Introduction
Attendance and performance at work are not only dependent on competence and knowledge; but a multitude of health-related, personal, and work-related factors also play an important role. Work loss and suboptimal work performance are not only a burden for the individuals themselves, but also for their colleagues and employers, their families and wider environment, and for society in general since both increase employer and healthcare costs [Reference Buist-Bouwman, De Graaf, Vollebergh and Ormel1,Reference De Graaf, Tuithof, Van Dorsselaer and Ten Have2]. Some well-known examples of sick-leave predictors are psychosomatic complaints, being unmarried, and low job control [Reference Duijts, Kant, Landeweerd and Swaen3,Reference Duijts, Kant, Swaen, van den Brandt and Zeegers4]. Another important factor that may influence absenteeism and presenteeism (i.e., missing work due to illness and going to work despite illness, respectively) is experiences with childhood trauma, but systematic studies on this topic are scarce. Possible, because the interval between cause and effect may be (very) long, this contributing factor may easily be overlooked. On top of that, the traumatic childhood events that eventually led to work impairment are usually not shared, out of shame, secrecy, and social taboo. Using data from the adverse childhood experiences study, Anda et al. [Reference Anda, Fleisher, Felitti, Edwards, Whitfield and Dube5] reported that individuals with a (self-reported) history of childhood emotional, physical, and/or sexual abuse were more likely to report problems with work absenteeism in adulthood. A more recent study also found evidence of an increased risk of permanent work disability at 55 years in persons with childhood adversity [Reference Fahy, Stansfeld, Smuk, Lain, Horst and Vickerstaff6]. It has been reported on more often that in persons with adverse childhood experiences more school absenteeism or overall poorer school performance was observed [Reference Stempel, Cox-Martin, Bronsert, Dickinson and Allison7,Reference Crouch, Radcliff, Hung and Bennett8]. To the best of our knowledge, however, the association between childhood trauma and presenteeism has as yet not been investigated.
How childhood trauma exactly results in absenteeism and presenteeism many years later has not been examined. However, it seems logical to assume that this can partly be accounted for by the large impact such early adverse events have on a person’s (developing) mental health. The WHO World Mental Health Surveys found strong associations between childhood adversities and overall adult psychopathology [Reference Kessler, McLaughlin, Green, Gruber, Sampson and Zaslavsky9]. Individuals reporting childhood trauma are about twice more likely to develop a depression and/or an anxiety disorder later in life [Reference De Venter, Demyttenaere and Bruffaerts10]. Literature reviews and meta-analyses have demonstrated that childhood trauma is associated with a range of psychological problems and psychiatric disorders, among which are depressive and anxiety disorders. Chen et al. [Reference Chen, Murad, Paras, Colbenson, Sattler and Goranson11] came to the conclusion that a history of sexual abuse correlates with an increased risk of a lifetime diagnosis of (one of) these disorders, which conclusion was confirmed in a recent umbrella review by Hailes et al. [Reference Hailes, Yu, Danese and Fazel12]. Furthermore, there is evidence of a relationship between nonsexual child maltreatment (i.e., physical abuse, emotional abuse, and/or neglect) and depressive and anxiety disorders [Reference Norman, Byambaa, De, Butchart, Scott and Vos13]. The effects of childhood trauma are not just limited to an increased initial vulnerability, but can also have a major impact on the clinical course of the two disorders, which notably is more serious, often chronic, and prognostically unfavorable course of the disease [14–16].
Mental-health problems, especially anxiety and/or depressive episodes, have consistently been linked to worse short- and long-term absenteeism and presenteeism outcomes among working people [Reference Buist-Bouwman, De Graaf, Vollebergh and Ormel1,Reference De Graaf, Tuithof, Van Dorsselaer and Ten Have2,17–22]. The large-scale Netherlands Study of Depression and Anxiety (NESDA) discerned that current depressive disorders were associated with a sevenfold higher risk of long-term work absence and a fivefold higher risk of impaired work performance, whereas current anxiety disorders increased these odds with around a factor two [Reference Plaisier, Beekman, Graaf, Smit, Dyck and Penninx23], which numbers were later confirmed by Hendriks et al. [Reference Hendriks, Spijker, Licht, Hardeveld, Graaf and Batelaan24] in their longitudinal study.
The purpose of the present study is to gain more insight into the association between childhood trauma and work functioning in terms of absenteeism as well as presenteeism, and to elucidate to what extent any associations can be accounted for by a depression and/or anxiety disorder. The results were obtained in an extensive sample of the Dutch working population participating in the NESDA study [Reference Penninx, Beekman, Smit, Zitman, Nolen and Spinhoven25].
Methods
Participants
We extracted our baseline data from NESDA, an ongoing Dutch longitudinal multisite cohort study for which 2,981 adults (aged 18–65 years) were recruited between 2004 and 2007. NESDA recruited from three different settings: the general population, primary care, and specialized mental healthcare and is designed to have a control group without a depressive or anxiety disorder. Participants with depressive and anxiety disorders were recruited in all three of these settings and thereby cover the entire range of mood psychopathology including both persons with and without treatment needs and experiences, whereas healthy individuals were only recruited in the general population and the general practices. The baseline assessments included a structured diagnostic psychiatric interview, after which all participants were monitored to investigate the course and consequences of depressive and anxiety disorders. Candidates not fluent in Dutch and those with a primary clinical diagnosis of psychotic, obsessive–compulsive, bipolar, post-traumatic stress or severe substance use disorder were excluded. The research protocol was approved by the ethics committees of the participating universities and written informed consent was obtained from all participants. For further details on NESDA, see Penninx et al. [Reference Penninx, Beekman, Smit, Zitman, Nolen and Spinhoven25].
In the present study, we describe and analyze the data of the 1,873 participants that had a paid job and worked at least 8 h a week at baseline, thereby excluding the 1,108 participants who were not employed or employed for less than 8 h a week.
Measurements
Work functioning
Work functioning was conceptualized as absenteeism and presenteeism, with absenteeism being assessed with the validated Health and Labor Questionnaire Short Form (SF-HLQ) [Reference Hakkaart-Van Roijen26,Reference Bouwmans, Jong, Timman, Zijlstra-Vlasveld, Feltz-Cornelis and Tan27]. Consistent with Plaisier et al. [Reference Plaisier, Beekman, Graaf, Smit, Dyck and Penninx23], absenteeism was computed by dividing the number of absent days from work during the past 6 months because of health problems by the number of actual work days per week. This variable did not meet normality assumptions and was therefore categorized into three categories: no absenteeism, short-term absenteeism (≤2 weeks), and long-term absenteeism (>2 weeks). This classification also enabled us to make a distinction between absence from work most probably due to self-limiting conditions such as common colds, and work absenteeism, where more chronic conditions were likely to be involved.
Presenteeism was assessed with the validated World Health Organization Disability Assessment Schedule II (WHODAS-II) [28,Reference Chwastiak and Von Korff29]. This 36-item instrument gauges health and disability and provides a work-functioning profile across six activity domains. For our study, we restricted ourselves to the four items of the work-activities domain: Difficulties in (a) day-to-day work, (b) doing most important work tasks well, (c) getting all work done, and (d) getting work done as quickly as needed. The respondents rated the items on a 5-point scale, ranging from 1 (None) to 5 (Extreme/Cannot do). The summated work domain scale has a range from 4 to 20, with higher scores reflecting a higher level of disability.
Childhood trauma
Childhood trauma was assessed retrospectively using the Childhood Trauma Interview (CTI), a structured interview for adults in the age range of 18–64 years first described in the Netherlands Mental Health Survey and Incidence Study (NEMESIS).
During their baseline assessment, all participants were asked whether they had experienced any kind of emotional neglect and/or psychological, physical, or sexual abuse before the age of 16. The definition of emotional neglect was lack of parental attention or support and ignorance of the participant’s problems and experiences as a child. Psychological abuse was defined as verbal abuse, undeserved punishment, subordination to siblings, and being emotionally blackmailed. Physical abuse was specified as being kicked or hit with hands or an object, being beaten up, or being physically abused in any other way, while sexual abuse was delineated as being sexually approached against one’s will in terms of being touched or made to touch someone in a sexual way. Participants answered “Yes” or “No” to each of the four forms of childhood trauma and were asked to give an indication of the frequency of the incidents on a 5-point scale (Once, Sometimes, Regularly, Often, and Very often). As was done in earlier studies [Reference Hovens, Wiersma, Giltay, Oppen, Spinhoven and Penninx30,Reference Wiersma, Hovens, Van Oppen, Giltay, Van Schaik and Beekman31], a cumulative childhood trauma index with sum scores from 0 to 8 was composed, where higher scores were indicative of more types and more incidents of adverse childhood trauma. The interrater reliability of the CTI is 0.73 to 1.0 [Reference Fink, Bernstein, Handelsman, Foote and Lovejoy32] and according to Spinhoven et al. [Reference Spinhoven, Penninx, Hickendorff, van Hemert, Bernstein and Elzinga33] there is a convergent correlation with the Childhood Trauma Questionnaire (CTQ) for trauma severity (r = 0.57–0.61). Data from the CTI was collected at wave 1 (baseline) and from the CTQ at wave 4 of the NESDA study (4 years after baseline).
Depressive and/or anxiety disorder
Our participants were screened for depressive and/or anxiety disorders at baseline by trained clinical staff using the validated Composite Interview Diagnostic Instrument (CIDI version 2.1) [34] that assesses the presence of major depressive disorder, dysthymic disorder, generalized anxiety disorder, panic disorder, agoraphobia, and/or social phobia. We took into account current (6-month recency) depressive, anxiety, and comorbid depressive-anxiety disorder, and remitted depressive and/or anxiety disorder (lifetime but not 6-month recency). The widely used CIDI classifies diagnoses according to DSM-IV criteria and has good overall reliability and validity [Reference Wittchen35].
Covariates
Because we specifically sought to describe the association between childhood trauma and work functioning, and the role of depressive and/or anxiety disorders in this relationship, we limited the number of potentially confounding covariates to sex, age, education (in years attained), the number of working hours per week, and the number of somatic conditions. The latter conditions were assessed with the chronic diseases interview, a 21-item face-to-face interview instrument that was designed for NESDA and provides a sum of self-reported heart disease, diabetes, stroke, cancer, high blood pressure, allergies, intestinal disorders, liver disease, epilepsy, chronic lung problems, injuries, and arthritis. In addition, participants could mention up to five chronic diseases that were not (yet) included in the list [Reference Penninx, Beekman, Smit, Zitman, Nolen and Spinhoven25]. Given the skewed distribution of this variable, it was dichotomized into (0) No somatic condition and (1) At least one somatic condition.
Statistical analyses
Sample characteristics are presented as means with standard deviations or numbers with percentages. Baseline characteristics were compared across the childhood trauma indices using analyses of variance for continuous variables and Pearson’s chi-squared tests for categorical variables.
We subsequently tested depressive and anxiety disorder more specifically as potential contributors to the childhood-trauma and work-functioning associations to see whether they would at least partially explain why adverse early-life events lead to work impairment later in life. To do so, we applied a single model to each outcome measure with four mediators (current depressive disorder, current anxiety disorder, current comorbid depression-anxiety, and remitted depressive and/or anxiety disorder). Because of the dichotomous nature of the mediators, we used probit regression for the mediating models. In the outcome model we applied ordinal probit regression for absenteeism (three categories) and linear regression for presenteeism. Mediation was investigated by directly testing the significance of the indirect effect of the independent variable (IV) on the dependent variable (DV) through the mediator (M), quantified as the product of the effects of the IV on M (a) and the effect of M on DV (b), partialling out the effects of the IV. The 95% percentile-based confidence intervals (CIs) were computed using bootstrap with 1,000 samples. As unstandardized coefficients in probit and linear models are on different scales, all effects were expressed as standardized effects [Reference MacKinnon and Dwyer36]. All models were adjusted for sex, age, education, number of working hours, and the presence of somatic conditions.
Descriptive analyses were conducted in SPPS version 26.0 (SPSS Inc, Chicago, IL) and mediation analyses using R. Statistical significance was set at p < .05.
Results
Baseline sample and outcome characteristics
Of the 1,873 working participants, 223 individuals (11.9%) were excluded due to missing data on work functioning and one additional person due to incompletion of the childhood trauma questionnaire, leaving 1,649 subjects (n = 584 male and n = 1,065 female) for analysis.
Table 1 shows the baseline demographic and clinical characteristics of the study sample according to childhood trauma level. The mean age of the participants was 40.7 years, 64.6% were women, and the mean years of attained education was 13 years. Childhood trauma was reported by 44.8 and 22.4% reported a remitted depressive and/or anxiety disorder, 13.6% a current depressive disorder, 19.0% a current anxiety disorder, 20.5% a current comorbid depressive-anxiety disorder, and 24.4% neither a current nor a remitted depressive or anxiety disorder. Of the study population, 47.4% had not been absent from work in the last 6 months, while 31.0% reported 0–2 weeks, and 21.6% more than 2 weeks of absenteeism. Significant differences were found between workers without childhood trauma, those with a childhood trauma index between 1 and 4 (mild), and those with scores between 5 and 8 (serious) regarding sex (p < 0.001), age (p = 0.002), educational level (p < 0.001), somatic health (p = 0.004), number of working hours (p = 0.023), and psychiatric status (p < 0.001). As for absenteeism (p < 0.001) and presenteeism (p < 0.001), those workers reporting the highest childhood trauma scored higher on these outcome measures. Of the workers who had not experienced any childhood trauma, 55% reported no absenteeism, which was 39.8% for those having experienced mild childhood trauma scores and 32.3% for those with high childhood trauma scores. Within the nontrauma group, short-term absenteeism was reported by 27.7%, within the mild trauma group 36.1%, and within the severe trauma group 31.7%. When looking at the long-term absenteeism reports, we noted even more pronounced increases: 17.3, 24.2, and 35.9%, respectively. Mean disability scores went from 7.64 in workers without childhood trauma, to 9.45 and 9.98 in workers with mild and higher childhood trauma scores, respectively.
Table 1. Baseline demographic and clinical characteristics according to childhood trauma scores at baseline (n = 1,649).
Abbreviations: CTI, Childhood Trauma Index; SD, standard deviation; %, within CTI score.
* p value based on chi-square statistics for categorical variables and analyses of variances for continuous variables (one-way ANOVA).
Mediation by psychopathology
Table 2 and Figure 1 display the main results of the mediation analyses in which childhood trauma and current and remitted depressive and/or anxiety disorders were entered with absenteeism and presenteeism as the dependent variables. After correction for age, gender, education, number of working hours, and somatic conditions, there is an overall significant, direct effect of the childhood trauma index on both work-functioning indices (p < 0.001). The significant indirect effects of current depression and current comorbid depression-anxiety indicate that these factors partially explain the association between the childhood trauma index and both absenteeism (p = 0.023 and p < 0.001, respectively) and presenteeism (p = 0.020 and p < 0.001, respectively). Current depressive disorder accounted for only a small fraction of the total effect on absenteeism (P M = 0.04 [indirect effect/total effect]), while the role of current comorbid depression-anxiety was larger (P M = 0.23). We found similar results for presenteeism: current depressive disorder: P M = 0.05; current comorbid depression-anxiety: P M = 0.29. Apparently, current anxiety disorder and remitted anxiety and/or depression play no significant mediating role in the childhood-trauma/work-functioning association.
Table 2. Explained variance of depression and anxiety in the relationship between childhood trauma index (IV) and work functioning (DV) in the working population (N = 1,649).
The childhood trauma score was considered a continuous variable. Models corrected for age, gender, education, number of working hours, and somatic conditions.
Abbreviations: CI, confidence interval.
Figure 1. Mediation model for childhood trauma on absenteeism and presenteeism through depressive and anxiety disorder. Note: β = standardized effects; *p < 0.001, **p < 0.01, ***p < 0.05; all models corrected for age, gender, education, somatic conditions, and number of working hours.
Discussion
This study expands upon prior literature, examining the relationship between childhood trauma and work absenteeism and presenteeism, and investigating contributing factors in a large cohort (NESDA) of the Dutch working population. Almost half (44.7%) of our participants reported having experienced some form of childhood trauma and we found a strong, significant relationship between exposure to early-life adversity and reduced work functioning, both in terms of absenteeism and presenteeism. Associations were most pronounced for long-term absenteeism, where we observed a marked increase proportional to childhood trauma level. We are the first to show that persons with childhood trauma have more presenteeism at the workplace and that the associations between childhood trauma and absenteeism and presenteeism were slightly but significantly accounted for by a current depressive disorder, while the role of current comorbid depression-anxiety in this relationship was more pronounced, explaining one quarter of the total effect in absenteeism as well as in presenteeism. A large part of the total effect remains unexplained, which means that the relationship between childhood trauma and reduced work functioning in adulthood goes beyond psychopathology. We, for instance, found no significant mediating role of current anxiety disorder and remitted depressive and/or anxiety disorder. The differences in the roles of current anxiety, depressive disorder, and comorbid disorder in this relationship are perhaps explained by differences in their course. Hovens et al. [Reference Hovens, Giltay, Van Hemert and Penninx15] showed that childhood trauma was associated with an elevated risk of the persistence of depression or a comorbid disorder, while this was not the case for the 2-year course of anxiety disorders. In our previous study [Reference De Venter, Van Den Eede, Pattyn, Wouters, Veltman and BWJH37] we found that childhood trauma was associated with chronicity of anxiety symptoms and occurrence of social phobia, but not with the persistence of panic disorder at 2-year follow-up. Moreover, depressive disorders usually have a more continuous and more persistent nature, with a higher risk of (long-term) work disability, whereas anxiety disorders tend to be more episodic. In addition, when people experience anxiety, they may still (prefer to) go to work because the workplace and work activities might distract them from their avoidance behaviors.
As to the methodological strengths of our study, we can point to the large representative cohort of working adults we were able to investigate. We, moreover, used validated, structured diagnostic interviews and well-defined clinical outcome measures (standardized self-report questionnaires), as well as a known mediation model to determine the role of depression and/or anxiety on work absenteeism and presenteeism.
When interpreting the results of our study, some caution is warranted. Importantly, we exclusively relied on retrospective self-reports of childhood trauma, which may give rise to memory bias. However, several studies concluded that this bias was too small to challenge the findings [38–40]. The CTI assessed type and frequency of the trauma and not duration and severity of the abuse. These latter objective characteristics are important predictors of psychopathology, as were more subjective characteristics (e.g., negative reactions to disclosure and feelings of guilt) [Reference Lange, Beurs, Dolan, Lachnit, Sjollema and Hanewald41]. We have to bear in mind that retrospective reports of childhood trauma identify other groups of individuals than prospective reports. We used an interview, and not a questionnaire, to assess a broad childhood trauma index, and not subtypes, which strengthens the concordance between retrospective and prospective measures [Reference JR, Reuben, Newbury and Danese42]. However, we have to take into account that not including subtypes of childhood trauma can also limit the results, since the test–retest reliability and agreement of retrospective self-reports on childhood abuse seems to be partly dependent on the type of childhood abuse. Wielaard et al. [Reference Wielaard, Stek, Comijs and Rhebergen43] found that the answer consistency on emotional neglect is significantly influenced by depression diagnosis and depression severity at baseline, but not the other types of childhood abuse. Moreover, patients with potential mental disorders more frequently reported severe and very severe childhood adversities than patients without a probable mental disorder [Reference Haüser, Hoffmann, Wolfe, Worthing, Stahl and Rothenberg44]. Also, given the long-time lapse between the adverse childhood events and our assessment of work functioning, we were not able to consider other potentially contributing factors in the period prior to the NESDA baseline wave, such as social and (work-related) environmental factors, or past and/or current medical/psychological treatments. We also should be very cautious with the concepts of mediation and mediators since we cannot assume a causal relationship between childhood trauma and the dependent variables we investigated on account of the cross-sectional nature of our approach. As a final remark we need to mention the exclusion of 223 individuals (11.9%) due to missing data on the independent or dependent variables.
Our findings have important clinical implications because they shed new light on the pathway from childhood trauma to adult work functioning. With many spending half of their waking hours at the workplace, employees ought to be supported by a proactive and effective management and offered preventive interventions, including the early recognition of reduced work quality, early support, and easily accessible and affordable treatment for any psychiatric disorder. Moreover, mental ill health is underreported in the sick-leave records because it often remains undetected or is dominated by or hidden behind somatic symptoms. Occupational coaches and medical staff need to take into account this potential vulnerability and the psychological origins of current symptoms and act accordingly when estimating the disability duration, potential replacement income and referral policy. Arguably these professionals are often not aware of the trauma the worker has experienced growing up because they do not have access to such sensitive information, but they could play a major role in the detection and recovery process by offering timely screening for psychopathology (e.g., depression and comorbidity) using widely-used and validated screening instruments. Ultimately, as a society, we have an important mission: to break the taboo around childhood maltreatment and mental-health problems and make help more accessible and readily available at as early a stage as possible.
Conclusion
We have shown that childhood trauma is directly associated with reduced work functioning in adult life, both in terms of absenteeism and presenteeism. The relationship is partly mediated by a current depressive and/or comorbid depressive-anxiety disorder. Apart from these psychopathology-related factors, there are indubitably other mechanisms that warrant attention since they may also play a part in the observed association. Childhood trauma is common and can have (delayed and prolonged) disruptive effects on workers’ physical and mental health and well-being. Adverse early-life experiences then merit serious attention from the business community and labor leaders, but far preferably, at an earlier stage from General Practitioners, mental-health services, and policy-makers at the government level.
Acknowledgments and Financial Support
The infrastructure for the NESDA study (www.nesda.nl) has been funded through the Geestkracht programme of the Netherlands Organization for Health Research and Development (Zon-Mw, grant number 10-000-1002) and by participating universities and mental health care organizations (Amsterdam University Medical Centers (location VUmc), GGZ in Geest, Leiden University Medical Center, University Medical Center Groningen, University of Groningen, Lentis, GGZ Friesland, GGZ Drenthe, Rob Giel Onderzoekscentrum). This project was partly funded by project number 5739 (bequest Cassiers) of the University of Antwerp.
Conflicts of Interest
Prof. Dr. Penninx has received—nonrelated—research funding from Jansen Research and Boehringer Ingelheim. Maud De Venter reports personal fees from Bequest Cassiers, during the conduct of the study. The authors Bernet Elzinga, Filip Van Den Eede, Kristien Wouters, Guido Van Hal, Dick Veltman, and Bernard Sabbe declare none.
Data Availability Statement
The data that support the findings of this study are property of the NESDA consortium, represented by the NESDA principal investigator (PI) prof. Dr. B. W. J. H. Penninx. The total NESDA dataset is available on the internal NESDA website. Restrictions apply to the availability of these data, which were used under license for this study (collaboration agreement between the University of Antwerp and the NESDA consortium).
Open access
Comments
No Comments have been published for this article.