Work-related eye injuries are important aetiologies of blindness and account for a large proportion of occupational injuries. Work-related eye injuries not only hurt workers and their families, but also cause huge losses in terms of productive power and social costs. The number of work-related eye injuries in the USA has increased over the past 25 years. In the 1980s, there were more than 2.4 million ocular injuries: approximately 1000 occurred in the workplace each day,1 and, the number doubled at the beginning of the 21st century.2 These incidences may be underestimated, inasmuch as many trivial ocular injuries are unreported. In developing countries, the incidence and severity of work-related eye injuries, as with other occupational injuries, is much worse than in highly developed countries, probably due to negligence regarding occupational health and safety. There have been no population-based studies or official statistical data on work-related eye injuries in Taiwan up until the time of this study; the burden of work-related eye injuries in Taiwan has not been described until now.

Work-related eye injuries are becoming important causes of vision impairment, following cataracts, glaucoma, and trachoma.3 Unlike blindness due to other causes, 90% of visual loss caused by eye injuries is considered preventable through the use of simple eye protective devices or preventive methods. Importantly, little information exists concerning the risk of and preventive factors for work-related eye injuries. Given the magnitude and scope of the public health problem presented by work-related eye injuries, the need for comprehensive prevention strategies is great. Developing such strategies begins with identifying the demographic data and determinants of work-related eye injuries. Most published studies of work-related eye injuries are cross-sectional and have been limited to descriptive epidemiology. These studies have usually focused on fixed factors of work-related eye injuries, such as age, gender, hazardous materials, or injury types.4^–6 With regard to potentially modifiable factors, most published studies have emphasised the importance of eye protection devices (EPDs) only, without quantifying the protective effect of EPDs. Few studies have documented other factors, such as work practice, work environment, or individual characteristics that may transiently increase or decrease the risk of work-related eye injuries.

A well-designed analytic study is necessary to explore the degree to which a modifiable factor might contribute to injury reduction if changed. We have conducted this retrospective study using a case-crossover design to explore the associations between modifiable factors and the occurrence of work-related eye injuries.

METHODS

Study subjects

Subjects were recruited over a 4-year period (2003–2006) from the ophthalmic clinics and emergency departments of seven medical centres located in northern, southern, eastern, and central Taiwan. To be eligible for the study, subjects must have experienced one or more physical, chemical, or biological injuries to the eyeball, adnexa, or other adjacent structures while working. After being treated, the eligible subjects were invited to participate in this study; all participants gave their written informed consent to their staff clinician. The study was approved by the Joint Institutional Review Board of the Public Health College, National Taiwan University.

Data collection

Case information on eligible subjects, including the time and date of the injury, the cause of the accident, results of the physical examination and laboratory findings, and the diagnosis (International Classification of Diseases, 9th Revision codes), was compiled after it had been verified by the clinician in charge. A trained interviewer then performed a telephone interview with the participant using a structured questionnaire to obtain the necessary information; the calls were placed in the evening and thus, were outside of the work and clinical environment. A study participant was considered unreachable following three unanswered telephone calls. The structured questionnaire was designed to collect demographic characteristics and detailed exposure information. At the start of the questionnaire, the date, time, and nature of the injury, as well as personal habits, occupation, and job experience of the study participant were verified. Questions assessing exposure to specific transient factors during the 60-minute time period before the accident (defined as the hazard period) were then recorded by the interviewer. The eight specific transient factors potentially at play during that 60-minute period were as follows: wearing EPDs; performing unfamiliar tasks; unskilful operation of tools or equipment; being distracted, rushed, or fatigued; under the influence of alcohol; and working in poor conditions such as an overheated, overcrowded, noisy, or poorly illuminated environment. Information on exposure to these eight factors during the control period, the same 60-minute time interval on the last work day prior to the injury, was also collected.

Case-crossover study

The case-crossover design is a useful methodology for studying the transient effect of brief exposure on the occurrence of a rare acute-onset event, including acute occupational traumatic injuries.7 In a case-crossover study, we compare the individual activity before the onset of the injury to the patient’s usual activities. Thus, both the “case” and “control” components are from the same individual. This design can reduce potential confounding factors related to the subject’s characteristics by using each individual as his/her own control in occupational injury studies. There are two statistical analyses: the matched-pair interval approach and usual-frequency analysis, which are commonly adopted for case-crossover studies. However, in consideration of the more significant “memory telescoping phenomenon” in the usual frequency approach,8 the matched-pair interval approach was used in this study to evaluate the effect of transient exposure to acute work-related eye injuries. According to experts’ opinions and our pilot study results, we defined the “case component” as a hazard period, a 60-minute time interval immediately before the injury, and the “control component” as a control period, the same 60-minute time interval on the last work day prior to the injury. The matched-pair interval approach follows standard methods of stratified analyses and a comparison was made between exposure to each transient factor during the hazard period and the control period, respectively.9 10

Statistical analysis

We estimated the odds ratios (ORs) and 95% CI for these highly stratified data by using a one-to-one matched case-control study formula from Rothman and Greenland.11 The Mantel–Haenszel OR is the ratio of the number of subjects unexposed in the control period but exposed in the hazard period, to the number of subjects unexposed in the hazard period but exposed in the control period.

RESULTS

Of all 435 eligible occupational work-related eye injury patients, 283 (65.1%) completed the interview. Their distribution in terms of age, gender, or type of injury was not different from that of the non-respondents. As shown in table 1, the mean age was 40.2 years (SD 13.1). Two hundred and fifty-nine (91.5%) of the subjects were male. The distribution of age and gender of interviewed subjects in this study was compatible with that of 184 disability-claimed cases from the Bureau of Labor Insurance (BLI), Taiwan in 2007. The major job categories of employees that were most liable to suffer work-related eye injuries were manufacturing (34.6%), construction (24.7%), agriculture (15.2%), the steel industry (11.3%), and services (8.5%). Compared with the disability-claimed data from BLI, Taiwan in 2007, the proportions of job categories in manufacturing and agriculture in this study were higher than those in BLI disability-claimed data; meanwhile, workers engaged in construction and services had a lower injured rate. In this study, 67.2% of all injured subjects were employed by small enterprises with less than 10 workers or were self-employed workers. In addition, 76.3% of those injured had more than 5 years of work experience. Pre-employment safety training was rare among these workers, with 84.1% of them having never received any job safety training before employment.

Among the 283 subjects, 125 (44.2%) suffered injuries to both eyes, while 158 (55.8%) were injured in one eye. The injuries occurred mostly at 11:00–12:00 and 15:00–17:00 during the work day (fig 1). Injury mechanisms and types among the 283 patients, as well as the severity of the work-related eye injuries are shown in table 2. Welding was the most common cause of a work-related eye injury (30.4%), followed by drilling/cutting (13.4%), splashing of chemicals (11.7%), hammering (10.2%), equipment cleaning/maintenance (7.1%), mowing (6.4%), grinding (6.0%), loading/unloading (4.9%), and falling (3.5%). The most common injury type was photokeratitis (33.2%), followed by eyeball penetration/laceration (22.3%), chemical burn (14.1%), corneal abrasion/foreign body (12.4%), ocular adnexa injury (9.2%), blunt trauma (4.9%), and scalding burn (3.9%). More than half of the injured subjects (55.2%) had visual acuity of less than 3/60, the World Health Organization definition of blindness, at the time of injury; and 35.7% of all injured cases needed hospitalisation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

The time of injury for work-related eye injuries among the 261 study subjects.

Table 3 is a two-by-two cross table of the exposure to eight transient factors during the control and hazard periods. The matched-pair interval approach was employed to evaluate the effects of the eight transient factors on work-related eye injuries. The ORs of work-related eye injuries were all significantly increased by the performance of unfamiliar tasks (OR = 57.0, 95% CI 14.1 to 230.7); unskilled operation of tools or equipment (OR = 24.0, 95% CI 3.2 to 177.4); being distracted (OR = 24.0, 95% CI 3.2 to 177.4), rushed (OR = 13.0, 95% CI 3.1 to 54.8), or fatigued (OR = 10.0, 95% CI 1.3 to 78.1); and working in a poor environment (OR = 4.3, 95% CI 1.4 to 12.6). The consumption of alcohol increased the OR slightly, but not to a statistically significant level (OR = 2.5, 95% CI 0.5 to 12.9). Wearing EPDs was the only protective factor found to reduce the OR for work-related eye injuries (OR = 0.4, 95% CI 0.2 to 0.7).

DISCUSSION

Work-related eye injuries not only result in economic losses for the industry, but also affect a considerable number of workers and their families. The preliminary results from this study suggest that specific modifiable factors could increase or decrease the risk of work-related eye injuries. Seven of the transient factors are from three domains: work practice (performing unfamiliar tasks, unskilled operation of tools or equipment, wearing EPDs), worker-related factors (being distracted, rushed, fatigued), and work environment (working in a poor environment). Unlike other case-control studies that identify fixed factors for work-related eye injuries, this study identifies factors that are potentially modifiable.

There are limited data that describe the general conditions correlated with work-related eye injuries in Taiwan. The disability-claimed data of work-related eye injuries were available since 2007 and there were only 184 claimed cases. Before that, the Labor Injury Surveillance System merged work-related eye injuries with other facial injuries (including eye, nose, ear, and mouth injuries). In addition, employers and even healthcare providers are not required to report eye injuries in Taiwan. This is the first report with multicentre cooperation in Taiwan, and the first to include patients recruited from emergency departments, ophthalmology outpatient departments and wards. Although the burden of work-related eye injuries could be underestimated, the results of this study more accurately reflect the general conditions of work-related eye injuries in Taiwan relative to previous studies.

These results, which are consistent with previous epidemiological studies,4 12^–15 indicate that males, young people, self-employed individuals, and small enterprise workers were most prone to eye injuries. Compared with 184 disability-claimed cases from BLI, Taiwan in 2007, the proportions of job categories in manufacturing and agriculture in this study were higher than those in BLI disability-claimed data; meanwhile, workers engaged in construction and services had a lower injured rate. However, these fixed individual factors of the study population could not be evaluated as risk factors with the case-crossover design.

The current study revealed that photokeratitis caused by blaze or radiation is a major cause (33.2%) of work-related eye injuries. This result is different from those of previous studies in which foreign bodies, corneal abrasion, blunt injury, or chemical burn were the major types of work-related eye injury reported.1 4 15 Ho et al indicated results similar to ours; they found that there was a strikingly high percentage of injury due to ultraviolet (UV) light radiation in Taiwan.14 Most patients who suffered photokeratitis with complaints of severe eye pain and tearing sought symptomatic treatment at emergency departments or ophthalmic clinics without being hospitalised. Due to the straightforward treatment of its symptoms, this injury is usually overlooked by clinicians and patients. However, repeated exposure to blaze or radiation causes chronic keratitis or cataracts and may eventually lead to visual impairment. This type of injury should be more consistently noted and emphasised.

From the viewpoint of occupational medicine and industrial hygiene, an occupational injury may involve several domains, such as factors having to do with work practice, work equipment, and the workers themselves, as well as those related to work environments. There is always a sequence of events that may lead to an injury; any transient alteration in this sequence or elimination of one of these factors in the injury chain will usually lead to or prevent an accident, respectively.16 This observation may explain why 63% of New Zealand labourers who wear EPDs as regulated still suffered eye injuries.17 The results of this study suggest that performing unfamiliar tasks, using unusual operating tools or equipment, being distracted, being rushed, being fatigued, and working in a poor environment play important roles in work-related eye injuries. There is little information available about the effects of these transient factors on work-related eye injuries, but a previous study elucidated the importance of these potential modifiable factors in the aetiology of acute hand injury.18

Performing unusual tasks and unskilled operation of equipment are probably the main causes of work-related eye injury. McGwin et al also indicated that performing an unusual task and using an unusual tool were transient risk factors in various types of occupational injuries.19 The incidences of these two modifiable factors were substantially different between the hazard and control periods and extremely high values were found for their relative risks. Since the vast majority of those injured were self-employed or employed in small enterprises, it may have been necessary for them to rotate to different jobs or operate various machines for which they had not been trained. In addition, there was little pre-employment training or demonstration for most of these injured subjects. These facts stress the need for enforced safety training before workers rotate jobs or operate unfamiliar equipment.

Being distracted or rushed also increases the relative risk of eye injury. Such distraction was primarily due to the worker’s talking, looking around, or thinking about something not related to the task. The workers were rushed mainly because of pressure from their supervisors or customers, and also because of their eagerness to leave work. The double peaks of injury time at 11:00–12:00, the time before lunch, and 13:00–17:00, the time before leaving work, suggest that this latter factor may be strongly associated with rushing.

This study also demonstrated that poor work environment also increased the OR for work-related eye injuries. Many enterprises, especially small enterprises only focused on how to cut costs in the manufacturing process, overlooking the potential loss of manpower due to a poor working climate or work environment. It is advisable to design appropriate safety training programmes for these high-risk workers. Such programmes must integrate awareness of hazards, improvement of job skills, management attitudes and practices, protective devices, safety climate, and factors that are important to creating a safe work environment. The employers should also be educated about their responsibility in terms of protective devices, engineering control, work environment layout, and administrative control.

It has been demonstrated that excessive alcohol use influences occupational injuries.20 However, the OR and relative risk of drinking alcohol were not statistically significant in this study. This finding was probably caused by recall bias of the interviewed subjects owing to personal unwillingness to answer the question. In addition, the hazard period for alcohol consumption may be longer than that which we defined in this study (60 minutes). It is possible that our conservative estimation of the hazard period led to misclassification bias and underestimation of the OR. A more refined exposure log could address this limitation in a future study.

Wearing EPDs is the only protection factor identified for work-related eye injuries in this study. A case-control study also showed that subjects wearing safety glasses regularly were less likely to have eye injuries.21 Another case-control study with 31 cases and 62 controls showed that the OR for eye injuries was reduced in workers who were wearing EPDs, but this difference was not statistically significant.22 Our study indicated that EPDs could reduce the risk of work-related eye injury by up to 60%. The authority in Taiwan has not enacted a rule to requiring the use of EPDs during work; a regulation for their compulsory use in the workplace is urgently needed.

More information about how to choose appropriate EPDs and how to wear EPDs correctly should be offered. Our study showed that only 18.4% of workers were wearing EPDs when injured. Most injured workers said that they did not know that they needed to wear EPDs while working. Some chose not to wear EPDs because of the inconvenience and discomfort of EPDs. Hunty and Sprivulis17 also commented that the workers’ willingness to wear EPDs is greatly influenced by specific factors, such as comfort, the clarity of the glasses, the price of the EPDs, rules set by the company, and awareness of the risks.

There were some limitations of our study. First, the definition of the hazard period is a key issue in the design of a case-crossover study. The hazard period can vary depending upon the nature and duration of the exposure. Misspecification of the hazard period is likely to bias the results. In this study, we defined a 60-minute period as the hazard period in accordance with expert opinions and our pilot study results. Most transient risk factors evaluated in this study influenced the risk of injury during the entire evaluated period, lasting from several seconds to 1 h; for this reason, a 60-minute period was selected. With regard to the potentially protective factor, EPD use, we found that photokeratitis was the most common injury type in our preliminary data. The hazardous effect of UV radiation was not immediate. The cumulative effect of UV radiation must be considered and 60 minutes was considered rational as the hazard period.

Second, we have employed the usual-frequency approach in order to evaluate the associations between these eight transient factors and work-related eye injuries. The relative risks and 95% confidence intervals for work-related eye injuries, evaluated using the usual-frequency approach, were larger than those evaluated using the matched-pair interval approach. Moreover, the protective effect of wearing EPDs became statistically insignificant. This finding could be attributed to a recall bias. The results of this study were obtained using retrospective data from a 4-year period, and the injury-to-interview period for some subjects was up to half of a year. This study may thus be vulnerable to the so-called “telescoping phenomenon” that the subject may overestimate exposure frequency close to the time of the injury and underestimate exposure frequency or person-time in the control period because of a memory lapse.8 This observation is particular relevant when recalling personal events more than 2 months prior. Memory telescoping is more significant for the usual-frequency approach and would bias the relative risks upwards. Shortening the injury-to-interview time interval in further studies would minimise the recall bias and make the results more stable and reliable.

Another potential limitation is a within-subject confounding factor due to the co-occurrence of transient risk factors in control and hazard periods; for example, when the injury occurs, a worker may be distracted and rushed at the same time. Complex interactions may determine the acute onset of an injury, and it is difficult to distinguish between them within individuals.

CONCLUSIONS

Our data conclude that work-related eye injuries are associated with the transient risk factors of performing unfamiliar tasks; unskilled operation of tools or equipments; being distracted, rushed, or fatigued; and working in a poor environment. Wearing EPDs protects against work-related eye injuries. As these transient factors are all potentially modifiable, controlling them in the workplace could effectively reduce the risk of work-related eye injuries. These findings should be considered in the design of appropriate safety training programmes in the future.