Smith, Peter M. PhD; Ibrahim-Dost, Jihan; Keegel, Tessa PhD; MacFarlane, Ewan PhD
Work-related injury and illness are an important public health concern. In Australia, the direct and indirect costs of work-related injury and illness for 2008 and 2009 have been estimated to be in excess of $60 billion, with the majority of this cost being borne by injured workers.1 A nontrivial proportion of work injuries can be attributed to employment in nondaytime shift schedules.2,3 Canadian studies have estimated the population-attributable risk associated with shiftwork on work injury to be between 12% and 14% for women and 6% to 8% for men.4,5 Shiftwork is defined as an employment system whereby workers rotate through set periods throughout the day, typically for companies or businesses that require 24-hour workforce. In 2005 and 2006, 16% of Australian workers were engaged in shiftwork arrangements.6 One of the primary pathways through which shiftwork results in an increased injury risk is via the disruptions in circadian rhythm due to the changes in the sleeping patterns associated with a shift-work schedule. This in turn leads to excess fatigue and a lower ability to maintain levels of attention and concentration at work.7–9
Studies have reported that the risk of injury associated with shiftwork, compared with nonshiftwork, differs for men and women, with the injury risk associated with shiftwork being higher among women than among men.4–6 This excess risk among female shiftworkers is brought about through both a lower risk of injury among non–shiftworking women than among non–shiftworking men, as well as a slightly increased risk of injury in shiftworking women compared with that in shiftworking men.4,6 A potential pathway that might lead to a particularly increased risk of injury among female shiftworkers is via domestic strain. Female labor market participants often have greater childcare and household responsibilities outside work than male labor market participants.10,11 The elevation of stress due to household or childcare activities and the decreased time for recovery and sleep have been linked to adverse consequences in the workforce.12 Specifically related to the risk of injury, increased domestic and childcare responsibilities may lead to less recovery and sleep among female shiftworkers, and subsequently increase their risk of work injury.
Despite the plausibility of this hypothesis, to date, the relationships between domestic and childcare responsibilities, shiftwork, and gender have not been explored. The objective of this study was to address this research gap by examining the relationship between shiftwork and work injury among men and women, taking into account the presence of dependent children in the household. We hypothesize, similar to previous work, that the relative risk of shiftwork on work injury will be greater among women than among men. We further hypothesize that gender differences in the relative risk associated with shiftwork will be greatest among the respondents living with dependent children, given that among these respondents, differences in nonwork commitments will be most pronounced. We also hypothesize that the relative risk associated with shiftwork will be greater among women with dependent children than among women without dependent children.
This study used data from the 2009–2010 Multipurpose Household Survey (MPHS) conducted by the Australian Bureau of Statistics. The MPHS is an annual supplement to the Australian Labour Force Survey, designed to provide statistics for a small number of labor, social, and economic topics.13 The sample for the MPHS consists of one-third of the dwellings in the outgoing panel of the Labour Force Survey each month (one-eighth of the Labour Force Survey sample is rotated out each month). Respondents from remote parts of Australia, members of the defense forces, certain diplomatic personal of overseas governments, overseas residents of Australia, and people living in nonprivate dwellings (eg, hotels) are excluded. The 2009–2010 MPHS covered six topics, one of which was work-related injuries. The sample asked questions on work-related injuries consisted of 28,554 respondents. The response rate for the MPHS was 97%.13
For the purpose of these analyses, we restricted our sample to non–self-employed respondents who had worked in the previous 12 months (n = 15,721). We also removed respondents who were not currently working and did not have a work injury in the previous 12 months (n = 1393; 8.9% of the original sample). We did this because respondents who were not working were not asked for information on the duration of their previous job or the number of hours worked. This did not allow hours of exposure to work to be estimated among this group, which is relevant when examining injury rates across age or gender groups.14 This left a final sample of 14,328 respondents. Analyses were conducted through the Remote Access Data Laboratory at the Australian Bureau of Statistics. Approval for the secondary analyses of the data was granted by the Monash University Human Research Ethics Committee.
Main Outcome: Work injury
The respondents were asked to report any work-related injury, illness, or disease that they first became aware of in the previous 12 months. This included work-related injuries that occurred while commuting to and from work, outside work but while on duty, or during work breaks.
Main Independent Variable: Shiftwork
Respondents were asked whether they worked under shift arrangements in their current job or whether the job in which they were injured was not the same as their current job. Shift arrangements were defined as where the daily hours of operation at the respondent's place of employment were split into at least two set work periods (shifts) for different groups of workers.
Main Moderator Variable: Presence of Dependent Children in the Household
The presence of dependent children in the household was derived from a variable describing the family composition of the respondent. Households where children younger than 15 years reside are defined as having dependent children. Using this information, respondents were classified as living with dependent children (yes or no).
Covariates included in the analyses included the age of the respondents (grouped), the marital status of the respondents (married or cohabitating vs single), whether they were born in an English-speaking country (as a proxy for mother tongue), whether they worked on a fixed-term contract (yes or no), the duration of their current job (less than 3 months, 3 to 12 months, and more than 12 months), the usual hours worked (less than 24 hours, 25 to 40 hours, 41 to 50 hours, and more than 50 hours); and their occupation (management and professional, sales and service occupations, other service occupations, trades occupations, and manufacturing and laboring occupations) on the basis of the Australian and New Zealand Occupational Classification System.15
Our initial sample totaled 14,328 respondents. Of this sample, 61 respondents (0.4%) with missing information on their hours of work or their occupation were removed from our sample, leaving a final sample of 14,267 respondents. We ran a series of regression models examining whether age, gender, shiftwork, dependent children, or work injury was associated with missing responses. No relationship was observed between any of these variables and increased probability of item-level nonresponse.
Initial descriptive analyses examined the percentage of work injuries across all study variables separately for men and women. Multivariate logistic regression models then examined the likelihood of work injury among shiftworkers (vs nonshiftworkers), adjusting for all covariates. Additional models were run separately for respondents with dependent children and those without dependent children. We then compared the difference in the association between shiftwork and work injury for men and women, in the full sample, and separately for the sample with dependent children and those without dependent children, using previously described methods to calculate the statistical difference between coefficients across regression models.16,17 We also examined the differences in the association between shiftwork and work injuries for respondents with dependent children and those without, separately for men and women. The results from post hoc tests of differences in estimates are similar to those that would be obtained from an interaction term between gender and shiftwork or dependent children and shiftwork being included in a regression model (results from interaction models are not presented in this article but are available from the authors on request). All analyses were weighted to account for the probability of selection to the initial survey and unit level nonresponse. Although the MPHS was administered by using a clustered survey design, it is currently not possible to adjust analyses for this design effect through the Remote Access Data Laboratory at the Australian Bureau of Statistics. It is, therefore, possible that the standard errors around our point estimates may be underestimated. As such, we recommend caution in interpreting the significance of variables with marginal statistical significance (P = 0.05).
Table 1 presents descriptive information on the percentage of work injuries across the study variables, stratified by gender. In total, 6.2% of women and 6.6% of men reported a work injury in the previous 12 months. Among both men and women, there seemed to be an elevated risk of work injury among those working in shifts, with the differences between the percentage of injuries among shiftworkers and that among nonshiftworkers slightly larger among female respondents (6.2% difference among women compared with 4.1% among men). Other variables associated with an elevated percentage of work injuries included trades occupations and occupations in manufacturing and laboring. Long work hours (more than 41 hours) were also associated with a high percentage of work injuries among female respondents. Being born in a non—English-speaking country was associated with a lower percentage of work injuries.
Table 2 presents the adjusted odds ratio for shiftwork and work injury stratified by gender, for all respondents, and separately for respondents with dependent children and those without. Among all respondents, the odds ratios associated with shiftwork were higher among female respondents (odds ratio, 2.43; 95% confidence interval, 1.95 to 3.03) than among male respondents (odds ratio, 1.56; 95% confidence interval, 1.26 to 1.94) (for difference in estimates, P = 0.005). Nevertheless, the difference in the risks associated with shiftwork across gender were amplified among respondents with dependent children (for difference in estimates, P = 0.001) and attenuated among respondents without dependent children (for difference in estimates, P = 0.14).
Table 3 presents the adjusted odds ratios for shiftwork and work injury for respondents with and without dependent children, presented separately for men and women. Note that the odds ratios in Table 3 are the same as those presented in Table 2, with the difference in these tables being the estimates examined for differences. Among male respondents, no differences in the association between shiftwork and work injury were present for respondents with dependent children versus those without dependent children (for difference, P = 0.80). Nevertheless, among female respondents, the association between shiftwork and work injury was stronger among women with dependent children than among those without dependent children (for difference, P = 0.02).
The objective of this study was to examine whether gender differences in the relationship between shiftwork and work injury were similar when taking into account the presence of dependent children in the household. We found that the risks associated with shiftwork and work injury were greater among women than among men; however, this difference in risk was most prominent among respondents with dependent children. Furthermore, we found that while among men no differences were present in the effects of shiftwork on work injury for those respondents with dependent children versus those without dependent children, the work injury risk associated with shiftwork was higher for women with dependent children than for those without.
These findings, however, should be interpreted by taking into account the following limitations. All our data are based on self-report. As such, some of our associations may be inflated because of common-method bias. There may be selective recall bias associated with work-injury, which may lead to some members of our sample being defined as not having an injury when in fact they did. In addition, the general work injury question may pick up some minor injuries. It should be noted, however, that only 6.4% of our sample reported a work injury. Previous self-reported data from Canada suggest that 2.5% of labor market participants have traumatic activity-limiting injuries over a 12-month period and 7.2% have activity-limiting repetitive movement activities.18 In addition, the rate of injuries in this sample is approximately five times as high as the rate of serious workers' compensation claims in Australia, which are those that involve deaths, permanent incapacities, or absences of 1 week or more from work.19 Taken together, these comparisons suggest that the percentage of injuries in our sample is consistent with injuries that would require some medical attention or result in some activity limitation. We were unable to explore whether particular types of shift schedules were more strongly associated with injury or influenced by children in the household than others. The previous study by Wong and colleagues5 detected the largest gender difference in work injury among rotating shifts, although both night shifts and rotating shifts were associated with a similar risk of injury among women. The absence of detailed information on the type of shift schedule in our sample likely biased the estimates associated with shiftwork and interactions between shiftwork and gender to null. Future studies should explore whether the presence of children has a differential effect on work injury risk among women working different shift schedules. In addition, we were unable to adjust our models for other important factors associated with work injury, including job tenure20 and workplace size.21,22 Nevertheless, our study also has a number of strengths, which include a large, generally representative sample of the Australian labor market, with a high response rate and a low percentage of item-level nonresponse. This allowed for one of the first examinations of the relationship between shiftwork and work injury across gender, taking into account childcare responsibilities.
Our finding of an increased risk associated with shiftwork, in general, but in particular among women relative to men is consistent with previous research in this area.2,4,5 The consistency of findings across these studies suggests a true elevated risk associated with shiftwork among women relative to men, independent of the source of injury information. The study by Wong and colleagues5 used the receipt of workers' compensation payments as a proxy for work injury, the study by Mustard and colleagues4 used accepted lost-time injury claims and emergency department records, and we have used self-reported work injuries in the previous 12 months. In this article, we have also documented that the elevated odds ratio associated with shiftwork among women compared with that among men is most pronounced among respondents with dependent children. In addition, among women, the odds ratio associated with shiftwork and work injury is higher among those with dependent children than among those without. Taken together, these findings support our hypothesis that increased childcare responsibilities may lead to less recovery and sleep among female shiftworkers and subsequently increase their risk of work injury compared with that in both male shiftworkers with dependent children and female shiftworkers without dependent children.
The Australian Bureau of Statistics estimates that 16% of the Australian labor market (approximately 1.4 million workers) work in shiftwork arrangements.23 In this article, we report that the risks associated with shiftwork are elevated among female labor market participants compared with those in their male counterparts. We have also demonstrated that the gender difference in the risks of shiftwork and work injury are exacerbated among respondents with dependent children. These findings demonstrate that the elevated risk of injury associated with shiftwork among women compared with that among men is likely to be a result of maternal responsibilities of women in the labor market. On the basis of our findings, we suggest that primary injury prevention efforts that target employees engaged in shiftwork need to take into account the gendered nonwork responsibilities of these employees in their design.
1. Safe Work Australia. Cost of Work-Related Injury and Disease for Australian Employers, Workers and the Community: 2008–09. Canberra, Australia: Safe Work Australia; 2012.
2. Dembe AE, Erickson JB, Delbos RG, Banks SM. Nonstandard shift schedules and the risk of job-related injuries. Scand J Work Environ Health. 2006;32:232–240.
3. Knutsson A. Health disorders of shift workers. Occup Med (Oxford, England). 2003;53:103–108.
4. Mustard CA, Chambers A, Bielecky A, McLeod CB, Smith PM. Work injury risk by time of day in two population-based data sources. Occup Environ Med. 2013;70:49–56.
5. Wong IS, McLeod CB, Demers PA. Shift work trends and risk of work injury among Canadian workers. Scand J Work Environ Health. 2011;37:54–61.
6. Safe Work Australia. The Impact of Shiftwork on Work-Related Injuries in Australia. Canberra, Australia: Safe Work Australia; 2009.
7. Folkard S, Lombardi DA, Spencer MB. Estimating the circadian rhythm in the risk of occupational injuries and accidents. Chronobiol Int. 2006;23:1181–1192.
8. Folkard S, Lombardi DA, Tucker PT. Shiftwork: safety, sleepiness and sleep. Ind Health. 2005;43:20–23.
9. Akerstedt T, Fredlund P, Gillberg M, Jannson B. A prospective study of fatal occupational accidents—relationship to sleeping difficulties and occupational factors. J Sleep Res. 2002;11:69–71.
10. Duxbury L, Higgins C. The 2001 National Work-Life Conflict Study. Ottawa, ON, Canada: Health Canada; 2002:1–99.
11. Berntsson L, Lundberg U, Krantz G. Gender differences in work-home interplay and symptom perception among Swedish white-collar employees. J Epidemiol Community Health. 2006;60:1070–1076.
12. Hammig O, Bauer G. Work-life imbalance and mental health among male and female employees in Switzerland. Int J Public Health. 2009;54:88–95.
13. Australian Bureau of Statistics. Technical Manual: Work Related Injuries, Expanded CURF, Australia 2009–10. Canberra, Australia: Commonwealth of Australia; 2011.
14. Breslin FC, Koehoorn M, Smith PM, Manno M. Age related differences in work injuries and permanent impairment: a comparison of workers' compensation claims among adolescents, young adults, and adults. Occup Environ Med. 2003;60:e10.
15. Australian Bureau of Statistics. ANZSCO—Australian and New Zealand Standard Classification of Occupations. Canberra, Australia: Australia Bureau of Statistics; 2005.
16. Allison PD. Comparing logit and probit coefficients across groups. Sociol Methods Res. 1999;28:186–208.
17. Austin PC, Hux JE. A brief note on overlapping confidence intervals. J Vasc Surg. 2002;36:194–195.
18. Smith P, Bielecky A, Mustard C. The relationship between chronic conditions and work-related injuries and repetitive strain injuries in Canada. J Occup Environ Med. 2012;54:841–846.
19. Safe Work Australia. Compendium of Workers' Compensation Statistics: Australia 2009–10. Canberra, Australia: Safe Work Australia; 2012.
20. Breslin FC, Smith PM. Trial by fire: a multivariate examination of the relation between job tenure and work injuries. Occup Environ Med. 2006;63:27–32.
21. Fabiano B, Curro F, Pastorino R. A study of the relationship between occupational injuries and firm size and type in the Italian industry. Saf Sci. 2004;42:587–600.
22. McVittie D, Banikin H, Brocklebank W. The effects of firm size on injury frequency in construction. Saf Sci. 1997;27:19–23.
23. Australian Bureau of Statistics. Shift workers. Aust Labour Market Stat. 2010;2010:20–24.
Copyright © 2013 by the American College of Occupational and Environmental Medicine