Article Text
Abstract
Objective We sought to examine the association between job tenure and lost-time claim rates over a 10-year period in Ontario, Canada.
Methods Data were obtained from workers’ compensation records and labour force survey data from 1999 to 2008. Claim rates were calculated for gender, age, industry, occupation, year and job tenure group. A multivariate analysis and examination of effect modification were performed. Differences in injury event and source of injury were also examined by job tenure.
Results Lost-time claim rates were significantly higher for workers with shorter job tenure, regardless of other factors. Claim rates for new workers differed by gender, age and industry, but remained relatively constant at an elevated rate over the observed time period.
Conclusions This study is the first to examine lost-time claim rates by job tenure over a time period during which overall claim rates generally declined. Claim rates did not show a convergence by job tenure. Findings highlight that new workers are still at elevated risk, and suggest the need for improved training, reducing exposures among new workers, promoting permanent employment, and monitoring work injury trends and risk factors.
Statistics from Altmetric.com
What this paper adds
-
What is known
-
Workplace injury accounts for a substantial source of illness burden and disability in working-age populations in Canada. Shorter job tenure is associated with an increased risk in occupational injury.
-
What this study adds
-
This study examines the associations between job tenure and lost-time claim rates over a 10-year period during which claim rates declined in many North American jurisdictions. Newly hired workers remained at an elevated work injury rate over the time period, suggesting that the factors which contributed to declining work injury rates over the past decade, may not have been as successful in reducing the risk for newly hired workers.
-
Policy implications
-
Our findings suggest the need for improved training for new workers, reducing exposures among new workers and promoting permanent employment.
Background
Workplace injury accounts for a substantial source of illness burden and disability in working-age populations in Canada.1 Previous research has examined the influence of being new to a job and work injury risk.2–6 Job tenure, as measured by length of service at a particular job, has been found to have an inverse relationship with work injury. That is, shorter job tenure is associated with an increased risk in occupational injury,2–5 with the relative risk (RR) decreasing with increasing job tenure.6 Furthermore, there is evidence that all workers who are new to a job have a similarly increased risk, regardless of factors such as age.7 ,8
New workers may be at greater risk for occupational injuries and illnesses due to lack of job experience/knowledge, inadequate safety training,9 exposure to more hazardous conditions10 and/or less accurate hazard appraisal.11 It has been suggested that changing employment structures in labour markets and the growth of precarious forms of work over the past decade have created more temporary employment and higher rates of job turnover.12 This in turn has resulted in a higher proportion of workers being exposed to the risks of being new to a job, making this an increasingly relevant issue to study.
In this study, we were interested in examining the degree to which work injury claims are associated with being new on the job after adjusting for potential confounders including age, gender, industry and occupation. These sociodemographic and work-related factors may confound the association between job tenure and work injury in several ways. Age is related to work experience13 and there are age differences in the rate of work injury.14 ,15 Young workers are often less experienced,16 and in addition, more likely to be employed in temporary, part-time or casual work.17 Occupation and industry may also be linked to job tenure. Labourers and those in manual jobs are often younger17 and have higher injury rates than those in professional jobs.18 Also, women are more likely to work in temporary jobs,19 and temporary workers may be at an increased risk of work injury related to greater inexperience and lack of safety training at the workplace.20
This study also sought to examine the extent to which age, gender and industry modified the job tenure/injury association. Past study findings suggest that these factors may act as effect modifiers in the relationship between job tenure and work injury. For instance, one study found that more than 50% of claims in the construction industry were by workers with less than 1 year of job tenure.5 Another study found that those working in manual jobs had higher injury rates among new workers than those in professional jobs.21 There is also evidence that suggests a difference in the injury risk of new workers by age6 and gender.13
Furthermore, a recent study by Siow et al22 found that new healthcare workers had increased risk of cuts and punctures, but a decreased risk of musculoskeletal injuries. Therefore, the interaction between job tenure and type of injury is important to examine as it may provide insight on the mechanisms causing injuries in new workers compared with those with longer job tenure.
Rates of compensable work injuries have been steadily declining for over a decade in Ontario.23 ,24 Specifically, recent data on workers’ compensation claims indicated that in several Canadian jurisdictions including Ontario, the rate of lost-time claims has declined substantially from 2000 to 2007.25 Furthermore, a recent study reported a convergence in workers’ compensation claim rates for youth and adult workers from 1991 to 2007 in Ontario.26 This notion led us to examine whether there were similar convergence patterns for new workers. To our knowledge, no published study has examined lost-time claims by job tenure over the last decade to see whether those with shorter job tenure benefited from an overall decline in work injury claims.
The objective of this study was to describe the association between job tenure and lost-time claim rates over a 10-year period (1999–2008) during which there was generally an overall decline in lost-time claims.23–26 To account for confounding demographic and work-related characteristics, claim rates were adjusted for age, gender, industry and occupation. We also examined the modifying effects of each age, gender and industry on the relationship between job tenure and lost-time claims. Interactions between the job tenure/work injury association by type of injury event, and by source of injury were also examined. We expect differences in the association across labour market subgroups, and hypothesise that there will be a stronger relationship between job tenure and work injury claims in men compared with women, among older workers compared with younger workers, and among those working in manual jobs compared with non-manual and mixed jobs.
Methods
This study used claims data from 1999 to 2008 from the Workplace Safety and Insurance Board of Ontario, which insures approximately 65%–70% of labour force participants and is the main provider of workers’ compensation in the province.27 The remaining 30%–35% not covered included those who are self-employed, domestic workers, federal government workers, the majority of the finance industry and workers associated with interprovincial commerce. Claims that result in one or more days of lost work, loss of wages, or permanent disability/impairment are considered to be lost-time claims. This study only selected lost-time claims from claimants who had no previous claim activity for at least 2 years preceding the injury date as a way to exclude claimants whose previous injury may increase their vulnerability to re-injury.
For each claim, information was available on gender, age (grouped), job tenure, occupation and industry sector (corresponding to the Canadian Standard Industrial Classification for Companies and Enterprises, 1980).28 Using the claim form completed by the employer, which includes both the employee date of hire and the injury date (date attributed to the injury, accident or illness), job tenure was calculated as the number of months between these two dates. Industry was further collapsed into two industrial groups: goods (such as agriculture, automotive, chemical processing, construction, forestry, manufacturing, mining, pulp and paper, and steel) and services (such as education, electrical utilities, food, healthcare, municipal, services and transportation). Occupation was also collapsed into manual, mixed and non-manual using a classification system that groups standard occupational codes by physical demand.29 This was done in order to ensure claim numerators matched estimated denominators of the annual number of workers and work hours for the insured Ontario workforce. Due to the group level nature of the data, each unique combination of independent variables (ie, gender, age, industry, occupation and job tenure) had a corresponding number of ‘events’ (ie, number of claims) and number of full-time equivalents that were used to calculate claim rates. Estimates of the number of workers and work hours for each subgroup (ie, denominators for claim rates) were derived from Statistics Canada's Canadian Labour Force Survey.30 Further details on the methodology for estimating denominators for the insured Ontario workforce can be found in a paper by Smith et al.31
A descriptive analysis included calculating claim rates using number of claims per 1000 full-time equivalents (2000 work hours=1 full-time equivalent) and RR for each predictor (age, gender, industry, occupation, year, job tenure). A negative binomial regression model was used to estimate the RRs and 95% CIs of lost-time claims by job tenure (grouped according to number of months at the job), using workers with 13+ months of job tenure as the reference group. A negative binomial model, which can be viewed as special case of the Poisson distribution that includes a random component,32 was used because it was determined in a previous study to better reflect the heterogeneity that exists in the true rates of injury.8 The multivariate analysis of first-time claim rates was performed by having gender, age, occupation, industry, calendar year and job tenure entered simultaneously as predictors.
Following analytic procedures outlined by Bailer and colleagues,33 we examined the degree of effect modification by conducting a series of separate regressions on each level of an explanatory variable (specifically, gender, age and industry), while adjusting for confounding variables. To further examine the modifying effects of gender, age and industry, we calculated claim rates per 1000 full-time equivalents by each gender, age and industry for each job tenure group, using direct standardisation techniques described by Hennekens and Buring.34 All analyses were conducted using SAS V.9.2.35
The interaction between job tenure and the type of event leading to an injury was examined descriptively using claim counts and percentages. Injury events were grouped into six different categories based on the Canadian Standards Association Z-795 coding system:36 bodily reaction or exertion, repetitive motion, contact with objects or equipment, falls, exposure to harmful substances/environments, and other events or exposures. The interaction between job tenure and the source of injury was similarly described. Injury source was grouped into ten categories: chemicals; containers; furniture and fixtures; machinery; parts and materials; persons, plants, animals or minerals (which includes, for example, personal bodily motion injuries, infectious and parasitic agents from living organisms, HIV infection, or exposure to metallic substances such as plutonium/radium which cause radiation injuries); structures and surfaces; tools and instruments; vehicles; and other source of injury.
After excluding workers with a previous injury in the last 2 years, we were left with a total of 950 906 first-time claims (from 1999 to 2008), which was approximately 60% of all the lost-time claims from 1999 to 2008. Of the 950 906 first-time claims, 614 387(64.6%) worked in companies with mandatory insurance coverage. We did not include industrial subsectors without mandatory claim reporting procedures because claim rate denominators could not be calculated for these industrial subgroups. Therefore, those industry groups with voluntary or mixed coverage or those covered under an alternate schedule (Schedule 2) with the Workplace Safety and Insurance Board of Ontario were excluded. This includes companies that opt to pay premiums in return for coverage, and companies that opt not to pay insurance premiums but personally cover the costs of accident claims filed by their employees. There is no administrative definition of what types of firms are likely to be covered under either category. Of these 614 387 lost-time claims, 313 (0.05%) were missing information on age of the worker, 137 (0.02%) were missing information on gender of the workers, 32 908(5.4%) were missing information on occupation, 978 (0.16%) were missing information on industry and 30 222 (4.9%) injury reports were missing information on job tenure. After removing those claims with one or more missing variables a sample of 549 829 injury reports with complete information remains.
Results
Table 1 presents unadjusted claim rates per 1000 full-time equivalents, claim and full-time equivalent counts, and RR by personal and work-related factors, year and job tenure categories for the period between 1999 and 2008. As expected, men had a higher injury rate than women, and those working in manual or mixed occupations had higher injury rates than those employed in non-manual occupations. There is also a decline in injury rates from 1999 to 2008. As well, injury rates declined with longer job tenure categories. Compared with workers employed at a job for 13 months or longer, workers in shorter job tenure categories showed higher RR of lost-time claims; most notably, a more than threefold greater risk for workers in their first month (RR=3.21; 95% CI 3.17 to 3.25).
Table 2 presents the adjusted RRs for lost-time claims by length of job tenure. All workers in job tenure groups of 12 months or less showed an elevated claims rate when compared with workers with 13+ month job tenure. To investigate the effect of job tenure, we examined the adjusted RRs for new workers when compared with those with 13+ months of job tenure. There was a consistent trend for increased risk in all worker subgroups for those with shorter job tenure when comparing the 3–4 months, 2 months and 1 month tenure groups. Men with shorter job tenure had a higher risk of claims than women in the same job tenure group. The elevated risk for men in their first month on the job was reduced with increasing job tenure. For instance, comparing the risk in men with 1 month tenure (RR=3.17; 95% CI 3.00 to 3.34) with those with 9–12 months tenure (RR=1.46; 95% CI 1.38 to 1.54), the risk had decreased by more than half with time on the job. For all job tenure categories, age groups older than 25 years old (25–34, 35–44, 45+) showed an elevated risk compared with young workers (15–19 and 20–24 years old). The oldest age group (45+ years old) had the highest risk in the first month on the job. The 25–34 and 35–44 age groups had the highest risk in the remaining job tenure categories. All occupations showed an elevated risk in new workers compared with those with 13+ months of job tenure, with manual jobs having the highest elevated risk in new workers.
To further examine the association between job tenure and work injury by gender, age and industry, figure 1 presents the adjusted claim rates per 1000 full-time equivalents for each subgroup and by job tenure group. Men showed a greater increase in claim rates when new to the job than women, particularly in their first month at a job. Young workers (15–19, 20–24 years old) new to a job had lower claim rates than adult workers (25+) who were new to a job, however, the claim rates for all age groups appear to converge with increasing job tenure. Workers employed in the goods industry had higher claim rates than those employed in the services industry, with rates converging once workers were employed at a job for more than 13 months.
As specified in our objectives, we were interested in examining whether there was convergence between different job tenure groups over a 10-year period during which there was an overall decline in work injury claims. Figure 2 presents the trends in adjusted lost-time claims from 1999 to 2008 by job tenure. Rates for all job tenure groups generally trended downward over the period, although those with shorter job tenure (ie, 1 month and 2 months) had steadily higher claim rates over time. With the exception of the 1-month tenure group, the differences in claims rates across the remaining job tenure groups narrowed by 2008. However, the claim rates for the 1-month tenure group remained substantially higher than those with a longer job tenure.
Table 3 shows the number and percentage of lost-time claims by the type of event that led to injury and by the source of the injury across job tenure groups. The percentage of claims occurring due to injuries from contact with objects/equipment, falls or exposure to harmful substances showed an inverse relationship with job tenure, where injuries of this kind decreased with more experience. In contrast, the percentage of claims due to body reaction/exertion increased with longer job tenure. When the source of injury was machinery, parts/materials, structures/surfaces or tools/equipment, there was a steady decrease in the percentage of claims with increasing job tenure. Conversely, as job tenure increased, a larger percentage of claims were made up of injuries due to persons, plants, animals or minerals.
Discussion
Lost-time claim rates for workers in their first year at a job were found to be significantly higher compared with those employed at a job for more than 1 year. Second, we found that the elevated claim rates for workers in their first year at a job remained after adjustment for any potential confounding due to gender, age, occupation and industry. This is consistent with previous research that found increased vulnerability of newly hired workers.8
In this study, the increased risk found for new workers was stronger among adult workers in their first month on the job than younger workers, consistent with a previous study that found the often cited increased injury risk of young workers was partially accounted for by job tenure.8 While young workers had lower claim rates than adult workers within the same job tenure group, claim rates appeared to converge with 1 year on the job. In contrast, a study of male railway workers found that younger age groups (<25 and 26–34-year-olds) had significantly higher injury incidence rates in their first 1–2 years of service compared with older age groups.6 We found that consistent with previous findings,8 ,13 new male workers had higher claim rates compared with new female workers, and workers employed in manual or mixed occupations had higher claim rates compared with those in non-manual occupations.
Potential explanations for our finding that adult workers (25+ years old) in their first month on the job had increased claim rates compared with younger workers may be partly due to a lesser capacity for these workers to be able to adapt quickly to changing work environments and tasks.37 It may also be due to increased physical vulnerabilities that tend to come with aging.38 Alternatively, if older workers have more previous work experience, then the observed elevated claim rates could be partly due to the cumulative effects of exposure to hazardous working conditions. Lastly, older workers may be differentially assigned to more hazardous working conditions than their younger counterparts due to the misperception that they have past work experience. The difference in the relationship between job tenure and work injury by age should continue to be examined over time in order to explore the effects of changing trends in the labour force.
The finding that the effect of job tenure differed by gender, where men had a higher claim rate than women in the same job tenure group, may be partially representative of the differing job tasks and exposures which are more predominantly associated with male labour force participation.39 We also found some differences in the type of events leading to injury and the source of injury by job tenure groups. As expected, events that lead to more acute injuries, such as contact with objects/equipment, falls and exposure to harmful substances, made up a greater percentage of claims among those with shorter job tenure. These injuries may be due to the elevated risk workers experience when they perform new and unusual tasks40 or the differential assignment of work tasks to workers in the start of their employment leading to greater hazard exposures. This finding is consistent with a recent study that found new healthcare workers had a decreased risk of musculoskeletal sprain and strain injuries, but an increased risk of cut and puncture injuries.22 Those with shorter job tenure also had a higher percentage of injuries occurring due to sources such as machinery, parts, equipment, tools or surfaces. This may be due to better knowledge of and experience with the tools and work environment by those who have been on the job longer.20 It may also be due to new workers being exposed or assigned to more hazardous working conditions than those who have been employed at a job longer.10 The mechanisms underlying these findings should be further assessed with a detailed examination of workplace hazards and work organisation and relationships as it relates to new workers.
The results of our study should be interpreted given the following limitations. There may be underreporting of injuries to compensation agencies that may result in an underestimation of the true injury rates among workers.41 It is also possible that differences in reporting of injuries to compensation agencies may be present across subgroups of workers. Another limitation is the categorisation of industry and occupation into broad groups of goods and services, and manual, mixed and non-manual, respectively. This was done in order to ensure claim numerators matched estimated workforce denominators. Maintaining detailed occupational information may be useful in providing further evidence on the role of the work tasks and exposures. Also, deriving job tenure using the date of hire and injury date might be inaccurate in some instances where workers changed positions within the same company during this period and had new tasks similar to being new to a job. This type of misspecification of job tenure, however, would lead to an underestimation of the influence of job tenure because new workers at a higher risk (those with a job change) would have been counted among the more experienced workers. Similarly, we were unable to take into account the potential confounding effects of previous job history. Last, while it is not uncommon for claimants to have multiple claims over time, in order to rule out the possibility that a previous injury might increase one's susceptibility for re-injury at a new job, workers with an injury in the previous 2 years were excluded; however, this limits the generalisability of the findings to workers without a previous injury (within the last 2 years).
Our study was the first to examine claim rates by job tenure groups over a 10-year period during which rates of lost-time claims declined in Canada. Our findings did not show a convergence in claim rates between job tenure groups; instead, we found that lost-time claim rates were significantly higher for workers with shorter job tenure compared with those employed at a job for more than 1 year. Therefore, this study makes an important contribution by demonstrating that the relationship between job tenure and lost-time claim rates has remained remarkably stable over the time period between 1999 and 2008. While there are many factors that have contributed to the declining rate in work injuries over the past decade, such as demographic changes in the labour force, increased attention to workplace safety and health management within the workplace, and larger scale prevention efforts,23 it seems these factors were not as successful in reducing the lost-time claim rates of newly hired workers.
Based on our findings, the need for more prevention efforts targeted at new workers is evident. Developing effective safety management systems for new workers may help alleviate some of the potential mechanisms underlying the risk for new workers. Prevention activities should involve employers and supervisors in order to create strategies at an organisational level to mitigate situations where new workers are placed at an increased risk of hazardous exposures. Our results provide policy makers in occupational health and safety with evidence to guide the targeting of resources and policy changes. On the basis of our results, we suggest promoting practices that reduce job turnover, encourage permanent employment and improved job security as one way to prevent the issue of injuries due to being new at a job. This is particularly important in industries where the level of temporary employment is increasing and turnover rates are high. We further suggest increasing new workers’ knowledge of their workplace by ensuring that new workers receive proper safety training, including information on workplace hazards and proper use of workplace machinery and equipment. Our results also highlight the need to further examine work injury rate trends and the factors influencing these trends.
Acknowledgments
The access to the data used for this study was provided by the Ontario Workplace Safety and Insurance Board (WSIB). Approval for the secondary data analyses was obtained through the University of Toronto Research Ethics Board, Health Sciences committee.
References
Footnotes
-
Contributors SM wrote and revised the paper. FCB designed the study and provided comments on the draft paper. MS cleaned and analysed the data, and provided comments on the draft paper. PS planned the statistical analysis, analysed the data and provided comments on the draft paper.
-
Funding PS is supported by a New Investigator Award from the Canadian Institutes of Health Research.
-
Competing interests None.
-
Ethics approval University of Toronto Research Ethics Board, Health Sciences Committee.
-
Provenance and peer review Not commissioned; externally peer reviewed.