Best practice intervention for post-traumatic stress disorder among transit workers

Abstract

BACKGROUND: Transportation industry workers are at high risk for exposure to traumatic incidents in the workplace. A considerable number of those exposed to such incidents will develop post-traumatic stress disorder (PTSD) symptoms, which leads to high rates of absenteeism and are costly to the public transit corporation and workplace safety compensation insurance.

OBJECTIVE: To determine whether the newly implemented Best Practice Intervention (BPI) provides superior outcomeswhen compared with Treatment-as-Usual (TAU) interventions in improving workers’ rates of return to work (RTW), decreasing duration of time lost from work and overall reduction in severity of PTSD symptoms 6 months after exposure.

METHODS: A sequential mixed methods approach was used with qualitative analysis followed by a pre-post intervention design. Sixty-two participants were recruited to the (TAU) phase of the study and 79 to the (BPI) phase.

RESULTS: Significant differences were observed between the TAU and BPI groups in number of lost work days (TAU: 20 days vs. BPI: 52 days, p = 0.02). PTSD symptoms decreased with time (MPPS score: 51.3 vs. 24.35; p < 0.001). One-fifth of the participants (21 %) did not return to work by the end of the 6 months follow-up period.

CONCLUSIONS: The study demonstrated the value of workplace interventions in improving awareness of psychological symptoms after exposure to a traumatic incident and the value of screening for PTSD symptoms.

Keywords

Workplace mental health acute psychological trauma post-traumatic stress disorder Best Practices Intervention public transportation

1 Introduction

Post-traumatic Stress Disorder (PTSD) is neither a population specific diagnosis nor limited to psychiatric patients or war veterans. Anyone exposed to a traumatic event is at risk of developing PTSD symptoms. Over the past decade, there has been growing recognition of the prevalence of work-relatedPost-traumatic Stress Disorder (PTSD) and the potential negative impact that PTSD symptomatology has on return to work [1]. The majority of workers will encounter at least one emotionally traumatic work incident at some point in their careers [2]. Workers in the transportation industry, especially the subway, train and bus operators, are at high-risk for exposure to workplace traumatic incidents [3]. Traumatic events in this occupational setting may include serious or fatal collisions, physical violence, person under train (PUT) incidents (most often suicides or attempted suicides) and verbal harassment. Results from a sample of 202 French train drivers exposed to a PUT incident show acute distress after the event, with more than half of drivers reporting medium to high distress in the hours and days following the incident [4]. A considerable number of those exposed to these events will develop PTSD symptoms. In addition to psychological distress, transportation workers involved in PUT accidents are more likely to have taken time off work soon after the incident and even up to 1 year post-event [5]. Negative interactions with passengers, including verbal and physical assaults or threats, can also be a stressor with transport workers[6]. In fact, the possibility of an assault was the most frequently cited stressor in a sample of 376 bus drivers [7].

The urban public transit system involved with this project is the third largest public transportationsystem in North America, with over 12,000 employees. According to Workplace Safety & Insurance Board (WSIB) data obtained by the Toronto Star the rate of PTSD is 4 times higher in public transit system employees compared to the urban police officers who patrol Toronto streets [8]. PTSD is the second leading cause of lost-time among transit system workers. Traumatic incident resulting in mental stress have increased 51% since 2002. From 2000 to 2005, nearly 200 bus, streetcar and subway operators were found to suffer from the disorder [8] missing an average of 49 days of work. In 2007, Occupational Injury data from the public transit system reports 633 incidents in the workplace resulting in mental stress, of which 160 resulted in time lost from work. According to the public transit system lost time injuries due to mental stress represented 26% of all lost time injuries in 2009 (Consentio V 2010, oral communication, 16th February).

The cost to the public transit system is also significant in terms of insurance claims. The Workplace Safety and Insurance Board of Ontario reported that the transit system has one of the greatest numbers of traumatic psychological claims of any employer in Ontario. The total time lost for approved psychological claims for this Ontario transit system was in excess of 3,100 days of work (Consentio V 2010, oral communication, 16th February), Given the high frequency of psychologically traumatizing incidents and the significant workplace costs of these incidents, novel methods of intervening with traumatized workers are needed to improve health outcomes and facilitate early return to work.

The specific aim of the Acute Psychological Trauma (APT) study was to implement a Best-Practice Intervention (BPI) for transit workers affected by an acute traumatic event and to improve workers” health and rates of recovery. The study was a 3-year, 3-stage study funded by the WSIB (see Fig. 1). The Treatment as Usual (TAU) phase examined the existing practice at the transit system following a workplace traumatic event, and explored factors that facilitate or hinder treatment seeking and return to work among traumatized workers. In the second stage, the Best Practice Intervention (BPI) was developed. The BPI broadened the existing Psychological Trauma Program (PTP), a provincial Workers’ Compensation Board (WCB) multidisciplinary assessment and treatment program for workers experiencing trauma-related psychological sequelae, and addressed the deficiencies identified by the workers in their interviews. The BPI developed for the transit workers included education and training of exposed workers, Occupation Health (OH) staff, and Employee and Family Assistance Program (EFAP) counselors; screening and surveillance for symptoms of PTSD: When appropriate, timely referral to established specialized mental health treatment program; and specialized return to work coordination. The final stage of the study was the evaluation of the BPI.

It was hypothesized that the newly implemented BPI would provide superior outcomes when compared with Treatment-as-Usual (TAU) by: 1. Improving workers’ rates of return to work (RTW); 2. Decreasing the duration of time lost from work; 3. Reducing the severity of PTSD symptoms 6 months after exposure to a work related traumatic event.

2 Methods

The sample was selected from among transitworkers who had experienced a traumatic incident in the workplace and had completed an Occupational Injury and Disease Report, which details physical injuries and/or emotional trauma, medical care, modifications in work related duties, and/or lost work time of the workers. To be eligible for the APT study, a worker must have taken time off from work after a workplace-related traumatic incident. Eligibletraumatic incidents included physical assaults (any act of violent contact between two individuals with the intent to cause harm), verbal and weapon related threats (any hostile gesture, act, or utterance made with the intent to intimidate or harm), operating a vehicle involved in a serious or fatal incident (i.e. vehicle/pedestrian contact), subway-related suicide and attempted suicide, witnessing street crime or crime against a third party (i.e. robbery, assault of patron), observing vehicle collisions and/or other workplace accidents involving serious injuries. Excluded from the study were workers who sustained severe physical injury and workers who reported emotional trauma caused by workplace conflicts or disputes between co-workers and/or management. Workers with a history of psychiatric disorders were not excluded from participating in the study.

2.1 Procedure for TAU and BPI

Details about the study were advertised in the various workplace divisions and workers were encouraged to self-refer. Staff in the OccupationalHealth and Claims Management department at the transit system contacted workers who had experienced a traumatic incident and completed a Workplace Insurance and Safety Board (WSIB) claim form (with injuries related to emotional trauma). Workers who verbally agreed to participate in the study were referred to the research team.

Two hundred fifty four workers were referred to the study by Occupational Health and Claims Management staff, 18 workers self-referred, 3 workers were referred by the Transit Union executives, and one worker was referred by WSIB. As shown in Fig. 2, of the 276 potential participants, 15 workers did not meet study criteria (e.g., industrial accident, personal stress), 82 refused to participate in the study (e.g., “too overwhelmed”, “not comfortable with information collected”), 38 workers could not be reached although voice messages were left for them, and 141 workers consented and participated in the study. Of 141 participants recruited to the study, 62 were included in the TAU phase and 79 in the BPI phase (see Fig. 2). Recruitment to the TAU phase commenced in May 2008 and completed in June 30th, 2009. Due to difficulties encountered in recruiting participants to the BPI phase, the recruitment period to the BPI exceeded that of the TAU, and extended from July 2009 and was completed in June 2011.

The workers met with the research coordinator who explained the study, and those who agreed to participate provided signed, informed consent. This first appointment was considered their baseline visit, and demographic information and baseline measures were collected. PTSD symptoms were assessed four weeks after the traumatic event with the Modified PTSD Symptom Scale (MPSS). Axis I and Axis II DSM disorders were also assessed with the Structured Clinical Interview for DSM-IV disorders(SCID I and SCID II). At 3 and 6 months post-incident, participants were again screened for PTSD symptoms with the MPSS and were asked about symptoms of depression and anxiety.

At the end of their baseline visit, TAU and BPI participants were given a binder with instructions to note the dates of any medical appointment (e.g., with a GP or a mental health professional), medications prescribed after the incident, and any referral to a health care provider. Participants were also asked to provide copies of any WSIB forms relating to the incident they have submitted during their participation in the study. These forms and the self-report information from the binders were used to record any appointments with mental health professionals in the data collection process.

2.2 Interventions

2.2.1 Treatment-as-Usual (TAU)

Treatment-as-Usual workers whose post-traumatic symptoms were above threshold on either the MPSS [27, 28] or the SCID I [16], or had met criteria for any other Axis I disorder were asked to provide signed consent to allow the research team to inform their family doctor in writing of such findings. No other interventions were provided to TAU workers since they were expected to seek and receive care from community care providers, and interact with the staff in the Occupational Health and Claims Management department at the transit system. Workers were referred to their family doctor who then proceeded with their usual care approach. Referrals to a psychologist or psychiatrist were made by the family doctor when deemed necessary.

2.2.2 Best Practice Intervention (BPI)

Participants in the BPI phase whose post-traumatic symptoms were assessed to be above threshold on either the MPSS or the SCID I, or they met criteria for any other Axis I disorder (e.g., mood or anxiety disorders) were referred to the specialized multidisciplinary Psychological Treatment Program for further assessment and clinical services.

The BPI was developed after the completion of the qualitative study [30]. The BPI extended the existing Psychological Trauma Program (PTP), a provincial Workers’ Compensation Board (WCB) multidisciplinary assessment and treatment program for workers experiencing trauma-related psychological sequelae, and addressed the gaps identified in the TAU by the transit system workers in their qualitative interviews. The BPI consisted of three components:

1. Educational programs for exposed workers and promotion of self-screening and help seeking. A standardized educational intervention for workers was developed to improve awareness of symptoms and familiarize workers with treatment options and destigmatize workers beliefs regarding the psychological sequelae of trauma [9, 10]. The Educational intervention included a PowerPoint presentation, a Response to Trauma Booklet, and a fact sheet for the workers and their families and loved ones describing traumatic incidents and common reactions, symptoms of PTSD, treatment options, and appropriate resources. Participants in the BPI phase were given the educational intervention within one month of being exposed to a traumatic incident [11]. The educational materials and screening tools were also disseminated to TTC Occupational Health (OH) staff, EFAP staff and others as a part of an occupational health educational initiative.

2. Referral to “evidence-based” multidisciplinary treatment program. The Psychological Trauma Program (PTP) is a multidisciplinary treatment program for injured workers who haveoccupational-related anxiety and mood disorders. The PTP provides comprehensive psychiatric and psychological assessment, treatment and disability management services. The treatment included “best practices” CBT interventions and/or pharmacologic interventions for Acute Stress Disorder (ASD) and PTSD were offered to those experiencing a well-defined level of severity of post-traumatic symptoms [12, 13]. The CBT treatment is provided by PhD-trained psychologists. In addition, the BPI included: Occupational therapy, physiotherapy, consulting psychiatric care, and return-to-work coordination (RTWc).

3. Specialized return to work strategies in collaboration with the transit company. To reduce barriers to reintegration and RTW, we developed an integrated and flexible system involving Occupational Health, the Amalgamated Transit Union, health care providers and the worker. This was overseen by the RTW coordinator at the PTP, who assessed the workers’ readiness to return to work using the “stages of change model” [14]. The RTW model utilized in the BPI was derived from a review of the existing literature [15] and based on the understanding that RTW is influenced by numerous clinical and non-clinical factors. The BPI return to work strategy, where appropriate, utilized the following methods: in vivo exposure therapy; job site RTW coordination meetings; reasonable accommodations and regular contact during RTW and post-RTW.

2.3 Participants

Overall, the mean age of participants was 45.1 (SD = 9.1) years old, with the greater majority being males (74.1%), married or common-law (64.7%), Caucasian (61.9%), and having post-secondary education (61.8%). The median time employed at the transit system was 4.4 years (IQR = 8 years). There were 59.7% participants who were bus or streetcar operators, 27.3% subway train operators, 10.1% fare collectors, and 1.4% maintenance or janitorialworkers. An examination of the socio-demographic profile of our sample reveals there were no significant statistical differences in characteristics of the participants in the TAU and BPI phases with the exception of length of employment and occupation (Table 1). For TAU participants, the median time of employment at the transit commission was 4 years (IQR = 5.5 years) compared to 6 years (IQR = 10.5 years) years for the BPI participants (p = 0.025). A greater proportion of BPI participants worked as subway train operators compared with TAU participants (34.7% vs. 19.7%; respectively, p = 0.05.

2.4 Diagnoses and psychopathology measures

The Structured Clinical Interview for DSM-IV [16, 17] was used to assess major current and lifetime psychiatric disorders (Axis I) and personality disorder diagnoses (Axis II). Other instruments used to measure psychopathology included: The Beck Depression Inventory [18, 19], to measure different aspects of depressive symptomatology; the Symptom Checklist 90-Revised (SCL-90R) [20 –22], to evaluate a broad range of psychological problems and symptoms of psychopathology; the Dissociative Experiences Scale (DES) [23] for dissociative symptoms; the Short Form Health Survey (SF-36) for overall subjective health status [24, 25] and the Multidimensional Health Locus of Control Scale (MHLC) was used to measure participants’ perceptions of their health outcomes. At baseline, data were also collected on lifetime and past year occurrences of trauma [26]. The aforementioned measures have been used extensively with clinical populations and have well-established reliability and validity.

2.5 PTSD symptoms measures

The MPSS is a 17-item self-report measure that assesses the 17 DSM-III-R symptoms of PTSD. This scale is a modification of the PTSD Symptom Scale [27]. Participants were asked to rate their PTSD symptoms based on their symptoms in the preceding two weeks. The 17 items are rated on a 4-point frequency (ranging from 0 = “not at all” to 3 = “5 or more times per week") and intensity scales (ranging from A = “not at all upsetting” to D = “extremely upsetting"). The MPSS can be used to make a preliminary determination of the diagnosis of PTSD using either DSM-III-R criteria or a frequency, severity, or total score cutoff scores. The total score for frequency is a sum of all the responses that may range from 0 to 51 while the total score for severity ranges from 0 to 68. The total symptom severity score on MPSS is calculated by summing up the scores for frequency and severity across all items [28]. The maximum possible total score on MPSS is 119 and a score of 46 or greater is considered positive for PTSD. The MPSS has demonstrated good internal consistency (Cronbach’s α= 0.98) and convergent validity with the Structured Clinical Interview (SCID) of the DSM-IV-TR PTSD module [29].

2.6 Return to work (RTW)

Data on return to work (RTW) were collected from the study participants’ self-report and WSIB forms to measure the number of days on sick-leave (continuous variable) and whether participants returned to full-time work or modified work during the 6 months follow-up period. Time to return to work was defined as the number of days between the onset of the sickness leave (1st day after the traumatic incident) and the return to full-time or modified work. Modified work includes partial work time or modified duties i.e. reassignment to other type of work.

2.7 Data analysis

2.7.1 Qualitative

We embedded a qualitative study in the TAU quantitative phase of the study. This complemented the development and implementation of the Best Practices Intervention (BPI) phase. For the qualitative study we recruited 20 participants from among the 62 TAU participants, and the findings are published elsewhere [30].

2.7.2 Quantitative

In all of the statistical analyses time (in days) to return to work and post-traumatic stress symptoms, were considered dependent variables while all other measures were considered independent variables. The duration of time to return to work started on the first day off work until the return to full-time or modified work. A participant was considered censored (i.e. the period of observations was cut off before the return to work) if he or she was lost to follow-up or did not return to work within the study follow-up period (i.e., 180 days). These participants were considered censored at the date of their last visit.

Some categories of type of traumatic incident, job type, and length of employment were pooled for analyses because of small cell counts. Two sample t-tests or Mann-Whitney U tests were conducted to compare the TAU vs. BPI participants with respect to continuous variables, and the chi-square or Fisher’s exact test were applied to compare proportions. Repeated measures analysis were conducted using the linear mixed-effects model (LME) framework to analyze the changes in the severity of PTSD symptoms, as measured by the MPSS, at 1, 3 and 6 months. The model included a group indicator (1 if BPI, 0 if TAU), time as a categorical variable (reference category was 1 month), and the interaction between group and time. Compound symmetry (CS) was used as the covariance structure. The CS on the residuals allowed for within-subject correlations across time. CS structure has constant variances across time and constant covariance between measurements. Contrast tests permitting post-hoc pair-wise comparisons between time and MPSS group estimated marginal means were performed using the Bonferonni approach.

A Cox Proportional Hazard regression analysis was used to determine factors significantly associated with the time to return to work. The Cox Proportional Hazard regression is concerned with studying the time between entry to the study and a subsequent event (i.e. return to work). Independent variables were grouped into 5 categories: (i) socio-demographics, (ii) job- related, (iii) health and physical functioning, (iv) psychological functioning, and (v) trauma. Correlations were calculated in order to identify variables covering similar aspects (i.e., correlation coefficient > 0.70). Bivariate analyses using Cox regression analyses were performed to determine the associations between independent variables and time to return to work (Table 3, Step 1).Within each category, variables associated at p-value of 0.25 [31] were considered for the multivariate Cox regression analysis Table 3, Step 2).The final model consisted of selection of the most significant correlates from step 2 in each of the 5 categories, and the BPI as a covariate. which subsequently were subjected to Cox regression analysis (enter method). The hazards ratio (HR) with 95% confidence interval (CI) is presented for all factors included in the final model.

All tests were 2-sided and p-values were considered significant if they did not exceed 0.05. All analyses were conducted using SPSS version 20 (IBM Corporation, SPSS^®).

2.8 Ethical consideration

The study was approved by the Research Ethics Boards at St. Michael’s Hospital and Centre for Addiction and Mental Health (CAMH). Participation in the study was voluntary and all participants were required to sign a Research Ethics Research Board approved consent form.

3 Results

A total of 276 workers were referred by the Occupational Health staff to the TAU and BPI: 15 did not meet study criteria, 82 refused to participate, 38 were lost to follow up, and 141 transit workers signed informed consents and were recruited to the study. Of the 141 participants, 126 (89.4%) completed the 6-month study. Of the 62 participants who were recruited to the Treatment as Usual (TAU) arm of the study; 58 (93.5%) completed the study, 3 (4.8%) withdrew, and 1 (1.6%) was lost to follow-up.Seventy-nine participants were recruited to the Best Practices Intervention (BPI) arm of the study; 68 (86.1%) completed the study, 2 (2.5%) withdrew, and 9 (11.4%) were lost to follow-up (see Fig. 2). When compared, there was a significant difference between the lower number of TAU than BPI participants lost to follow-up (p = 0.043).

3.1 Occupational injury

The majority of traumatic incidents (51.8%) were physical or verbal assault incidents, 24.8% were subway-related suicides or attempted suicides, and 15.6% involved being directly in an accident. As indicated in Table 2, there were significant statistical differences between the TAU and BPI in the type of traumatic incident they experienced. While the majority of TAU participants (62.9%) experienced physical (38.7%) or verbal (24.2%) assaults, nearly a third of the BPI participants (30.4%) were involved in a subway-related suicide or suicide attempt.

3.2 Diagnoses and psychopathology

TAU and BPI significantly differed on the occurrence of Axis-I diagnoses compared to the TAU a greater proportion of BPI participants met criteria for Major Depressive Episode 17.7% vs. 44.7%, respectively, p = 0.001), Panic disorder (8.1% vs. 21.1%, p = 0.004), and PTSD (38.7% vs. 55.3%, respectively, p=0.04). There were no significant differences between TAU and BPI in other Axis I or Axis-II diagnoses, severity of depression (BDI), dissociative symptoms (DES), perceptions of locus of control (MLOC), social support (PIYL), psychological problems and symptoms of psychopathology (SCL-90R). However, there were differences in pain perception with TAU reporting experiencing greater bodily pain than participants in the BPI (71.4 (SD = 23.1) vs. 60.5 (SD = 29.8), p = 0.023, respectively) and general health perception (SF: General health perception – BPI 63.8(SD = 21.8) vs. TAU 54.9 (SD = 22.8), p = 0.033).

3.3 PTSD symptoms

Of the 79 BPI participants, 27 (34.2%) did not meet criteria for referral for treatment assessment at PTP and were treated by community providers, and 44 (55.7%) were referred for treatment assessment at the specialized multidisciplinary treatment program. An additional 8 (10%) of the participants met criteria for referral but declined the offer of referral to specialized treatment as they had their own mental health provider.

Among all participants, at the 1 month assessment, MPSS scores ranged were between 0–112 with median score being 54 and 55.2% had MPSS score above the cut-off of score of 46. At the 3 months scores were between 0–103, with median score of 30 and 37.4% continued to have scores above the cutoff score of 46. At the 6 months assessment scores were between 0–102, with median score of 13 and 19.2% continued to have scores above the cutoff score of 46.

There were no statistically significant differences in change over time in MPSS scores between TAU (estimated marginal means (SE) = 48.89 (3.92); 34.78 (3.95), and 25.08 (3.95), at 1, 3, and 6 months, respectively), and BPI (estimated marginal means (SE) = 53.69 (3.51), 36.58 (3.64), and 23.54 (3.68 at 1, 3, and 6 months, respectively) (p = 0.347). However, both groups experienced significant statistical decrease in PTSD symptoms with time (p < 0.001). Additional post hoc contrast analyses using Bonferroni pairwise comparisons procedures adjusting for multiple comparisons with overall alpha level < 0.05 indicated statistically significant decrease in the MPSS estimated marginal mean scores between the 1 and the 3 months assessment interview (p < 0.001), and between the 3 and 6 months interview scores (p < 0.001) (M_1month = 51.29, SE = 2.63, M_3months = 35.68, SE = 2.69M_6months = 24.31, SE = 2.70).

3.4 Return to work

Nearly ten-percent (9.8%) of the participants in the TAU returned to work immediately after the traumatic incident compared to 4% of the BPI participants. By one-month follow up 55.74% of the TAU participants had returned to work compared to 45.3% of the BPI participants. By 3 months follow-up 75.4% of TAU participants had returned to work compared to 60.0% of the BPI participants. By 6 months follow-up interviews 13.1% of the TAU participants had not returned to work compared to 26.7% of BPI participants. In all 28 (20.6%) study participants had not returned to work by their 6 months appointment.

Among all participants, the median time to RTW was 30 days (SE = 5.83; CI 95% = 18.57–41.43). The median time to return to work among the TAU participants was 20 days (SE = 11.16; 95% CI = .000–41.87). The survival curve illustrates that at 20 days, 49.2% of TAU participants had not RTW. For the BPI, the median time to RTW was 52 days (SE = 14.42; 95% CI = 23.73–80.27). The survival curve illustrates that at 52 days, 48.0% of BPI participants had not RTW. The Kaplan–Meier curves and results of the log rank tests indicate that there was a significant difference between the survival curves in each of the treatment arms of the study (Log rank Chi-square 5.46; df = 1; p = 0.02). Figure 3 displays results from a Kaplan-Meier survival analysis comparing the length of time to RTW by treatment group (TAU or BPI).

3.5 RTW – Modified duties

Of the 53 TAU participants who had returned to work within the study period, 22 (36.1%) returned to modified duties and 31(58.5%) returned to full-time work. Three participants (4.9%) returned to modified duties; however they did not resumed their regular duties within the study period. In the BPI arm of the study, 18 (29.3%) participants returned to modified duties, 4 participants (5.3%) returned to modified duties however did not resume regular duties during the study period, and 38 (50.7%) of the participants returned to full-time work. There was no significant statistical difference in the number of TAU and BPI participants who had returned to modified duties after a traumatic incident (p = 0.103).

3.6 RTW and type of traumatic event

Additional Kaplan-Meier survival analyses were performed to compare the duration of time to return to work by study arm (TAU or BPI) and type of traumatic incident experienced. The estimated median duration of time to return to work after a suicide in the TAU was 50 days (SE = 12.11), 95% CI = 26.26 –73.74) compared to BPI 140 days (SE = 70.27), 95% CI = 2.26–277.73). The Kaplan-Meier curves and the results of the log rank tests indicate that there is no significant difference between the survival curves in each of the treatment arms of the study after exposure to a suicide (Log rank Chi-square = 3.181, df = 1, p = 0.075).

A similar analysis was conducted to examine the duration of time to return to work after an assault and accident. The estimated median duration of absence from work after an assault in the TAU was 10 days (SE = 3.69; 95% CI = 2.76 –17.24 compared to BPI 22 days (SE = 9.48; 95% CI = 3.43 –40.57). The Kaplan-Meier plot revealed very similar survival curves (log-rank p = 0.277) for the TAU and BPI after exposure to an assault.

The estimated median duration of time to return to work after an accident in the TAU was 18 days (SE = 9.53; 95% CI = 0.00 –36.67) compared to BPI 44 days (SE = 37.12; 95% CI = 0.00 –116.76). The Kaplan-Meier survival curves and the results of the log rank tests indicated that there were no statistical significant differences (p = 0.415) in the duration of time to return to work after exposure to an accident between TAU and BPI arms of the study.

As described in the data analysis section several Cox regression analyses were performed to assess the effect of potential explanatory variables on the duration of time to return to work. A maximum of four variables (the 4 strongest variables) within each category were entered into the analysis; however, if variables were covering similar aspects, as evident by their bivariate correlation coefficient (>0.70), only one of the variables was included in the analysis. Since the Global SCL-90R score was highly correlated with our outcome variable MPSS at 1 month (Pearson’s coefficient = 0.71) it was not included in the final model.

The final multivariate Cox regression included a selection of the most significant variables from each of the following 3 categories: Health & physical functioning, psychological functioning, type of trauma experienced. The variables included in the model of analysis were: MPSS at 1 month, role limitations due to emotional problems, emotional well-being, social functioning, and suicide. The study arm was included as a covariate. The multivariate Cox regression with unadjusted and adjusted hazard ratios and 95% confidence intervals (CI) was performed in Wald survival analysis.

For binary variables, a hazard ratio of less than 1 indicates the variable was beneficial and was associated with a shorter duration of sick leave while a hazard ratio greater than 1 indicates an adverse variable associated with increased risk for a longer duration of sick leave time [32]. For continuous variables, a hazard ratio less than 1 indicates the variable was associated with a longer time to return to work while a hazard ratio greater than 1 indicates a shorter time to returning to work [33]. Table 4 displays the unadjusted and adjusted hazard ratios for the covariates entered in the Cox regression analysis.

The hazard ratio shows that participants in theBPI were less likely to RTW. Exposure to a suicide(HR = 0.621, p = 0.049) and a higher number of PTSD symptomatology one month (HR = 0.344, p = <0.001) following the traumatic incident adversely influenced return to work. Participants who had not returned to work by the end of the 6 months follow-up period reported a significantly greater number of post-traumatic stress symptoms at the 1, 3, 6 months follow-up interviews (MPSS at 1 Month: 79.98 (SD = 16.01) vs. 43.75(SD = 30.72), p < 0.001. MPSS at 3 Month: 65.11(SD = 21.51) vs. 27.49 (SD = 28.03), p < 0.001; MPSS at 6 Month: 43.85 (SD=27.85) vs. 18.41 (SD = 24.00), p < 0.001). Nearly a third of the 34 participants who had experienced a suicide (10/34; 29.4%) did not returned to work at the end of the 6 months follow up period compared to 11 of the 61(18.3%) participants who experienced an assault, and 6 of the 29 participants (20.7%) who were involved in a motor vehicleaccident.

4 Discussion

The APT study is one of the first studies carried out in the workplace attempting to measure and improve lost-time from work due to psychological trauma. The intervention was focused on a large public transit system, where exposure to traumatic incidents is frequent and unavoidable, and was intended to improve awareness of psychological symptoms, promote treatment-seeking, enhance accessibility to work-focused psychological treatment and improve recovery and lost time from work.

The study findings suggested that systematic screening and individualized education did improve awareness of symptoms and increased treatment seeking rates in the BPI versus the TAU (79% vs. 69 %), which is similar to findings from other workplace interventions such as the APRAND programme [34]. Education was provided through posters and distributed articles although promoting the BPI was restricted by ethics protocols and the absence of regularly used email accounts, phones, offices, or other common spaces by operators. Due to the anxiety-provoking nature of psychological trauma, reaching potential participants for screening was difficult, with many simply not returning phone calls after agreeing to be referred by occupational health. Education appeared to be helpful in reducing stigma regarding symptoms and familiarizing workers with treatment options, as many had no prior contact with mental health professionals. Participants were also provided with a package for their family as family members often encouraged participation in treatment. A small number of the participants self-referred to the study in the context of workplace disputes and as a form of malingering, but they were largely self-referrals, suggesting that routine screening and education is a more effective intervention than relying solely on self-identification.

In terms of recovery from PTSD symptoms, no statistically significant differences were found between BPI and TAU PTSD symptom scores over the 1, 3 and 6 months follow up. However, both arms of the study demonstrated a statistically significant decrease in PTSD symptoms over time. Also, the study finding with regards to number of lost work days was contrary to our expected findings. Our results demonstrated that TAU participants lost an average of 10 days from work while the BPI participants lost on average 21 days from work thus the BPI did not offer an advantage in terms of significantly reducing the lost days from work. However, the study was hampered by low recruitment numbers particularly in the BPI arm due to loss of union support for the intervention. The lack of an adequate sample may have lessened the ability to find differences between the two treatment protocols. Lost time from work was also observed to be influenced by several non-clinical factors that may have made it difficult to demonstrate differences between the two interventions. Suitable participants were not always efficiently triaged due to inconsistencies in reporting of incidents and time needed to obtain consent for the study. Also, untimely staff shortages occurred at the PTP, which was the primary provider of care during the BPI, leading to increased wait times needed for assessment and treatment provision. Lastly, existing labour relations and policies imposed significant restrictions on what constitutes suitable work accommodations for workers with work limitations. Accommodating work was not readily available and several cases were subject to dispute by the employer or the Workers’ Compensation Board.

Exposure to a suicide and a higher number of PTSD symptoms, which indicates more severe PTSD; one month following the traumatic incident adversely influenced return to work. These findings suggest that certain workers may be at greater risk for delays in returning to work and those employees who have certain types of traumatic exposure, such as witnessing a suicide, or with high levels of PTSD symptoms at one month might require more specialized services. However, the treatment after an acute psychologically traumatic event should be targeted at the needs of the particular individual, thus one standardized approach will not be appropriate for all affected employees. Individual differences in vulnerability and resilience and differences in the nature, intensity and duration of trauma may be important factors in determining the person’s course after exposure to a traumatic incident. These factors appear important to consider in the assessment and management of workplace trauma [35]. This finding is inferred from the study results that suggested that the specialized mental health services were not needed by all affected workers and many employees were adequately served by the services provided through their family physician or community providers.

4.1 Limitations

There are several limitations that must be considered when attempting to ascertain the validity and generalizability of our findings. The results of this study must be interpreted with caution considering significant confounders such as co-interventions introduced in the course of the study, the impact of stigma on help seeking, and the declining union support [36]. A randomized controlled trial was not acceptable to the employer who did not want to restrict access to the BPI once it was implemented. The study was also subject to significant sample bias as only a limited number of workers subject to trauma actually entered the study. It might be argued that the workers who responded to this study constituted a self-selected sample that may not be truly representative of transit worker in general. This was a significant issue during the BPI phase when workers were actively discouraged by their union representatives from participating in the study. The union’s support for the study appeared to have been driven by factors external to study but nonetheless had an impact on recruitment. Overall, the study was underpowered due to low recruitment and could not adequately control for the non-clinical confounders previously noted. This appears to be intrinsic to studyinginterventions in a workplace. The implementation issues discussed highlight the importance of strong political support and adequate resources when undertaking mental health research in a workplace setting.

The APT study did provide insights into the nature of workplace psychological trauma and some implications for the clinical care of traumatized workers. Most participants in the TAU and BPI did improve and ultimately returned to work. Screening was effective in identifying those with symptoms and at higher risk for long-term disability. Participants who had not returned to work by the end of the 6 months follow-up period reported a significantly greater number of post-traumatic stress symptoms at the 1, 3, 6 months follow-up interviews and a greater proportion of those exposed to a suicide did not return to work at 6 months. As discussed above, the highest need groups ideally should be directed to interdisciplinary care. Unfortunately, differences in care did not improve RTW rates at 6 months but the treatment approach may be a differentiating factor for a period of follow up greater than 6 months and more severe cases. The need for flexible work accommodations to meet the needs of traumatized workers was reinforced by the experience of return-to-work of the study participants. This flexibility requires strong communication and stable relationships between thetreating professionals, the employees, unions and the employer personnel.

Lastly, factors associated with relapse were not evaluated but are of great interest for long-termworkers returning to same or similar positions after a traumatic incident. This may be particularly relevant for emergency services, banks and heavy industry where exposure to work-related traumatic events is also unavoidable.

5 Conclusions

Overall, the APT study demonstrated the value of workplace interventions in improving awareness of psychological symptoms after trauma exposure and the value of screening for PTSD symptoms among exposed workers [11]. The study did not show a benefit from developing a BPI versus TAU in terms of reducing symptoms over follow up or the number of lost work days. However, these findings point to transit workers with exposure to a suicide or a higher number of PTSD symptoms one month after the event as having a much greater risk for delays in returning to work.

Conflict of interest

None to declare.

Footnotes

Acknowledgments

This research project was funded by a grant from the Workplace Safety and Insurance Board Research Advisory Council.

References

Strauser

. Trauma symptomatology: Implications for return to work. Work: A Journal of Prevention, Assessment and Rehabilitation 2008;31(2):245–252.

Beaton

, Murphy

, Johnson

, Pike

, Cornell

. Coping Responses and Posttraumatic Stress Symptomatology in Urban Fire Service Personnel. Journal of Traumatic Stress 1999;2:293–308.

McFarlane

, Bryant

. Post-traumatic stress disorder in occupational settings: Anticipating and managing the risk. Occupational Medicine 2007;57(6):404–410.

Limosin

, Loze

, Cothereau

, Beaurepaire

, Payan

, Conso

, Rouillon

. A prospective study of the psychological effects of “person under train” incidents on drivers Journal of Psychiatric Research 2006;40(8):755–761.

Theorell

, Leymann

, Jodko

, Konarski

, Norbeck

, Eneroth

. “Person under train” incidents: Medical consequences for subway drivers. Psychosomatic Medicine 1992;54(4):480–488.

Evans

. Working on the hot seat: Urban bus operators. Accident Analysis & Prevention 1994;26(2):181–193.

Duffy

, McGoldrick

. Stress and the bus driver in the UK transport industry. Work & Stress 1990;4(1):17–27.

Bruser

, Kalinowski

. TTC Driver’s In Crisis. The Toronto Star. 2008 Jan 21;Sect A: 1(col. 1). Available from: http://www.thestar.com/News/GTA/article/295876.

Weisaeth

. Acute posttraumatic stress: Nonacceptance of early intervention. The Journal of Clinical Psychiatry. Supplement 2001;62(17):35–40.

10.

Stuart

. Stigma and Work. HealthcarePaper 2004;5(2):100–111.

11.

Bance

. Evaluation of an Educational Intervention for Employees Exposed to Workplace Trauma (MSc thesis), Toronto: University of Toronto;2011.

12.

National Collaborating Centre for Mental Health. Clinical Guideline 26. Post-Traumatic Stress Disorder (PTSD): The Management of PTSD in Adults and Children in Primary and Secondary Care. National Institute for Clinical Excellence; 2005. Available from: www.nice.org.uk/CG026NICEguideline.

13.

American Psychiatric Association (2004). Diagnostic and statistical manual of mental disorders. DSM-IV-TR ed. Washington, DC.

14.

Franche

, Krause

. Readiness for return to work following injury or illness: Conceptualizing the interpersonal impact of health care, workplace, and insurance factors. Journal of Occupational Rehabilitation 2002;12(4):233–256.

15.

Bender

, Farvolden

. Depression and the workplace: A progress report. Current psychiatry reports 2008;10(1):73–79.

16.

First

, Gibbon

, Spitzer

, Williams

JBW

, Benjamin

. User’s guide for the structured clinical interview for DSM-IV Axis II personality disorders. New York, NY; New York State Psychiatric Institute, Biometrics Research; 1997.

17.

First

, Spitzer

, Gibbon

, Williams

JBW

. Structured Clinical Interview for Axis I DSM-IV Disorders- Patient Edition (SCID-I/P, Research Version). New York, NY: New York State Psychiatric Institute, Biometrics Research Department; 2002.

18.

Beck

, Ward

, Mendelson

, Mock

, Erbaugh

. An inventory for measuring depression. Archives of General Psychiatry 1961;4(6):561–571.

19.

Beck

, Steer

, Brown

. Manual for Beck Depression Inventory. 2nd ed. San Antonio, TX: Psychological Corporation; 1996.

20.

Derogatis

, Cleary

. Factorial invariance across gender for the primary symptom dimensions of the SCL-90. British Journal of Social and Clinical Psychology 1977;16(4):347–356.

21.

Derogatis

, Cleary

. Confirmation of the dimensional structure of the SCL 90-R: A study in construct validation. Journal of Clinical Psychology 1977;33:981–989.

22.

Derogatis

. BSI Brief Symptom Inventory. Administration, Scoring, and Procedures Manual. 4th Ed. Minneapolis, MN: National Computer Systems; 1993.

23.

Bernstein

, Putnam

. Development, reliability, and validity of a dissociation scale. The Journal of Nervous and Mental Disease 1986;174(12):727–735.

24.

Hays

, Sherbourne

, Mazel

. User’s Manual for Medical Outcomes Study (MOS) Core measures of health-related quality of life. Santa Monica, CA: RAND Corporation; 1996. Available from: www.rand.org.

25.

Ware

JR JE

, Sherbourne

. The MOS 36-item short-form health survey (SF-36): I. Conceptual framework and item selection. Medical Care 1992;473–483.

26.

Stein

, Walker

, Hazen

, Forde

. Full and partial posttraumatic stress disorder: Findings from a community survey. American Journal of Psychiatry 1997;154(8):1114–1119.

27.

Foa

, Riggs

, Dancu

, Rothbaum

. Reliability and validity of a brief instrument for assessing post-traumatic stress disorder. Journal of Traumatic Stress 1993;6(4):459–473.

28.

Kolts

, Robinson

, Tracy

. The relationship of sociotropy and autonomy to posttraumatic cognitions and PTSD symptomatology in trauma survivors. Journal of Clinical Psychology 2004;60(1):53–63.

29.

Falsetti

, Resnick

, Resick

, Kilpatrick

. The Modified PTSD Symptom Scale: A brief self-report measure of posttraumatic stress disorder. The Behaviour Therapist 1993;16:161–162.

30.

Bance

, Links

, Strike

, Bender

, Eynan

, Bergmans

, Hall

, O’Grady

, Antony

. Help-seeking in transit workers exposed to acute psychological trauma: A qualitative analysis. Work: A Journal of Prevention, Assessment and Rehabilitation 2014;48(1):3–10.

31.

Hosmer

, Lemeshow

. Applied logistic regression. New York: John Wiley & Sons; 1989.

32.

Lang

, Secic

. How to Report Statistics in Medicine. Philadelphia, PA: American College of Physicians; 1997.

33.

Steenstra

, Koopman

, Knol

, Kat

, Bongers

, de Vet

. Prognostic factors for duration of sick leave due to low-back pain in Dutch health care professionals. Journal of Occupational Rehabilitation 2005;15(4):591–605.

34.

Godard

, Chevalier

, Lecrubier

, Lahon

. APRAND programme: An intervention to prevent relapses of anxiety and depressive disorders: First results of a medical health promotion intervention in a population of employees. European Psychiatry 2006;21(7):451–459.

35.

Lopez

. Posttraumatic stress disorder and occupational performance: Building resilience and fostering occupational adaptation. Work: A Journal of Prevention, Assessment and Rehabilitation 2011;38(1):33–38.

36.

Links

, Bender

, Eynan

, O’Grady

, Shah

. Facilitators and Barriers to Doing Workplace Mental Health Research: Case Study of Acute Psychological Trauma in a Public Transit System. WORK: A Journal of Prevention, Assessment, and Rehabilitation 2016.