The preceding chapters have described a set of objectives and a shared vision for a “smart” occupational safety and health (OSH) surveillance system and have provided an overview of the current status of occupational safety and health surveillance as well as recommendations to strengthen those efforts. This chapter discusses ongoing efforts that offer promise for improving occupational health and safety surveillance. Seven areas are focused on:
- Exploration and implementation of a household survey,
- Use of electronic health records,
- Coding of occupational information,
- Electronic reporting initiatives,
- Use of workers’ compensation data,
- Leveraging existing surveys and data systems, and
- Improving occupational hazard and exposure surveillance.
Each topic is explored briefly and that description is followed by the committee’s recommendation on that issue.
This chapter focuses on a variety of activities that are already under way and that constitute opportunities for improving OSH surveillance. In the next chapter, the committee introduces the enabling components and emerging methods that may further enhance our ability to achieve a smarter system for OSH surveillance in the future.
The Bureau of Labor Statistics (BLS) is exploring the feasibility of conducting a nationwide Household Survey of Occupational Injuries and Illnesses (HSOII) with the goals of better understanding the needs of workers, employers, and the safety and health community and addressing the current undercount of occupational illnesses and injuries (see Chapter 4). The HSOII would contact workers directly—outside of the employer-employee relationship—and would ask questions similar to the establishment-based (i.e., employer-based) Survey of Occupational Injuries and Illnesses (SOII) to provide additional data and allow for comparability (Monaco, 2016). The primary goal of the new survey would be to produce as complete as possible a measure (counts and rates) of occupational injuries and acute illnesses in the U.S. economy by capturing all workers including self-employed, contract and “gig” workers (defined in Chapter 1), household workers, migrant laborers, and immigrant workers. Additionally, the new survey would identify gaps in the estimates derived from the SOII. The household survey under development has been designed on a statistically valid platform that will provide better demographic data as well as present the opportunity to ask occupational safety and health questions directly of employees, facilitate special studies, allow for rotating topics and questions, and obtain improved descriptions of acute events. Nonresponse will likely be a challenge, thus BLS has proposed several methods to improve response rates for the household survey, including the use of dialing protocols that rotate calls through different times of day and days of week, increasing the number of attempts of contact, and maximizing use of highly-trained interviewers (https://www.reginfo.gov/public/do/DownloadDocument?objectID=71013400).
This approach would build on BLS’s successful record of collaborating on the collection of complicated and sensitive information using population surveys, such as its collaboration with the Bureau of Census on the Current Population Survey (CPS). The basic infrastructure for such a survey is already in place and would provide results that would be consistent over time and across states, as well as being statistically valid and providing a measure of reliability. As currently envisioned in a recent report produced for BLS, the HSOII would not be sufficiently robust to allow examination of findings that are state specific. The target population being considered for an HSOII is workers age 16 years and older, with a worker being anyone who worked in at least 1 of the prior 52 weeks (NORC, 2016b).
BLS contracted with the National Opinion Research Center (NORC) to develop recommendations for survey design options intended to meet the requirements of such a survey regarding sample representativeness, data quality, timeliness, and cost (NORC, 2016a,b). NORC provided three
options for a HSOII (NORC, 2016b). The one that is considered most economical—and assessed by NORC as meeting the full requirements—is the option that uses “supplemental questions on occupational safety and health following the CPS Annual Social and Economic Survey (sometimes referred to as the CPS March Supplement) for those sample persons identified as meeting the HSOII eligibility requirements” (NORC, 2016b). About 3 percent of CPS respondents would be asked the full HSOII questionnaire under this option (NORC, 2016b).1 A second option would add questions to the June or July CPS when there is no major supplement, but this option is slightly costlier than the first as it entails slightly more screening and respondent burden. A third option would use the ACS respondents as a sampling frame, and then administer the full HSOII separately from the ACS. This option offers more flexibility in targeting selected industries and occupations, not possible with the first two options that add questions to the CPS, but at a higher cost, due to increased respondent burden and additional fixed costs of administering a separate survey rather than adding to an existing survey.
According to NORC’s estimates, the first two options based on adding questions to the CPS would produce 51,000 to 57,000 completed interviews within a $1 million budget, while the third option would yield less than 50,000 completed interviews within the same budget (NORC, 2016b). The third option, therefore, represents a trade-off between greater flexibility in targeting the HSOII versus a higher cost per completed interview. Before selecting one of these options, BLS plans to conduct a pilot survey on a smaller, nationally representative sample of about 5,000 workers to assess their occupational injuries and illnesses, with the goal to have results available in 2018 (Monaco, 2016). In this pilot, the screening questions in the core CPS will serve to identify the subset of respondents to receive the full HSOII questionnaire—those aged 16+, working, and who reported in the core CPS that they had sustained a work-related illness or injury.
A similar effort to assess workplace injuries by self-report has been carried out in the United Kingdom: the Labour Force Survey (see Box 6-1). The United Kingdom’s Health and Safety Executive (HSE) uses two sources of data to assess occupational injury and illness: (1) the Reporting of Injuries, Diseases and Dangerous Occurrences Regulations (RIDDOR), the equivalent of the SOII, and (2) the Labour Force Survey, akin to the proposed
1 In addition to the ~57,000 CPS ACES households, of which ~65% of the individuals in the sample aged 16+ are estimated to be working, consideration is being given to include the ASEC supplemental samples of 6,500 Hispanic HUs and 19,000 CHIP HUs (NORC, 2016b). Based upon CPS response rates and ACS worker rates, it is estimated that this inclusion would add an additional 23,400 sample persons eligible for the HSOII (NORC, 2016b). Consequently, the sample yield estimated to receive the full set of HSOII questions should substantially exceed 4,000 (NORC, 2016b).
HSOII. These two systems are considered complementary. The HSE does not attempt to integrate the two directly to arrive at a single annual estimate of occupational injury and illness. Rather they present both rates along with analysis of determinants of the rates based on data available from each of the systems.
As with the United Kingdom’s two surveys, the working design for the HSOII targets acute injuries and illnesses in a manner meant to parallel the objectives of the SOII. Despite being significantly challenged in terms of enhancing the details of injuries as currently represented in the SOII, the HSOII will serve a critically important function by improving the count
and hence the understanding of the burden of injuries in populations not covered by the current SOII. The secondary goal—using this approach to develop a better understanding of the undercount—may be a greater challenge to fulfill. Recordkeeping requirements from the Occupational Safety and Health Administration (OSHA) provide the basis for the SOII and include parameters such as days lost, treatment beyond first aid, and others. BLS has determined that these parameters are currently not well understood by employers, therefore, there are likely to be significant challenges to formulating survey questions on these aspects of an injury in such a way that they are understood by individual survey respondents.
Two additional gaps recognized in the 1987 National Research Council (NRC) report could potentially be addressed in the household survey:
- The almost complete absence of information about chronic diseases that are either uniquely caused by work or that include work as one of several important factors, and
- The need to develop and enhance information about hazards at work.
Opportunities for collecting self-reported information on chronic conditions was discussed in Chapter 4. Although the quality of information on disease that is self-reported is always potentially problematic, the National Center for Health Statistics (NCHS) has been successful in collecting and reporting on reliable information from respondents asked about their health. For example, in the ongoing National Health Interview Survey (NHIS), conditions such as low back pain, carpal tunnel syndrome, eye-nose-throat irritation, and skin conditions have been reliably reported by respondents, as has information about doctor-diagnosed diseases such as diabetes and high blood pressure. With insights from the NCHS experience regarding disease and chronic health conditions, BLS could consider including these questions in the HSOII.
BLS would also need to consider the HSOII’s potential to collect information about work environment hazards that have not been available nationally since the National Institute for Occupational Safety and Health (NIOSH) stopped conducting the National Occupational Hazard/Exposure Surveys in the early 1980s (see Chapter 4 and the section on “approaches to hazard and exposure surveillance” later in this chapter). The HSOII presents an opportunity to collect self-reports of important and prevalent working conditions, including exposure to physical hazards (exposure to noise, dust, chemicals, or infectious agents; lifting heavy loads; and repetitive hand movements), work intensity (working at speed and to tight deadlines, not having enough time to do the job, frequent disruptive interruptions, pace determinants and interdependency, and emotional
demands), types of working relationships (shift work, working hours, second jobs, working alone, and contract or on-demand work), and skills and discretion (cognitive dimension of work, decision latitude, organizational participation, access to training, use of technology at work, and teamwork). Such information has been collected and tracked successfully in the European Working Conditions Survey approximately every 5 years since 19912 (Eurofound, 2017a). Work organization factors currently have been collected in NHIS supplements on occupational health in 2010 and 2015 (NCHS, 2016), and the importance of collecting exposure information on psychosocial factors has been provided by Australian researchers concerned with important factors at work that affect mental health (LaMontagne et al., 2016). Enhancement of HSOII by questions about health conditions and disease and work exposures would require further careful investigation concerning respondent burden and feasibility.
The results of the household survey (HSOII) and employer-based survey (SOII) would need to be disseminated together as they would offer complementary insights. A nationwide HSOII expanded to include health conditions and diseases as well as exposure to work hazards would need to occur at least every 5 years or more frequently if feasible. Some health conditions that are well known to be caused by work (e.g., silicosis and lung cancer) have long latency periods and so do not appear in SOII. Many other health conditions are not uniquely caused by work but work is one important factor in the development and evolution of disease. While it would be desirable to have these conditions reported and tracked annually, a survey interval of 5 years would add immeasurably to the knowledge base of the distribution and determinants of these conditions.
Conclusion: A household survey on occupational injuries and illnesses would provide data needed to provide more comprehensive surveillance data that will include important information currently lacking in the SOII. The committee finds that the HSOII will serve the BLS objectives to provide
- Greater accuracy by capturing data on all workers in the U.S. economy;
- A statistically valid platform to ask occupational safety and health questions through special studies, rotating topics and questions, better demographic data, and improved description of the event; and
- Self-reported data on occupational injuries or illnesses that provide
2 The European Foundation for the Improvement of Living and Working Conditions’s (Eurofound’s) 2015 European Working Conditions Survey was administered in 35 countries to nearly 44,000 subjects through computer-assisted interviews conducted in 49 languages (Eurofound 2017a,b).
- a complement to the employer-based survey (SOII) so that together the two sources offer broader insights to prevention.
In addition to these advantages, experience from the United Kingdom and from the European Foundation for the Improvement of Living and Working Conditions [Eurofound] provide strong evidence that the HSOII approach presents an excellent opportunity to collect information that has been missing from any routine national surveys:
- Reports of chronic conditions or diseases and their relationship to work, and
- Reports of a wide variety of work hazards that cover many conditions that are considered important in affecting the acute and chronic health of the workforce.
Recommendation D: BLS should place priority on implementing its plan for a household survey of nonfatal occupational injury and illnesses (HSOII). With the assistance of NIOSH and the Centers for Disease Control and Prevention (CDC), BLS should also expand this effort to include a periodic nationwide household survey to identify and track reports of occupational exposures and should determine how best to identify and track chronic work-related illnesses.
In the near term:
- BLS should survey occupational injuries and acute illnesses (as in SOII) in a nationally representative sample of the entire working population including those who are self-employed or engaged in temporary contract work.
In the longer term:
- To address the inadequacies of current surveillance tools, BLS should
- Seek assistance from NIOSH to enhance the HSOII survey scope by assessing occupational exposures and risks in a manner like that used in the Eurofound Working Conditions Survey.
- Questions should be included to capture exposure determinants and work characteristics with sufficient details on industry, occupation, work organizational characteristics, and working relationships in a way that supports the development of a flexible job exposure matrix and supports integration of newly available or ancillary data.
- Seek assistance from NCHS and NIOSH to address currently
- inadequate information on chronic disease and work by determining whether self-report of illnesses and chronic conditions is best tracked by inserting occupational information into the NHIS or inserting chronic illness questions into the HSOII. Part of this consideration should include the determination of whether a sample of retirees and those not working due to disability should be part of the HSOII.
- BLS should prepare and implement a specific plan for routine analysis, interpretation, and preparation of a report on the findings from the HSOII along with a plan for dissemination and appropriate database access by researchers and the public.
Passage of the 2009 Health Information Technology for Economic and Clinical Health (HITECH) Act has led to increased adoption of electronic health records (EHRs) by both individual clinicians and health care provider organizations. With the goal of ensuring the meaningful use of the EHR, the Office of the National Coordinator for Health Information Technology (ONC) is guiding a process that provides financial incentives so that EHR designers and vendors will develop new EHRs, and augment older ones, to meet a broad range of requirements that enhance the role of information technology in supporting clinical and population health (HealthIT, 2017). The meaningful use criteria established by the ONC are designed to result in better clinical outcomes, improved population health outcomes including reduction in health disparities, increased transparency and efficiency, empowered individuals, and more robust and integrated research data available through health data systems.
The increasing adoption of the EHR and the establishment of meaningful use criteria offer an unprecedented opportunity to improve data capture of the impact of work and work exposures on individual health. Advances in capturing these data could improve occupational health practice and are expected to inform general medical practice by more effectively placing health conditions in the context of work. This context is particularly important when accounting for the impact of work on a health condition, even if the condition was not caused by work (for example, a diabetic working rotating shifts needs to plan to adjust insulin management to the pattern of working hours). Readily accessible occupational data in health records also can inform community-health needs assessments. Every 3 years, under the Patient Protection and Affordable Care Act (P.L. 111-148), tax-exempt 501(c)(3) hospitals are required to conduct a Community Health Needs Assessment to develop and adopt an implementation plan to address an identified population health need. Local health departments seeking national
accreditation must also conduct a Community Health Needs Assessment every 5 years as part of their strategic planning.
NIOSH has been engaged in efforts to promote the inclusion of occupational information for health in EHRs (MacKenzie et al., 2016). Inclusion of occupational data in the EHR offers promise for improved clinical care but also would enhance surveillance of occupational disease and injury to improve population health and address health disparities through several routes. First, data from the EHR could be used to improve documentation of injuries related to work. Second, the data in the EHR could potentially be used to identify when occupation is partially or wholly responsible for illness or disease (i.e., asthma in a machinist or chronic obstructive pulmonary disease in a miner), especially chronic conditions. And third, the EHR data could be used to identify work or work-exposure relationships with new or emerging patterns of disease. Inclusion of occupational information in the EHR is critically important to increasing the availability of this information in other public health data systems used for surveillance such as cancer registries, trauma registries, and emergency department data sets such as the National Electronic Injury Surveillance Systems—Occupational Supplement (NEISS Work) that rely on obtaining this information from medical records.
Shortly after the enactment of the HITECH Act, NIOSH began an effort to establish meaningful use requirements for the collection of occupational data through an EHR. As part of this effort, NIOSH asked the National Academies to analyze the potential benefits of including occupational data in the EHR and how technical challenges for the effective incorporation of occupational data in the EHR could be overcome. The ensuing report of the Institute of Medicine (IOM, 2011) noted that, from the perspective of public health or occupational health surveillance, linking occupation data to patient care data provides an opportunity to evaluate injuries, illnesses, and health status in relation to work in the populations receiving care. The report recommended NIOSH focus initially on developing feasible means to incorporate an appropriate level of data on occupation, industry, and work-relatedness into the EHR and subsequently consider what efforts were needed to enhance the value and use of occupational data that would be available in the EHR in the future.
In response to the recommendations in that report, NIOSH undertook a series of projects to build support for the capture of occupational data using EHR systems primarily focused on demonstrating the feasibility of capturing these data in the record, modifying or developing guidance regarding software systems for efficient management and retrieval of occupational data in the electronic record, and developing model systems of clinical decision support for the specific examples of asthma, musculoskeletal disorders,
The occupational data that minimally meet the needs identified in the IOM report are current occupation and industry (useful for acute occupational injury, short-latency occupational illnesses, and management of current medical conditions) and usual occupation and industry (necessary for evaluating occupational illnesses of long latency) (IOM, 2011).
Once in the EHR, current occupational information and, more importantly, the development over time of full occupational histories will permit examination of specific common conditions to seek important signals for an occupational factor. Once noted, these can be tracked, hot spots can be identified, and formal etiologic studies can be planned to advance knowledge and prevention.
A related concern is determining the optimal time and method for collecting occupational data during a clinical encounter. Currently collection generally occurs at the intake interview, if at all. Ensuring that occupational data are accurately entered in the EHR will require tools and training. Pilot efforts suggest that personnel can be effectively encouraged to collect these data for most jobs (NIOSH, 2017). As an alternative, NIOSH is assessing the feasibility of having patients enter occupational data themselves at the time of intake. This approach will most likely require a dictionary of job titles or industries in a drop-down list to facilitate selecting a standard term from a pick list, although it may also be possible to assign codes automatically to narratives entered by patients.
Regardless, for occupational data to be most useful, they must be linked to the clinical record when they are collected. NIOSH has thus focused its attention on interoperability, or the ability to move structured data between systems in a manner that uses standards that ensure that the data can be used by the recipient system. Such interoperability is required for effective data movement among patient care domains as well as between clinical domains and public health systems. Such efforts also pertain to systems that automatically encode data captured as free text and can utilize methods that provide common language translations of those codes for use by other systems (see section on “coding of occupational data” in this chapter).
NIOSH is also exploring what additional data would need to be collected about work. It has framed an information model, which could further support clinical care, population health, and public health activities by including information about work hours, schedule, job duties, and exposures (NIOSH, 2017).
Finally, for data collected in an EHR to be useful for public health surveillance, they must be accessible by the pertinent agencies. One possible mechanism is electronic case reporting, wherein logic is encoded in
the EHR to identify cases of reportable conditions and then to report the cases automatically to the appropriate public health agency. Efforts are well under way in the United States to define a technical framework and data elements, including occupation, for electronic case reporting (MacKenzie et al., 2016).
Summary and Conclusion
The 2011 IOM report provides extensive documentation about the value of routinely capturing occupational data in an EHR (IOM, 2011). Those findings are reinforced by work done since that report. Currently, the major impediments to progress are feasibility and the availability of resources.
In 2015, the ONC acknowledged that occupational data captured in the EHR can benefit patient care and population health, but noted that data standards and software tools for capturing these data in the EHR were too immature at that time to establish a certification criterion for the capture of occupational data (HHS, 2015a). The ONC has indicated that it will monitor the development of these tools and standards for future rulemaking (HHS, 2015b).
Conclusion: Routine inclusion of occupational data in the EHR is required for improved diagnosis and treatment of work-related conditions; for better-informed management of health conditions that are affected by work circumstances; for enhanced understanding of community health needs and resources; and for local, state, and national surveillance of occupationally related conditions. Further efforts are critical to ensure that industry and occupation are included in the EHR meaningful use data.
Data relevant to occupational health and safety surveillance have three basic dimensions: occupation, industry, and a description of the case characteristics of an occupational injury, illness, or fatality.
For occupation and industry, the official coding schemes or controlled terminology used by federal statistical agencies are the Standard Occupational Classification (SOC) for occupation and the North American Industry Classification System (NAICS) for industry. The U.S. Census Bureau has a related less detailed industry and occupation coding system, derived from SOC and NAICS, for purposes of coding data provided by individuals. The Census occupation codes and Census industry codes are mapped to the SOC and NAICS codes (see Box 6-2).
Workers’ compensation insurance rating firms and agencies use classification systems that combine elements of occupation and industry. These
classifications are called “manual classes,” with the most commonly used systems across the United States being those from the National Council of Compensation Insurance (an example from New York can be found at New York Workers’ Compensation Board (NY WCB, 2017); also see Box 6-2 and discussion under workers’ compensation section below).
Coding of type of work and workplace is complemented by classification systems to characterize the occupational injury or illness. Injury and illness classification is an important concern—the classification will likely vary based on who provides the information (health care provider, employer, or worker), what data are available (narrative text with injury descriptions, and physician diagnoses), when it is provided during the course of care, and the quality of the coders.
Injury or illness events are coded in three different ways depending on
the source of reporting. There is no clear mapping between the terms or codes in the three systems (see Box 6-3).
Occupation and industry are often recorded as free text in death certificates, some birth certificates, and in several national surveys (e.g., NHIS, the Behavioral Risk Factor Surveillance System [BRFSS], the National Health and Nutrition Examination Survey [NHANES]), cancer registry reports, and clinical records). Currently, there is no single or universal standard rule or system that institutes a uniform approach to recording occupational data in these documents. For example, the U.S. standard certificate of death includes a section where the funeral director is asked to record the “decedent’s usual occupation” (“indicate type of work done during most of working life; do not use retired”) and “kind of business/industry.” NIOSH
has developed guidance for funeral home directors concerning completion of this section (NIOSH, 2012). The effectiveness of this guidance in improving quality and detail of occupational information across the 50 states has not been evaluated. However, the form of the question about work and the instructions for completing this section both vary among the 50 states as not all states follow the U.S. standard. Regardless, the information on occupation and industry in death certificates is present more than 95 percent of the time and has proved useful for surveillance and research (NIOSH, 2012).
Use of these data for surveillance requires that data written as free text be translated into codes appropriate for the various record types. The process of extracting structured data (or codes) from unstructured data (or free text) requires the following:
- The existence of one or more defined coding system(s),
- Trained individuals and/or software to extract the codes from the free text, and possibly
- The incorporation of the coding system(s) and software into a larger system (e.g., an EHR).
Several coding software systems have been designed to extract a common set of codes for occupation and industry from free text. NIOSH has developed and continues to enhance the NIOSH Industry and Occupation Computerized Coding System (NIOCCS), a coding software package designed to extract injury and occupation codes from free-text data (NIOSH, 2016a). This web-based system is designed to map free text to Bureau of Census codes for industry and occupation with a crosswalk function mapping those codes to NAICS and SOC codes, the major coding systems for most record types. Developed initially to code occupation and industry information on death certificates, NIOCCS has been expanded to other data sources. Thus far, NIOSH has been able to demonstrate moderate success with approximately 60 to 70 percent accuracy in coding industry and occupation by NIOCSS compared with coding by trained coders using a variety of different record types (NIOSH, 2016b). NIOCCS v.2 is currently available for public use and allows for single record or batch processing and for automatic as well as computer assisted coding. Continuous improvement has been documented and the NIOCSS software version 3.0 was scheduled to be released as of the release of this report (January 2018). At least one study that used NIOCCS to code cancer registry occupational information proved less successful than that reported by NIOSH’s work (Weiss et al., 2015).
At the University of California, Los Angeles, the California Health Interview Survey is using NIOCCS coding in providing public search services to compare health demographic and insurance topics by industry and
occupation. The new industry and occupation indicators in the California Health Interview Survey are coded with the help of the NIOCCS (UCLA, 2016). As described elsewhere, NIOSH is using NIOCCS to code BRFSS data on industry and occupation using interview records from 26 states. NIOSH is also starting new projects on coding cancer registry information from a small number of states and piloting use of real-time coding for death certificates in a sample of a few states. Building on the same knowledge base, NIOSH is adapting this approach for electronic health records to better serve clinician needs by preserving more of the rich details in occupation and industry titles with real-time coding.
BLS has also been developing software for coding occupational data in surveys, most notably the SOII. Previous research and ongoing monitoring of the SOII coding effort indicates that the software they developed is assigning codes as accurately or more accurately, on average, than their Office of Safety and Health Statistics’ human coders.3
NIOSH and BLS have also been involved in coding research on other data elements relevant to surveillance: nature of injury, body part and event, result of injury, and source of injury codes (Measure, 2014; Bertke et al., 2016).
In addition to efforts by NIOSH and BLS to develop coding software, many academic groups around the world have also developed and evaluated tools for coding occupational data captured as free text from a variety of sources (Nanda et al., 2016; Marucci-Wellman et al., 2017). There does not appear to be adequate coordination of these activities across agencies and researchers.
Conclusion: Several different coding systems are in use by agencies and other entities to record occupation and industry as well as injury and illness events. These systems have evolved to serve different objectives in different circumstances. Surveillance of conditions that are attributed appropriately to work requires an effort to pursue coordination across agencies and other entities collecting relevant data. Examples include accurate and large-scale coding of occupational data of all types in EHRs, survey responses, death certificates, workers’ compensation records, and related records that can be useful for surveillance of occupational conditions.
In 2016, OSHA issued a new electronic reporting rule requiring certain employers to submit establishment-level injury and illness data to OSHA (OSHA, 2017a). While the agency has required employers to keep injury
3 BLS responses to the NAS OSH Surveillance Committee, August 19, 2016.
and illness records since 1971, for most employers, this information has only been available at the workplace, limiting the utility of the data. The new rule provides a much-enhanced source of injury and illnesses data that can be used for effective targeting of interventions and prevention efforts as well as compliance activity focused on hazardous industries, workplaces, exposures, and high-risk groups. Furthermore, these data are not currently available to agencies or the public from other surveys. This employer-based system also provides new opportunities to conduct outreach and build tools and provide assistance to employers to identify and address hazards at individual worksites.
For decades OSHA has utilized injury and illness data to help target its enforcement, compliance assistance, and other activities. With limited resources and staff—there are fewer than 2,000 federal and state OSHA inspectors responsible for overseeing the safety and health of over 140 million workers at nearly 8 million workplaces (AFL-CIO, 2016)—both federal OSHA and the state plans have sought to target their efforts on the most hazardous industries and employers and on the most serious and widespread hazards. But the agency’s targeting and priority-setting efforts have been hindered by a lack of data, particularly establishment-level information, to evaluate the hazards and risks at individual worksites. While BLS gathers establishment-level injury and illness data for a representative sample of employers through the SOII, the information that is published or public includes summary estimates based on the experience of the sampled companies (weighted to be representative). No establishment-level data are made available to OSHA or the public, due to the BLS policy of maintaining confidentiality of the data in all surveys it conducts.
In 1995, OSHA launched its own initiative to collect injury and illness information from certain establishments. Under the initiative, called the OSHA Data Initiative, OSHA annually collected summary injury and illness information from approximately 80,000 establishments in selected high-hazard industries. This information was used to generate injury rates for individual establishments, with the establishments that showed the highest days-away-from-work injury rates placed on OSHA’s site-specific targeting list for inspections. In more recent years OSHA made the data available through a search function on its website. The initiative was suspended in 2012 due to a combination of a reduced budget and OSHA’s intention to replace it with expanded electronic reporting under the injury reporting rule.
In May 2016, OSHA issued a new rule requiring certain employers to report electronically injury and illness information required under OSHA recordkeeping regulations annually (29 CFR 1904) to OSHA. Under the new rule OSHA will be receiving establishment-specific injury and illness data from more than 460,000 worksites on an annual basis. The rule will provide injury counts and rates for all covered worksites and, for larger
establishments, detailed case and demographic information on all injury cases, unlike the BLS SOII, which only collects detailed information on cases resulting in days away from work. OSHA plans to make much of the data collected from the electronic injury reports available on its website after scrubbing personally identifiable information and information restricted from disclosure under federal law. The Mine Safety and Health Administration (MSHA) has publicly posted injury and compliance results from all U.S. mines since the 1990s.
This new rule now requires electronic submission of relevant injury and illness reports that employers had already been required to maintain, but only onsite. Establishments with 20-249 employees in industries with historically high injury and illness rates will now be electronically submitting information from the OSHA Form 300A–Summary of Work-Related Injuries and Illnesses.4 The electronic reporting rule covers more industries than were covered by the earlier OSHA Data Initiative, including many more industries in the service sector, which will provide valuable data for surveillance and intervention in this growing sector of the economy. For example, hospitals and ambulatory health care facilities, which both have high injury rates, are required to report injuries and illnesses to OSHA. All establishments with 250 or more employees that are covered by OSHA recordkeeping rules are required to submit information from the OSHA Form 300A, as well as the more detailed information being maintained already that is on the OSHA Form 300 (Log of Work-Related Injuries and Illnesses) and Form 301 (Injury and Illness Incident Report).5,6
The new rule will provide an extensive new data source regarding injury and illness that can be used by OSHA, NIOSH, state agencies, employers, workers, and researchers for a range of surveillance and prevention purposes. OSHA estimates that 34,000 larger establishments (≥250 employees) and 431,000 smaller establishments (with 20-249 employees) will provide information on the numbers of fatalities and injuries and illnesses
4 Industries covered include agriculture, fishing, and forestry; utilities; construction; manufacturing; wholesale trade; and other industries with an average rate of days away from work, job transfer, or restriction of 2.0/100 employees or higher for 2011, 2012, and 2013 (OSHA, 2017b).
5 The OSHA Form 300A provides an annual summary including information on the number of injury and illness cases, days away from work, and employment (numbers of employees and hours worked). The OSHA Form 300 provides a listing of each recordable injury and illness and includes information on the employee, the job/activity, the injury, incident, and days away from work or restricted activity. Form 301 is the incident report for each individual case and provides more detailed information on the employee, medical treatment, job activity, nature, source, and events and exposure related to the case.
6 The new regulation also includes provisions to prohibit retaliation against workers who report injuries and policies and practices that discourage the reporting of injuries (29 CFR 1904.35 and 1904.36).
along with employment information that can be utilized to generate injury rates (OSHA, 2016). OSHA estimates that 34,000 logs and 700,000 injury incident reports will be submitted annually. As a benchmark, according to BLS, the SOII receives data from 240,000 establishments and data on 300,000 days-away-from-work cases (BLS, 2013).
The universe of establishments and cases covered by the OSHA and BLS collections are different. OSHA is collecting reports from all establishments with 20 or more employees in designated industries. For larger establishments (≥250 employees) OSHA will be collecting detailed data on all individual injury and illness case reports. BLS collects reports of a sample of establishments in each industry sector and a sample of case reports on days away from work cases in order to generate statistically valid estimates across all industry sectors. As described in Chapter 4, for all other injury cases (e.g., cases resulting in job restriction or job transfer), which represent 70 percent of all cases, BLS does not collect detailed case-level information in the SOII.
In addition, the information collected and available under the electronic reporting rule holds potential value for employers, workers, public health agencies, researchers, and others. Employers will be able to use the information to compare their experience with others in the industry. Workers will be able to have ready access to an employer’s injury reports prior to seeking employment and while employed to assess the safety record of the employer. Public health agencies will be able to determine if there are types of injuries or illnesses occurring in the workplaces of particular industries. Public health departments will be able to initiate intervention efforts, including educational efforts and adjustments to public health standards in industries such as health care facilities, food establishments, or schools, which are regulated by the states. And researchers will have ready access to a large database of injury information to assist them with better characterizing high risks as well as assessing the effectiveness of interventions (O’Halloran et al., 2017).
The electronic reporting initiative also provides an opportunity to create a new avenue for expanding and targeting outreach to employers, particularly smaller employers, to assist them with hazard identification and prevention efforts. The agency could provide automatic feedback or reports to employers on how their injury rates compare with others in the industry. In addition, the agency would need to provide software and other tools and materials to employers to help them analyze their injury reports. Such feedback might be implemented as part of the Injury Tracking Application that OSHA is designing to collect occupational injury data directly from employers.
Among concerns raised about the new electronic reporting requirement, two are relevant to its potential role for occupational safety and health sur-
veillance: duplication of reporting and questions regarding OSHA’s capacity to utilize the substantial amount of new information it will be receiving. The OSHA electronic reporting rule and the BLS SOII could require some employers to submit the same information twice to the Department of Labor—once to OSHA and once to BLS. BLS estimated that there is a 40 percent overlap between the two reporting requirements (Monaco, 2016). OSHA and BLS are collaborating on the implementation of OSHA’s electronic reporting rule so that BLS can use the data received by OSHA in the annual SOII. Such collaboration and coordination is critical to provide both agencies with the data they need while avoiding duplicate reporting requirements for employers. Furthermore, OSHA will have access to detailed data not available to the agency from the BLS-SOII efforts—data useful for prioritizing program efforts for targeting inspections and for efforts to support employers in compliance. The committee notes that currently there are some differences between the data included on the OSHA and BLS reporting forms that will need to be reconciled. Most specifically, the OSHA forms do not include race and ethnicity information for individual cases, which is an optional field on the BLS form.
Historically, OSHA’s capacity to utilize data for enforcement and other purposes has been limited and concerns have been raised about the agency’s ability to effectively utilize the data collected under this initiative. Similar concerns were noted in the 1987 NRC report when the issue of OSHA collection of employer injury and illness data was reviewed. As one response to the 1987 report, OSHA developed and implemented the OSHA Data Initiative, collecting summary injury data from 80,000 employers and utilizing the information for inspection targeting. As OSHA develops its system to receive and manage the electronic records, the agency will need to consult with BLS to assure that coding of the newly available case and demographic data that are submitted to OSHA as free text is compatible with the current methods that are in use for the BLS-SOII.
Conclusion: The OSHA electronic reporting rule will serve a key role by providing data essential for injury and illness surveillance not available from the SOII. These data are useful for targeting interventions and prevention efforts that focus on hazardous industries, workplaces, and exposures as well as high-risk groups. The rule also provides new opportunities to conduct outreach and to provide tools and assistance to employers who need to identify and address hazards at individual worksites.
Coordination and integration of data-collection efforts by OSHA and BLS will prevent duplication of reporting by some employers to both agencies which otherwise may undermine support for this new initiative. New data tools, including development of off-the-shelf software for use by employers or tools for OSHA to provide feedback directly to employers, will also be important in building support for this new initiative. Increased
collaboration among OSHA, BLS, NIOSH, and state agencies will ensure the maximum use of this important new data source on work-related injuries and illnesses.
Recommendation E: OSHA, in conjunction with BLS, NIOSH, state agencies, and other stakeholders, should develop plans to maximize the effectiveness and utility of OSHA’s new electronic reporting initiative for surveillance. These should include plans to provide ongoing analysis and dissemination of these data and to minimize duplication of reporting by employers.
In the near term:
- To avoid duplicate reporting, OSHA and BLS should integrate data-collection efforts so that employers selected in the annual BLS sample for SOII but reporting electronically to OSHA need not make separate reports to BLS. This will require that a unified reporting form include requiring race and ethnicity in submitted case reports.
- OSHA should provide timely and automatic feedback to employers that provides comparative information specific to the employer and others in that industry.
- OSHA should develop a publicly available and easily searchable injury and illness database based on the electronic reports.
In the longer term:
- OSHA and NIOSH should work with stakeholders to develop software and other tools and materials that facilitate further establishment-level analysis of injury data with specific attention to enabling effective use by employers as well as others to identify hazards and job-specific issues for prevention. With experience from participants in this electronic reporting, OSHA should explore feasibility to expand electronic reporting to all employers required to maintain OSHA logs.
Historically, the state and national occupational health community has recognized the value of workers’ compensation data despite their limitations (Utterback and Schnorr, 2010, 2013). For the most part, the data are useful at the state level where the workers’ compensation claims data provide an extensive data source for case-based surveillance programs. Claims for specific conditions can be identified through queries of electronic databases, resulting in rapid and efficient case ascertainment. Claims data
allow contact information for follow-back surveys to capture additional information about exposures, controls, and medical outcomes. Use of unique data sets within states that are the sole provider of workers’ compensation insurance (exclusive state-funded states) has led to the characterization of conditions typically difficult to identify in other surveillance systems. For example, Washington State workers’ compensation data allow characterization of specific work-related musculoskeletal disorders using ICD-9-based case definitions available in medical billing data (NRC and IOM, 2001; Silverstein et al., 2002; Marcum and Adams, 2017). Similarly, Liberty Mutual Insurance uses workers’ compensation data to characterize the most common and costly types of workplace injuries, publishing the annual Liberty Mutual Workplace Safety Index (Marucci-Wellman et al., 2015). The California’s workers’ compensation information system similarly is building a comprehensive database of injury, lost-time cases, and medical care provision used in many system evaluation efforts and to promote targeting of prevention efforts (Das et al., 2012; Joe et al., 2014).
Workers’ compensation data arise out of a system of oversight and case management. Data collected by state labor and workers’ compensation agencies typically address four distinct phases of the workers’ compensation case: the injury, the medical care, the financial compensation process, and the claims process. Injuries that become claims, particularly “lost-time” claims, beget much recordkeeping. A complete insurance record on a workers’ compensation case includes information about the occupational injury or illness, about the provision of timely and adequate medical care, and about the payments for financial compensation for short- or long-term disability. A further record may be developed if there are disputes over how an injury occurred, whether it was on the job, the severity of the injury, or disputes over which employer holds responsibility.
Thus, workers’ compensation records are uniquely positioned to collect data not generally provided by other injury and illness surveillance sources. These include data on injury and illness severity (e.g., number of office visits, hospitalizations, and days of compensated lost work time), direct costs of the components of medical care (e.g., office visits, procedures, medications, and physical rehabilitation), and financial benefits to the injured workers for time loss or permanent disability from work and/or benefits to their survivors. Furthermore, the collection of detailed medical billing data and medical records associated with care allows research on medical outcomes, and the appropriateness and quality of care (Prang et al., 2016).
Workers’ compensation data are usually analyzed to determine how to price an insurance product for a single employer rather than using the data to focus on ways to lower the rate of injury and the associated social and economic costs that accompany occupational injuries. But in combination with other prevention-oriented surveillance tools, workers’ compensation
records hold promise to improve understanding and prevention of occupational injuries.
Due to limited coverage of illnesses that also occur in the general population (e.g., lung cancer), illnesses with long latency even when an occupational cause is likely (e.g., mesothelioma), and the difficulties associated with workers’ compensation claims for illnesses, almost all workers’ compensation claims are for injuries, musculoskeletal disorders, and acute illnesses. The potential use of workers’ compensation data collated across states is also limited by different eligibility criteria in different states, changes over time in the medical management and work restriction policies of worker-related injuries, and varying data availability in different states (some states only computerize lost-work-time claims). Within a given state, the data have the potential to provide estimates of the magnitude of occupational injuries and select illnesses, trends in these conditions, emerging problems, and local variation in injury, and, when matched with other public health data, they can offer the expanded capacity to provide a more contextual and complete understanding of the landscape of occupational injury and some acute illnesses.
Researchers have increasingly used workers’ compensation data for understanding injuries in specific occupations, industries, or trades, such as studies in law enforcement officers (Holloway-Beth et al., 2016), the trucking industry (Smith and Williams, 2014), and machine-related injuries in metal fabrication businesses (Yamin et al., 2016). This surveillance approach focusing on multiple types of injuries and health outcomes of an industry facilitates the development of prevention partnerships with affected unions and employers (e.g., TIRES, 2017).
The utility of workers’ compensation data for public health surveillance, however, has been constrained for many reasons, among them administrative barriers to accessing data, data in unusable or burdensome formats for efficient use, and limited investment of public health resources to develop the technical expertise to use these data. Current advances in technology, electronic reporting and billing of workers’ compensation claims data, development of standardized formats, and promotion of the value of workers’ compensation data to the public health surveillance community are making workers’ compensation data more desirable and usable for public health surveillance. Workers’ compensation information is currently used in 3 of the 22 occupational health indicators recommended by the Council of State and Territorial Epidemiologists (CSTE, 2017). Workers’ compensation data have also been used extensively in several states for surveillance of work-related injuries, including musculoskeletal disorders, as part of multisource surveillance (Kica and Rosenman, 2014; Largo and Rosenman, 2015) or where a state agency is the sole insurer for workers’ compensation, most notably, Ohio and Washington.
In the United States, workers’ compensation is primarily a state-based program7 that provides no-fault medical care with no deductible or copay, and provides partial wage replacement (indemnity benefits) to workers injured on the job. The state systems vary markedly across the United States despite the recommendations for national minimum standards across state systems made in 1972 (National Commission on State Workmen’s Compensation Laws, 1972). Each of the 50 states plus the District of Columbia has compensation systems, with varying eligibility requirements, benefit levels, dispute-resolution systems, and types of data that are collected for purposes other than claims administration. The U.S. Department of Labor ceased tracking the number of minimum standards met by each state in 1984. Such information is needed to explore using similar data from more than one state. For the most part,8 these state programs are broadly based and mandatory for workers who are employees; however, they typically do not cover independent contractors, freelancers, or other self-employed persons. This distinction is especially important as increasing numbers of workers are considered part of the growing “gig economy.” As an experience-rated insurance system, workers’ compensation can theoretically provide incentives to employers to reduce their costs by making places of employment safer and healthier.
In many states, workers’ compensation data are collected in a structured regulation-driven data system. Use of these data for direct national reporting and injury surveillance, however, is challenging. There is little uniformity among the states about what data and which transactions (and which versions of the transactions) are transmitted from claims administrators to the state. Some states mandate the reporting of claims using the Electronic Data Interchange standard9 while others make reporting voluntary, while still others have no reporting rules to the state.
In addition to the inability to agree on a set of mandatory national
7 There are also several federal programs covering specific categories of workers: federal workers (Federal Employees’ Compensation Act); longshore and harbor workers employed in maritime employment upon the navigable waters of the United States; coal miners who qualify for programs to compensate for black lung disease (miners totally disabled by pneumoconiosis arising out of coal mine employment), and to survivors of coal miners when deaths are attributable to the disease; and current or former employees (or their survivors) of the Department of Energy, its predecessor agencies, and certain of its vendors, contractors and subcontractors, who were diagnosed with a radiogenic cancer, chronic beryllium disease, beryllium sensitivity, or chronic silicosis, as a result of exposure to radiation, beryllium, or silica while employed at covered facilities (the Energy Employees Occupational Illness Compensation Program Act).
8 Texas allows employers to “opt out” of mandatory workers’ compensation coverage.
9 International Association of Industrial Accident Boards and Commissions’ Electronic Data Interchange Claims Standards “are used by claims administrators to report workers’ compensation first report of injury and subsequent report of injury claims data to U.S. jurisdictions” (IAIABC, 2017).
standards, state workers’ compensation agencies have proved unwilling (and maybe unable) to standardize across systems, share data with other jurisdictions, or share with any central body. In most states, in contrast to voluntarily collected data under the BLS SOII system, there is no collection of race and ethnicity information. Other challenges to using workers’ compensation data in national surveillance efforts are challenges to comparability: differing scopes of coverage or varying waiting period before temporary total disability benefit receipt begins or when injuries need to be reported to administrative agencies.
A different type of barrier to the use of workers’ compensation data is underreporting of work-related injuries and recognized occupational illnesses to these systems (Rosenman et al., 2000; Azaroff et al., 2002; Fan et al., 2006). Reasons for underreporting are numerous, including worker and health care provider’s limited understanding of the available worker compensation benefits due an injured worker, characteristics and severity of the injury or illness, the administrative burden filing a claim imposes upon the worker and health care providers, availability and access to alternative health care benefits, individual worker characteristics, the workers’ employment relationships to the workplace, and disincentives placed upon the health care provider and worker by the employer to claim workers’ compensation benefits.
Variation exists in the underreporting of worker-related injury and illness, benefit eligibility restrictions imposed by varying state statutes, and regulations and case law across U.S. states. This limits the value of workers’ compensation data for interstate comparisons of occupational injury and illness rates (Bonauto et al., 2010). The potential variation in underreporting based on individual and employment characteristics may limit the value of workers’ compensation as a singular source of data for establishing safety and health priorities within a state (Fan et al., 2006).
Even with their limitations, workers’ compensation insurance data include unique information not available such as medical and wage replacement cost data and information on medical care provided as well as work outcomes. Cost data can be categorized and reported in conjunction with claim incidence rates to help target high hazard and high economic cost industry groups (Bonauto et al., 2006). Medical and disability outcome data provide opportunities for surveillance and research. Developing measures to track the quality of care provided to injured workers (e.g., use of opioids for noncancer pain in workers’ compensation) or measures to assess the timeliness of care in the treatment of common occupational conditions are unexplored areas for the occupational health surveillance community. In Australia, a Compensation Research Database is being assembled to study the influence of compensation system processes and practices on health and health-related outcomes (Prang et al., 2016).
While most federal systems use standard industry and occupation classification systems (i.e., NAICS for industry and the SOC system for occupation; or Bureau of Census Classification for industry and occupation), workers’ compensation offers data classification systems that supplement these traditional classification systems (Spector et al., 2014). The most common are “risk classes” or “manual class codes” (see earlier discussion) that group workers based on similar risk for financial loss. Job tasks within an occupation or industry vary, and risk classes by design attempt to capture these differences. For example, employees in a grocery store are cashiers, butchers, and delivery drivers, all of whom would be identified by a common establishment industry code but each would be separated in workers’ compensation data with individual class codes assigned. Strategic use of manual class codes may be useful to identify specific high-hazard job tasks occurring within an otherwise seemingly low-risk industry or occupation.
Opportunities for Incorporating Insurance Data and Improving Workers’ Compensation Data for Occupational Safety and Health Surveillance
Insurers can both contribute to and benefit from the increased amount of risk information that comes from improved surveillance. Insurers can combine their own information on site-specific hazards from safety engineering and industrial hygiene resources within their companies, with improved information from broader industry focused exposure surveillance and industry-wide injury statistics. The aggregation of claims information in risk classifications by national, regional, and state specific insurance rating organizations is used to identify emerging hazards and outcomes for companies purchasing workers’ compensation coverage. In the past, in some specific industries and workplaces, workers’ compensation and health insurance information has been combined to indicate types of injuries and illnesses that would not otherwise become apparent in workers’ compensation claims alone. Re-insurers are in good position to look more globally at such emerging risks as they might take on coverage of industrial hazards that are not apparent to individual insurers.
The facilitation of claims data sharing among state workers’ compensation systems, OSHA programs, workers’ compensation insurers, and health departments may be useful in targeting scarce governmental and private resources for health and safety, including both governmental enforcement and consultation services, and insurer loss control functions. Surveillance findings could also be disseminated broadly through workers’ compensation insurers to their insured employer clients as they emerge. The NIOSH Center for Workers’ Compensation Studies (described below) can be an important bridge between these different entities.
A few methodological research areas could help improve the use of
workers’ compensation data for OSH surveillance. BLS and NIOSH have engaged in research toward coding of both injury narratives and text fields that document occupation and industry. This work could also contribute to crosswalks between distinct systems in the standardized OSH surveillance systems and workers’ compensation categorization methods used in the insurance industry, such as by the NCCI and the International Association of Industrial Accidents Boards and Commissions.
In 2012, NIOSH established the Center for Workers’ Compensation Studies (CWCS) (Utterback and Schnorr, 2013). According to NIOSH, this center focuses on three areas: “(1) expanding use of state-level workers’ compensation claims data for research and prevention, (2) identifying and communicating interventions most effective at preventing injury and illness, and (3) encouraging collaborations between the public health and workers’ compensation communities” (NIOSH, 2016c). The program has begun to “build the capacity of states to use workers’ compensation claims data for prevention purposes through grants, partnerships, and technical assistance” (NIOSH, 2016c). It also “evaluates approaches to preventing illness and injury by working with workers’ compensation insurers” (NIOSH, 2016c). Through information distributed to insurers, state workers’ compensation bureaus, and state departments of health, the center has focused on best practices for treatment of illness and injury, data analysis and denominator considerations, and issues related to return to work. The center is currently working with five states—Ohio, California, Massachusetts, Michigan, and Tennessee—to use claims data toward focusing efforts of injury prevention in high-risk industries (Wurzelbacher et al., 2016).
The center’s orientation is particularly important given the lack of injury prevention–oriented research in most state workers’ compensation agencies. The center is working to bring state public health epidemiologists associated with OSH surveillance programs together with workers’ compensation data experts to focus the data on primary, secondary, and tertiary prevention of work-related injury and illness.
Conclusion: Workers’ compensation data have potential value for surveillance of occupational injuries at the state level and have been specifically used to this end in three states (Michigan, Ohio, and Washington). In 2015, NIOSH established the CWCS to promote use of workers’ compensation data to improve workplace safety and health. The Center works to develop new methods for coding, analyzing, and disseminating workers’ compensation data, foster new research collaborations, and share best surveillance and research practices across state agencies, researchers, and insurance companies using these data. Because workers’ compensation laws and eligibility requirements vary by state, surveillance findings based on workers’ compensation records cannot be readily compared across states. The Department of Labor no longer tracks how well each state meets mini-
mum standards. Surveillance use would be enhanced if this tracking were resumed and made public, and possibly new minimum standards would apply to all states considered.
Recommendation F: NIOSH with assistance from OSHA should explore and promote the expanded use of workers’ compensation data for occupational injury and illness surveillance and the development of surveillance for consequences of injury and illness outcomes, including return to work and disability.
In the near term:
- NIOSH should organize an advisory group of workers’ compensation data experts to advise both the NIOSH Center for Workers’ Compensation Studies and interested states concerning their use of workers’ compensation data for surveillance and research.
- NIOSH should encourage states to expand the use of workers’ compensation information beyond the Council of States and Territorial Epidemiologists (CSTE) occupational health indicators. Specifically, the agency should work through the state surveillance cooperative agreements to develop and enhance use of workers’ compensation data for state-based occupational injury and illness surveillance and prevention activities.
In the longer term:
- NIOSH and OSHA should collaborate with states to pursue the development of surveillance systems that capture cost of work-related injury and illness, measure work-related disability and return-to-work outcomes, and assess the adequacy of benefits administered through workers’ compensation insurance programs.
NIOSH’s surveillance strategy places importance on leveraging existing surveys and data systems. Such efforts provide a valuable addition to the limited resources that NIOSH has available to dedicate specifically to occupational safety and health surveillance. In some examples, such as the National Health Interview Survey described in Chapter 4, additional questions concerning work-relatedness of conditions, common workplace exposures, and other issues related to work have been incorporated in the survey. NEISS-Work, also described in Chapter 4, is another example of collecting more extensive information about work and work-relatedness in an existing surveillance system. Listed in Table 6-1 are some examples
TABLE 6-1 Examples of Surveys and Studies Leveraged to Generate Occupational Health and Safety Information
|Survey||OSH Objective||NIOSH Engagement|
|National Birth Defects Prevention Study (NBDPS)||Examine potential associations between role of work (parental jobs and workplace exposures) and birth defects and other pregnancy outcomes (Lupo et al., 2012).||Partnering with the NBDPS to examine the role of work on birth defects and other pregnancy outcomes. From this study, associations have been found between several parental jobs and workplace exposures and specific birth defects.|
|National Ambulatory Medical Care Survey Asthma Supplement||Learn more about work-related asthma management strategies and track improvements.||Examine work-related asthma management strategies to measure barriers to work-related asthma management strategies, and acceptance of the guidelines by health care practitioners.|
|National Crime Victimization Survey (NCVS)||Provide annual information on burden and track changes in patterns of work-related violence.||Improve data on work-related violence collected through the NCVS.|
|National Health and Nutrition Examination Survey (NHANES)||Examine and track work associations with documented health conditions—initial emphasis on respiratory conditions using spirometry results.||Provide the technical support to ensure high-quality spirometry data to study obstructive lung disease and to collect data on workplace exposure for occupational surveillance purposes.|
|Fatality Analysis Reporting System (FARS)||Measure and track work-related motor vehicle accident fatalities.||Collaborate with the National Highway Traffic Safety Administration and BLS to link CFOI and FARS data for comprehensive data on fatal occupational crashes across industries and vehicle types.|
of national surveys that NIOSH is taking advantage of, as opportunity and resources allow, to incorporate additional questions and generate new occupational safety and health information. An important advantage of collecting information about occupational health and safety within broader public health data sources is that, unlike employer-based reporting systems, this approach allows for assessment of the contribution of work to the overall problem under investigation (e.g., asthma prevalence and incidence of violence). It also offers opportunities to collaborate with other public health
programs to develop more comprehensive approaches to understanding and addressing multifactorial public health problems. These efforts suggest that NIOSH, working in collaboration with other agencies, could take further advantage of existing surveys and data sources to go beyond filling information gaps in order to help meet specific surveillance objectives.
There are, however, significant challenges in getting multiple work-related questions into existing surveys and studies that have been designed for other purposes. These include competing public health or other agency priorities and increasing pressure to limit lengths of surveys to address declining response rates and to reduce costs. Failure to recognize the importance of assessing the impact of work on health within the broader public health community tasked with addressing other public health problems is also common. Ultimately, final decisions about survey content are beyond the control of NIOSH or state occupational health programs and continued (ongoing or periodic) data collection is accordingly unpredictable.
NIOSH is also collaborating with other federal agencies and state partners to promote routine collection and coding of basic information about industry and occupation in existing health surveys and other public health surveillance systems. Collection of this information not only increases potential use of these data sources for OSH surveillance and research, but in many instances, it can also enhance practice in other public health domains by providing information about patterns of health outcomes and determinants of health (e.g., prevalence of smoking behaviors and access to preventive services, and access to health insurance) in relation to work. Given the many industry and occupation categories, this information is generally collected in surveys as narrative text, unlike most other survey variables, which imposes a substantial coding burden. Automated approaches to assigning standardized industry and occupation codes to narrative text are therefore essential to gain acceptance for collecting these data elements and realizing the potential of utilizing the existing surveys and data systems (see above section on coding and further discussion in Chapter 7). Advances in electronic coding of industry and occupation described elsewhere in this report demonstrate much promise in meeting this challenge and have facilitated some successes described below. Further support for maintaining and enhancing these tools is needed.
Occupation and industry are core sociodemographic variables collected in the decennial census and other population and economic surveys including, among others, the Current Population Survey and the American Community Survey. Given the connectivity of one’s occupation to an individual’s level of education, it is also an important component that impacts an individual’s economic status. The importance of documenting health status in relation to type of employment has a long history in the United States. In 1847, Massachusetts was the first state to establish a death registration
system including information about occupation on the state death certificate, with subsequent annual reports on mortality by occupation. Today work is widely recognized as an important social determinant of health, having both direct impacts though the physical and psychosocial environment as well as indirect impacts through access to economic and health resources (Wilkinson and Marmot, 2003; An et al., 2011). Information about industry and occupation are currently collected and coded in most federal health surveys, including NHIS, NHANES, the Medical Expenditure Panel Survey (MEPS), and the National Survey on Drug Use and Health, which gathers information about substance use and dependence or abuse. Industry and occupation are particularly salient sociodemographic measures in national health care surveys such as the MEPS when determining an individual’s access to health insurance and the comprehensiveness of the coverage, given the primary source of coverage for a substantial representation of the population is employment based.
In light of the importance of the inclusion of these measures in such health-related surveys, an impending analytic challenge may develop as a consequence of the current revisions under way in NHIS which is considering collecting industry and occupation information only on a rotating basis (NCHS, 2016). Furthermore, there are also some relevant nationwide health surveys and many CDC public health surveillance systems in which these data elements are not collected, thus limiting the usefulness of these systems for identifying potential cases of work-related disease and characterizing patterns of health outcomes under surveillance in relation to occupation and industry. In some instances, such as CDC’s Pregnancy Risk Assessment Monitoring System (PRAMS), the employment information (maternal occupation and industry) is not collected at all although it is collected on birth certificates in some states.10 There are other information sources, such as state and local cancer registries, where the industry and occupation data are collected and maintained (when present in the medical records) at the state level, but they are not routinely coded and not reported centrally to the National Program of Cancer Registries at CDC which funds state cancer registries in 45 states and uses aggregated state data to generate the official federal statistics on cancer incidence.
NIOSH has had some success in recent years working with other CDC programs and state partners to promote collection of industry and occupation information in additional data sources used for public health surveillance. A major critical initiative is NIOSH’s ongoing work to incorporate industry and occupation information in electronic health records (described
10 The CDC PRAMS survey, conducted in partnership with state health departments, collects information annually on approximately 83 percent of all U.S. births specifically regarding experiences and behaviors before, during, and soon after pregnancy (CDC, 2017b).
above). NIOSH has under consideration a pilot effort with the National Center for Health Statistics to code industry and occupation data on death certificates from 17 states in real time, which, if successful, will provide the opportunity to extend the effort and analyze mortality patterns by industry and occupation for all 50 states.
NIOSH is currently compiling and coding the industry and occupation data from six state cancer registries and will conduct analysis to assess associations between cancer incidence and industry and occupation in aggregate and stratified by state. Additionally, NIOSH has provided support to pilot the collection of maternal industry and occupation in the PRAMS survey in five states. Box 6-4 describes another initiative working with the states on the Behavioral Risk Factor Surveillance System that is illustrative of both the opportunities and the challenges in incorporating occupational information in public health data systems.
The CSTE has recommended that “occupational and industry and other work information as appropriate be included within CDC surveillance systems where feasible” (CSTE, 2014b). Also, the National Committee on Vital and Health Statistics’ Subcommittee on Population Health, charged with recommending minimum standards for measures of socioeconomic status for federal health surveys, has recommended to the Department of Health and Human Services (HHS) that occupation and industry be collected as socioeconomic variables (also referred to as demographic variables) in all federal health surveys (NCVHS, 2012).
For public health surveillance systems that rely on data from health care providers and administrative data from the health care system, lack of information about occupation and industry in medical records is an underlying obstacle. Concurrent efforts are therefore needed to promote collection of industry and occupation in EHRs, as well as to develop automated approaches to coding industry and occupation (see above recommendations) to realize a 21st-century vision in which industry and occupation information is routinely collected and coded in relevant public health surveillance systems.
Conclusion: Occupation and industry are demographic variables that describe core features relevant to adults and are characteristics of individuals essential to understand fully health and the factors that influence it. These variables are considered core demographic variables in the decennial census but are not currently treated as such in systems that collect information on health. Public health information on health and disease among the adult working age population is gathered through surveys or collected in data systems primarily located in the Centers for Disease Control and Prevention’s National Center for Health Statistics and in the Agency for Healthcare Research and Quality (AHRQ). On occasion these have included information on occupation and industry that allow characterization of health by work characteristics. The inclusion of occupation and industry information as a core demographic variable in systems designed to inform the nation about adult health would add important information to guide disease and injury prevention and delivery of health care in the population.
Recommendation G: HHS should designate industry and occupation as core demographic variables collected in federal health surveys, as well as in other relevant public health surveillance systems, and foster collaboration between NIOSH and other CDC centers in maximizing the surveillance benefits of including industry and occupation in these surveys and surveillance systems.
In the near term:
- HHS should re-establish industry and occupation as core demographic variables in all federal health surveys.
- CDC surveillance programs, as they proceed with their state partners to streamline and harmonize data across systems, should work with NIOSH to identify appropriate processes for collecting and coding occupational and industry data.
- NIOSH with assistance from CDC should explore and prioritize public health surveys that can be used to enhance occupational health surveillance objectives by collecting relevant occupational information.
In the longer term:
- To promote proper analysis of surveillance data NIOSH should develop methods and training materials on approaches to basic as well as new and creative use of occupation and industry data and on the selection and use of appropriate labor force denominators.
A key component that is largely missing from U.S. occupational safety and health surveillance is collection and analysis of data on occupational hazards and exposures. Hazard and exposure data are leading indicators for anticipating and preventing work-related chronic disease and, to a lesser extent, acute disease and injuries. As noted in Chapter 4, useful information about hazards and hazardous exposures requires the identification of the hazard and the assessment of the parameters regarding the exposure (i.e., duration and intensity).
Hazard and exposure surveillance provides several benefits not achievable in surveillance of health outcomes, at least for certain classes of occupational risk. First, and perhaps most obvious, is that exposure must precede an adverse outcome, thus, at least in theory, supporting the timely mitigation of the risk prior to injury or illness. For most conditions of interest, there are irreversible consequences, from lost productivity and income
due to days lost at work due to an injury, to development of a chronic disabling disease or death; effective intervention requires assessing the risk prior to the effects occurring. While disease surveillance may be effectively used to prevent future harm, only exposure or hazard surveillance fulfills the primary prevention of occupational ill health.
Related to the timeliness of hazard and exposure surveillance is the opportunity to provide timely feedback to employers interested in maintaining the effectiveness of exposure controls—both for prevention purposes and to ensure compliance with government regulations. As discussed in Box 6-5 a silica exposure surveillance scheme has demonstrated the effectiveness of rapid feedback and benchmarking of exposure monitoring results in further reducing workplace exposures to silica. Collection of such routine exposure data can be integrated into an employer’s occupational health program to ensure compliance with regulatory limits, as well as being part of a government agency’s routine collection of compliance information. Thus, a surveillance system can be built as part of other programmatic goals and does not have to be solely dedicated to surveillance.
The second compelling argument for exposure surveillance is in the context of multifactorial chronic diseases, such as cardiovascular disease, obesity, chronic lung disease, and musculoskeletal disorders. For those diseases with established exposure-response relationships, the availability
of exposure data allows for identification of workplaces with excess risk, and thus opportunities to reduce the occurrence of these multifactorial conditions.
On a population basis, the attributable fraction of a specific condition can be derived based on a comparison of risk in an “exposed” population compared to an unexposed control population. Such comparisons are usually made based on work in a specific industry and/or occupation and are not well informed by the actual probability and level of exposure within those populations. Thus, even this population-level estimate is based on exposure assumptions and is usually fraught with misclassification error. The error can be controlled only to the degree that specific exposure information is available on a population.
A compelling example of using the attributable risk to conduct risk assessment based on the population distribution of exposure is illustrated in the estimation of the future burden of work-related cancer in the United Kingdom in Box 6-6. This powerful approach to understanding the risk of chronic multifactorial disease is the key strength of a robust exposure surveillance system.
The uses of a robust exposure surveillance system are many. The primary use is, of course, the rapid identification of emerging issues or trends in risks and the use of such information for workplace interventions. The collection and analysis of exposure or hazard data, however, can also be used for epidemiological studies, both identifying new risks and refining the quantification of risk for those already identified. Exposure data are also instrumental in conducting risk assessments to understand how and where population risks of disease will occur, estimating the cost burden associated with such risks. Finally, studies based on good quantitative exposure data are needed to formulate policies to prevent or mitigate health impacts and the costs associated with alternative policy choices.
Approaches to Hazard and Exposure Surveillance
Several approaches to hazard and exposure surveillance (described below) have been or are currently under way through the work of NIOSH, OSHA, and international organizations. Learning from those experiences and building on them will set in place a systematic approach to hazard and exposure surveillance with the potential to greatly improve worker safety and health. Data for such surveillance come from a wide variety of sources. Figure 6-1 shows a Venn diagram of the many sources of data and the overlap among the systems that collect these data for occupational exposure surveillance. It provides an overview of the roles and relationships for many of the systems further discussed in this chapter.
As with injury and illness outcomes, exposure information is derived
from multiple sources, including individual, public agency, or employer. Figure 6-1 illustrates how these various exposure data sources relate to each other. While the three circles representing public agencies, individuals, and employers each have multiple sources within them, the totality of the data derived from the three major types of exposure information also may intersect. Public Agencies: OSHA, its state-based affiliates, and MSHA collect exposure monitoring data during regulatory or consultation inspections. NIOSH, through Health Hazard Evaluations and other research-oriented studies, collect substantial exposure monitoring data. Not included but potentially relevant is the hazard information contained in the Department of Labor’s Occupational Network Database (O*NET) and Occupational Requirements Survey (ORS). Employers: Employers, or consultants working on their behalf, routinely monitor exposures among their workforce,
though primarily for regulated agents. A small fraction of the monitoring done by employers may also be reported to the agencies regulating them. Individuals: While individuals generally do not conduct exposure monitoring, they contribute important exposure-related information through various population based surveys such as the BRFSS and NHIS. Not included is the self-reported hazard information in the NSF-NIOSH Quality of Worklife Survey (QWS). Expansion on the types of exposure information collected using the Household Survey of Occupational Injuries and Illnesses could be integrated with exposure measurement databases, and thus the overlap of the individual and public agency data sources. Also missing from this diagram is one additional source of exposure information: biological monitoring results which may be collected through employers or other means and collected by analytical laboratories.
Workplace-Based Direct Observation Surveys
From 1972 to 1974 NIOSH conducted the National Occupational Hazard Survey (NOHS) in approximately 5,000 workplaces (Frazier, 1983). The surveys included a walk-through inspection in which engineers were supposed to observe “every plant process and every employee,” making estimates of the numbers of workers exposed full time and part time to various substances and collecting other information, including data on whether engineering controls had been implemented or personal protective equipment was required for specific exposures. A similar survey, the National Occupational Exposure Survey (NOES), was conducted by NIOSH from 1981 to 1983. Neither of these surveys included any industrial hygiene sampling. Despite the ground-breaking utility of these surveys at the time, no follow-up surveys of a similar nature have been conducted since, largely because of the expense.
Direct observation of exposure conditions within a representative sample of workplaces offers benefits that include
- Characterization of both the use and exposure potential of multiple hazards at the same time, while also characterizing the number of individuals or probability of exposure occurring to the workforce.
- Obtaining a distribution of the frequency, intensity, and duration of such exposures, thus providing a complete distribution of both hazard and exposure, and allowing direct calculation of risks.
However, some risks are not easily observed, and within any workplace there may be too many hazards for an individual observer to identify as many risks may occur in highly incidental conditions and be difficult to observe on any specific day. Selection of a sample of representative workplaces is increasingly difficult due to the mobile and temporary nature of work organization in many industries. The biggest challenge for this approach is its complexity and expense.
Exposure Measurement Databases
Quantitative measurement of exposure intensity using personal sampling methods provides the “gold standard” for industrial hygiene monitoring and standards compliance and provides the most accurate means of evaluating current exposure levels. Because most regulatory compliance is based on quantitative exposure measurement, both industry and government agencies collect a large number of exposure measurements in a wide range of industries. There is powerful potential for compiling these
routinely collected quantitative measurements into databases, which could then be used for surveillance activities.
For instance, the compilation of MSHA data on silica or coal dust during various mining activities has allowed for active surveillance of conditions and risks, intervention at locations with high exposures, and for epidemiology studies of dust-related health conditions. Additionally, OSHA includes quantitative industrial hygiene measurements in their publicly available Integrated Management Information System (IMIS). Inspector-collected samples that were analyzed in OSHA’s Salt Lake City laboratory were included in the Chemical Exposure Health Data (CEHD), made publicly available since 2010 as part of the OSHA Information System. At present the CEHD contains quantitative industrial hygiene measurements and related information on the analyses for Salt Lake City laboratory samples only. The CEHD and IMIS have a significant degree of overlap (about 50 percent) but each data set contains a substantial amount of unique data and the IMIS may include an underrepresentation of nondetectable results (Lavoue et al., 2013a,b). The IMIS database has provided the opportunity for monitoring exposure levels of a few key agents in the workplace throughout the United States. However, the limitations of such data and the challenges in collecting all the relevant information (e.g., the distribution of those exposures among the working population, the frequency of their encounter, the duration of their occurrence, and the wide range of potential agents to be measured) need to be recognized.
Validated biomonitoring methods have been developed for some exposures, such as lead in blood, metabolites of selected pesticides in urine, and cadmium in urine. Biomonitoring has the advantage of circumventing limitations of environmental monitoring, such as the effectiveness of personal protective equipment used by the worker, but has several challenges including potential invasion of privacy and sensitivity to the timing of the sample. Nevertheless, blood lead surveillance, both based on OSHA requirements for monitoring for lead-exposed workers and reporting of blood lead results by analytic laboratories, has proven highly effective in identifying high-risk conditions and stimulating personal and workplace interventions where needed (see Chapter 4).
Job Exposure Matrices
Classical job exposure matrices (JEMs) were developed as relatively crude and static matrices of jobs with assigned exposures, usually nominal
exposure based on self-report or expert judgment; however, JEM methodology has continued to be refined to allow for incorporation of quantitative measurements, refined job categories, information on the duration of exposure, etc. Two examples are worth noting: FINJEM, a comprehensive JEM based on Finnish industry information that allows for periodic updates and includes quantitative measurements (Kauppinen et al., 2014), and CAREX, focused on workplace carcinogens and first developed to cover all European Union member states (Kauppinen et al., 2000) and later tailored to Canada and numerous other countries (also see Chapter 5). Both systems may be used for exposure estimation of populations where an independent data set for duration within each job category and time period is available. Estimates derived in this way are useful for surveillance (see Chapter 5, Figure 5-1 and Box 6-6) and conducting epidemiological or risk-assessment studies for chronic disease. These JEMs include physical agents, chemical agents, biologic agents, physiologic and ergonomic factors, as well as psychosocial factors. As a consequence, the disease risks that can be monitored range from cancer to mental illnesses.
The use of job exposure matrices in surveillance activities could be limited by the extent to which the matrix information can be updated to reflect current conditions. However, the combination of a well-established population-specific JEM with ongoing collection and integration of exposure measurements could prove a powerful surveillance approach.
Given the importance of exposure duration in the estimation of risk, it is vital that sources of information about time spent in various exposure scenarios are also addressed. Questionnaires and administrative data can both be used to estimate risk, assuming the categories of activity addressed can be linked in some way to exposures. Such data may be linkable to exposure level using a JEM, for instance.
Other Sources of Generic Job Exposure
O*NET is a publicly available online database that describes occupational features across U.S. job titles (O*NET, 2017). The database is continually updated by surveying a broad range of workers from each occupation. O*NET has been used to estimate workplace physical and psychosocial exposures and organizational characteristics. The data from a generic job exposure system offers the potential to impute hazard presence associated with a job. This, in turn would allow identification and tracking of potential occupational risks that otherwise have escaped consideration due to missing data or resource constraints on direct collection of job exposure information in the field (Cifuentes et al., 2010). For example, Gardner and colleagues (2010) noted that “job title–based exposure estimates from O*NET and self-reported and observer-rated exposures showed moderate
to good levels of agreement for some upper extremity exposures, including lifting, forceful grip, use of vibrating tools, and wrist bending” (Cifuentes et al., 2007; Evanoff et al., 2014). O*NET also provides job information that could be useful to evaluate psychosocial working conditions especially for the demand/control and effort/reward models (Cifuentes et al., 2007; d’Errico et al., 2007; Boyer et al., 2009; Meyer et al., 2011). Further validation of these data is necessary to determine the utility of the O*NET databases for surveillance.
While O*NET has been partially validated, the Bureau of Labor Statistics, National Compensation Survey (NCS) program has initiated a refined “Occupational Requirements Survey” (ORS) with additional detail about the prevalence and level of physical demands and environmental exposures (BLS, 2014b, 2017b). While designed for the Social Security Administration (SSA) to describe occupation requirements that will assist the agency in eligibility determinations for Social Security Disability Insurance and Supplemental Security Income disability benefits for applicants, the ORS may prove a more useful source for JEM information for some prevalent exposures. For example, this database may supplement the O*NET with a more nuanced description of some of the cognitive and mental requirements for a job as well as adding information about the duration of specific physical demands and environmental exposures associated with jobs. (See the discussion on challenges of using survey data for exposure assessment in the next section on questionnaire surveys.)
Questionnaire surveys of working populations are another powerful and underappreciated approach to exposure surveillance. Survey responses are well suited to characterizing a population’s distribution of exposure duration, and may provide an estimate of intensity of the exposure. In addition, questionnaires offer the opportunity to collect data on multiple hazards, including those that are difficult to measure (e.g., musculoskeletal stressors) and those which can only be directly ascertained through individual experience (e.g., psychosocial stressors).
Some general idea of the prevalence of psychosocial stressors is available from existing surveys (NHIS, BRFSS, and QWS (NIOSH, 2013). NHIS and BRFSS have sufficient sample size to provide a reasonable level of detail on occupation but are fairly limited in detail on the stressors. The QWS is a special module assessing the quality of work life in America that has been added to the General Social Survey, a biannual, nationally representative, personal interview survey of U.S. households conducted by the National Opinion Research Center and funded by the National Science Foundation.
The module has been fielded every four years.11 The QWS provides excellent detail on the stressors but the sample is small and does not provide sufficient detail about occupation. Experience with these, however, would need to inform approaches to survey of these stressors that could produce sufficient detail on both occupation and stressors to enhance surveillance of work and poor mental health.
Questionnaire surveys also offer respondents the opportunity to report on their “usual” exposure experience, while individual measurements observe only the condition at a specific time or day, and thus either miss less frequent conditions or require large sample sizes to capture them. Employee responses are valid when it comes to most of the questions on these surveys because they deal with attitudes or experiences.
An additional key advantage of questionnaire surveys is that they are relatively inexpensive to conduct. Furthermore, surveys on exposures can be added into other ongoing population-based surveys and/or conducted at regular intervals to provide information on changes in work conditions over time. Challenges in using questionnaires include reporting bias, employee awareness of exposures (e.g., low-level gaseous exposures may not be detected), and lack of quantitative information on the intensity of exposure. It may also be challenging to generalize across occupations.
Since 1991, Eurofound has been monitoring working conditions in Europe through its European Working Conditions Survey (Eurofound, 2017a,b). The survey aims to measure working conditions across European countries, identify groups at risk, and highlight concerns and progress, with the aim of contributing to European polices that would improve job quality (Eurofund, 2017b). In 2015, the sixth European Working Conditions Survey was conducted across 35 European countries and interviewed approximately 44,000 total employees and self-employed workers (Eurofund, 2017b). Workers were asked (in their native language) a range of questions concerning employment status, work organization, learning and training, working time duration and organization, physical and psychosocial risk factors, health and safety, work-life balance, worker participation, earnings and financial security, as well as work and health and trends in these exposures were reported.
Several of the systems discussed above have potential for contributing to a comprehensive exposure and hazard surveillance. A system, properly structured, maintained, and funded, would fill important gaps in our ability to identify, monitor, and intervene on the myriad work-related risks,
especially those risks to long-term health of workers. Thus, the committee suggests that a comprehensive hazard and exposure system be developed and implemented, incorporating the strengths of the various approaches identified above, to collect, analyze, and distribute work-related risk information on a population basis. The creation of such a system will need to be undertaken with sensitivity to the need to protect the confidentiality of employees who participate in worksite monitoring. This issue has arisen in the past in settings where employees have expressed concern about risks to their employment relationship when negative exposure data are reported. This issue is one that employers, employees, and the agencies will need to bear in mind during the evolution of exposure surveillance activities.
The foundation of such a system would be a nationally representative survey of the working population in which many different types of work-related risks could be assessed for their prevalence, duration, and, crudely, for intensity. Only through such a system could workers in nonstandard employment arrangements be fully represented and avoid the challenges and limitations of employer-based surveys. In addition to classical dusts, chemicals, and physical and biological agents, emerging sources of risk such as psychosocial dimensions, shiftwork, and other organizational structures could be adequately described.
Such a population-based survey could be a stand-alone survey conducted by NIOSH, a part of the proposed HSOII survey, or as a periodic supplement to existing surveys such as the NHIS. It may be possible to collect a large sample at long intervals, e.g., every 10 years, or a smaller sample, appropriately stratified, at more frequent intervals, if the data can be integrated to provide an ongoing representative picture. Such a survey could be modeled on the European Working Conditions Survey, which has provided very substantial information for European policy makers and researchers.
Missing from a survey based on self-reported responses are quantitative exposure-level data, and risk factors that are not easily perceived by respondents. As discussed above, quantitative surveys of exposures at work are infeasible and limited for other reasons; however, ongoing exposure measurement activity by both government regulators and regulated industries provide the potential for development of exposure-level information, at least for selected highly regulated agents, and in some types of workplaces.
As described above, the compilation of MSHA particulate and noise-level information, and OSHA’s integration of measurements into the IMIS database, has provided researchers the opportunity to extract and analyze exposure levels and trends while also investigating certain biases inherent in these compliance-based databases. Important limitations in these systems come from missing contextual information associated with the measurement data. In addition, employers collect a substantial amount of exposure-level information either independently or through use of consulting services.
These data rarely find themselves included in publicly available data resources due to privacy concerns. OSHA would need to explore ways these routinely collected data could be systematically collected and integrated with OSHA data while protecting the legal rights of the contributing workplaces. A precedent for this may be found in regulations concerning reporting injury and illness through the OSHA logs.
If these two systems—a periodic comprehensive population-based survey of working conditions and a publicly available compilation of exposure measurement data—could be linked together through common variables (e.g., industry, occupation, location, organizational type and size, etc.) the nation would begin developing a clear picture of exposures and, thus, work-related chronic disease risk and needs for disease prevention activities.
Each of these approaches offers opportunities for improving worker health through providing more detailed data on workplace hazards and hazardous exposures. Taking several concrete steps toward implementation would begin building a powerful system of risk identification and reduction. OSHA could enhance its IMIS by working with other regulatory agencies which already collect, or require collection of, exposure measurements including MSHA, DOE, state OSHA programs, and OSHA consultation programs. Federal research efforts (e.g., NIOSH health hazard evaluations) could also be included. In addition, research entities, especially those with federal funding, could be required to contribute their data to this system.
A major limitation to the utility of current IMIS data is missing data elements. The database design and data-collection tools would need to be updated to take advantage of modern informatics systems, in order to automate the entry of those elements which can be automated and reduce the data-entry requirements for compliance officers or inspectors. Use of a tablet-based system, integrated with laboratory and inspection-level data, could greatly enhance the comprehensiveness and completeness of the measurement database.
In the long term, a potential source of data would be from those workplaces required to collect exposure measurements in compliance with specific OSHA regulations. OSHA would need to consider amending its regulations to require such employers to contribute these data to the system. Provided to OSHA these data could be organized and analyzed to the benefit of the reporting employers. Feedback to individual employers could be provided that interpret the results in relative terms (benchmarking) and absolute terms (e.g., presence or frequency of exposures > Action Level).
Furthermore, OSHA could explore the development of a more comprehensive inspection observation tool, which could be used to describe the presence of multiple hazards in a workplace, and the distribution of exposure to these key agents. A predefined list of hazards could be developed, and inspectors could quickly estimate the number of employees
potentially exposed, and the duration and frequency of exposure. This type of comprehensive data could be collected relatively easily through a tablet-based system. The burden on inspectors using such a tool would have to be addressed.
Conclusion: The elements of a comprehensive exposure surveillance system can be largely achieved by building upon the exposure-related self-reported data envisioned in the expanded HSOII. These data would provide population distributions of exposure to all types of hazards (chemical, physical, biological, ergonomic, physiological, psychosocial, and work organization), among workers in all types of working arrangements, and would include duration and frequency components of exposure. However, they would be limited in their ability to provide quantitative levels. The exposure intensity information could be integrated through the use of measurements, such as those collected in the OSHA IMIS. IMIS needs to be significantly enhanced to provide a more complete database of routinely collected measurement data.
Recommendation H: NIOSH, in consultation with OSHA, should place priority on developing a comprehensive approach for exposure surveillance. The objective should be to build systematically a comprehensive and continuously updated database of risks and exposures that provides the basis for estimating work-related acute and chronic health conditions for prevention.
In the near term:
- NIOSH should fully exploit the existing OSHA exposure databases by cleaning and integrating all available data sources to make them useful for surveillance purposes, taking proper account of the database limitations.
As an intermediate goal:
- NIOSH, in collaboration with OSHA and other agencies as appropriate, should construct an integrated exposure database to include the multiple sources of exposure measurement data already available, specifically MSHA’s MSIS, Department of Energy and Nuclear Regulatory Commission personal exposure data, and relevant data from others conducting research with federal funds.
In the long term:
- NIOSH should link the integrated exposure database with the comprehensive survey data obtained in the recommended expanded HSOII
- (Recommendation D) and new data from any characterization of exposures from targeted industry-specific assessments.
- NIOSH and OSHA should explore the feasibility of receiving employer-mandated exposure sample results after considering the reliability and quality of those measurements. The agencies should work with stakeholders to develop software and other tools and to facilitate establishment-level analysis of exposure data along with benchmarking.
This chapter has examined specific promising developments that, if strengthened and sustained, could significantly bolster occupational safety and health surveillance. The household survey has the potential to provide data about work from the perspective of the employees and thus supplement the current employer-based surveys. By including occupation and work-related information in electronic health records, meaningful links can be made between health and occupation with in-depth information on the health conditions. New methods for coding relevant to occupational health can expand the range of useful information by allowing the use of free-text documents as well as standardizing and assessing current efforts to extract and codify such information using emerging computational methods. Also pertinent are new employer-based electronic reporting initiatives that will provide greater data for tracking work-related illness and injury targeting and thereby enhancing prevention and treatment efforts. Worker’s compensation data and programs contribute valuable occupational safety and health information, and enhanced use of those systems could help move surveillance efforts forward. Additionally, a number of public health surveys could provide valuable work-related information if questions or modules were added. Efforts by NIOSH to strengthen the leveraging of existing systems is key to cost-effective means of enhancing occupational safety and health surveillance. Development of a more comprehensive system for periodic collection and analysis of exposure-related information would greatly enhance the nation’s ability to anticipate work-related health risks, especially for chronic diseases, which are less easily identified as being occupational in origin. Great strides could be made in this direction by improving the comprehensiveness and completeness of the exposure measurement databases already existing, and by incorporating exposure-related questions into a periodic population-based survey. Standardized, coordinated, and enhanced surveys, tools, and programs will allow occupational safety and health surveillance to become the fully functioning system that is needed to improve the health and well-being of workers. These promising developments will benefit from the research and new technologies discussed in Chapter 7.
AFL-CIO. 2016. Death on the job: The toll of neglect. A national and state-by-state profile of worker safety and health in the United States, April. Available online at https://aflcio.org/sites/default/files/2017-03/1647_DOTJ2016_0.pdf (accessed April 27, 2017).
Allen, A., L. Welch, K. Kirkland, D. Trout, and S. Baron. 2017. Development of a diabetes mellitus knowledge resource for clinical decision support assisting primary care physicians with work-related issues. Journal of Occupational and Environmental Medicine 59(11):e236-e239.
An, J. A., P. Braveman, M. Dekker, S. Egerter, and R. Grossman-Kahn. 2011. Work, Workplaces, and Health. Princeton, NJ: Robert Wood Johnson Foundation. Available online at http://www.rwjf.org/content/dam/farm/reports/issue_briefs/2011/rwjf70459 (accessed June 26, 2017).
Azaroff, L. S., C. Levenstein, and D. H. Wegman. 2002. Occupational injury and illness surveillance: Conceptual filters explain underreporting. American Journal of Public Health 92(9):1421-1429.
Baron, S., M. S. Filios, S. Marovich, D. Chase, and J. S. Ash. 2017. Recognition of the relationship between patients’ work and health: A qualitative evaluation of the need for clinical decision support (CDS) for worker health in five primary care practices. Journal of Occupational and Environmental Medicine 59(11):e245-e250.
Bertke, S. J., A. R. Meyers, S. J. Wurzelbacher, A. Measure, M. P. Lampl, and D. Robins. 2016. Comparison of methods for autocoding causation of injury narratives. Accident, Analysis, and Prevention 88:117-123.
BLS (Bureau of Labor Statistics). 2013. Support statement, survey of occupational injuries and illnesses, 2013. Available online at https://www.reginfo.gov/public/do/PRAViewDocument?ref_nbr=201511-1220-004 (accessed January 30, 2017).
BLS. 2014a. Occupational injury and illness classification manual. Available online at https://www.bls.gov/iif/oshoiics.htm (accessed June 1, 2017).
BLS, 2014b. Occupational Requirements Survey (ORS) consolidated FY 2014 feasibility tests summary report. Available online at https://www.ssa.gov/disabilityresearch/documents/fy14_test%20summary.pdf (accessed November 8, 2017).
BLS. 2017a. Standard Occupational Classification. Available online at https://www.bls.gov/soc (accessed June 1, 2017).
BLS. 2017b. Occupational Requirements Survey. Available online at https://www.bls.gov/ors/home.htm (accessed November 8, 2017).
Bonauto, D., B. Silverstein, D. Adams, and M. Foley. 2006. Prioritizing industries for occupational injury and illness prevention and research, Washington State Workers’ compensation claims, 1999-2003. Journal of Occupational and Environmental Medicine 48(8):840-851.
Bonauto, D. K., J. Z. Fan, T. W. Largo, K. D. Rosenman, M. K. Green, J. K. Walters, B. L. Materna, J. Flattery, T. St. Louis, L. Yu, S. Fang, L. K. Davis, D. J. Valiante, K. R. Cummings, J. J. Hellsten, and S. L. Prosperie. 2010. Proportion of workers who were work-injured and payment by workers’ compensation systems—10 states, 2007. Morbidity and Mortality Weekly Report 59(29):897-900.
Boyer, J., M. Galizzi, M. Cifuentes, A. d’Errico, R. Gore, L. Punnett, and C. Slatin. 2009. Promoting Healthy Safe Employment (PHASE) in Healthcare Team. Ergonomic and socioeconomic risk factors for hospital workers’ compensation injury claims. American Journal of Industrial Medicine 52(7):551-562.
CDC (Centers for Disease Control and Prevention). 2016a. Cardiovascular health status by occupational group—21 states, 2013. Morbidity and Mortality Weekly Report 65:794-798.
CDC. 2016b. Asthma among employed adults, by industry and occupation—21 states, 2013. Morbidity and Mortality Weekly Report 65:1325-1331.
CDC. 2017a. About OIICS. Available online at https://wwwn.cdc.gov/wisards/oiics/About.aspx (accessed June 1, 2017).
CDC. 2017b. What is PRAMS. Available online at https://www.cdc.gov/prams/index.htm (accessed June 12, 2017).
Cifuentes, M., J. Boyer, R. Gore, A. d’Errico, J. Tessler, P. Scollin, D. Lerner, D. Kriebel, L. Punnett, C. Slatin, and PHASE in Healthcare Research Team. 2007. Inter-method agreement between O*NET and survey measures of psychosocial exposure among healthcare industry employees. American Journal of Industrial Medicine 50(7):545-553.
Cifuentes, M., J. Boyer, D. A. Lombardi, and L. Punnett. 2010. Use of O*NET as a job exposure matrix: A literature review. American Journal of Industrial Medicine 53(9):898-914.
CSTE (Council of State and Territorial Epidemiologists). 2014a. 2013 National Assessment of Epidemiology Capacity: Findings and Recommendations. Available online at http://www.cste2.org/2013eca/CSTEEpidemiologyCapacityAssessment2014-final2.pdf (accessed November 8, 2017).
CSTE. 2014b. Inclusion of work information as data elements in CDC surveillance systems. Available online at http://c.ymcdn.com/sites/www.cste.org/resource/resmgr/2014PS/14_OH_02upd.pdf (accessed December 4, 2017).
CSTE. 2017. Occupational health indicators. Available online at http://www.cste.org/?OHIndicators (accessed June 8, 2017).
Das, R., J. McNary, K. Fitzsimmons, D. Dobraca, K. Cummings, J. Mohle-Boetani, C. Wheeler, A. McDowell, Y. Iossifova, R. Bailey, K. Kreiss, and B. Materna. 2012. Occupational coccidioidomycosis in California: Outbreak investigation, respirator recommendations, and surveillance findings. Journal of Occupational and Environmental Medicine 54(5):564-571.
d’Errico, A., L. Punnett, M. Cifuentes, J. Boyer, J. Tessler, R. Gore, P. Scollin, C. Slatin, and Promoting Healthy and Safe Employment in Healthcare Research Team. 2007. Hospital injury rates in relation to socioeconomic status and working conditions. Occupational and Environmental Medicine 64(5):325-333.
Eurofound (European Foundation for the Improvement of Living and Working Conditions). 2017a. Sixth European Working Conditions Survey: 2015. Available online at https://www.eurofound.europa.eu/surveys/european-working-conditions-surveys/sixth-european-working-conditions-survey-2015 (accessed June 1, 2017).
Eurofound. 2017b. Sixth European Working Conditions Survey: Technical report. Available online at https://www.eurofound.europa.eu/sites/default/files/ef_survey/field_ef_documents/6th_ewcs_-_technical_report.pdf (accesssed June 1, 2017).
Evanoff, B., A. Zeringue, A. Franzblau, and A. M. Dale. 2014. Using job-title-based physical exposures from O*NET in an epidemiological study of carpal tunnel syndrome. Human Factors 56(1):166-177.
Fan, Z. J., D. K. Bonauto, M. P. Foley, and B. A. Silverstein. 2006. Underreporting of work-related injury or illness to workers’ compensation: Individual and industry factors. Journal of Occupational and Environmental Medicine 48(9):914-922.
Filios, M. S., E. Storey, S. Baron, G. B. Luensman, and R. N. Shiffman. 2017. Enhancing worker health through clinical decision support (CDS): An introduction to a compilation. Journal of Occupational and Environmental Medicine 59(11):e227-e230.
Frazier, T. 1983. NIOSH Occupational health and hazard surveillance systems. Journal of Toxicology: Clinical Toxicology 21(1-2):201-209.
Gardner, B. T., D. A. Lombardi, A. M. Dale, A. Franzblau, and B. A. Evanoff. 2010. Reliability of job-title based physical work exposures for the upper extremity: Comparison to self-reported and observed exposure estimates. Occupational and Environmental Medicine 67(8):538-547.
Harber, P., C. A. Redlich, S. Hines, M. S. Filios, and E. Storey. 2017. Recommendations for a clinical decision support system for work-related asthma in primary care settings. Journal of Occupational and Environmental Medicine 59(11):e231-e235.
HealthIT. 2017. Meaningful use definition & objectives. Available online at https://www.healthit.gov/providers-professionals/meaningful-use-definition-objectives (accessed June 12, 2017).
HHS (Department of Health and Human Services). 2015a. 2015 Edition Health Information Technology (Health IT) Certification Criteria, 2015 Edition Base Electronic Health Record (EHR) Definition, and ONC Health IT Certification Program Modifications, 80 Federal Register 16804-16921, March 30, 2015.
HHS. 2015b. “2015 Edition Health Information Technology (Health IT) Certification Criteria, 2015 Edition Base Electronic Health Record (EHR) Definition, and ONC Health IT Certification Program Modifications; Final Rule”, 80 Federal Register 62602-62759, October 16, 2015.
Holloway-Beth, A., L. Forst, S. Freels, S. Brandt-Rauf, and L. Friedman. 2016. Occupational injury surveillance among law enforcement officers using workers’ compensation data, Illinois 1980 to 2008. Journal of Occupational and Environmental Medicine 58(6):594-600.
HSE (Health and Safety Executive). 2017. Data sources. Available online at http://www.hse.gov.uk/statistics/sources.htm (accessed June 12, 2017).
Hutchings, S., J. Cherrie, M. van Tongeren, and L. Rushton. 2012. Intervening to reduce the future burden of occupational cancer in Britain: What could work. Cancer Prevention Research 5:1213-1222.
IAIABC (International Association of Industrial Accident Boards and Commissions). 2017. EDI claims standards. Available online at https://www.iaiabc.org/iaiabc/EDI_Claims.asp (accessed December 21, 2017).
IOM (Institute of Medicine). 2011. Incorporating Occupational Information in Electronic Health Records: Letter Report. Washington, DC: The National Academies Press.
Joe, L., R. Roisman, S. Beckman, M. Jones, J. Beckman, M. Frederick, and R. Harrison. 2014. Using multiple data sets for public health tracking of work-related injuries and illnesses in California. American Journal of Industrial Medicine 57(10):1110-1119.
Kauppinen, T., J. Toikkanen, D. Pedersen, R. Young, W. Ahrens, P. Boffetta, J. Hansen, H. Krumout, J. Maqueda Blasco, D. Mirabelli, V. de la Orden-Rivera, B. Pannett, N. Plato, A. Savela, R. Vincent, and M. Kogevinas. 2000. Occupational exposure to carcinogens in the European Union. Occupational and Environmental Medicine 57:10-18.
Kauppinen, T., S. Uuksulainen, A. Saalo, I. Mäkinen, and E. Pukkala. 2014. Use of the Finnish Information System on Occupational Exposure (FINJEM) in epidemiologic, surveillance, and other applications. Annals of Occupational Hygiene 58(3):380-396.
Kica, J., and K. D. Rosenman. 2014. Multi-source surveillance system for work-related skull fractures. Journal of Safety Research 51:49-56.
LA DH (Louisiana Department of Health). 2016. Louisiana service worker wellness report, results from the Behavioral Risk Factor Surveillance System 2013 and 2014. Available online at http://dhh.louisiana.gov/assets/oph/Center-PHI/BRFSS/data/LA_Service_Worker_Wellness_Report_BRFSS_2016_FINAL.pdf (accessed June 12, 2017).
LaMontagne, A.D., A. Milner L. Krnjacki, M. Schlichthorst, A. Kavanagh, K. Page, and J, Pirkis. 2016. Psychosocial job quality, mental health, and subjective wellbeing: A cross-sectional analysis of the baseline wave of the Australian Longitudinal Study on Male Health. BMC Public Health 16(Suppl 3):1049. doi: 10.1186/s12889-016-3701-x.
Largo, T. W., and K. D. Rosenman. 2015. Surveillance of work-related amputations in Michigan using multiple data sources: Results for 2006-2012. Occupational and Environmental Medicine 72:171-176.
Lavoue, J., M. Friesen, and I. Burstyn. 2013a. Workplace measurements by the U.S. Occupational Safety and Health Administration since 1979: Descriptive analysis and potential uses for exposure assessment. Annals of Occupational Hygiene 57(1):77-97.
Lavoue, J., M. Friesen, and I. Burstyn. 2013b. Workplace measurements by the U.S. Occupational Safety and Health Administration since 1979: Descriptive analysis and potential uses for exposure assessment. Annals of Occupational Hygiene 57(5):681-683.
Lupo, P. J., E. Symanski, P. H. Langlois, C. C. Lawson, S. Malik, S. M. Gilboa, L. J. Lee, A. J. Agopian, T. A. Desrosiers, M. A. Waters, P. A. Romitti, A. Correa, G. M. Shaw, L. E. Mitchell, and the National Birth Defects Prevention Study. 2012. Maternal occupational exposure to polycyclic aromatic hydrocarbons and congenital heart defects among offspring in the National Birth Defects Prevention Study. Birth Defects Research Part A: Clinical and Molecular Teratology 94:875-881.
MA DPH (Massachusetts Department of Public Health). 2016. Putting data to work, 23 health indicators by occupation and industry: Findings from the Massachusetts Behavioral Risk Factor Surveillance System, 2012-2013. Available online at http://www.mass.gov/eohhs/docs/dph/occupational-health/full-report-health-indicators-15.pdf (accessed June 12, 2017).
MacKenzie, W. R., A. J. Davidsony, A. Wiesenthal, J. P. Engel, K. Turner, L. Conn, S. J. Becker, S. Moffatt, S. L. Groseclose, J. Jellison, J. Stinn, N. Y. Garrett, L. Helmus, B. Harmon, C. L. Richards, J. R. Lumpkin, and M. F. Iademarco. 2016. The promise of electronic case reporting. Public Health Reports 131(6):742-746.
Marcum, J., and D. Adams. 2017. Work-related musculoskeletal disorder surveillance using the Washington State workers’ compensation system: Recent declines and patterns by industry, 1999-2013. American Journal of Industrial Medicine 60(5):457-471.
Marucci-Wellman, H. R., T. K. Courtney, H. L. Corns, G. S. Sorock, B. S. Webster, R. Wasiak, Y. I. Noy, S. Matz, and T. B. Leamon. 2015. The direct cost burden of 13 years of disabling workplace injuries in the U.S. (1998-2010): Findings from the Liberty Mutual Workplace Safety Index. Journal of Safety Research 55:53-62.
Marucci-Wellman, H. R., H. L. Corns, and M. R. Lehto. 2017. Classifying injury narratives of large administrative data bases for surveillance. Accident, Analysis, and Prevention 98:359-371.
McLellan, R. K., N. S. Haas, R. P. Kownacki, G. S. Pransky, J. B. Talmage, and M. Dreger. 2017. Using electronic health records and clinical decision support to provide return-to-work guidance for primary care practitioners for patients with low back pain. Journal of Occupational and Environmental Medicine 59(11):e240-e244.
Measure, A. C. 2014. Automated coding of worker injury narratives. Available online at https://www.bls.gov/iif/measure_autocoding.pdf (accessed June 1, 2017).
Meyer, J. D., M. Cifuentes, and N. Warren. 2011. Association of self-rated physical health and incident hypertension with O*NET factors: Validation using a representative national survey. Journal of Occupational and Environmental Medicine 53(2):139-145.
Monaco, K. 2016. Presentation to Committee on Developing a Smarter National Surveillance System for Occupational Safety and Health in the 21st Century, the National Academies of Sciences, Engineering, and Medicine, June 15, 2016, Washington, DC.
Nanda, G., K. M. Grattan, M. T. Chu, L. K. Davis, and M. R. Lehto. 2016. Bayesian decision support for coding occupational injury data. Journal of Safety Research 57:71-82.
National Commission on State Workmen’s Compensation Laws. 1972. Report of the National Commission on State Workmen’s Compensation Laws. Available online at http://workerscompresources.com/wp-content/uploads/2012/11/Introduction-Summary.pdf (accessed June 8, 2017).
NCCI (National Council on Compensation Insurance). 2017. NCCI codes. Available online at https://classcodes.com/ncci-codes (accessed June 1, 2017).
NCHS (National Center for Health Statistics). 2016. NHIS supplements and co-sponsors. https://www.cdc.gov/nchs/nhis/data-questionnaires-documentation.htm (accessed June 1, 2017).
NCVHS (National Committee on Vital and Health Statistics). 2012. Development of Standards for the Collection of Socioeconomic Status in Health Surveys Conducted by the Department of Health and Human Services. Letter to Kathleen Sebelius, Secretary, HHS, from Justine M. Carr, Chairperson, National Committee on Vital and Health Statistics, Washington, DC. June 22, 2012. Available online at https://www.ncvhs.hhs.gov/wp-content/uploads/2014/05/120622lt.pdf (accessed November 8, 2017).
NH HSS (New Hampshire Department of Health and Human Services). 2015. Utilization of the NH Behavioral Risk Factor Surveillance System (BRFSS) to Better Understand Underreporting of Work-related Injuries. Division of Public Health Services. Available online at https://iod.unh.edu/sites/default/files/media/NHOHSP/Pubs/utilizationnhbehavioralrisk.pdf (accessed June 12, 2017).
NM DH (New Mexico Department of Health). 2014. Occupation and health in New Mexico. Available online at https://nmhealth.org/data/view/report/162/ (accessed June 12, 2017).
NIOSH. 2012. Guidelines for Reporting Occupation and Industry on Death Certificates. DHHS (NIOSH) Publication No. 2012-149. Available online at https://www.cdc.gov/niosh/docs/2012-149/pdfs/2012-149.pdf (accessed April 27, 2017).
NIOSH. 2013. Quality of worklife questionnaire. Available online at https://www.cdc.gov/niosh/topics/stress/qwlquest.html (accessed April 17, 2017).
NIOSH. 2016a. About NIOCCS. Available online at https://www.cdc.gov/niosh/topics/coding/overview.html (accessed April 17, 2017).
NIOSH. 2016b. How NIOCCS Works. Available online at https://www.cdc.gov/niosh/topics/coding/how.html (accessed April 17, 2017).
NIOSH. 2016c. NIOSH Center for Workers’ Compensation Studies. May. Available online at https://stacks.cdc.gov/view/cdc/39893 (accessed December 21, 2017).
NIOSH. 2017. Responses to questions from the National Academies committee about recommendations from the 1987 National Research Council report, January 24, 2017. Available through the National Academies Public Records Office.
NORC (National Opinion Research Center). 2016a. Survey design and questionnaire for a Household Survey on Occupational Injuries and Illnesses: Report on suitability of existing surveys and frames (Deliverable #5). Available online at https://www.bls.gov/iif/norc-report-on-the-suitability-of-existing-surveys-and-frames.pdf (accessed June 1, 2017).
NORC. 2016b. Survey design and questionnaire for a Household Survey on Occupational Injuries and Illnesses: Household survey design report (Deliverable #9). Available online at https://www.bls.gov/iif/norc-household-survey-design-report.pdf (accessed June 1, 2017).
NRC (National Research Council). 1987. Counting Injuries and Illnesses in the Workplace: Proposals for a Better System. Washington, DC: National Academy Press.
NRC and IOM (National Research Council and Institute of Medicine). 2001. Musculoskeletal Disorders and the Workplace: Low Back and Upper Extremities. Washington, DC: National Academy Press.
NY WCB (New York Workers’ Compensation Board). 2017. Description of manual classification code. Available online at http://www.wcb.ny.gov/content/ebiz/eclaims/ReqTables/ManualClassificationCodes_DN0059.pdf (accessed June 1, 2017).
O’Halloran, A. C., P. J. Lu, W. W. Williams, P. Schumacher, A. Sussell, J. Birdsey, W. L. Boal, M. H. Sweeney, S. E. Luckhaupt, C. L. Black, and T. A. Santibanez. 2017. Influenza vaccination among workers—21 U.S. states, 2013. American Journal of Infection Control 45:410.
O*NET. 2017. O*NET online. Available online at https://www.onetonline.org (accessed June 12, 2017).
OSHA. 2016. Improve tracking of workplace injuries and illnesses. Federal Register 81(92): 29624-29693.
OSHA. 2017a. Final rule issued to improve tracking of workplace injuries and illnesses. Available online at https://www.osha.gov/recordkeeping/finalrule (accessed June 12, 2017).
OSHA. 2017b. Establishments in the following industries with 20 to 249 employees must submit injury and illness summary (Form 300A) data to OSHA electronically. Available online at https://www.osha.gov/recordkeeping/NAICScodesforelectronicsubmission.html. (accessed February 20, 2017).
Prang, K.-H., B. Hassani-Mahmooei, and A. Collie. 2016. Compensation Research Database: Population-based injury data for surveillance, linkage and mining. BMC Research Notes 9:456.
Rosenman, K. D., J. C. Gardiner, J. Wang, J. Biddle, A. Hogan, M. J. Reilly, K. Roberts, and E. Welch. 2000. Why most workers with occupational repetitive trauma do not file for workers’ compensation. Journal of Occupational and Environmental Medicine 42(1):25-34.
Rushton, L., S. Hutchings, L. Fortunato, C. Young, G. S. Evans, T. Brown, R. Bevan, R. Slack, P. Holmes, S. Bagga, J. W. Cherrie, and M. van Tongeren. 2012. Occupational cancer burden in Great Britain. British Journal of Cancer 107:S3-S7.
Shockey, T. M., and A. G. Wheaton. 2017. Short sleep duration by occupation group—29 states, 2013-2014. Morbidity and Mortality Weekly Report 66:207-213.
Silverstein, B., E. Viikari-Juntura, and J. Kalat. 2002. Use of a prevention index to identify industries at high risk for work-related musculoskeletal disorders of the neck, back, and upper extremity in Washington State, 1990-1998. American Journal of Industrial Medicine 41(3):149-169.
Smith, C. K., and J. Williams. 2014. Work related injuries in Washington State’s trucking industry, by industry sector and occupation. Accident, Analysis, and Prevention 65:63-71.
Spector, J. T., J. Krenz, E. Rauser, and D. K. Bonauto. 2014. Heat-related illness in Washington State agriculture and forestry sectors. American Journal of Industrial Medicine 57(8):881-895.
Towle, M., R. Tolliver, A. G. Bui, A. Warner, and M. Van Dyke. 2015. Adding industry and occupation questions to the Behavioral Risk Factor Surveillance System: New opportunities in public health surveillance. Public Health Reports 130:153-160.
TIRES. 2017. TIRES prevents injuries in the trucking industry. Available online at http://www.lni.wa.gov/Safety/Research/Trucking (accessed June 8, 2017).
UCLA (University of California Los Angeles). 2016. New on AskCHIS! Health, insurance data by industry and occupation. Center for Health Policy Research News, September 14. Available online at http://healthpolicy.ucla.edu/newsroom/press-releases/pages/details.aspx?NewsID=254 (accessed June 1, 2017).
U.S. Census Bureau. 2017a. North American Industry Classification System. Available online at https://www.census.gov/eos/www/naics (accessed June 1, 2017).
U.S. Census Bureau. 2017b. Methodology. Available online at https://www.census.gov/people/io/methodology (accessed June 1, 2017).
UT HD (Utah Department of Health). 2015. Utah health status update: Industry and occupation impact on health. Available online at http://health.utah.gov/opha/publications/hsu/1509_IndOcc.pdf (accessed June 12, 2017).
Utterback, D. F., and T. M. Schnorr, eds. 2010. Use of Workers’ Compensation Data for Occupational Injury and Illness Prevention: Proceedings from September 2009 Workshop. Available online at https://www.cdc.gov/niosh/docs/2010-152/pdfs/2010-152.pdf (accessed June 8, 2017).
Utterback, D. F., and T. M. Schnorr, eds. 2013. Use of Workers’ Compensation Data for Occupational Safety and Health: Proceedings from June 2012 Workshop, DHHS/CDC/NIOSH, Publication No. 2013-147. Available online at https://www.cdc.gov/NIOSH/docs/2013-147/pdfs/2013-147.pdf (accessed June 8, 2017).
van Tongeren, M., A. S. Jimenez, S. J. Hutchings, L. MacCalman, L. Rushton, and J. W. Cherrie. 2012. Occupational cancer in Britain: Exposure assessment methodology. British Journal of Cancer 107:S18-S26.
VT DH (Vermont Department of Health). 2016. General worksite wellness—Data brief (4 of 4), 2014 Vermont Behavioral Risk Factor Surveillance System. Available online at http://www.healthvermont.gov/sites/default/files/documents/2016/12/data_brief_201611_worksite_genl.pdf (accessed June 12, 2017).
WA DLI (Washington State Department of Labor and Industries). 2016. BRFSS 2012-2014: Demographic and employment characteristics of cell phone only respondents. Available online at http://www.lni.wa.gov/safety/research/files/96_1_2016brfss.pdf (accessed June 12, 2017).
WCIO (Workers Compensation Insurance Organizations). 2017. Injury description table. Available online at https://www.wcio.org/Document%20Library/InjuryDescriptionTablePage.aspx (accessed June 1, 2017).
Weiss, N. S., S. P. Cooper, C. Socias, R. A. Weiss, and V. W. Chen. 2015. Coding of Central Cancer Registry industry and occupation information: The Texas and Louisiana experiences. Journal of Registry Management 42(3):103-110.
WHO (World Health Organization). 2016. Classifications. Available online at http://www.who.int/classifications/icd/en (accessed June 1, 2017).
Wilkinson, R., and M. Marmot. 2003. Social Determinants of Health: The Solid Facts, 2nd ed. Copenhagen: WHO. Available online at http://www.euro.who.int/__data/assets/pdf_file/0005/98438/e81384.pdf (accessed June 19, 2017).
Wurzelbacher, S. J., I. S. Al-Tarawneh, A. R. Meyers, P. T. Bushnell, M. P. Lampl, D. C. Robins, C. Y. Tseng, C. Wei, S. J. Bertke, J. A. Raudabaugh, T. M. Haviland, and T. M. Schnorr. 2016. Development of methods for using workers’ compensation data for surveillance and prevention of occupational injuries among state-insured private employers in Ohio. American Journal of Industrial Medicine 59(12):1087-1104.
Yamin, S. C., A. Bejan, D. L. Parker, M. Xi, and L. M. Brosseau. 2016. Analysis of workers’ compensation claims data for machine-related injuries in metal fabrication businesses. American Journal of Industrial Medicine 59(8):656-664.
Zilaout, H., J. Vlaanderen, R. Houba, and H. Kromhout. 2017. 15 years of monitoring occupational exposure to respirable dust and quartz within the European industrial minerals sector. International Journal of Hygiene and Environmental Health 220(5):810-819.
This page intentionally left blank.