Acute Chemical Incidents Surveillance — Hazardous Substances Emergency Events Surveillance, Nine States, 1999–2008

Maureen F. Orr, MS1

Jennifer Wu, MS1

Sue L. Sloop, PhD2

1Division of Toxicology and Human Health Sciences, Agency for Toxic Substances and Disease Registry, CDC

2Office of Information Services, Centers for Medicare & Medicaid Services

Abstract

Problem/Condition: Although they are infrequent, acute chemical incidents (i.e., uncontrolled or illegal release or threatened release of hazardous substances lasting <72 hours) with mass casualties or extraordinary levels of damage or disruption severely affecting the population, infrastructure, environment, and economy occur, and thousands of less damaging chemical incidents occur annually. Surveillance data enable public health and safety professionals to better understand the patterns and causes of these incidents, which can improve prevention efforts and preparation for future incidents.

Reporting Period: 1999-2008.

Description of System: The Hazardous Substances Emergency Events Surveillance (HSEES) system was operated by the Agency for Toxic Substances and Disease Registry (ATSDR) during January 1991-September 2009 to describe the public health consequences of chemical releases and to develop activities aimed at reducing the harm. This report provides a historical overview of HSEES and summarizes incidents from the nine states (Colorado, Iowa, Minnesota, New York, North Carolina, Oregon, Texas, Washington, and Wisconsin) that participated in HSEES during its last 10 full years of data collection (1999-2008).

Results: During 1999-2008, a total of 57,975 chemical incidents occurred: 41,993 (72%) occurred at fixed facilities, and 15,981 (28%) were transportation related. Chemical manufacturing (NAICS 325) (23%) was the industry with the most incidents; however, the number of chemical incidents in chemical manufacturing decreased substantially over time (R2 = 0.78), whereas the educational services category (R2 = 0.65) and crop production category (R2 = 0.61) had a consistently increasing trend. The most common contributing factors for an incident were equipment failure (n = 22,535, 48% of incidents) and human error (n = 16,534, 36%). The most frequently released chemical was ammonia 3,366 (6%). Almost 60% of all incidents occurred in two states, Texas and New York. A decreasing trend occurred in the number of incidents in Texas, Wisconsin, and Colorado, and an increasing trend occurred in Minnesota.

Interpretation: Although chemical manufacturing accounted for the largest percentage of incidents in HSEES, the number of chemical incidents over time decreased substantially for this industry while heightened awareness and prevention measures were being implemented. However, incidents in educational services and crop production settings increased. Trends in incidents and number of incidents varied by state. Only a certain few chemicals, sectors, and areas were found to be related to the majority of incidents and injured persons. Equipment failure and human error, both common casual factors, are preventable.

Public Health Implications: The findings in this collection of surveillance summaries underscore the need for educational institutions and the general public to receive more focused outreach. In addition, the select few chemicals and industries that result in numerous incidents can be the focus of prevention activities. The data in these surveillance summaries show that equipment maintenance, as well as training to prevent human error, could alleviate many of the incidents; NTSIP has begun work in these areas. State surveillance allows a state to identify its problem areas and industries and chemicals for prevention and preparedness. Beginning in 2010, ATSDR replaced HSEES with the National Toxic Substance Incidents Program (NTSIP) to expand on the work of HSEES. NTSIP helps states to collect surveillance data and to promote cost-effective, proactive measures such as converting to an inherently safer design, developing geographic mapping of chemically vulnerable areas, and adopting the principles of green chemistry (design of chemical products and processes that reduce or eliminate the generation of hazardous substances). Because the more populous states such as New York and Texas had the most incidents, areas with high population density should be carefully assessed for preparedness and prevention measures. NTSIP develops estimated incident numbers for states that do not collect data to help with state and national planning. NTSIP also collects more detailed data on chemical incidents with mass casualties. HSEES and NTSIP data can be used by public and environmental health and safety practitioners, worker representatives, emergency planners, preparedness coordinators, industries, emergency responders, and others to prepare for and prevent chemical incidents and injuries.

Introduction

In spite of efforts to improve chemical safety, chemical incidents continue to occur, and educational efforts and interventions are needed. The U.S. Environmental Protection Agency (EPA) and the U.S. Department of Labor's Occupational Safety and Health Administration (OSHA) reported the results of an investigation involving several large incidents and found recurring causes, including inadequate process hazards analysis, use of inappropriate or poorly designed equipment, and inadequate indications of process conditions (1). In addition, other incidents were preceded by a series of similar incidents, incidents that narrowly avoided causing mass casualties, or low-level failures, indicating the need for more attention to implementation of lessons and more thorough company investigation of low-level failures and incidents that could have caused numerous injuries and deaths (1).

In 2007, an explosion at the British Petroleum (BP) refinery in Texas City, Texas, resulted in 15 deaths and injured 170 people. This incident spurred the U.S. Chemical Safety Board (CSB) to recommend an independent review panel, the findings of which are commonly referred to as the Baker Panel Report. The panel urged companies to regularly and thoroughly evaluate their safety culture, the performance of their process safety management systems, and their corporate safety oversight for possible improvements. The panel noted that complacency results in chemical incidents with injuries and deaths (2). In 2009, OSHA announced $87,430,000 in proposed penalties for BP Products, North America, for its failure to correct potential hazards faced by employees (3). A similar explosion involving ammonium nitrate occurred 6 years later and approximately 250 miles away in West, Texas, at the West Fertilizer Company. Fifteen people died, and hundreds were injured. On June 27, 2013, the CSB chairperson testified to the U.S. Senate Committee on Environment and Public Works that current U.S. standards have safety gaps. Despite what is known about the dangers of ammonium nitrate, no federal, state, or local rules restrict the storage of large amounts of ammonium nitrate near homes, schools, or hospitals (4).

Other incidents could have caused mass casualties, such as the 2008 incident at the Bayer Crop Science facility in Institute, West Virginia. During that incident, an out-of-control chemical reaction occurred inside a 4,500-gallon pressure vessel, causing it to explode and resulting in a fire that burned for approximately 4 hours. Although two deaths and eight injuries occurred, the potential for hundreds or thousands of additional deaths and injuries existed. According to the CSB incident report, methyl isocyanate (MIC) could have been released during that incident. MIC is the same chemical that killed thousands and permanently injured many more in the worst industrial incident in history, which occurred in Bhopal, India, in 1984 (5). Because of the potential for a similar incident in the United States, the U.S. Congress provided funding for CSB to commission a study by the National Academy of Science (NAS) on the feasibility of implementing safer alternative chemicals to MIC and processes at this plant.

Bayer no longer produces MIC at the plant. However, as a result of the NAS study, CSB found that the chemical industry could benefit from incorporating the principles of inherently safer design to effectively eliminate or reduce hazards, prevent accidents, and protect nearby communities (6). Inherently safer design and green chemistry (design of chemical products and processes that reduce or eliminate the generation of hazardous substances) have become increasingly accepted as preventive measures by industry organizations. The Presidential Green Chemistry Challenge Awards recognize chemical technologies that incorporate the principles of green chemistry into chemical design, manufacture, and use. EPA sponsors the Presidential Green Chemistry Challenge Awards in partnership with the American Chemical Society Green Chemistry Institute and other members of the chemical community, including industry, trade associations, academic institutions, and other government agencies (7).

Several different agencies provide oversight for chemical incidents in the United States, including EPA, OSHA, CSB, and the National Transportation Safety Board (NTSB). The National Institute of Environmental Health Sciences Worker Training Program supports training of thousands of workers in hazardous waste operations and emergency response (8). The Agency for Toxic Substances and Disease Registry (ATSDR) is a nonregulatory agency that serves the public by using the best science, taking responsive public health actions, and providing trusted health information to prevent harmful exposures and diseases related to toxic substances (9). During 1990-2009, ATSDR funded surveillance of hazardous substance releases in up to 17 states and two other countries (India and Poland) through the Hazardous Substances Emergency Events Surveillance (HSEES) system (10), an active, state-based surveillance system to systematically collect data on thousands of incidents for evidence-based preparedness and prevention efforts. Although HSEES collected the most complete information on the public health impacts of chemical releases, including worksite, transportation, and private property incidents, not all states were included.

This report summarizes the incidents occurring in selected states participating in HSEES during 1999-2008 and is a part of a comprehensive collection of surveillance summaries (11). Public and environmental health and safety practitioners, worker representatives, emergency planners, preparedness coordinators, industries, emergency responders, and others who prepare for or respond to chemical incidents can use the findings in this report to prepare for and prevent chemical incidents and injuries.

Methods

This report is based on data reported to HSEES by health departments in nine states (Colorado, Iowa, Minnesota, New York, North Carolina, Oregon, Texas, Washington, and Wisconsin) that participated in HSEES during its last 10 complete calendar years of data collection, 1999-2008. Data from 2009 were not included because several states ended data collection mid-year.

Description of Data System

Since 1991, ATSDR and CDC have supported state health departments with actively gathering information about chemical incidents. Funded states negotiated formal or informal data sharing agreements with the federal, state, and local agencies that are routinely notified when hazardous substances emergencies occur. Data sources included but were not limited to the National Response Center (NRC), the U.S. Department of Transportation (DOT) Hazardous Materials Incident Reporting System (HMIRS), police and fire departments, environmental health agencies, poison control centers, various emergency response agencies. The media also served as a resource for identifying incidents. For each incident, information was collected about the location, industry, substances released, contributing factors, injured persons, injuries, and evacuations. States entered the data into a standardized ATSDR-provided online system from which the HSEES database was constructed (10).

HSEES Case Definition

During the surveillance period, a hazardous substances emergency incident was defined as an acute uncontrolled or illegal release or threatened release of hazardous substances lasting <72 hours. Threatened releases were defined as imminent releases that did not occur but led to an action (e.g., an evacuation) that could have affected the health of employees, emergency responders, or members of the general public. In accordance with the Comprehensive Environmental Response, Compensation, and Liability Act, commonly known as Superfund (12) (the authorizing legislation for ATSDR), incidents in which the only substance released was petroleum (e.g. crude oil, gasoline, or mineral spirits) were excluded. Incidents involving petroleum and another qualifying hazardous substance were included.

Case Definition Changes

Beginning in 2006, ATSDR changed the inclusion and exclusion criteria for an HSEES incident to improve the uniformity of reporting among states and reduce investigation of incidents that had minimal public health impact. Before 2006, the case definition included any release of a hazardous substance in an amount that was required by federal, state, or local law to be cleaned up. State and local laws could vary and could be more stringent than federal laws. In 2006, the definition was changed to fixed amounts for all states. Beginning in 2006, an incident qualified for inclusion if the amount released was >10 lbs or 1 gallon or any amount of a substance on the HSEES mandatory reporting list (13). The HSEES mandatory list was compiled from the highly hazardous substance lists from other agencies, such as those of EPA and the U.S. Department of Homeland Security (DHS), with additions of substances shown to be hazardous in HSEES. Also in 2006, reports of smokestack emissions above permitted values of carbon monoxide, sulfur oxides, or nitrogen oxides were excluded because they rarely resulted in a substantial public health incident and because of the sheer number of incidents.

Variable Definitions

An incident was considered to be related to transportation if it occurred during surface, air, pipeline, or water transport of hazardous substances or before the substance was totally unloaded from a vehicle or vessel. All other incidents were considered fixed-facility incidents.

HSEES defined an injured person as a person who experienced at least one documented acute (i.e., occurring in <24 hours) adverse health effect or who died as a consequence of the incident; injured persons must have had at least one injury type or symptom, and up to seven could be listed (10).

Contributing factors consisted of primary (root) causes and secondary (contributing) causes. Each incident could have one of each (i.e., up to two) factors.

Analyses

Descriptive data analysis on incidents by industry, factor, chemical, and state were performed. The data were then plotted to look for trends over time. The coefficient of determination (R2) value was used to determine goodness of fit for trend. Ranging from 0 to 1, a higher R2 value denotes that the variable of interest was increasing or decreasing at a steady rate over time. A perfectly linear trend would have an R2 value of 1. The statistical analyses were performed using statistical software.

Businesses were categorized using the U.S. Census Bureau's 3-digit North American Industry Classification System (NAICS) for 2002 (14). For the chemical-specific analysis, only incidents in which one substance was released were included (n = 54,989, 94.8% of incidents). The top 10 substances released were examined. To determine whether the 2006 reporting guidelines had an effect on the top 10 substances, the totals for the 3 years before (2003-2005) and after (2006-2008) the changes for each substance were measured as a percent change. Contributing factors were examined for the last 8 years only, corresponding to the dates the list of factors was expanded.

Results

During 1999-2008, a total of 57,975 chemical incidents occurred. A total of 41,993 (72%) were fixed-facility incidents and 15,981 (28%) were transportation related.

Industries

Of the 57,975 incidents that occurred in the 10-year period, 61% were reported from five industries: chemical manufacturing (NAICS 325) (23%), truck transportation (NAICS 484) (20%), petroleum and coal products manufacturing (NAICS 324) (10%), utilities (NAICS 221) (6%), and nondurable good merchant wholesalers (NAICS 424) (2%) (Figure) (Table 1). Although the chemical manufacturing category accounted for the largest percentage of incidents in HSEES, the number of chemical incidents decreased substantially over time for this industry (R2 = 0.78). Certain industry groups had an increasing number of chemical incidents over time. The educational services category (R2 = 0.65) and crop production (R2 = 0.61) category had a consistently increasing trend (Table 1).

Contributing Factors

Each incident could have up to two contributing factors listed. Of the 64,270 reported contributing factors for 46,489 incidents, the most common contributing factors associated with a chemical incident reported in HSEES were equipment failure (n = 22,535, 48% of incidents) and human error (n = 16,534, 36%). Other commonly cited factors included improper filling, loading, or packaging (n = 6,551, 14%) and system or process upset (any glitch in the system that upsets the process; the problem has to be specific to the facility) (n = 4,092, 9%) (Table 2).

Contributing factors that could cause smokestack emissions, such as system or process upset, system startup or shutdown, performing maintenance, and power failure or other electrical problems, decreased in 2006 when the reporting change went into effect that excluded smokestack emissions of carbon monoxide, sulfur oxides, or nitrogen oxides. Fires also decreased (Table 2).

Chemicals

A total of 54,989 (95%) incidents occurred that involved release of one substance. Ammonia was the chemical most often involved in HSEES single-substance incidents, with ammonia released in 3,366 (6%) incidents. Other frequently released chemicals included paint not otherwise specified; alkaline hydroxides (sodium hydroxide and potassium hydroxide), sulfuric acid, mercury, hydrochloric acid, carbon monoxide, ethylene glycol (antifreeze), nitrogen oxide, and sulfur dioxide (Table 3). These top 10 substances involved 28% of all single-substance incidents.

A comparison of the 3-year period before and after the change to the inclusion criteria took place in 2006 (2003-2005 vs. 2006-2008) indicates that incidents involving the substances that became subject to mandatory reporting, regardless of the amount released, increased (i.e., ammonia, alkaline hydroxide, sulfuric acid, mercury, and carbon monoxide >50 parts per million [ppm]) (Table 3). Some decreases occurred among the substances that were newly excluded: smokestack emissions of nitrogen oxide (441 fewer cases, a 94% decrease) and sulfur dioxide (198 fewer cases, a 69% decrease). Conversely, the frequency of incidents involving paint (not otherwise specified), and ethylene glycol increased, presumably because their reportable quantities were decreased to 10 lbs or 1 gallon after the 2006 change in reporting guidelines.

Distribution by State

Almost 60% of all incidents occurred in two states, Texas and New York (Table 4). A decreasing trend occurred in the number of incidents in Texas, Wisconsin, and Colorado, and an increasing trend occurred in Minnesota. However, no trends were associated with the case definition changes in 2006.

Discussion

The change in the case definition in 2006 appears to have had expected effects on the data. Contributing factors that would be expected to cause smokestack emissions of carbon monoxide, sulfur oxides, or nitrogen oxides such as system or process upset, system startup or shutdown, performing maintenance, and power failure or other electrical problems, decreased in 2006. Incidents increased involving the substances that became subject to mandatory reporting (i.e., ammonia, alkaline hydroxide, sulfuric acid, mercury, and carbon monoxide >50 ppm). Substantial decreases occurred in the newly excluded substances: smokestack emissions of nitrogen oxide (441 fewer cases, a 94% decrease) and sulfur dioxide (198 fewer cases, a 69% decrease). Conversely, the frequency of incidents involving paint (not otherwise specified) and ethylene glycol increased, presumably because their reportable quantities were higher than 10 lbs or 1 gallon previously. However, no trends in state distribution were associated with the case definition changes.

A slight decreasing trend (R2 = 0.30) in incidents over the 10 years has been reported in another analysis in this collection of surveillance summaries (15). Incidents in the industry with the highest number of incidents, chemical manufacturing, decreased substantially over the period, possibly as a result of HSEES and others agencies such as OSHA and CSB and industry group outreach. However, the lack of an overall decrease in HSEES injuries and an increase in deaths reported (16) suggest a need to continue to evaluate trends and direct outreach. Five industries accounted for almost one third of all injured persons: truck transportation, educational services, chemical manufacturing, utilities, and food manufacturing (17). The injuries of many persons injured in truck transportation incidents were unrelated to the chemical release. The numerous injuries that occurred in educational institutions were surprising, and the finding is concerning because children are more susceptible to environmental hazards (17).

The five most commonly released chemicals associated with injured persons (carbon monoxide, ammonia, chlorine, hydrochloric acid, and sulfuric acid) were the five chemicals most commonly released by the five industries with the most chemical incidents resulting in injuries (18). This is not an unexpected finding because these are extremely hazardous, fairly ubiquitous chemicals. Chlorine, although not one of the 10 most commonly released chemicals, was one of the five chemicals most commonly associated with injury because of its hazardous properties. Given this finding, outreach that focuses on these chemicals is likely to have a substantial effect on reducing morbidity and mortality.

States varied in their number of incidents, with highly populous Texas and New York having the most. This might be a result of the increased number of incidents that occur in population centers. In addition, the top chemicals and industries varied by state (19). This finding highlights the importance of using a state-based surveillance system, which provides important data for each state to prioritize planning and prevention strategies.

As reported in another analysis, approximately one fourth (26%) of all incidents resulted in at least one public health action (e.g., evacuation, decontamination, shelter-in-place order, road or area closure, or environmental sampling) (20). Although necessary, these actions can be stressful, disruptive, and costly. Therefore, ATSDR and participating states formed partnerships to link agencies responsible for responding to these incidents (e.g., state environmental departments, state health departments, and other state and local agencies emergency agencies) to increase situational awareness and state emergency notification for chemical emergencies.

Every year, participating states submitted outreach plans with logic models to ATSDR with several planned activities that were substantiated by their data and had measurable effects. Details on these activities were published previously (21). HSEES data were used by local and state emergency planners (EPA regional response teams, regional hazardous materials teams, state emergency management offices, local emergency planning committees, DHS, law enforcement, and chemical incident response teams) to identify chemicals, industries, and locations at high risk for involvement in a chemical release. HSEES data were used as a source of data for case scenario drills. HSEES data were used to support documentation for legislation, particularly for illegal methamphetamine laboratories and mercury bans in schools. Minnesota passed legislation banning the sale of mercury thermometers in 2001 and passed methamphetamine laboratory ordinances in many counties, which substantially reduced the number of illegal laboratories. New York provided data on methamphetamine laboratories for a governor's program bill that became law in 2005. In 2004, a law banning use of elemental mercury in all primary and secondary schools in New York required the development and dissemination of informational materials, which were developed in partnership with HSEES staff in New York. Iowa methamphetamine data were used by the governor to promote a new law restricting the sale of pseudoephedrine in 2005. Reports show a reduction of as much as 90% in methamphetamine laboratory incidences after the Iowa law was enacted. In Oregon, HSEES data were used to support state legislation to decrease the availability of precursor chemicals used in methamphetamine laboratories, which resulted in a 95% reduction in the number of laboratories during 2003-2007. In Wisconsin, HSEES data were used to demonstrate the large percentage of incidents that resulted in injuries but involved less than the state-reportable quantities, when the state legislature had a proposal to raise the state-reportable quantities. In addition, ATSDR and the state health departments have collaborated on numerous journal articles and presentations, which can be found on the ATSDR HSEES website (available at http://www.atsdr.cdc.gov/HS/HSEES).

The chemical incidents and injury prevention stakeholders include persons in areas such as labor, industry, academia, public safety, other state and federal agencies, and nongovernmental organizations. ATSDR convened stakeholders during 2004-2007 to gather input on making improvements to HSEES. As a result, in 2010, NTSIP replaced HSEES (22). NTSIP has a streamlined incident database as suggested by participating states and added petroleum incidents as suggested by stakeholders. In addition, NTSIP takes a multifaceted approach to incident surveillance and response. Specifically, stakeholders stated that national data were needed; therefore, HSEES and NTSIP data are used in collaboration with other federal national incident databases (the DOT and NRC incident databases) to provide national incident estimates for national planning purposes. NTSIP states now focus more on promoting progressive practices, such as green chemistry or inherently safer technology and other hazard reduction strategies. The Assessment of Chemical Exposure (ACE) feature of NTSIP was designed after stakeholders suggested that more extensive data collection was necessary for certain incidents with mass casualties, and sharing of lessons learned should be increased. ACE provides a tool kit, or if necessary a public health response team, to rapidly assess the public health effects of a mass casualty chemical incident and develop recommendations for prevention and preparedness. An ACE investigation might result in the identification and formation of a cohort of exposed persons who need to be monitored to assess long-term health consequences of the exposure.

Limitations

The findings in this report are subject to at least six limitations. First, despite the attempts to make the case definition the same among states, results are not comparable between states because reporting to HSEES was voluntary and data sources varied by state. Second, results from these nine states might not be representative of the entire United States. Third, inconsistencies within and across states likely exist because of reporting capacity (e.g., staffing) or local requirements varied. Specifically, certain states and localities had more stringent reporting regulations than the federal regulations or had more resources to conduct surveillance, possibly resulting in more reported incidents. These factors might have influenced the quality and number of reports or level of detail provided about the incidents. Fourth, changes in reporting guidelines in 2006 had an effect on some of the trends specifically chemicals and factors. Fifth, because some incidents are difficult to identify, such as carbon monoxide or illegal methamphetamine chemical incidents, which often happen in private homes, they might be underreported, with a bias toward those that cause injuries. Finally, incidents that occurred in the transportation and warehousing industries often might be related to motor vehicle crashes, and the associated injuries might be related to the trauma of the crash rather than the chemical release.

Conclusion

The limited number of industries, chemicals, and geographical areas associated with most of the HSEES incidents and injuries could be the main focus of future prevention activities. Improved maintenance of equipment and training of workers and the public should result in decreases in incidents. Because of several large industrial incidents, such as the BP refinery explosion and the BP oil spill in the Gulf of Mexico, the chemical manufacturing and oil refining industry has come under tighter scrutiny in recent years. On August 1, 2013, shortly after the large ammonia nitrate explosion that resulted in 15 deaths, 160 persons injured, and damage to or total destruction of numerous buildings in the town of West, Texas, an executive order was issued: Improving Chemical Facility Safety and Security (23). According to the order, in coordination with owners and operators, executive departments and agencies with regulatory authority need to take additional measures to improve chemical facility safety and security, which will better coordinate federal efforts toward reducing large industrial incidents. However, data indicate that nonindustrial sectors as well as transportation sectors (17) are responsible for a large number of incidents and injured persons, and the percentage of injuries among members of the public is increasing (16). Outreach for these sectors is still needed.

Because of the large number of persons injured in educational institutions and the increasing number of incidents in this sector, ATSDR is evaluating preexisting prevention practices, including green cleaning and green purchasing, chemical cleanout programs, school laboratory education, and mercury and pesticide restrictions. ATSDR is using information from this evaluation to develop a report that synthesizes effective elimination strategies and policies to share with stakeholders to promote a more unified, evidence-based approach to preventing chemical incidents in U.S. schools (24). NTSIP also has focused on educating pool owners and operators about safety practices, including state fact sheets (25) and a collaboration with the CDC Healthy Swimming Program and the American Chemistry Council to produce a video (26) and posters (27,28). These are just two examples of many NTSIP initiatives to reduce chemical injuries in educational institutions and among members of the general public. In May 2014, ATSDR published a report using NTSIP data pertaining to pool chemical releases (29) and the health consequences associated with them, which were cited as baseline data in CDC's Model Aquatic Health Code (MAHC). MAHC is a voluntary guidance document based on science and best practices that can help local and state authorities make swimming and other water activities healthier and safer (30). NTSIP continues to help states to collect surveillance data and to promote cost-effective, proactive measures, such as adopting the principles of green chemistry, converting to inherently safer design, and developing geographic mapping of chemically vulnerable areas.

References

1. Belke JC. Recurring causes of recent chemical accidents. Washington, DC: US Environmental Protection Agency, Plant Maintenance Resource Center; 1998. Available at http://www.plant-maintenance.com/articles/ccps.shtml.
2. BP. The report of the BP U.S. refineries independent safety review panel. London, England: BP; 2007. Available at http://www.csb.gov/assets/1/19/Baker_panel_report1.pdf.
3. Occupational Safety and Health Administration. U.S. Department of Labor's OSHA issues record-breaking fines to BP. Washington, DC: US Department of Labor, Occupational Safety and Health Administration; 2009. Available at https://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=NEWS_RELEASES&p_id=16674.
4. US Chemical Safety Board. Preliminary findings of the U.S. Chemical Safety Board from its investigation of the West fertilizer explosion and fire. Washington, DC: US Chemical Safety Board; 2014. Available at http://www.csb.gov/assets/1/19/West_Preliminary_Findings.pdf.
5. US Chemical Safety Board. CSB report on 2008 Bayer CropScience explosion. Washington, DC: US Chemical Safety Board; 2011. Available at http://www.csb.gov/csb-issues-report-on-2008-bayer-cropscience-explosion-finds-multiple-deficiencies-led-to-runaway-chemical-reaction-recommends-state-create-chemical-plant-oversight-regulation.
6. US Chemical Safety Board. Inherently safer: the future of risk reduction [video]. Washington, DC: US Chemical Safety Board; 2012. Available at http://www.csb.gov/videos/inherently-safer-the-future-of-risk-reduction.
7. US Environmental Protection Agency. Green chemistry [website]. Washington, DC: US Environmental Protection Agency; 2015. Available at http://www2.epa.gov/green-chemistry.
8. National Institute of Environmental Health Sciences. HAZMAT safety and training. Research Triangle Park, NC: National Institutes of Health, National Institute of Environmental Health Sciences. Available at http://www.niehs.nih.gov/careers/hazmat.
9. Agency for Toxic Substances and Disease Registry. Home page. Atlanta, GA: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, CDC. Available at http://www.atsdr.cdc.gov.
10. Agency for Toxic Substances and Disease Registry. Hazardous Substances Emergency Events Surveillance: biennial report 2007-2008. Atlanta, GA: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, CDC. Available at http://www.atsdr.cdc.gov/HS/HSEES/annual2008.html.
11. CDC. Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. MMWR Surveill Summ 2015;64(No. SS-2).
12. Comprehensive Environmental Response, Compensation, and Liability Act of 1980, Pub. L. No. 95 510 (Dec. 11, 1980), as amended by the Superfund Amendments and Reauthorization Act of 1986, Pub. L. No. 99 499 (Oct. 17, 1986), 42 U.S.C. 9604(i).
13. Agency for Toxic Substances and Disease Registry. National Toxic Substance Incidents Program (NTSIP) training manual: appendices I-III. Atlanta, GA: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, CDC; 2010. Available at http://www.atsdr.cdc.gov/NTSIP/Documentation/UserManual.pdf.
14. US Census Bureau. 2002 North American Industry Classification System (NAICS). Washington, DC: US Census Bureau; 2013. Available at http://www.census.gov/cgi-bin/sssd/naics/naicsrch?chart=2002.
15. Ruckart PZ, Orr MF. Temporal trends of acute chemical incidents and injuries-Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. In: CDC. Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. MMWR Surveill Summ 2015;64(No. SS-2).
16. Duncan MA, Wu J, Neu MC, Orr MF. Persons injured during acute chemical incidents-Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. In: CDC. Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. MMWR Surveill Summ 2015;64(No. SS-2).
17. Anderson AR, Wu J. Surveillance of the top five industries resulting in injured persons-Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. In: CDC. Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. MMWR Surveill Summ 2015;64(No. SS-2).
18. Anderson AR, Wu J. Surveillance of the top five chemicals resulting in injured persons-Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. In: CDC. Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. MMWR Surveill Summ 2015;64(No. SS-2).
19. Young R, Pallos L. Geographic distribution of acute chemical incidents-Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. In: CDC. Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. MMWR Surveill Summ 2015;64(No. SS-2).
20. Melnikova N, Wu J, Orr MF. Public health response to acute chemical incidents-Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. In: CDC. Hazardous Substances Emergency Events Surveillance, nine states, 1999-2008. MMWR Surveill Summ 2015;64(No. SS-2).
21. Mary Kay O'Connor Process Safety Center. Developing a roadmap for the future of Hazardous Substance Incidents Surveillance. College Station, TX: Mary Kay O'Connor Process Safety Center, Texas A&M University System; 2009. Available at http://pscfiles.tamu.edu/library/center-publications/white-papers-and-position-statements/Developing%20a%20Roadmap%20for%20the%20Future%20of%20National%20Hazardous%20Substances%20Incident%20Surveillance.pdf.
22. Agency for Toxic Substances and Disease Registry. National Toxic Substance Incidents Program. Atlanta, GA: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, CDC; 2015. Available at http://www.atsdr.cdc.gov/ntsip.
23. The White House. Executive order-improving chemical facility safety and security. Washington, DC: The White House; 2013. Available at http://www.whitehouse.gov/the-press-office/2013/08/01/executive-order-improving-chemical-facility-safety-and-security.
24. Agency for Toxic Substances and Disease Registry. Reducing chemical accidents involving pesticides, mercury, cleaning products, and science labs in schools. Atlanta, GA: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry, CDC; 2015. Available at http://atsdr-dev.cdc.gov/ntsip/docs/Reducing_Chemicals_in_Schools.pdf.
25. New York State Department of Health. Swimming pool chemicals: storage, handling and emergency response. Albany, NY: New York State Department of Health; 2014. Available at http://www.health.ny.gov/environmental/chemicals/pool_chems.
26. American Chemistry Council. Pool chemical safety [video]. Washington, DC: American Chemistry Council; 2013. Available at https://www.youtube.com/watch?v=f-T6czL0cKM.
27. CDC. Pool chemical safety: use [poster]. Atlanta, GA: US Department of Health and Human Services, CDC. Available at http://www.cdc.gov/healthywater/pdf/swimming/resources/chemicalsafety/pool-chemical-safety-poster.pdf.
28. CDC. Pool chemical safety: storage [poster]. Atlanta, GA: US Department of Health and Human Services, CDC. Available at http://www.cdc.gov/healthywater/pdf/swimming/resources/chemicalsafety/pool_chemical_storage.pdf.
29. Anderson AR, Welles WL, Drew J, Orr MF. The distribution and public health consequences of releases of chemicals intended for pool use in 17 states, 2001-2009. J Environ Health 2014;76:10-5.
30. CDC. The Model Aquatic Health Code (MAHC): a model public swimming pool and spa code. Atlanta, GA: US Department of Health and Human Services, CDC. Available at http://www.cdc.gov/healthywater/swimming/pools/mahc.