharge its responsibilities and make progress in achieving its prevention mission:
1. Investigating either the root causes of pertinent accidental releases or evaluating hazards leading to such releases that may require additional societal responses.
2. Disseminating investigative findings and corrective recommendations among pertinent stakeholders.
3. Conducting outreach and educational efforts aimed at having managers, industry employees, educators and government agencies translate the Board's investigative findings and recommendations into effective prevention.
4. Tracking recommendations to completion and informing the public on progress made in implementing such recommendations by the parties to whom the recommendations were made.
The Board faces a classic management problem: How should it distribute its resources among and within these programs to maximize the prevention of accidental chemical releases pertinent to the Board's responsibilities?
Pertinent accidental releases
The law spells out significant classes of releases that are not directly pertinent, such as those from offshore facilities, or releases that have no significant acute effects. However, a great deal of insight into what types of releases are pertinent to the Board's responsibilities can be obtained by closely examining the Clean Air Act's mission statement:
"... [to] prevent the accidental release and to minimize the consequences of any such release of any substance listed pursuant to Paragraph (3) or any other extremely hazardous substance." (Paragraph (3) concerns the language used in the Act and defines terms such as "stationary source" - see below.)
Three of the key terms used to describe the type of accidents covered under the mission statement - accidental release, regulated substance and stationary source - are defined in Section 112(r)(2) as follows:
"The term `accidental release' means an unanticipated emission of a regulated substance or other extremely hazardous substance into the ambient air from a stationary source."
"The term `regulated substance' means a substance listed under Paragraph (3)."
"The term `stationary source' means any buildings, structures, equipment, installations or substance-emitting stationary activities (i) which belong to the same industrial group, (ii) which are located on one or more contiguous properties, (iii) which are under the control of the same person (or persons under common control), and (iv) from which an accidental release may occur."
The term "consequences" is not explicitly defined, but the body of the law emphasizes "acute" exposures causing, or having the potential to cause "death, injury or serious adverse effects to human health or the environment."
The meaning of "extremely hazardous substances," is also ambiguous in Section 112(r). However, as used in this section, the term includes at least all of the of approximately 360 substances listed as extremely hazardous in Section 302 of the Emergency Planning and Community Right-to-Know Act of 1986 (EPCRA). The qualifier "at least" has been inserted because, on occasion, the Senate report on Section 112(r) and the law itself use the term to cover "any substance which causes death or serious injury because of its acute toxic effect or as the result of explosion or fire, or which causes substantial property damage by blast, fire, corrosion or other reaction."
While the bulk of the Board's investigations are likely to focus on accidental releases of regulated or defined "extremely hazardous substances," there are clearly instances when the Board will need to conduct an investigation of an accidental release of a substance that is neither a defined Section 302 extremely hazardous substance, nor is regulated under the EPA or OSHA rules. Such investigations will need to be done to discharge the Board's obligation to ascertain whether the hazard and/or substance should be recommended for regulation or for listing as extremely hazardous by EPA.
The list of chemicals covered under the original RMP rule was established to include most historical accidental releases, while minimizing the economic cost of the program. Unless Congress explicitly decides against such actions, similar adjustments to the list of covered substances will be made in the future as a result of EPA and OSHA actions supported by Board recommendations. This is particularly true if such releases affect the public or are a significant fraction of the total releases resulting from accidental release of Paragraph (3)-regulated substances.
Minimum metrics
General considerations - Congress and the public want to know whether the Board and its sister agencies (e.g., OSHA and EPA), as well industry, are making progress in preventing the type of accidental chemical releases that led to the current prevention laws and regulations. The Board needs to be able to answer the following two questions:
* Are the incidence rate and frequency of pertinent accidental releases going up or down? (prevention)
* Are the consequences of such accidents increasing or decreasing? (mitigation)
Moreover, while the resources entrusted to the Board by Congress are significant, they allow for investigating only a small fraction of the releases that are pertinent to even the narrowest view of the Board's duties. For the most part, Board investigations will cover accidental releases of Paragraph (3)-regulated and "extremely hazardous" substances that are certainly within its jurisdiction. However, as the language of Section 112(r)(6) indicates, the Board's range also extends to any accidental release of a substance in any type of fixed facility that might be a candidate for regulation or listing by EPA or OSHA, since the consequences or circumstances of the release may indicate that an investigation is required for the Board to discharge its responsibilities in generating recommendations to these agencies.
Therefore, the Board also needs metrics to help decide upon which industry/process/substance combinations it would be most cost-effective to concentrate its tools and activities resources. Given these considerations, the author believes that the Board requires two types of metrics to achieve its prevention mission:
1. Metrics capable of tracking industry's progress in regard to the frequency, incidence rate and consequences of pertinent accidental releases.
2. Metrics for guiding cost-effective deployment of the Board's resources aimed at prevention.
Tracking progress in reducing releases
Examples of metrics that might be considered for tracking progress in the overall prevention of accidental releases and/or mitigating their consequences are:
1. Total accident frequency - the number of pertinent accidents in a specified time period
2. Total accident incidence rate - the number of pertinent accidents per year divided by the total number of pertinent processes operating in a specified time period
3. Total frequency or incidence rates for specified accident consequences:
a. employee fatalities
b. employee injuries
c. public injuries
d. public sheltering-in-place and evacuations
e. property damage
f. ecosystem impacts
4. Other possible metrics that could offer useful insights:
a. OSHA Occupational Injury and Illness (011) rate for the facility
b. number of Board recommendations adopted within a specified period
Metrics for guiding resource deployment
Examples of some possible metrics for guiding deployment are:
1. Frequency and incidence rates of accidents within the Board's purview by:
a. North American Industrial Classification (NAIC) code
b. substance released
c. pound of covered substance potentially released under the RMP worst-case scenario
d. type of process involved in the accidental release (e.g., batch, semi-batch, continuous, etc.)
2. Consequences of Board-pertinent accidents by NAIC code and/or substance expressed as:
a. worker fatalities and injuries from accidental releases
b. public injuries, sheltering-in-place and evacuations
c. property damage
d. ecosystem impacts
There are obvious problems in relying solely on the frequency of accidental releases to track the progress in preventing them, since the number of accidents occurring in a fixed period can change purely because the number of facilities or the level at which the facilities operate changes. For this reason, the Board should try to use incidence rates adjusted for the level of operation to track progress in preventing accidental releases.
Both the frequency and incidence rates by the NAIC code and the covered substance can be useful in deciding upon the deployment of prevention resources among different classes of facilities. One needs to consider that the law of diminishing returns probably also operates in regard to prevention efforts.
Consider two classes of facilities with the same accident frequency F events/yr. Class A has x number of facilities, the other, Class B, has 2x. Therefore, Class A has an incidence rate of F/x per year and Class B an incidence rate of F/2x or one-half that of Class A. All things being equal, the law of diminishing returns would say that one should expect a greater degree of accident reduction for a given prevention investment in Class A vs. Class B.
Problems with existing data
There are difficulties in generating the Board's metrics using existing databases. Many existing federal-government systems contain data on accidental releases that are pertinent to the Board. Perhaps the most prominent of these are the:
* Incident Reporting Information System (IRIS), from the National Response Center
* Emergency Response Notification System (ERNS), operated by the EPA
* Integrated Management Information System (MIS), kept by OSHA
* Hazardous Material Incident Reporting System (HMIRS), from the U.S. Dept. of Transportation
* National Fire Incident Reporting System (NFIRS), recorded by the U.S. Fire Administration
* Hazardous Substances Emergency Events Surveillance (HSEES), from the Agency for Toxic Substances and Disease Registry.
* Risk Management Accident Reporting System (RMP*INFO), EPA's online database.
A number of studies of these seven systems have been made by EPA (1), the Board and others. Based on these, the following conclusions apply to all but the last database:
1. The accidental-release data in these systems cannot be used to calculate incidence rates because the reporting population (denominator) is not defined.
2. The quality, completeness and appropriateness of the data for a substantial part of the accidental-release reports are not adequate. For example, substantive deficiencies variously exist in regard to the:
* entry of NAIC or Standard Industrial Classification (SIC) codes
* identity of the substance being processed or released
* type of process and facility where the release occurred
* consequences of the release
* mechanisms for establishing the completeness of reporting on incidents pertinent to the Board
* reports on accidental releases of some substances and releases pertinent to the Board are excluded, eg., HSEES excludes hydrocarbons and RMP*INFO has no reports on nonregulated EPCRA-Section 302 extremely hazardous substances or some accidents covered under OSHA's PSM standard
* many events not pertinent to the Board's responsibilities are included and are not identifiable as such
* multiple reports on the same incident are often not identifiable
* focus of the system is on accidental releases other than those from fixed facilities (HMIRS).
3. The best data and system available to the Board for pertinent accidental releases is RMP*INFO. This system was established to collect the 5-yr accident history on processes covered under the EPA's RMP rule:
* All facilities covered under this system must report the number of defined accidental releases or their absence. This establishes the sample population at risk (denominator), the number of defined accidental releases which occur in this atrisk population (numerator) and, thus, one can calculate an incidence rate. The requirement that each facility that does not report any accidental releases certify "`zero releases" decreases the likelihood of omitted accident reports.
* All substances, consequences and facilities involved in the reported accidents are pertinent to the Board's responsibilities.
* Substances and facilities involved in accidents are clearly identified.
* Data are readily available to the Board and public online.
* The associated RMP report contains data on demographics, NAIC codes, some data on what the facility believes is the proximate cause of the accident and links to financial information. These are important datum for research on possible links between accidental releases and factors such as, organizational culture, financial business results, management systems, etc.
4. While the RMP*INFO system has most of the data needed to generate the minimum metrics needed by the Board, it has a few significant limitations. Examples include:
* Complete reporting of all accident data from all facilities to RMP*INFO is required only every 5 yr, even though accidents must be investigated and recorded by facilities as they occur.
* Explicit data are not available on how long the covered process has been in active operation during the 5-yr filing interval.
* Data are not available on all accidental releases pertinent to the Board's responsibilities, e.g., fewer than 200 extremely hazardous substances are covered under the EPA RMP and OSHA PSM rules.
5. EPCRA Title III is an additional possible source of data that could be important in generating the needed metrics by the Board. EPCRA Section 304 requires:
* Immediate notification of accidental releases of Section 302(a) extremely hazardous substances to the State Emergency Planning Committee (SERC) or Local Emergency Planning Committee (LEPC) having jurisdiction over the facility that had the release.
* Confirmation of the initial notification and further specified information on the release must be supplied in writing "as soon as practicable" after the occurrence (Section 304(c)).
* Unfortunately, there are no requirements that the written reports must contain important data needed by the Board to generate its minimum metrics. Examples are: NAIC/SIC codes, property or environmental damage, the total number of facilities handling Section 302 substances within the state or local community, etc.
* While most occurrences of accidental releases pertinent to the Board's responsibilities would be contained in the data required to be reported to either RMP*INFO and/or the SERCs/LEPCs, most states do not appear to have effective systems for collecting, storing and making the reported data available online.
* Finally, there appears to be no requirement that EPCRA Section 302(a) releases be reported to either the National Response Center, EPA or, for that matter, to any national government agency.
Based on the current state of existing accidental-release-- data collection systems, the author concludes that these systems cannot be used as is or in combination to generate the minimum metrics that the author believes the Board requires.
Generating the minimum metrics
The Board needs metrics now and certainly not 10 years from now. Therefore, any proposal for obtaining the proper data should be practical, which, in the author's opinion, requires that it meet the following boundary conditions:
1. Provides data needed to generate the minimum metrics.
2. Receives strong stakeholder support and acceptance.
3. Is achievable within 3 years after the Board commits to a specific proposal.
4. Appears acceptable to Congress, the Board's sister agencies, and the Office of Management of the Budget (OMB).
Based on informal conversations with a variety of parties, the author proposes such a practical approach that would use the existing EPA RMP*INFO system with essentially two changes. More specifically, rulemaking would require:
* Annual, rather than 5-yr, reporting of the data on accidental releases that RMP-covered facilities must presently submit to the RMP*INFO system under 40 CFR Part 68.
* Approximate information on the fraction of the reporting year that each covered process was in operation (produced some level of commercial product).
Annual reporting of the two new suggested elements to RMP*INFO should not be costly, since the majority of facilities file electronically, accidents must presently be investigated immediately and rough estimates of past operating times are available.
This proposal would not, of course, furnish data on all accidental releases pertinent to the Board's responsibilities. The validity of metrics developed under this proposal rests on the likely assumption, discussed above, that RMP coverage mirrors the larger universe pertinent to the Board's responsibilities. This assumption could be tested by cross-checking with events captured in representative SERC and LEPC records, and conducting studies to relate accidents contained in the database with those reported elsewhere, eg., in the news media.
Other useful data and metrics
Stakeholders have expressed the need for additional data elements over and above those presently collected in RMP*INFO. There is a reasonable chance that some of these requests would meet the practicality criteria and not delay adoption of the two changes that the author believes are vital for satisfying the Board's minimum metric needs.
Examples of potentially desirable practical data other than those required to meet minimum the Board's metrics include:
* Information on whether the covered process involves batch, continuous or storage operations. (Does one class of operations pose more risk than the others?)
* Inclusion of data on OSHA OR for the covered process or the facility housing the process. (Is there a correlation between OR and accidental release frequency? Has the 50-80% reduction in OR projected by OSHA after implementation of the OSHA PSM occurred?)
* The census track number. (Are there correlations between community demographics and accidental releases, e.g., environmental justice issues.)
* OSHA facility number, which would enable crosschecks between OSHA reports and accidental releases.
* Stock-exchange ticker symbol of parent company to help unravel problems with Dunn and Bradstreet identifiers. (Studies relating accidental release data to corporate culture and financial status.)
* The Board and its partners should also consider and make plans to evaluate other opportunities to generate or embrace any practical proposals put forward by stakeholders.
Also, there are other desirable changes in data collection that may not meet the criterion of practicality. The author wants to emphasize his belief that the Board and its sister agencies should focus on proceeding rapidly toward accomplishing the two practical changes in the reporting requirements of the RMP rule that are the key to obtaining the minimum metrics. However, as noted above, there may be opportunities to "bundle in" additional improvements without affecting the speed with which the changes needed would be accomplished (for minimum metrics).
Further, any discussions on short-term goals can also be an opportunity to evoke and illuminate longer term needs. This can be accomplished without compromising the achievement of the Board's other stakeholder's minimum goals.
Typical examples of data that might be useful, but do not appear to meet the stated practicality criteria are:
1. Report of accidental releases from processes covered only under the OSHA PSM standard to RMP*INFO. (This would increase the representiveness of the population of facilities that would report on pertinent accidental releases and might provide better insight into the prevention value of both EPA's RMP and OSHA's PSM rules.)
2. Identification of equipment, instruments and controls whose malfunction or failure was directly involved in an accidental release. (This would allow industry to move more rapidly in detecting significant failure patterns and take corrective actions.)
Next steps
The Board should proceed along the following lines:
1. Partner with EPA and OSHA in preparing a draft advance notice of proposed rulemaking (ANPR) along the lines put forward here.
2. Use the draft ANPR to search for stakeholder inputs on other changes in present RMP*INFO reporting elements that add value and are compatible with the stated criteria for practicality.
3. Modify the draft ANPR based on stakeholder inputs.
4. Move forward to formal rulemaking if a consensus emerges on the need and appropriateness of the modified ANPR.
[Sidebar]
This article reflects the author's opinions and should not be viewed as the policy of the U.S. Chemical Safety and Hazard Investigation Board. The author acknowledges contributions from Anna Johnson and Daniel Horowitz of the Board staff, who analyzed the contents of the databases referenced here.
[Reference]
Literature Cited
[Reference]
1. "User's Guide to Federal Accidental Release Databases," U.S. Environmental Protection Agency, Washington, DC; Office of Solid Waste and Emergency Response, National Response Team, Washington, DC. NTIS Order Number: PB96-134291 (1995).
[Author Affiliation]
ISADORE ROSENTHAL,
U.S. CHEMICAL SAFETY AND
HAZARD INVESTIGATION
BOARD
[Author Affiliation]
ISADORE (IRV) ROSENTHAL is a member of the National Chemical Safety and Hazards Investigation Board (2175 K Street, N. W., Suite 400, Washington, DC 20037; Phone: (202) 261-768o; Fax: (202) 943-9723; E-mail: Rosenthal@board.gov). He joined the Board in 1998. Previously, he was employed at Rohm and Haas for 38 years in a variety of research, development, business unit and corporate staff positions. For his last 13 years there, Rosenthal served as corporate director of safety, health, environmental affairs and product integrity. After his retirement, he joined the Wharton Risk Management and Decisions Processes Center as a senior research fellow. His areas of research were in the management of risks associated with lowprobability high-consequence accidents, market-based alternatives to government regulation of industrial risks, and the methodology of risk assessment. Rosenthal has been a member the Technical Advisory Committee of the Environmental Management Division of the Los Alamos National Laboratory, the OECD Risk Terminology Project Steering Group, and the EPA Accident Prevention Subcommittee of the Clean Air Act Advisory Committee. He holds a BA from New York Univ., an MS from Purdue Univ. and a PhD from Penn State (the latter two in physical chemistry), where he was also a post-doctoral fellow. He is a member of AIChE.
No comments:
Post a Comment