Workers’ compensation claim counts and rates by injury event/exposure among state-insured private employers in Ohio, 2007–2017
Introduction
Workers’ compensation (WC) in the United States involves large, state-governed, administrative data collection systems. WC claims data include coded fields for the injured worker’s occupation/industry, injury cause, part of body, and nature of injury, as well as unstructured narratives that describe how the incident occurred and the worker’s occupation. States use a variety of coding systems, and codes for each claim are manually generated by a combination of employers, claims administrators, insurers and/or state WC bureaus based on free text descriptions of how the incident occurred, diagnoses, and other claims information.
WC systems have been successfully utilized by several states for occupational sentinel surveillance, research, and employer worksite follow-up. Recent studies demonstrated that large state datasets of WC claims can be successfully linked to state employment data to examine claim counts and rates by industry and cause of injury (Anderson et al., 2013, Harrison et al., 2019, Massachusetts Department of Industrial Accidents et al., 2020, Michigan Department of Health and Human Services, 2020, Taylor et al., 2020, Wurzelbacher et al., 2016). Several of these studies were developed through cooperative agreements funded by the National Institute for Occupational Safety and Health (NIOSH) awarded to the states of California, Massachusetts, Michigan, Ohio, and Tennessee.
Although there are challenges with WC claim data quality and completeness, other studies have determined that machine learning techniques (Bertke et al., 2016, Lehto et al., 2009, Marucci-Wellman et al., 2011, Marucci-Wellman et al., 2015, Marucci-Wellman et al., 2017, Measure, 2014) can be successfully applied to the incident narratives in WC claims and other similar data to gather prevention insights. For example, Liberty Mutual uses such techniques to publish their popular Safety Index (Liberty Mutual Research Institute, 2020). The US Bureau of Labor Statistics (BLS) also uses these approaches to analyze data from the Survey of Occupational Injuries and Illnesses (SOII). Researchers from NIOSH and the Ohio Bureau of Workers’ Compensation (OHBWC) applied these techniques to identify ergonomic and safety priorities within many specific industries (Meyers et al., 2018). Machine learning approaches are constantly evolving, and NIOSH recently sponsored a crowd-sourcing competition to improve methods for coding free text narratives in WC claims and other similar data (NIOSH Blog, 2020). NIOSH and other partners have shared machine learning programs with various WC bureaus, insurers and employers, but these approaches are not yet in widespread use in WC systems. The current analyses fill a gap in the literature to demonstrate how the next iteration of machine learning methods can enable the coding of claims to a more detailed causation level in large WC datasets to encourage wider use.
The main purpose of this study was to analyze patterns in OHBWC-insured private employer WC claims to identify specific high-risk industries by detailed cause of injury. Another purpose was to summarize the leading examples of safety/health programs and interventions by cause types for higher-risk industries identified in the data. This demonstrated the opportunities for use and potential impact of the detailed analysis of claims data.
Section snippets
Methods and materials
The OHBWC is the largest state-run WC system in the United States. From 2007 to 2017, the OHBWC provided WC insurance for two-thirds of Ohio’s workers. The remaining one-third of workers were either employed by a small number of larger employers (usually >500 employees) authorized to self-insure or worked at other employers (e.g., sole proprietorships) exempt from OHBWC coverage. Employers can self-insure if they demonstrate the following: strong financial stability, an organizational plan for
Demographics
From 2007 to 2017, 140,780 accepted LT claims and 633,373 accepted MO claims (94.7% of the 817,103 accepted claims) for OHBWC-insured private employers could be reliably matched to UI data on NAICS and employee count and were included in subsequent event/exposure rate analyses. For a proportion of claims (11,098 claims, 1.3%), a reliable employee count could not be determined for the corresponding policy/year or quarter, and these claims are not included in reported analyses. For another
Machine learning comparisons
This study used an algorithm based on logistic regression (Bertke et al., 2016) to determine 1- and 2-digit OIICS event/exposures (up to forty-six categories). Overall, findings were consistent with a prior study (Meyers et al., 2018) that used an earlier algorithm based on Bayesian analyses to determine three broad categories of causation (Bertke et al., 2012). Additionally, more recent published data (National Institute for Occupational Safety and Health (NIOSH), 2020) indicated that shares
Limitations
As with other data sources, there are several limitations associated with the use of WC claims data for OSH surveillance. This includes underreporting, which differs by industry, especially for illnesses (Azaroff et al., 2002, Azaroff et al., 2013, Biddle et al., 1998, Fan et al., 2006, Lipscomb et al., 2009, Rosenman et al., 2000, Scherzer and Wolfe, 2008, Sears et al., 2013, Shannon and Lowe, 2002). There are also still a limited (although growing) number of comparable datasets because there
Conclusions
Although counts and rates of WC LT claims declined for all injury types among OHBWC-insured private employers from 2007 to 2017, the relative ranking of leading injury event/exposures remained largely unchanged. The majority of claims were due to three main event/exposure types: Overexertion and Bodily Reaction; Falls, Slips, and Trips; and Contact with Objects and Equipment. The other event/exposures in decreasing order of LT frequency include Transportation Incidents; Exposure to Harmful
Practical applications
Employers can use these data to benchmark their safety and health performance against industry peers and develop data-driven plans for prevention. OSH practitioners and researchers can also use these data to understand industry differences in the level and mix of risks, as well as industry trends, and to tailor safety, health, and disability prevention services and research. WC bureaus, regulators, insurers, and employers can use the open source machine learning algorithms and methods to code
Authors’ contributions
- •
SJW, ARM, MPL, PTB, SJB, and DCR contributed to study conception and design.
- •
DCR, MPL, CYT, SJW, and ARM contributed to data acquisition.
- •
SJW, CYT, DCR, SJB, ARM, MPL, and SJN contributed to data analysis.
- •
SJW, ARM, PTB, MPL, SJB, DCR, CYT, and SJN contributed to data interpretation.
- •
SJW, SJB, ARM, CYT, DCR, and MPL contributed to methods development.
- •
CYT, DCR, SJW, SJB, and ARM were key contributors for data management.
- •
SJW and ARM coordinated the manual coding of claims for the training and quality
Funding
- •
This research was supported by intramural National Institute for Occupational Safety and Health (NIOSH) funds. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Institution and Ethics approval and informed consent
- •
This work was performed at both the National Institute for Occupational Safety and Health (NIOSH) and the Ohio Bureau of Workers’ Compensation. This study was internally reviewed by NIOSH and it was determined that it did not constitute human subjects research. Rather, the study involved the analysis of coded and previously-collected WC administrative claims data.
Disclaimer
- •
The findings and conclusions in this report are those of the authors and do not necessarily represent the official positions of the National Institute for Occupational Safety and Health nor the Ohio Bureau of Workers’ Compensation.
Research data for this article
- •
To prevent the identification of individual workers or employers, claims-level data cannot be shared. Extensive aggregate claims data are shared in this article.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The authors would like to acknowledge the following for their contributions: Ibraheem S. Al‐Tarawneh, Jennifer L. Bell, Xiangyi Duan, Jean Geiman, Denise Giglio, Edward F. Krieg Jr., Nhut Van Nguyen, Jill A. Raudabaugh, Lisa M. Thomas, and Shelby Zuchowski.
Steve Wurzelbacher directs the Center for Workers’ Compensation Studies at the National Institute for Occupational Safety and Health. In this role, he coordinates workers’ compensation claim analyses, exposure assessment research, safety/health intervention effectiveness studies, and health services research with public and private sector partners. Steve has worked in the safety/health field since 1998, as both a researcher and a practitioner. Steve earned a PhD in Occupational Safety and
References (147)
- et al.
Evaluation of an in-vehicle monitoring system (IVMS) to reduce risky driving behaviors in commercial drivers: Comparison of in-cab warning lights and supervisory coaching with videos of driving behavior
Journal of Safety Research
(2017) - et al.
Development and evaluation of a Naïve Bayesian model for coding causation of workers’ compensation claims
Journal of Safety Research
(2012) - et al.
Comparison of methods for auto-coding causation of injury narratives
Accident Analysis & Prevention
(2016) - et al.
Use of a video monitoring approach to reduce at-risk driving behaviors in commercial vehicle operations
Transportation Research Part F: Traffic Psychology and Behaviour
(2011) - et al.
On-board safety monitoring systems for driving: Review, knowledge gaps, and framework
Journal of Safety Research
(2012) - et al.
Use of mechanical patient lifts decreased musculoskeletal symptoms and injuries among health care employees
Injury Prevention
(2004) Effectiveness of interventions to prevent work related eye injuries
American Journal of Preventive Medicine
(2000)- et al.
Evaluation of the effectiveness of occupational injury prevention programs at the company level
Safety Science
(2013) - et al.
Classifying injury narratives of large administrative databases for surveillance—A practical approach combining machine learning ensembles and human review
Accident Analysis & Prevention
(2017) - et al.
A practical tool for public health surveillance: Semi-automated coding of short injury narratives from large administrative databases using Naïve Bayes algorithms
Accident Analysis & Prevention
(2015)
Workers’ compensation insurer risk control systems: Opportunities for public health collaborations
Journal of Safety Research
Comparative analyses of workers' compensation claims of injury among temporary and permanent employed workers in Ohio
American Journal of Industrial Medicine
Systematic literature review on the effects of occupational safety and health (OSH) interventions at the workplace
Scandinavian Journal of Work, Environment & Health
Barriers to use of employees’ compensation for patient care at Massachusetts community health centers
Health Services Research
Occupational injury and illness surveillance: Conceptual filters explain underreporting
American Journal of Public Health
Standardizing industrial hygiene data collection forms used by workers' compensation insurers
Journal of Occupational and Environmental Hygiene
Effectiveness of a no-cost-to-workers, slip-resistant footwear program for reducing slipping-related injuries in food service workers: A cluster randomized trial
Scandinavian Journal of Work, Environment & Health
Evaluation of a comprehensive slip, trip and fall prevention program for hospital employees
Ergonomics
What percentage of employees with work-related illnesses receive employees’ compensation benefits?
Journal of Occupational and Environmental Medicine
Occupational heat illness in Washington State, 1995–2005
American Journal of Industrial Medicine
Gloves and dermal exposure to chemicals: Proposals for evaluating workplace effectiveness
Annals of Occupational Hygiene
An evaluation of a “best practices” musculoskeletal injury prevention program in nursing homes
Injury Prevention
Underreporting of work-related injury or illness to employees’ compensation: Individual and industry factors
Journal of Occupational and Environmental Medicine
Factors underlying observed injury rate differences between temporary workers and permanent peers
American Journal of Industrial Medicine
Effectiveness of community health workers for promoting use of safety eyewear by Latino farm workers
American Journal of Industrial Medicine
Development and evaluation of an exposure control efficacy library (ECEL)
Annals of Occupational Hygiene
The effect of ergonomic interventions in healthcare facilities on musculoskeletal disorders
American Journal of Industrial Medicine
Risk factors for injury among veterinarians
Epidemiology
Injured workers’ underreporting in the health care industry: An analysis using quantitative, qualitative, and observational data
Industrial Relations: A Journal of Economy and Society
The epidemic of violence against healthcare workers
Journal of Occupational and Environmental Medicine
Workplace violence in healthcare settings: Risk factors and protective strategies
Rehabilitation Nurse
Workplace violence injury in 106 US hospitals participating in the Occupational Health Safety Network (OHSN), 2012–2015
American Journal of Industrial Medicine
California workers’ compensation surveillance
Cited by (8)
Construction industry workers’ compensation injury claims due to slips, trips, and falls – Ohio, 2010–2017
2023, Journal of Safety ResearchOn clustering levels of a hierarchical categorical risk factor
2024, Annals of Actuarial ScienceOccupational injuries and illnesses among law enforcement officers, 2001−2019: Findings from the Ohio Bureau of Workers' Compensation
2023, American Journal of Industrial MedicineTemporal trends in occupational injuries treated in US emergency departments, 2012–2019
2023, Injury EpidemiologyWorkers' Compensation Costs for Occupational Hearing Loss Claims in the United States, 2009-2013
2023, Seminars in Hearing
Steve Wurzelbacher directs the Center for Workers’ Compensation Studies at the National Institute for Occupational Safety and Health. In this role, he coordinates workers’ compensation claim analyses, exposure assessment research, safety/health intervention effectiveness studies, and health services research with public and private sector partners. Steve has worked in the safety/health field since 1998, as both a researcher and a practitioner. Steve earned a PhD in Occupational Safety and Ergonomics from the University of Cincinnati, a BS in Chemical Science from Xavier University, is a Certified Professional Ergonomist (CPE), and holds the Associate in Risk Management (ARM) designation.
Alysha R. Meyers, PhD, CPE is an epidemiologist and ergonomist dedicated to promoting musculoskeletal health. Since 2010, Dr. Meyers has been an epidemiologist at NIOSH, working in the NIOSH Center for Workers’ Compensation Studies since its inception in 2013. Her research interests include using workers’ compensation data for occupational safety and health, preventing work-related musculoskeletal disorders, and Total Worker Health®. As a Principle Investigator at NIOSH, she is responsible for overseeing, implementing and evaluating occupational health epidemiologic studies. Dr. Meyers earned a PhD in Occupational and Environmental Health from the University of Iowa.
Michael Lampl is the Director of Research & Training at the Ohio Bureau of Workers’ Compensation (BWC), Division of Safety and Hygiene. Mike’s BS is in Industrial & Systems Engineering from the Ohio State University and his MS is in Occupational Health from the Medical College of Ohio. He is a Certified Professional Ergonomist. Mike has worked in the occupational safety and health field both in private industry and at BWC since 1993.
Tim Bushnell is an economist in the Economic Research and Support Office at the National Institute for Occupational Safety and Health. He also serves as economist in the NIOSH Center for Workers’ Compensation Studies, and as a coordinator for the Healthy Work Design and Well-being program. His research interests include costs of work-related illness and injury, industry differences in illness and injury rates, and impacts of work schedules, prevention measures, and work arrangements. He holds a PhD in economics from Michigan State University, specializing in industrial organization, and a master’s in public administration from the Harvard Kennedy School of Government.
Stephen Bertke is a Mathematical Statistician at the National Institute for Occupational Safety and Health. He assists with many projects within NIOSH analyzing data related to workplace safety and health. In particular, he is the lead statistician within the Center for Workers’ Compensation Studies where he advices on methods and study designs of workers’ compensation studies. He holds a PhD in mathematics with a concentration in statistics from University of Cincinnati.
David Robins has been a Management Analyst for the Ohio Bureau of Workers’ Compensation Division of Safety and Hygiene since 2004 and has been with the Bureau since 1990. David gained an Associates of Applied Science in Microcomputing Technology in 2001 and has years of experience in developing data reports, SQL queries, and data report applications. David has worked on innumerable projects to coalesce multiple data-sources to analyze and summarize large data sets.
Chih-Yu Tseng has worked for the National Institute for Occupational Safety and Health (NIOSH) since 1998 and has been involved in Center for Workers’ Compensation Studies (CWCS) projects since 2013. She became a full-time member of the CWCS team in 2015. She earned her MS degree in Statistics from The Ohio State University. Prior to joining NIOSH, she was a Research Data Analyst at the University of Cincinnati, Department of Environmental Health, Epidemiology and Biostatistics Division. As an IT Specialist, Chih-Yu provides programming, database and data visualization support to research projects.
Steven Naber is the Analytics Manager with the Ohio Bureau of Workers’ Compensation (BWC) Division of Safety and Hygiene. In his role, he oversees analysis of workers’ compensation policy and claims data for the purposes of assisting BWC in identifying and applying resources to employers and for occupational safety research. Steve earned a PhD in Mathematical Statistics from The Ohio State University with a specialty in geostatistics and has spent his career as a statistical consultant for federal and state governmental agencies, universities, and private industries before taking his current position at the Ohio BWC.