Using State Agency Reports to Augment Ohio’s Agricultural Injury Surveillance Efforts

ABSTRACT

Background

Agriculture is a hazardous industry with undocumented injury events. Credible surveillance measures are critical for this industry, especially to guide injury prevention programs with targeted recommendations for specific commodity groups and populations. This multi-phase study explored the feasibility for two state agency databases, the Ohio Bureau of Workers’ Compensation (BWC) Program and the Emergency Medical Services Incident Reporting System (EMSIRS), to augment the state’s Bureau of Labor Statistics (BLS) annual reports.

Methods

BWC data described injury claims in agricultural workplaces from 1999 to 2008. State EMSIRS data described the types of medical emergencies for which EMS services were requested to Ohio farms in 2013–2014. Descriptive analyses were performed on each distinctive source.

Results

Over 14,000 BWC claims were analyzed, with primary nature of injury identified as sprains and strains of bodily extremities; falls were the most common cause of injury. The EMSIRS data provided 1,376 cases, where EMS services were requested to Ohio farms at injury onset. Some cases had possibility to be excluded in CFOI or employment claims data, with 24% patients 65 years and older and 6% children 13 years and younger. The primary cause of injury was falls, and the highest reported injury type was blunt trauma.

Conclusions

Both BWC and EMSIRS databases showed the potential to enhance Ohio’s agricultural surveillance data with viable information not found in previously used systems. Each agency database had its own merits to further clarify and quantify morbidity. When used together, these sources enrich surveillance statistics to describe Ohio’s agricultural injury incidents.

KEYWORDS:

• Agriculture, workers’ compensations
• emergency medical services
• injury surveillance
• occupational incidents

Introduction

While agriculture is cited as a hazardous industry for both fatal and nonfatal events, these incidents may be underrepresented due to the broad and diverse nature of the workplace.^1,2 Many federal surveillance systems have definitions and limitations that exclude agricultural workers. The Bureau of Labor Statistics excludes workers below age 16, unpaid family members, undocumented and seasonal workers, or self-employed laborers. The Occupational Safety and Health Administration (OSHA) exempts farms with fewer than 11 employees from filing injury reports, making it difficult to calculate agricultural injuries that occur on these OSHA exempted farms.³ Likewise, youth injuries are difficult to quantify based on the non-routine reporting mechanisms available to practitioners to submit data, the limited surveys available in the public domain to collect data, and access to occupational data for computing injury rates.⁴

At the last U.S. Census (2017) Ohio reported 77,805 farms, with 87.4% classified as family or individual operator, 6.1% in partnership, 4.3% owned as corporations, and 2.2% labeled as estate/trust, prison farm, and miscellaneous land use.⁵ Nearly fourteen million acres of land are used for agricultural purposes, with an average size of 179 acres per farm. Approximately 78% of Ohio’s farms range in size of 1–179 acres, indicating Ohio has many small farm operations that utilize family-based or small-sized labor forces.

Beyond the complexity of defining the workforce, additional surveillance challenges lie in the diverse nature of the industry. The U.S. Department of Agriculture (USDA) defines a farm as “any place from which $1,000 or more of agricultural products were produced and sold, or normally would have been sold, during the year.”⁶ Injury surveillance is challenging on these farm worksites especially when many are physically and functionally intertwined with the homesite. In addition, incidents can take place off the property, as in the case when farm machinery shares the roads when traveling between fields; and during agritourism events where visitors participate in farm activities. Furthermore, equipment injuries and fatalities may not be counted when they occur in settings beyond the typical workplace, for example, when tractors are used by hobby farmers at places of residence, or at campgrounds, parks, and other public spaces not deemed as farm-use.⁷

Data reporting systems vary too. Variations within numerator data (i.e. number of cases) and denominator data (i.e. the studied population) occur depending upon the data source. Documented inconsistencies in and between these reporting systems make comparison and interpretation of the data difficult, time-consuming, and therefore, costly.^1,4

Background

Multiple approaches are used within the US to quantify the burden of injury for agricultural populations. For example, in Northeast Texas, the regional trauma registry has been used in conjunction with census data and spatial analyses to identify geographic areas of high agricultural-related injury risk.⁸ In North Carolina, emergency medical data collected from hospital emergency departments, has been used to describe farm-related injury prevalence, severity, and the demographic information of those with farm-related injuries.^9,10

The twelve Centers for Agricultural Safety and Health represent a major NIOSH (National Institute for Occupational Safety and Health) effort to protect the health and safety of workers in the agriculture, forestry, and fishing (AgFF) sector. Each center conducts its own surveillance research within its defined region. Two examples of these efforts include a survey conducted by researchers at the Central States Center for Agricultural Safety and Health, who found livestock operators had higher odds of suffering an injury to their chest/trunk, finger, and leg/knee/hip compared to their counterparts in crops and mixed production; and full-time farmers had a 2.11 times higher odds of back injury.¹¹ Research staff at the National Children’s Center for Rural and Agricultural Health and Safety developed an interactive collection of agriculturally related news and media reports to provide states and the public an accessible agricultural injury database, specific for youth injuries.¹²

Other successful approaches to quantify morbidity and mortality data have utilized Workers Compensation databases.^13–17 Labor reports generated by compensation claims serve as viable sources of injury data and have been used within states and industry sectors to quantify agricultural occupational injury.

While such surveillance efforts are necessary and commendable, many farm-related injury incidents remain unreported.^1,7,18 Continuous work is needed across the nation to describe the occurrence of, and factors associated with, agricultural injuries.

Ohio agricultural injury surveillance in Ohio

Ohio is not a state included in the NIOSH regional Agricultural Center surveillance initiatives; any agricultural surveillance systems for Ohio are maintained locally. Prior to 2014, the Ohio Department of Health (ODH) was responsible for collecting, analyzing, and reporting Census of Fatal Occupational Injuries (CFOI) and Survey of Occupational Injuries and Illnesses (SOII) data to the Bureau of Labor Statistics (BLS). Since 2014, the Ohio Bureau of Workers Compensation (BWC) agency serves Ohio’s federal reporting obligation.¹⁹ In addition to CFOI reports, Ohio BWC is a sole source provider for state employers carrying workforce insurance.

In 1960, the Agricultural Safety and Health Program at The Ohio State University (OSU) launched the Farm Fatality and Injury Database of Ohio (FFIDO). Research staff continue the early efforts to oversee a state surveillance system by searching death certificates, hospital records, physician records, newspapers, Google alerts, department of transportation’s crash reports, and personal testimonies to capture injury incidents, both fatal and non-fatal.²⁰ Each case is independently reviewed to ensure it is an agriculturally related incident. Annually, FFIDO data are shared with the state’s BLS reporting agencies for inclusion in state-wide reports.

Despite the dedicated attempts, some fatality cases may continue to be missed, or not linked to production agriculture. Even more likely, non-fatal incidents and occupational illnesses are missed due to the non-reporting allowance for personal and/or small farm injuries. Understanding such omissive nuisances, a multi-phase study was launched as student-based feasibility projects to explore existing or new databases suitable to collect agricultural morbidity data that were not utilized or available prior to 2011. Conducted in different years, the purpose of each independent research project was to identify, code, analyze, and describe data from state agency databases to determine their potential to enhance agricultural injury and illness cases, with the goal to identify sources that contribute to the better understanding of morbidity and mortality in Ohio’s agricultural population.

Phase 1: bureau of workers’ compensation database

Ohio law mandates all employers carry insurance for their workers. Unless the operators are self-insured, the primary insurance source is Ohio’s Bureau of Workers’ Compensation (BWC) program. Four US states, including Ohio, have workers’ compensation programs that are sole source providers, meaning they do not allow for workplace compensation insurance through private companies or other competitive sources. Ohio employers must pay into a state-funded government program.²¹

The transition of BLS reporting obligations within our state agencies created opportunities for greater involvement for Ohio BWC to collect and report agricultural data. In June 2011, OSU’s safety researchers received approval to access de-identified injury claims submitted for compensation when the injured persons were performing agricultural-related work. These data were evaluated to assess injury type and location of occurrence for claims identified as occupational agricultural injuries.

Methods

Ohio BWC supplied all injury claim data reported for years 1999–2008. Data for the claims were derived from First Report of Injury forms and additional follow-up between claims representatives and injured workers.

Claims data were aggregated to remove personal identifiers in compliance with the Health Insurance Portability and Accountability Act (HIPAA). The data fields supplied to the research team included International Classification of Disease version 9 (ICD-9) codes for most limiting injury, worker age at injury, worker gender, and time lost from work. A systematic review of each claim was conducted to discern if the claim was deemed a valid, agricultural-related claim.

Analyses

All agricultural-related BWC claims were searched based on injury event coding. Using this method, all injuries involving agricultural production are captured, regardless of primary industry of the employer. Using the Barell Body Region by Nature of Injury Diagnosis Matrix,²² body region and nature of injury were established for 89.3% of total claims examined (n = 12,814). This method of injury classification from ICD-9 codes has been used by such agencies as the Centers for Disease Control and Prevention as well as other researchers.^23,24 Frequencies and percents were used to summarize the results.

Results

The initial BWC data query yielded 18,688 claims from the farm-related occupational groups. After screening for agricultural-validity, the size of the claim pool was decreased by 24%, yielding a total of 14,344 claims for the 10-year period. The average age of the injured worker was 35.2 years. As age increased, claim frequency decreased; 10% (n = 1,489) of the claims included workers ≤19 years of age. Male workers made up 75.6% of injuries. Of the 14,344 injury claims identified, nursery employees and drivers made up most injuries (24%) and their average cost per injury was $6,623. Other categories with higher ranked claims included poultry and egg producers and their drivers, florists and their drivers, and field crops and their drivers.

Table 1 illustrates the injury site as reported on the BWC claim. Extremities were the most common place of injury (63.5%), followed by other head, face and neck (11.1%), torso (9.4%), vertebral column (9.0%), system wide and late effects (5.4%), and traumatic brain injury (1.0%).

Table 1. Ohio Bureau of workers’ compensation reported injury type.

Injury Type	Number of Claims	% Total
Sprains & Strains	3,267	26.9%
Contusion/Superficial	2,675	22.0%
Open Wound	2,458	20.3%
Fracture	1,212	10.0%
Crush	175	1.4%
Burns	149	1.2%
Dislocation	143	1.2%
Internal	106	0.9%
Amputations	37	0.3%
Nerves	23	0.2%
Unspecified	4	0.0%
Blood Vessels	1	0.0%
Not Covered by Barell	1,888	15.6%
Total	12138	100%

Sprains, strains, and contusions in extremities were the most common type of injury. The proportion of medical-only versus lost-time [at work] claims also showed that most of the injuries sustained by agricultural workers were of minor severity. However, during any particular day included in the study data, an average of 192 workers were off the job because of the extent of their injury. A total of 37 deaths were reported.

Overall, the BWC claims data provided additional context to the types of injuries on Ohio farms, specific for injury type, claim cost, commodity group, and time needed before returning to work. These morbidity data provided opportunity to calculate annual injury rates and better target prevention programs for injury type and agricultural commodity group.

Phase 2: emergency medical services incident reporting system database

The Emergency Medical Services Incident Reporting System (EMSIRS) is maintained by the Ohio Department of Public Safety, Division of Emergency Medical Services (EMS) to collect statistics on operational runs of Ohio’s emergency medical, fire, and emergency transportation services. Ohio law mandates all EMS services be reported regardless of injury outcome or locale where services are rendered.

The first EMSIRS annual report was available in 2011. Beyond utilizing the system to improve EMS performance, the report provides descriptive information about incidents across the state, patient population, and benchmarks for patient care.²⁵ Of interest to OSU surveillance researchers studying agricultural injuries, “farm” was an identifiable variable within their database. The EMSIRS data dictionary defines a farm as a location that “includes farm buildings and land under cultivation,” and the definition excludes farmhouses and home premises that exist on the farm. Having access to such data provides an opportunity for researchers to explore the EMS service calls made to farms, regardless of treatment outcome. More importantly, this data has the potential to report injury and illness statistics for farm populations, regardless of employment status, including undocumented workers and children.

Methods

State-level EMS summary statistics were requested from the Ohio Department of Public Safety (ODPS) for years 2011–2014. This time period represents the first 4-year data was submitted to the database by local EMS personnel. As with any new system, reporting errors and missing data fields were numerous, resulting in ODPS statisticians recommending not to use data from years 2011–2012. The summary statistics for 20,132,014 were deemed usable and were de-identified for use in the feasibility study. Data fields selected from the EMSIRS definition handbook and supplied to the research team included EMS service location to “farm,” county, year, age, gender, cause of injury, injury type, provider impression of the health issue, transport service received, patient disposition, anatomic location of complaint, organ system complaint, factors affecting delivery of care, alcohol and drug indicators, transport destination, and destination determination. Injury type was categorized in EMSIRS as being blunt (nonpenetrating trauma due to physical impact), penetrating trauma, burns, other, or no injury present. EMSIRS statisticians provided a prepared report comparing EMS service location to “farm” compared to “all other locations.”

Analyses

Data from EMSIR for years 2013 and 2014 were combined and summarized using frequency and percent for each variable. Variables of interest were computed and reported by state and county level.

Results

Within the 2-year period, there were a total of 1,976,197 EMS calls reported to all locations in Ohio. Of this number, 0.07% (n = 1,376) were made to farms. The average patient age receiving EMS services was 47.8 (SD 23.5). Persons 65 years and older comprised 24% (n = 328) of the calls, while youth patients between 14 and 18 years of age comprised 10% (n = 136) and children 13 years and younger comprised 6% (n = 82). Patient genders were male (55%), female (36%), and unknown/unreported (9%). When EMS responded to a farm location, an injury was present 51% of the time.

The most common cause of farm-related injury was falls (23.6%), machinery incidents (6.3%), and being struck by a blunt or thrown object (5.8%). Injury type is quantified in Table 2, with the highest reported injury being blunt trauma. Types of “other injury” included chemical poisoning, electrocution, bites, excessive heat, fire and flame, accidental firearm injuries, self-inflicted firearm injuries, machinery accidents, mechanical suffocation, smoke inhalation, being struck by a blunt object, and non-traffic motor vehicle incidents. Additional injury descriptors available in this database include injury by anatomic location and patient disposition, which describes if patient was treated and released, transported to medical facility, or found dead at the scene. Of the total on-farm requests, 3.4% (n = 47) required a facility that had a specialty resource center to treat trauma, burns, or pediatric care. Proximity to medical facility was the primary reason for facility choice.

Table 2. Type of injury reported by Ohio Emergency Medical Services (EMS) requests, 2013–2014.

Injury Type	EMS Requests to All Locations	EMS Requests to Farms
Blunt	229,293 (11.60%)	513 (37.28%)
Penetrating	19,253 (0.97%)	43 (3.13%)
Burns	3,181 (0.16%)	15 (1.09%)
Other Injury Type	431,749 (21.85%)	146 (10.61%)
No Injury Present	1,292,721(65.42%)	659 (47.89%)
Total	1,976,197 (100%)	1,376 (100%)

From the prepared comparison data provided by agency statisticians, there were other notable data fields that could be explored in future studies. For instance, 4.5% (n = 62) EMS runs to farm locations experienced a delayed response due to distance compared to 1.4% (n = 27,646) of EMS runs on all other locations. EMS providers reported “unable to find victim” for 10% (n = 134) cases. Refusal of care by patients was reported in 7.0% cases (n = 96).

The EMSIRS data provided community-reported information from any air or ground medical units responding to an EMS call by a public or private EMS organization. These service calls responded to cases at injury onset on the farm, which included working and non-working populations. The data also included information about the medical services not reported elsewhere in the healthcare system, including incidents when the patient refused care or sought care out of the state.

Discussion

Prior to these initial studies, BWC and EMSIRS data had not been explored for their feasibility to augment agricultural surveillance efforts within Ohio. Through this multi-phase project, two state agencies were identified as potential sources to enhance current agricultural data collection efforts, including their limitations and benefits.

The limitations within each system included improper data entry and injury classification coding discrepancies; however, such limitations are not new and are prevalent in many surveillance systems.^1,7,26–28 Data entry competency, along with regular training, is recommended at all levels of the reporting and analysis process to ensure accurate surveillance outputs. Variability exists in how each agency inputs the data, including the omission of qualitative narrative about the injury incident, which oftentimes enriches the context for how the injury occurred. Currently, the EMSIRS “narrative” box on the EMS report form is an optional reporting data field. A higher-level limitation to data discrepancy is the non-reporting requirements of family and small farm labor forces, which minimizes a comprehensive record of agricultural claim data. This is a similar limitation to state-reported CFOI data.

The benefits of each system are their inherent ability to enrich the FFIDO database with cases not otherwise accounted for by family and small-farm exempted employers, who are not mandated to report injuries. Through this feasibility study, the information contained in these two state agency systems complemented the existing BLS fatality data (a source aimed to provide state-based fatality data) to classify cases in similar ICD-9 formats for research comparison and reporting purposes. The BWC and EMSIRS sources filled gaps in quantifying and describing morbidity statistics for the state of Ohio, which was missing from the BLS-generated data.

In comparison to other occupational surveillance data, agricultural cases are substantially smaller subsets of the workforce population. Yet in one of the most hazardous industries of the United States, understanding injury type and causation can lead to better administration of services, research, and safety programming. While Ohio’s agricultural and food production contributes $20 billion to the state’s economy,²⁹ agricultural injury statistics are not a focus of state agency personnel managing the BWC and EMSIRS databases. The feasibility project brought about an awareness with each respective agency for the scope and magnitude of agricultural-related injuries, which has led to additional collaborations between these state agencies and the OSU Agricultural Safety and Health program.

The primary benefit of this enhanced relationship allows for resource sharing and strengthens injury prevention efforts within the state. The state-agency reports added over 2,000 agricultural cases annually to the FFFIDO database, of which approximately 1,400 were BWC claims cases and an estimated 700 were EMS service calls. In the years following the feasibility studies, OSU researchers continued to request these agency data reports. The cost to add such reports is a relatively small burden for state agency staff who respond to public information requests. The OSU Ag Safety staff need approximately 10 calendar days to further review, and report agency reports into the FFIDO-generated reporting system. This is considered a minimal cost for the return on investment for having foundational knowledge of state-based surveillance data to guide recommendations for workplace practices to reduce the risk and occurrence of negative health outcomes for Ohio agriculturalists.

Conclusion

This feasibility study explored the BWC and EMSIRS databases to identify viable data to enhance agricultural injury surveillance in Ohio. Having informed systems to document injury events allows Ohio to come closer to the goals set in the NIOSH National Occupational Research Agenda (NORA) of 2018 to improve surveillance of injury and illness in the agricultural industry.³⁰ When used with BLS data reports, these additional agency reports improved the understanding for the prevalence, severity, and source of agricultural incidents. Ohio surveillance data are utilized by OSU program staff for news media, awareness programs, and occupational trainings; data are also reported publicly for others to utilize in their community, school, and agribusiness association meetings. Collectively, the data are used to guide outreach programs, identify effective intervention projects, initiate industry partnerships build collaborations with allied health organizations, seek external funding, and set future research agendas. The goal is to positively impact the health and safety environment of agricultural workplaces.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Department of Food, Agricultural, and Biological Engineering in the College of Food, Agricultural and Environmental Sciences, at The Ohio State University.