Income Changes Due to Disability Ratings and Participation in Economic Activities Caused by Industrial Accidents: A Population-Based Study of Data from the Fourth Panel Study of Workers’ Compensation Insurance (PSWCI)

Industrial accidents cost a huge amount of money, but they also have negative consequences in many respects. We analyzed the data of the first to fourth panel study of workers’ compensation insurance (PSWCI). Repeated measures ANOVA was used to compare the annual income before and after the industrial accident, and a general linear model was used to identify changes in income due to disability ratings and participation in economic activities. The wages before the industrial accident and the annual income varied among the disabilities ratings. In addition, for affected workers, the average income during four years post-accident was lower than the average income before the accident. Regression analysis to see changes in income after the industrial accident showed that the group with a disability rating of 11–14 and no injuries had a suffered a greater income decrease than those with a disability rating of 1–3, and the unemployment group saw a greater decrease in income than the employment group. Workers who were affected by industrial accidents received lower incomes than before the accident, and even considering different disability ratings, there was a greater decrease in income among the unemployed group than in the working group.


Introduction
According to the Korean Ministry of Employment and Labor (2017), 18.4 million workers from approximately 2.5 million businesses were eligible for the Workers' Compensation Insurance. Among them the number of victims was 90,656, and accident rate was 0.49% [1]. This was a 0.01% decrease from that of 2015, but due to the increased number of business sites and employees, the coverage of industrial insurance continued to increase [2]. In 2015, the manufacturing industry constituted 29.12% of total work-related injuries and illnesses and the construction industry 29.54% [3]. Industrial accidents both directly and indirectly cause significant losses to the government, business owners,

Study Design and Participants
We used the data from the first to fourth panel study of workers' compensation insurance (PSWCI) organized by the Korean Workers' Compensation and Welfare Service. The PSWCI was conducted to objectively examine the effects of the policy system created to aid industrial workers and to help provide proper medical care and assistance in their returning to economic activities. In 2012, the 82,493 workers who had completed medical care were selected as the population and stratified by gender, age, region, disability grade, and recovery service, resulting in 2000 samples. Currently, four trials of the survey have been conducted on the sample population. The surveys were conducted by computer-assisted personal interviewing (CAPI), a one-to-one and face-to-face survey guided by professional interviewers [27][28][29][30][31][32].
In this study, the analysis data were collected by referring to the PSWCI and using the following procedure. After the fourth survey, 1660 out of the initial 2000 samples were retained, but only 1588 had continuously participated in all four surveys. Of those who had participated in all four surveys, the final 1538 samples were selected excluding 50 industrial workers who had not participated in the survey or had descriptive variables that were difficult to estimate or were unobserved.

Sociodemographic Characteristics
Age was considered as a continuous variable and the groups were divided by five from "<30" to "≥60" in increments of 10 years. Marital status was divided into three groups: "Not Married", "Married", and "Others (separated, divorced, widowed, etc.)". Education level was categorized by eight groups in the PSWCI but we re-grouped into three: "Less than high school", "High school", and "College or above".

Occupational Characteristics
The industrial and occupational classifications were categorized according to the Korean Standard Industrial Classification (KSCI) and the Korean Standard Occupational Classification (KSCO) respectively. Based on the International Classification of Standards and Industry (ISCI), the KSCI groups the industrial and occupational classifications into three, accounting for number of cases of work-related injuries and illnesses. As a result, according to KSCI method, the industrial and occupational classifications were categorized into "Manufacturing", "Construction", and "Others". The KSCO is based on the International Standard Classification of Occupations (ISCO) and the samples were categorized by "White collar", "Service" (service sector workers including sales), and "Blue collar (including others etc.)".
Participation in economic activities was divided into six groups, including returned to work, re-employment, self-employment, unpaid family worker, unemployment, and non-economic activities. Among the six groups, the first four groups-such as the return to the original job, re-employment, self-employment, and unpaid family worker-were considered "economic activity participants" while the latter two groups, including unemployment and non-economic activities, were considered "economic activity non-participants".
For accurate analysis, those who had been employed for four consecutive years were classified as "Returned to Work" while those who had not been employed for four consecutive years were classified as the "Non-RTW". Those who had repeatedly been employed and unemployed were classified as "Others".

Main Outcome Variables
The income before the industrial accident was estimated by the answer to the question "What is the average monthly wage at the workplace (job) where the industrial accident has occurred?" This was then multiplied by 12 to calculate annual income. When the recovery period ranged from one to two years, we estimated the 2012 annual income by multiplying the 2011 pre-accident income by 1.0376, the average pay increase of all workers in 2012. When the recovery period was longer than two years, however, it was impossible to estimate the pre-accident income, which was therefore excluded. The post-accident income was estimated to be the sum of the individual's earned income and non-work income. Non-work income included industry-related income, i.e., lump sum compensation and disability pensions, excluding property income and private income prior to the accident or unrelated to the industrial accident. The average wage increase rate reported by the Ministry of Employment and Labor was used to compare the income before and after the industrial accident and was all adjusted to fit the wages of 2015 [34].

Statistical Analyses
The comparison of the social and professional characteristics and the mean income before the industrial accident of the subjects was made using the t-test and ANOVA. Repeated measures ANOVA was used to compare the annual income before and after the industrial accident, and in addition, the average income before the accident was compared with the average income over the four-year period after the accident. A paired t-test and the repeated measures ANOVA were performed to compare the pre-accident income and the average income of four years after the accident with respect to the general characteristics of the subjects. A general linear model was used to identify changes in income due to disability ratings and participation in economic activities. The dependent variable was defined as the average of the first to fourth-year income excluding the income prior to the industrial accident. All analyses were completed using the SAS statistical package version 9.4 (SAS Institute, Cary, NC, USA).

Results
The general characteristics of the survey subjects are as shown in (Table 1). The average annual income before the industrial accident was 26.82 million KRW. There were more men participating in the survey than women as men composed 82.8% (1274 men) of the total survey personnel. A large proportion of participants were in their 50s. The income prior to the accidents was relatively higher for men (p < 0.0001). Income differences before the industrial accident were present according to age (p < 0.0001) and disability ratings (p = 0.0017). Pre-accident income was higher in the group that participated in economic activities after the industrial accident than the group that stayed unemployed (p < 0.0001).
The wages before the industrial accident and the annual income varied among the disability ratings (p < 0.0001) ( Table 2). Incomes for the first year after the industrial accident were higher than those before the accident for all workers who were affected, except those who did not suffer from any disabilities ("None"). The second-year income dropped sharply but showed a gradual increase in following years. In addition, for affected workers, the average income during four years post-accident was lower than the income before the accident (p = 0.0005). Income changes were also observed in the earned income and non-work income respectively (Supplementary Materials).
In the case of employed population, those with disability ratings between one and seven showed higher annual and average income for four years than income before the industrial accident (p < 0.0001). However, the unemployment group and others showed lower annual (p < 0.0001, p < 0.0001) and average income (p = 0.0004, p = 0.0003) ( Table 3). Table 4 shows the difference in pre-accident income and the average income over the four years after the industrial accident. The total income after the accident decreased significantly compared to the previous gross income (p < 0.0001), and with the exception of the age group "30-40", the four-year average income was lower than the income before the accident. In order to identify the income changes before and after the accident, a regression analysis was performed with the dependent variable as a four-year average of post-accident income minus the income before the industrial accident and the independent variables as general characteristics, employment status, disability grade, and others (Table 5). Adjusting for all covariates, income for men showed a greater decrease after the accident than that of women (p = 0.0090), and manufacturing industries presented a greater decrease than the construction industry (p < 0.0001). Compared to the disability rating one through three group, the rest showed reduced income and there were statistically significant decreases in both the disability rating 11 through 14 group and no disability group (p = 0.0321, p = 0.0200). The unemployment group had significantly decreased income compared to those who participated in economic activities (p < 0.0001).  * A four-year status of economic activity (return to the original job, re-employment, self-employment, and unpaid family worker). † A four-year status of non-economic activity (unemployment and non-economic activity). ‡ Status of economic activities and non-economic activities. § Analyses were done by using repeated measures ANOVA. || Analyses were done annual income before industrial accident and the average of the first to fourth-year annual income by using repeated measures ANOVA. * All adjusted to fit the wages of 2015. † Analyses were done by using repeated measures ANOVA. ‡ Analyses were done by using paired t-test.

Discussion
Industrial accidents not only impair, if not completely eliminate, the worker's work capacity but they may also cause physical and mental disabilities. This can limit labor market participation, reduce annual income, and therefore threaten household livelihoods [35][36][37]. It has long been criticized that those who are compensated due to industrial accidents through the compensation system show more unfavorable outcomes than those uncompensated [38]. This study was conducted to identify the changes in income before and after the industrial accidents of affected workers whose recovery periods had ended. The income before the industrial accident is not related to industrial accident characteristics and therefore shows similar pattern as the general income characteristics. Nevertheless, we believe that the reason work injury patients have lower pre-accident incomes than work-related illness patients is that physical injuries occur more in the service sector and blue collar jobs than in white-collar job environments. Furthermore, the reason the pre-accident income of the victims with no disabilities after an industrial accident is lower than that of those who were injured is that many of the uninjured belong to the service sector where accidents caused are usually minor. According to the 2016 analysis on the status of industrial accidents, accident rate for white-collar workers was 8.8%, the blue-collar 83.7%, and the service sector 7.3%. The death toll among the service sector and the sales sector were 0.4% and 0.2%, respectively, which is minimal [1]. Under Korea's Industrial Accident Compensation Insurance Act, disability compensation is provided depending on the disability rating. Disability pensions are provided for Grades 1-3, or recipients can be paid half the pension amounts for the first four years in advance. Recipients for Grades 4-7 can choose between a "pension" option and "lump sum compensation" option. All recipients for Grades 8-14 are paid lump sum compensations [33]. The change in income before and after an industrial disaster was classified by disability rating to show that first-year income after the termination of medical care increased from the income prior to accident for all workers excluding the uninjured. This is believed to be related to the lump sum payment made on the day of the accident. Without the lump sum payment, the income in the second year rapidly decreases. This reduction in income can also be attributed to continued unemployment in the second year. Accordingly, the reason for gradual increase in income over time is most likely due to reemployment. We believe that the reason the income of those with no disabilities decrease after an industrial accident is because the injured cannot return to their original job or find another job even though they do not have any disabilities. In addition, the average income over four years post-accident decreased compared to the income prior to the industrial accident (Table 2). This result is similar to that of an existing research [34], and we believe that this is due to insufficient compensation or difficulties in employment after industrial accidents. We chose the four-year average income to compare with the pre-accident income since the fourth-year income reflects future income trends and the first-year income includes lump sum payment effect and thus deemed inadequate for this study. Those with an up to Grade 7 disability rating, who are eligible for the disability pension plan, showed higher annual and four-year average income than the pre-accident income. However, the unemployed, even if they were eligible for pensions, saw a decrease in both annual and four-year average income after the first year (Table 3). This means that compensation is sufficient for the employed but not for the unemployed. The Ohio Workers' Compensation for industrial accidents is provided depending on the severity and types of wage loss and injury. In cases where wage losses make up two-thirds of total losses, the compensation varies depending on whether workers return to work; two-thirds of average weekly wage is compensated for up to 52 weeks for those returning to work and 200 weeks for those not returning [39]. American Medical Associations (AMA)'s Evaluation of Permanent Disability guide is used to determine the rating of permanent disabilities of workers from industrial accidents in California. The compensation varies depending on the employment coefficient and on the number of employees, and whether there was a legal employment offer; employees who are offered proper employment receive two-thirds of the average weekly wage minus 15% and those not offered receive two-thirds of the average weekly wage plus 15% [40]. In summary, the economic status of workers declines after industrial accidents, regardless of the degree of disability. Even those who usually have less difficulty in finding jobs-such as disability ranking 8 to 10, 11 to 14, and no disability-saw a decrease in income. There are a few limitations with this study. Since the survey was conducted at a specific period, the study does not effectively remove retrospective bias. Also, even though this was a panel survey, due to the excluded personnel, it was not weighted. Finally, the disability benefit that was given in the form of a lump sum or prepayment may have caused the first-year income to be higher than that of following years. Nevertheless, this study has strength in that it is Korea's first panel study on workers' compensation insurance, includes a relatively large sample, and has a sufficient tracking period to see income changes.

Conclusions
In conclusion, workers who were affected by industrial accidents received lower incomes than before the accident, and even considering the different disability ratings, there was a greater decrease in income among the unemployed group than in the working group. Therefore, efforts should be made to improve the return-to-work ratio after industrial accidents. It is also necessary to consider different disability compensation measures depending on the return-to-work status of the workers.
Supplementary Materials: The following are available online at http://www.mdpi.com/1660-4601/15/11/2478/ s1, Table S1: Relationship between disability rating and total earned annual income changes before and after industrial accident, Table S2: Relationship between disability rating and total non-work annual income changes before and after industrial accident.