The Centers for Disease Control and Prevention estimate that job-related eye injuries occur frequently, with about 800 000 cases reported annually, of which 70% are ocular traumas due to objects contacting the eye.1 Ocular injuries can be serious in nature, and have been recognised to be a significant source of workplace trauma. According to the National Institute for Occupational Safety and Health, approximately 2000 workers in the USA seek treatment for work-related eye injuries daily, and nearly 100 of these cases result in missing one or more days of work.2 It is estimated that as many as 50 000 eye injury cases result in impaired vision annually.3 Unfortunately the prevalence of work-related eye injuries remains significant, even though it has been estimated that 90% of all such injuries are preventable with proper use of safety equipment.45

The majority of job-related ocular injury studies have used hospital or emergency department records; evaluated particular industries such as construction and agriculture, or specific occupations such as carpentry and welding; or relied on individual case reports.6^–17 While these investigations have and continue to contribute valuable information to assess occupational eye injury risk, there are some inherent limitations to these methods. For example, emergency department records are restricted to individual treatment centres and thereby do not provide a large enough sample to make inferences with respect to employee populations in general. Hospitals often do not treat injury cases except for the most serious incidents, and thus data on less serious incidents are not collected. Both industry-based and occupational-specific studies are valuable for examining their specific areas, but are unable to be directly compared to other industries and occupations, and thus make the prioritisation of preventive measures difficult to assess.

An alternative method of evaluating occupational eye injuries that has proven highly valuable has been the analysis of workers’ compensation claims. The use of state level workers’ compensation claim data allows for the examination of population-based injury information that often includes such important variables as claimant demographics, cost and indemnification information, injury cause and nature data, and allows for the comparison and analysis of risk at the occupational and industrial levels.18 Statewide workers’ compensation analysis has been previously utilised for assessing ocular injury claims from Rhode Island and West Virginia.1819 This study contributes to the literature on work-related eye injuries by examining workers’ compensation claims in Kentucky for a 10-year period between 1994 and 2003. Kentucky’s workers’ compensation programme is especially useful as it covers an estimated 1.7 million employees and close to 80 000 employers, thereby providing a large sample by which risk factors and areas most in need of prevention interventions can be identified. Kentucky is also particularly useful because all injuries in which one or more days off are taken as the result of workplace injuries are required to be reported. Indemnity or lump-sum payments are made for all workplace injuries when 7 or more days of work are lost; in cases where the injured person takes 14 or more days of work off for recovery, lost wage compensation is made retrospectively from the first day of lost work. Of note, legislation was also passed in the state in the year 2000, imposing significantly higher awards for workplace injuries. These penalties included doubling the amount of monetary fines for violations of Kentucky safety regulations and/or OSHA violations. Caps that were previously in place for money that could be paid to plaintiffs showing injury due to workplace violations were also raised. Thus this analysis is able to assess injuries and costs for both the pre- and post-legislative intervention period. For these reasons, the Kentucky data is both sensitive at detecting eye injuries, even in cases where the severity of injury was minor, and allows for the ability to assess potential effects arising from safety legislation.

METHODS

Data characteristics

This study used workers’ compensation claim data on eye injuries provided by Kentucky’s Office of Workers’ claims for the period 1994–2003. The dataset included information claimant age, gender, occupation, industry, body part affected, lump-sum and indemnity payments. The claims costs were tracked through 2004 and represent the accumulated claim costs during this time. Costs are rounded to the nearest dollar. By the end of the 10-year period, over 99% of all accepted claims examined were closed and the cost data complete. Eye injuries, as well as their cause and nature, were identified using the International Association of Industrial Accident Board and Commissions coding system. Only claims that reported the primary body part injured as eye (code 14) were included in this investigation.

Denominator estimation

While Kentucky’s workers’ compensation data itself contains no information on employment levels from which claim rates by year could be derived, yearly employment levels for Kentucky workers who were not self-employed were obtained from the Bureau of Labor Statistics Quarterly Census of Employment and Wages (QCEW) for the years 1994–2003. The QCEW contains information on all employees for whom employers must pay unemployment insurance tax. Therefore, the QCEW allows the population of all workers (except those that are self-employed who are not eligible for workers’ compensation benefits) to be derived. However, the QCEW contains no demographic data such as the worker’s age or gender, so estimates for specific claimant demographics, industrial and occupational level analyses had to be estimated.

To estimate employment levels for the aforementioned groups, a random sample of Kentucky workers was obtained using the US Bureau of the Census’ Current Population Survey (CPS) monthly outgoing rotation group (CPS–MORG), which is a monthly survey of approximately 50 000 households used by the US government to primarily assess monthly unemployment rates. All monthly outgoing rotation groups for the years 1994–2003 are used. The sample was restricted to individuals who reported residing in Kentucky at the time of the interview. Data from self-employed workers was excluded as they are not reported in Kentucky’s workers’ compensation claim reports. The final CPS–MORG sample used in this study contains 18 669 Kentucky workers. Sampling weights contained in the CPS–MORG data were then used to calculate employment estimates for different demographic groups.

Rate computation and risk analysis

Injury rates per 10 000 workers for Kentucky during the 1994–2003 period were derived using the formula R = 10 000×(I/N), where I is the number of eye injury claims over the 1994–2003 period and N equals the number of Kentucky worker-years for all workers in Kentucky obtained from the QCEW. Injury rates by year were calculated using the formula R_y = 10 000×(I_y/N_y), where I_y is the number of eye injury claims that occurred in a given year y and N_y equals the number of Kentucky employees who worked in a given year y, y = 1994, …, 2003. Since the yearly injury rates involved no estimation there were no associated confidence intervals.

When eye injury rates per 10 000 for the j-th demographic group of workers were estimated, the formula Rj = 10 000×(I_j/Np_j) was used, where I_j was the number of claims over the 1994–2003 period for the j-th group of workers, N equalled the number of Kentucky worker-years for all employees in Kentucky obtained from the QCEW, and p_j was the estimated fraction of workers in group j that was derived from CPS–MORG. Ninety-five per cent CI estimates were based on standard error estimates of R_j, which was calculated by applying the delta method.719 Logistic regression was used to assess the risk of factors associated with the eye claims that resulted in positive indemnification or lump-sum payments. All statistical analyses were performed using Stata V.10.0 software.

RESULTS

Demographic analysis

From 1994 to 2003, there were 10 545 ocular injury workers’ compensation claims filed in the state of Kentucky. During this period, the highest number of claims was 1342 for 1994; the lowest was 617 for 2003. Because unemployment insurance data was able to provide exact denominator information, exact rates were able to be calculated. Over all years, the eye injury claim rate averaged 6.29 per 10 000 workers, with the highest rate of 8.68 per 10 000 found for 1994, varying between the range of 6.80 (found in 1997) to 8.68 until 1999, and then dropping sharply after 1999, from 4.97 in the year 2000 to the lowest rate of 3.60 per 10 000 workers in 2003. Figure 1 presents the workers’ compensation eye injury claim rate for all years. A comparison of workers’ compensation rates for all injuries in Kentucky showed a similar pattern, with 1994 having a rate of 326.71 claims per 10 000 workers, and varying within a narrow range between 1994 and 1999, until a sharp and steady drop in rates occurred in 2000 from 227.89 to 187.78 per 10 000 workers in 2003.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Kentucky eye injury workers’ compensation claims 1994–2003; claim rates per 10 000 workers by year.

A demographic analysis revealed that the majority of claims (8361, 79.3%) were filed by men; 1914 (20.7%) were filed by women. Analysis by age found that workers aged 25–39 years were the largest group of claimants, having filed 5204 (49.4%) claims, followed by workers aged 40–54 years, with 2479 (23.5%) claims. Using the CPS, the claim rate was estimated to be 10.01 per 10 000 workers for men (95% CI 9.87 to 10.15) and 2.38 for women (95% CI 2.34 to 2.41). The claim rates estimated for workers by age found that those aged 25–39 years had the highest rate of 12.30 (95% CI 12.00 to 12.60) per 10 000 workers, followed by 15–24 year-olds with 7.34 (95% CI 7.13 to 7.55), 40–54 year-olds with 5.51 (95% CI 5.38 to 5.65), 55–64 year olds with 1.09 (95% CI 1.06 to 1.12), and those aged ⩾65 years with 0.43 (95% CI 0.41 to 0.46).

Claim costs and injury characteristics

Of the 10 545 claims, the average cost per claim (rounded up to the nearest dollar) was approximately $331 (SD $3428), representing a total cost of slightly over $3 480 000 over the 10-year period examined. It should be noted that the median total claim cost was $0, as the majority of claims engendered no costs. Of all claims, 2.3% resulted in either indemnity or lump-sum payment. Of the claims requiring compensatory remediation, 199 were paid as a lump-sum settlement and averaged $15 332 (SD $17 976) per claim, while 53 were adjudicated indemnification costs averaging $8248 (SD $11 380) per claim. The claim amounts paid to men averaged $381 per claim and those to women averaged $106 per claim. The highest total cost by age group was for workers aged 55–64 years, averaging $606 per claim. By year, the lowest average claim amount of $193 occurred in 1999; the highest average claim amount of $654 was recorded for 2001. When the claim costs were analysed by pre- and post-legislation years, it was found that although the claim rate dropped significantly post-legislation, when claims engendered payment they became more expensive on average, with the post-legislation claims that engendered payment averaging $16 500, and the pre-legislation claims garnering payment averaging $12 128 per claim. Figure 2 provides an annual breakdown and comparison of average claim costs by total claim amount, average lump-sum amounts, and indemnity amounts by year.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Kentucky eye injury workers’ compensation claims 1994–2003; breakdown of claim cost averages by year.

The causes and natures of injury were examined. The most frequent cause of eye injuries were “foreign body in eye”, with 4920 (44.7%) claims. The highest total claim amounts were associated with “fall or slip from liquid or grease”, averaging $3577 per claim. Table 1 presents a detailed analysis of the causes of injuries by frequency, percentage, and lump-sum, indemnity, and total cost claim averages. The most common natures of eye injury were “foreign body” with 5236 (49.7%) claims, “burns” with 2135 (20.3%) claims, and “laceration” with 1764 claims. One particular nature of injury of interest in this study was “vision loss” as it directly implies a direct consequence of injury to the eye. Of all claimants, 63 (0.6%) cited vision loss as their nature of injury, of which 46 (73%) were male and 17 (27%) were female.

Industry and occupation analysis

When the claims were analysed by claimant major occupation, the highest number was found to be “precision, production, and craft” workers comprising 2763 (26.2%) of the total claims. Using the CPS to estimate claim rates by 10 000 for major occupations, it was found that “handlers, helpers, and labourers” had the highest rate of 13.46 (95% CI 12.86 to 14.07). The greatest average cost per claim of $767 was for workers in “transportation and moving equipment” occupations. However, because the majority of claims did not result in a payment, the frequency and means of costs were also calculated using only claims which resulted in positive amounts; the occupation with the largest number of claims with positive costs was “precision, production and craft”, with 103 (3.7%) of the 2763 claims having positive costs (mean cost $14 967).

Analysis by claimant major industry found that the highest number of claims from all worker industries was filed by those in the manufacturing/durable goods industry, with 2475 (23.5%) claims. The construction industry had the highest estimated claim rate of 19.95 (95% CI 18.73 to 21.17) claims. The most expensive average total claim cost was found for the mining industry with an average cost of $1149 per claim. Among all industries, construction had the highest frequency of claims, with 32 (7.2%) of their 443 claims engendering costs. Table 2 provides the full cost data for major occupation and industry.

Risk analysis of positive claim costs

To examine what demographic and work-related attributes increased the odds of an eye injury claim resulting in a positive workers’ compensation cost, controlling for other potentially confounding factors, a multivariate logistic regression model was estimated. This is especially important since claims incurring costs necessarily mean claimants needed a week or more for recovery. Claims by women were significantly less likely than men to have positive claim amounts (odds ratio (OR) 0.52; 95% CI 0.32 to 0.85; p = 0.009). The logistic regression also determined the ORs of positive claim amounts for year, industry, occupation and cause (see table 3).

DISCUSSION

The initial finding that over 10 000 incidents of occupational eye injuries occurred in Kentucky over a 10-year period shows that ocular trauma in the workplace remains a serious problem. While the per-worker rate of claims was substantially lower than other statewide workers’ compensation analyses reported previously in West Virginia and Rhode Island, such variation is not uncommon given the differing injury reporting requirements between states.1718 Of note, the average rate of 6.29 claims per 10 000 workers, with payments indicating that 2.3% of all eye injuries required a week or more recovery time, highlights the significance of the problem. Moreover, it should be emphasised that while the majority of claims did not result in lost-time payment, because the eye is an especially vulnerable organ, what may determine whether an injurious event is minor or a highly serious event may simply be predicated on a small variance in the nature of the event itself. For this reason, even those injuries that did not engender substantial costs or lost time should be viewed with the same degree of seriousness as those requiring significant medical attention and recovery time. The fact that in 63 claims some degree of vision loss was reported, effectively makes this point.

Some of the results were consistent with those of other eye injury investigations. For example, the finding that men have over three times the eye injury rate as women has been regularly reported by studies employing both workers’ compensation analysis and other methods of surveillance.171820^–23 In this study, claims by men were also found to be more expensive and more likely to have associated positive costs than claims by women. While on the one hand the higher incidence rate among men may be attributed to a greater numbers of men holding jobs in which eye injuries are more likely, the logistic regression, having controlled for claimant occupation and industry, did not support this being the reason why the injuries were more severe between the genders. It has been postulated that such gender differences were attributable to variations in work exposures in the past, and thus using other methods to confirm if this were the case would be valuable in the future.24

One positive discovery was a significant decline in the trend of eye injuries beginning in the year 2000. In this year, Kentucky passed a comprehensive piece of legislation, House Bill 992, that doubled financial penalties against employers for safety violations shown to result in workplace injury, raised the cap on the amount claimants could collect, and augmented indemnification benefits by taking employee age and education level into account. It is important to note that beginning in the year 2000, when an increase in penalties for safety violations resulting in injury were enacted, there was concurrently a significant decline in the rate of eye injuries that resulted in costs, while the average cost of injuries with positive costs was higher, which is consistent with the enforcement of higher penalty and monetary cap amounts.

Additionally, there was a marked decline in indemnity payments relative to an increased proportion of lump-sum payments, hence indicating that most of the injury claims involving lost work time were being settled as opposed to adjudicated. Of the cases settled, two claims of particularly high outlying value, averaging about $100 000 each, were settled in the year 2001, and thus account for some of the particular jump in average payment for that year. These findings are consistent with many previous workers’ compensation studies which have shown that when penalties for safety violations were increased, a decrease in worker injury rates were found to occur in tandem, as was the case for occupational hearing loss in Oregon and amputations in Kentucky, and in a recent study of mandated eye protection in Norway’s metal industry, which demonstrated over a 14-fold drop in ocular injuries following legislation.25 However, such studies have not all found such uniform results as following the increased reporting of work-related hearing loss in Washington State.2627 Thus, while no direct causal relationship can be established, the increase in average penalty amounts for safety violations, accompanied by both a simultaneous decrease in eye injury claim rates and workers’ compensation claims overall in the state, provides strong support that stricter legislative measures may have a positive impact on worker safety, and should be considered by state policymakers as a means of decreasing eye injuries among workers in the future.

Another important avenue for eye injury prevention lies in the education of both employers and employees. While the findings of this study concurred with previous studies showing that the most hazardous industries for eye injuries were manufacturing, construction, and mining, eye injuries occurred across a wide variety of occupations and industries that are not typically thought of as ones at high risk for eye trauma. To make interventions for reducing eye injuries successful, workplace safety managers must first educate themselves of what the most their problematic areas are, and then develop appropriate protocols by which to address them. For example, in this study, it was found that the highest average costs per eye claim came from incidents involving slips and falls due to grease, which is not generally an area that first comes to mind when thinking about eye injury prevention (in contrast to, for example, protective goggles for welders). Knowing this, and that traditional eye protection methods may be impractical (eg, wearing protective goggles in establishments with greasy floors), employers, in recognising this as a hazard, can attempt to address the actual problem (keeping the floors grease-free), which in turn would reduce serious eye injuries in the future, as well as potentially decrease other injuries that may stem from the same hazard.

Once such areas are identified, and interventions enacted, subsequently obtaining employee compliance to adhere to such measures is necessary for success. While in some cases where employees are closely monitored, and the enforcement of mandatory rules could suffice, additionally a growing body of literature is demonstrating that aligning employee beliefs through the development of safety cultures can synergistically work to improve intervention efficacy.28 These cultures are characterised by a clear and sincere commitment by management to the well-being of their employees, combined with open communication and intensive training.2930 It is essential that managers realise that eliciting motivation from their workers to utilise protective measures and understanding correct protocols through training are necessary for eye injury reduction. Such approaches proved successful among metal workers in Italy, in a recent case–control study in Taiwan, and among Latino farm workers in Michigan and Illinois.31^–33 Understanding that much, if not all, of the costs of extra training and communication can offset the costs associated with workers’ compensation claims, especially when the cost of such claims is made more expensive through legislation like Kentucky’s, could be a particularly persuasive rationale for employers to adopt a holistic approach to improving workplace safety through the interest alignment that accompanies the implementation of a safety culture mindset. As there is a stream of research on safety culture that is emerging in the occupational injury literature, specific applications of such research to assess how such changes affect ocular injury rates should be conducted in the future to assess the efficacy with regard to improvements realised and areas in need of extra or specific eye protection interventions.

There are several limitations to this study of which readers should be aware. First, because the state of Kentucky does not collect information on medical costs associated with workplace injuries, the total cost of eye injuries reported is underestimated. Additionally, because Kentucky does not require incident reporting unless one day or more of work is lost due an injurious event, the actual number of eye injuries during the period examined is likely underreported. Previous work has demonstrated that workers’ compensation analysis of workplace injuries underestimates the true number of actual injuries in the workplace, and thus, the total injuries in this paper should not be viewed as a measure of injury prevalence.34 In addition, self-employed workers, such as individual contractors, plumbers, and other such occupations were not required to be covered by workers’ compensation, and hence, data on eye injuries from these individuals were also not able to be accounted for in this analysis; on the other hand, as these groups were excluded when running the CPS for denominator purposes, the rates were adjusted to account for these missing subgroups. Importantly, however, as the 10 545 injuries found are an underestimation of both the total incidence of eye injuries and associated costs, the problem of occupational eye injury may be higher than the analysis demonstrates and in even more need of continued study and intervention. Because the codes of injury causation in the data were sometimes not entirely specific, precise interventions for the future prevention of such injuries were impossible to identify and recommend. Future research into the exact causes of eye injury these areas would be valuable for reducing such risk in the future.