1. Introduction
It has been reported that persons with severe traumatic injuries are more vulnerable to become depressed than the rest of the population [
1,
2,
3,
4]. The reverse situation has also been demonstrated: individuals with pre-existing mental health problems or depressive symptoms show an increased likelihood of physical disability, or traumatic injury [
5,
6,
7,
8]. One study documented a bidirectional relationship between unintentional injury and depression, such that individuals who are exposed to more traumatic injuries are at increased risk of developing depression, and more severe depression is associated with a higher likelihood of subsequent traumatic injuries [
9].
Depression and injury are common health problems in the working population [
10,
11]. Suggestive evidence exists for each direction of the association between occupational injury and depression; workers with occupational injury have been reported to be more likely to be depressed [
12,
13,
14,
15,
16], and workers with depressive symptoms or psychosocial job stress have exhibited an increased likelihood of injury at work [
17,
18,
19]. However, the empirical literature is limited by a lack of longitudinal studies in examining a reciprocal link between injury and depression in the workplace setting. In assessing the association between occupational injury and depression, methodological concerns regarding confounding and reverse causation are important because the positive findings may not be attributable to depression or injury per se, but may instead be created spuriously from confounding or reverse causation. Most previous cross-sectional studies have failed to address these concerns adequately. Only one study examined the bidirectionality of the relationship between traumatic injuries and depression in a prospective setting, but it was neither about workplace injury nor differentiated the effect by gender [
9].
Previous studies often neglect gender specificity by adjusting for or matching gender in their analysis [
20]. It has been shown that women have higher prevalence of depression than men [
10,
21], while men are more vulnerable to injury than are women [
22,
23]. However, no study has yet investigated simultaneously how the bidirectional association between occupational injury and depression differs by gender in a prospective setting. Therefore, the present study aimed to assess the bidirectionality and gender differences in the longitudinal association between work-related injuries and depression within the same cohort. Each direction of the association between occupational injury and depression was contrasted and the pattern of association by gender was examined for differences, using data from a nationally representative survey.
4. Discussion
This study analyzed whether there is a bidirectional association between occupational injury and depression within the same cohort of the US working population. Our findings suggest that a modest association does exist, but the association differs by gender. Male workers showed significantly higher risk of getting depressed after occupational injury, but female workers did not. Female workers with depression faced a greater risk of being injured on the job as compared to those without depression, but this association was not significant among male workers, as revealed in the multivariate model.
The directional strength of the association between depression and occupational injury was opposite in men and women. In other words, injured male workers show higher odds of getting depressed than non-injured male workers, while depressed female workers show higher odds of getting injured at the workplace than non-depressed female workers. The only study reporting a bidirectional relationship between occupational injuries and depression suggested a stronger effect in women [
9]. However, it was not directly comparable to our study, because it combined the data on males and females and then compared the effect of depression/injury on both genders. Gender differences in the association between workplace injury and depression appear to be very complex. It appears to be a product of both differences in depression between male and female workers as well as in injury characteristics. Men and women are exposed to different working conditions and they react differently. The different levels of severity or nature of injury may also explain the differential impact of occupational injury and depression by gender.
Several studies have reported an association between occupational injury and depression in one direction or the other [
9,
18,
29,
33]. Risk factors for post-injury depression in our study were consistent with the previous studies: female gender, injury severity, and higher levels of perceived stress or pain [
14,
34,
35]. Although female workers tend to get depressed more after an injury, our subgroup analysis by gender revealed a relatively higher impact of occupational injury on depression in males. A previous study from that used linked WC and medical insurance data found that the before- and after-injury increase was larger for men although fewer men were treated for post-injury depression [
36]. This may suggest that men are more deeply affected by occupational injuries than are women. Men had greater sensitivity than women to the depressogenic effects of financial, occupational, and legal stress [
37,
38].
Our finding suggests that the psychological stress associated with the WC process may lead to elevated depression among male workers. We hypothesized that the distress related to litigation may act as an intermediate factor in the pathway from occupational injury to depression for male workers. A few studies have suggested that the very process of dealing with WC has a deleterious effect on the post-injury trajectory of injured workers [
39,
40]. In a Canadian study, the experience of litigation substantially explained the level of depression in workers who had suffered mild-to-moderate traumatic brain injury [
41]. WC payments was a crucial factor in explaining male workers’ depression after an occupational injury, but this did not hold true for female workers. For both genders, more severe injury, musculoskeletal injury and longer time after the injury were significant predictors for post-injury depression.
Our finding that depression was associated with higher risk of occupational injury occurrence among females is consistent with the previous study [
17]. Peele reported that depression may serve as a precursor to occupational injury for women. In our study, taking anti-depressant medications yield significantly higher odds (OR = 1.43, 95% CI: 1.16, 1.75) for occupational injury among female workers. Like anti-depressants, increasing duration of depression and number of depression episodes showed a similar pattern of higher risks for occupational injury among females, but not among males.
Most previous studies examining the relationship between anti-depressants and risk of injury have been confined to falls or fractures in the elderly [
5,
6,
42], or traffic injuries in drivers [
43]. Only few studies reported the psychotropic effects of medication and the risk of injury on the job, but their results were mixed [
29]. In a case-control study conducted on the British population [
18], anti-depressants were found to be related to a higher risk of occupational injury, and there were only marginal differences between gender in terms of this risk. A Canadian study using a national health survey [
9] reported no association between anti-depressants and injury risk. The relatively higher impact of antidepressants on workplace injury risk in women implies that anti-depressants may play a different role in the association of depression and occupational injury between genders. Women seem to be more vulnerable to the detrimental effects of anti-depressants, especially in terms of occupational injury risk. This might be explained, at least partially, by different pharmacokinetics [
44,
45] and different occupational risk factors among men and women [
46], including the fact that women more often occupy part-time jobs and that they potentially face different levels of exposure to danger and stress on the job.
Several limitations of the present study need to be considered. First, the dataset did not include information on several potential confounding variables, including family history, other stressful life events outside the workplace, detailed job descriptions, and psychosocial conditions in the workplace. Lack of this data limited the exploration of the potential mechanisms underlying the observed gender difference in the association between occupational injury and depression, and vice versa. Second, the information from the MEPS was self-reported. The self-reported depression/injury measure, if under-reported, was more likely to attenuate the observed association between occupational injury and depression. In sensitivity analysis of excluding participants with the self-reported mental health status of being “poor”, the association between occupational injury and depression was attenuated in both directions, regardless of gender (
Supplementary Materials Table S1). Self-reported measures may also introduce recall bias, although the intervals between the rounds in this study were relatively short. Additionally, it was verified that the error rate for coding medical conditions based on the ICD-9 codes did not exceed 2.5% [
47]. Therefore, it is unlikely that our results were biased by the self-reported data on medical conditions on the MEPS. Third, history of injury or depression was assessed based on data from Round 1. There was an interval of 4–5 months between subsequent interview rounds. A longer time period may have been necessary, especially for a history of depression to be evident. Additionally, due to these intervals, we were not able to capture outcomes soon after the onset of their injury/depression. Forth, there were not enough cases for male injury after depression for Analysis 2, which may lead to inadequate power to detect a relationship should one actually exist. The interpretation of results should be cautious accordingly. Finally, the attrition rate of the MEPS may not be random. The initial response rate on the MEPS was over 85%, but 30% of the respondents had been lost by the fifth round. This could lead to a potential bias, although the AHRQ reported that it had no evidence of potential non-response bias attributable to survey attrition on resultant national estimates of health care cost [
48].
Despite these limitations, this study adds to the evidence on gender difference and directionality in the association between occupational injury and depression, within the same cohort, using a nationally representative population-based survey. By using the longitudinal feature of the MEPS—its repeated measures of occupational injury and depression over time—we were able to establish the temporality of the association between occupational injury and depression, and vice versa. The rich information for all medically treated or related health conditions enabled us to take into account comorbidity, activity limitation, and severity of injury or depression, which can affect the association between workplace injury and depression. In addition, by adjusting for potential confounders that may mask the true effect of depression or injury, and by excluding cases with pre-existing injury and depression at baseline, we were able to take into account the potential bias due to reverse causation.