Abstract
BACKGROUND:
Posttraumatic stress injury (PTSI) is a term used to describe a range of psychiatric difficulties which arise following exposure to a psychologically traumatic event. The impact of being diagnosed with multiple psychiatric conditions on the return-to-work (RTW) outcomes of individuals with PTSI has not been adequately researched.
OBJECTIVE:
The current study examined whether the presence of two or more psychiatric conditions occurring simultaneously is predictive of RTW outcomes in workers with PTSI.
METHOD:
A population-based cohort design was conducted using archival data from injured workers admitted to a PTSI rehabilitation program. Differences in RTW outcomes and demographic, administrative, and clinical variables were compared between individuals with single and multiple psychiatric diagnoses. A range of variables were entered into a multivariable logistic regression model predicting RTW.
RESULTS:
The final logistic regression model indicated workers had higher odds of RTW if they had a single psychiatric diagnosis (Adjusted Odds Ratio (AOR) 2.20), non-elevated scores on a measure of traumatic stress (AOR 1.85), and reported higher self-perceived readiness to RTW (AOR 1.24).
CONCLUSION:
Being diagnosed with multiple psychiatric conditions appears to be associated with more negative RTW outcomes following PTSI rehabilitation.
Keywords
Introduction
Work-related injuries are a prevalent issue impacting the lives of Canadians, with over 270,000 workers’ compensation claims in Canada due to a work-related injury in 2019 alone (Tucker & Keefe, 2022). Although physical injuries are common, psychological injuries also frequently occur in the workplace. In fact, the Mental Health Commission of Canada (MHCC, 2019) reports that roughly 30% of all long-term workplace disability claims arise due to mental health-related illnesses. Further, in a one-year period, workplace disability claims associated with mental illness cost approximately $6 billion (CDN) to Canadian employers due to absenteeism and high turnover rates (MHCC, 2019). Due to the high frequency and cost of mental illness-related disability claims, research is needed regarding mental health conditions arising in the workplace.
One cause of, or contributor to, mental health conditions in the workplace is psychologically traumatic events. This refers to the direct or indirect exposure to or threat of a stressful event such as sexual violence, a physically harmful accident, a motor vehicle accident, or an assault, among others (Canadian Institute for Public Safety Research and Treatment (CIPSRT), 2019). Mental health difficulties arising from psychologically traumatic events are known as posttraumatic stress injury (PTSI; CIPSRT, 2019). PTSI is a non-medical term used to describe varying degrees of psychiatric symptoms that, in some cases, may reach the threshold of one or more psychiatric diagnoses, including but not limited to posttraumatic stress disorder (PTSD; CIPSRT, 2019). Previous research suggests that less than 50% of workers with PTSI successfully return to work (RTW) following PTSI rehabilitation (Gross et al., 2021). As such, research examining factors associated with RTW outcomes in individuals with PTSI is needed. Factors such as injury duration before admission to rehabilitation, presence of concurrent physical and psychological injuries, the intensity of the rehabilitation program, and work status at the time of admission have been found to be predictive of RTW outcomes in workers with PTSI (Gross et al., 2021). However, additional research is required to examine a broader range of factors that may be predictive of RTW outcomes in individuals suffering from PTSI.
One factor of interest that has not been examined extensively in workers with PTSI is how being diagnosed with more than one psychiatric condition impacts RTW outcomes. PTSI is associated with various symptoms and functional disturbances, which are linked to PTSD, but also associated with psychiatric diagnoses such as depressive or anxiety disorders (CIPSRT, 2019; Di Nota et al., 2021); thus, the occurrence of multiple psychiatric conditions are of particular interest in individuals with PTSI. This is significant as being diagnosed with various combinations of two or more psychiatric conditions is associated with a wide range of adverse signs, symptoms, and outcomes in samples, including veteran and general clinical populations (Knowles et al., 2019; McCauley et al., 2012; Nichter et al., 2019a; Nichter et al., 2019b; Post et al., 2011; Schäfer & Najavits, 2007; Zayfert et al., 2005). This includes but is not limited to, increased suicidal ideation, lower quality of life ratings, decreased cognitive functioning (Nichter et al., 2019a), decreased physical functioning (Nichter et al., 2019b; Zayfert et al., 2005), increased interpersonal problems (Schäfer & Najavits, 2007), higher severity of psychiatric symptoms (Knowles et al., 2019), and worse adherence to treatment interventions (McCauley et al., 2012). Despite research documenting the adverse impacts associated with multiple psychiatric conditions, limited research has examined how they impact individuals receiving PTSI rehabilitation in an RTW context.
This study aimed to examine if being simultaneously diagnosed with more than one psychiatric condition is predictive of RTW outcomes following rehabilitation for PTSI. This study focused on injured workers across Alberta being treated through the Workers’ Compensation Board of Alberta’s (WCB-Alberta) PTSI rehabilitation program. To address the objective of this study, we formulated two key research questions: Are individuals with multiple psychiatric diagnoses less likely to RTW following a PTSI rehabilitation program compared to individuals with only a single psychiatric diagnosis? How do individuals with multiple psychiatric diagnoses differ from those with a single psychiatric diagnosis in terms of demographic, occupational, injury-related, and clinical variables at program intake?
Method
Design
We conducted a population-based cohort study using anonymized data collected from individuals admitted into the WCB-Alberta PTSI program to examine how multiple psychiatric diagnoses impact RTW outcomes. The database included administrative data provided by WCB-Alberta and was supplemented by additional data extracted from clinical files detailing the injured workers’ symptoms, history of exposure to psychologically traumatic events, and scores on clinical measures of anxiety, depression, and traumatic stress. Two trained researchers conducted data extraction from supplementary claimant files, and all data underwent a double data verification process where an alternative data collector verified 10% of the claims collected. The University of Alberta Research Ethics Board approved this research on January 21st, 2022 (Pro00117127).
Sample
The merged database contains information on all injured workers (n = 773) admitted to the WCB-Alberta PTSI rehabilitation program across the province of Alberta between January 2017 and August 2019. Differentiating individuals with single and multiple psychiatric diagnoses, we categorized individuals into two groups. The single psychiatric diagnosis group (n = 627) was not significantly different from the multiple diagnoses group (n = 146) on demographic variables such as age, gender, and level of education (Table 1).
Demographic and administrative variables of injured workers admitted to a traumatic psychological injury vocational rehabilitation program between the years 2017–2019
Demographic and administrative variables of injured workers admitted to a traumatic psychological injury vocational rehabilitation program between the years 2017–2019
aIndicates that median was used instead of mean due to normality concerns. bP-value represents the significance level from the Kruskal-Wallis test. cCohen’s d effect size measurement (small d = 0.20, medium d = 0.50, and large d = 0.80). vCramer’s V effect size measurement (weak φc > 0.05, moderate φc > 0.10, strong φc > 0.15, and very strong φc > 0.25). pPearson’s r effect size measurement (small = 0.1–0.3, medium = 0.3–0.5, large > 0.5).
The primary outcome measure examined was the claimant’s RTW/fit to work (FTW) status at program discharge. There were no significant differences between the RTW, FTW and Failure to RTW groups on demographic variables such as age, gender and level of education (see Supplementary Table 1). Differences between the RTW, FTW and Failure to RTW groups for clinical measures were identified (see Supplementary Table 2). For logistic regression analysis, RTW and FTW were grouped and labelled as RTW, as both outcomes indicate that an individual has recovered to a point where they are considered able to work. Statistical comparison between the RTW and FTW groups also revealed that they were not significantly different across key variables (i.e., variables included in the final regression model), except for lower self-reported readiness to RTW within the FTW group (p < 0.001). RTW, in this study, is defined as the ability to return to the claimant’s previous place of occupation, performing the same or modified tasks and maintaining the same or modified hours upon program discharge. The RTW rate is a common measure used in workers’ compensation settings to measure the effectiveness of rehabilitation (Lagerveld et al., 2012; Pachoud, 2010).
Breakdown of groups (single versus multiple psychiatric diagnoses)
Breakdown of groups (single versus multiple psychiatric diagnoses)
The WCB-Alberta offers PTSI rehabilitation programs to address PTSI resulting from work-related accidents (Rose, 2006; WCB-Alberta, 2018). This treatment program provides a graded approach to vocational rehabilitation, beginning with a screening assessment conducted by a registered psychologist that assesses workers on clinical factors inhibiting RTW. This screen is used to determine the severity of the injury and assign workers to one of three levels of treatment (Rose, 2006; WCB-Alberta, 2018). The PTSI level one program consists of 1–2 appointments per week of 1-on-1 psychotherapy with a registered psychologist. The PTSI level two program consists of 1-on-1 psychotherapy with a registered psychologist 1–2 times per week and 1–2 treatments per week with an occupational therapist, targeting reintegration into society and the workforce. Finally, the PTSI level three program consists of multidisciplinary rehabilitation services such as individual psychotherapy, group therapy, physiotherapy for comorbid musculoskeletal injuries, and occupational therapy to target workplace reintegration. Treatment at this level is carried out 4 or more times per week for roughly 5 hours daily. The primary goal of all three levels of the PTSI rehabilitation program is to facilitate RTW for individuals who have developed a psychological injury following exposure to a potentially psychologically traumatic event in the workplace (Rose, 2006).
Psychiatric diagnostic categories
Injured workers were provided with a psychiatric diagnosis from the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association, 2013) upon referral to the PTSI program by a qualified professional (psychologist, provisional psychologist, psychiatrist, general practitioner). The treating psychologist at WCB-Alberta then confirmed psychiatric diagnoses upon admission to the PTSI program based on an intake assessment/interview which included the administration of a range of clinical measures (Beck Depression Inventory (BDI-II), Beck Anxiety Inventory (BAI), Trauma Symptom Inventory-2 (TSI-2)). Individuals admitted to the PTSI program primarily met the criteria for a Trauma- and Stressor-Related psychiatric diagnosis; however, a subset also met the diagnostic criteria for additional psychiatric diagnoses. Psychiatric diagnoses were categorized based on their DSM-5 diagnostic category (i.e., major depressive disorder was categorized under depressive disorders, generalized anxiety disorder was categorized under anxiety disorders, PTSD was categorized under trauma- and stressor-related disorders). Workers were further categorized based on whether they had single or multiple psychiatric diagnoses (see Table 2). Individuals with a trauma- and stressor-related diagnosis, as well as a psychiatric diagnosis from another DSM-5 diagnostic category, comprised the multiple diagnoses group, while individuals with only a trauma- and stressor-related diagnosis comprised the single diagnosis group. The two groups were compared to one another on outcome measures.
Measures
Descriptive (independent) variables
The dataset obtained from WCB-Alberta contained a range of descriptive variables, including demographic factors (e.g., age, gender, education level), occupational factors (e.g., occupational category, job attachment status, return-to-work outcome), and injury-related variables (e.g., the period between accident and assessment, previous claim history). Demographic variables such as gender and education level were collected through self-report measures administered by WCB-Alberta treatment providers at the time of program assessment, while other demographic and administrative variables (e.g., PTSI rehabilitation level, accident type, occupation) were collected from a WCB-Alberta administrative database.
Furthermore, data concerning potential risk factors that could impact treatment outcomes (e.g., self-reported history of psychologically traumatic events, risk of suicide, self-reported problematic substance use) and clinical measures of depression, anxiety, and posttraumatic stress (e.g., BDI-II, BAI, TSI-2), were extracted from worker intake reports administered at the time of admission to the PTSI program. History of psychologically traumatic events (i.e., does the worker have a history of exposure to psychologically traumatic events prior to the claim-related incident) and problematic substance use patterns were self-reported and recorded by the rehabilitation provider conducting the intake assessment. Clinical measures extracted include the following:
WCB-alberta’s psychology log
The Psychology Log is a 7-item measure used by WCB-Alberta to assess workers on various psychological variables. The Psychology Log was administered at intake to the PTSI rehabilitation program. The Psychology Log is rated on a 0–10 numerical scale and reports on injured workers’ self-reported current pain intensity, stress, energy, readiness to RTW, mood, the average estimated hours of sleep nightly, and ability to apply relaxation skills. The Psychology Log has been found to predict RTW outcomes in workers with PTSI (Gross et al., 2021).
Trauma symptom inventory-2
The TSI-2 is a 136-item self-report questionnaire that assesses symptoms of psychological trauma (Briere, 2011). The TSI-2 is measured on a 0–3 Likert scale, with higher scores indicating a higher severity of trauma-related symptoms. The TSI-2 is scored according to four overarching factors and 12 clinical subscales. The 4 factors are: 1) general trauma symptoms; 2) externalization; 3) self-disturbance; and 4) somatization. The trauma factor contains defensive avoidance, intrusive experiences, dissociation, and anxious arousal subscales. The externalization factor contains anger, tension reduction behaviour, sexual disturbance, and suicidality subscales. The self-disturbance factor includes insecure attachment, depression, and impaired self-reference subscales. Finally, the somatization factor consists of only the somatic preoccupation subscale (Briere, 2011).
Each subscale is comprised of 10 items and scored to produce categorical classifications of not clinically significant, problematic, and clinically elevated (Briere, 2011). For the current study, problematic and clinically elevated scores were grouped together, as both indicate the elevation of the subscale/factor. The measure contains two validity scales to detect false reporting, malingering, and random responding, which are scored to produce classifications of valid, atypical, and invalid (Briere, 2011). For the current study, results from the measure were removed if the validity scales produced a categorical rating of invalid. Data from the TSI-2 were included in the database if the validity scales produced a rating of valid or atypical. Responses that produced a rating of atypical remained in the database because research has demonstrated that a score of atypical on the TSI-2 can commonly arise due to genuine response patterns indicative of psychological distress (Ales & Erdodi, 2022). The TSI-2 has demonstrated good factor validity and strong measurement properties in clinical and research settings and has been successfully used in previous research to assess trauma symptomology in respondents during treatment and rehabilitation services (Every-Palmer et al., 2019; Godbout et al., 2016; Nilsson et al., 2018).
Beck depression inventory-II
The BDI-II is a 21-item self-report measure of depression with high internal consistency, external validity, and test-retest reliability in clinical and research settings (El-Den et al., 2018; Garcia-Batista et al., 2018). Each item is rated from 0–3 on a Likert scale, producing a total score out of 63 (Beck et al., 1996). Total scores correspond to one of four severity categories: 0–13 indicates minimal depression, 14–19 indicates mild depression, 20–28 indicates moderate depression, and 29–63 indicates severe depression (Beck et al., 1996). The BDI-II is a widely used measure that has been successfully applied in previous research to assess symptoms of depression in a variety of settings with various populations (Garcia-Batista et al., 2018; Iniesta et al., 2018; Naidu et al., 2019).
Beck anxiety inventory
The BAI is a widely used 21-item self-report anxiety measure with adequate to strong psychometric properties, including moderate-high ratings of internal consistency, inter-rater reliability, and test-retest reliability (Bardhoshi et al., 2016; Beck et al., 1988; Pang et al., 2019). The 21 items are rated individually on a 0–3 Likert scale, resulting in a total ranging from 0–63, with high scores indicating greater levels of anxiety-related symptomatology. Total scores correspond to one of four categories indicating varying severities: 0–7 indicates minimal anxiety, 8–15 indicates mild anxiety, 16–25 indicates moderate anxiety, and 26–63 indicates severe anxiety (Beck et al., 1988; Rector & Arnold, 2006). Previous research has successfully used the BAI in various clinical and research settings as a valid and reliable measure of anxiety (Bardhoshi et al., 2016; Pang et al., 2019).
Data analysis
Descriptive analysis
The merged database was initially cleaned and examined for missing data. Appropriate descriptive statistics were calculated for the full sample (n = 773) as well as independently for the single (n = 627) and multiple diagnoses groups (n = 146). Independent sample t tests (for continuous variables) and Chi-squared tests of independence (for categorical variables) were used to identify significant differences between the two groups. Median values were displayed for continuous variables that violated the normality assumption, and non-parametric tests were used to compare medians. The statistical analysis was conducted with IBM’s SPSS Statistics (Version 28).
Missing data analysis
A proportion of individuals had incomplete data on the clinical variables (BAI, BDI-II, TSI-2, and Psychology Log), with 211 (27.3%) workers missing at least one of these measures. Data were found not to be missing at random. Workers with missing data were less likely to work as Public Safety Personnel (PSP; 20.8% versus 38.1%, p < 0.001), have a history of psychologically traumatic events (57.8% versus 74.6%, p < 0.001), and display self-reported problematic substance use patterns (12.4% versus 25.4%, p < 0.001).
Alpha level and effect size interpretation
To adjust for inflated type I error, without introducing the degree of type II error that a Bonferroni correction might add, the alpha level, used to detect statistical significance in our descriptive comparisons, was set to < 0.001. For analysis conducted using t tests, the effect size measure Cohen’s d was used. The criteria we used to classify Cohen’s d effect sizes were small (d = 0.20), medium (d = 0.50), and large (d = 0.80; Cohen, 1988). The eta-squared measure of effect size was used for analysis using one-way Analysis of Variance (ANOVA), and the criteria we used to classify eta-squared measures of effect size were small (η2 = 0.01), medium (η2 = 0.06), and large (η2 = 0.14; Cohen, 1988). For categorical variables compared using Chi-squared tests of association Cramer’s V measure of effect size was applied using the criteria of weak (φc > 0.05), moderate (φc > 0.10), strong (φc > 0.15), and very strong (φc > 0.25; Akoglu, 2018). Finally, for non-parametric comparisons, Pearson’s r was used as a measure of effect size, with the criteria small (r = 0.10), medium (r = 0.30), and large (r = 0.50; Brydges, 2019).
Logistic regression analysis
We conducted a multivariable logistic regression analysis modelling RTW outcome to determine if being diagnosed with multiple psychiatric conditions is predictive of RTW outcome. For the logistic regression model, RTW and FTW were grouped and labelled as RTW. We used a risk-factor modelling strategy, with the multiple psychiatric diagnoses variable always forced into the model (Hosmer et al., 2013). Initially, a univariate screen was conducted to identify significant differences between the RTW and not FTW groups (Hosmer et al., 2013). Variables at a significance level of p < 0.01 were selected for the multivariable logistic regression analysis. Variables not reaching a significance of p < 0.10 were removed from the regression model, other than critical demographic variables such as age and gender, which were forced into the final model. The alpha level used to detect statistical significance for the regression model was set to p < 0.05. The relevant assumptions for logistic regression analyses (e.g., normality, collinearity, and linearity at the logit) were assessed and met. Relevant interactions were also tested.
Results
Characteristics of workers with single and multiple psychiatric diagnoses
There were numerous statistical differences in occupational and injury-related variables when comparing the single and multiple psychiatric diagnoses groups (see Table 1). Notably, fewer individuals in the multiple diagnoses group reported co-occurring physical injuries (26.7% versus 43.3%, p < 0.001), and more individuals in the multiple psychiatric diagnoses group reported a higher degree of self-reported problematic substance use (30.1% versus 18.2%, p < 0.001). Further, the median program length and time between accident and assessment for program intake were significantly longer for individuals with multiple psychiatric conditions (113.0 versus 91.0 days, p < 0.001 and 155.5 versus 76.0 days, p < 0.001, respectively).
Significant differences in clinical measures were also observed between the single and multiple diagnoses groups (see Table 3). The multiple psychiatric diagnoses group was significantly more likely to report severe scores on the BDI-II (69.2% % versus 49.5% %, p < 0.001) and reported elevated scores on the TSI-2 trauma (58.9% versus 44.0%, p < 0.001), self-disturbance (31.5% versus 17.2%, p < 0.001), and externalization factor scales (24.0% versus 12.9%, p < 0.001), and on numerous TSI-2 subscales (see Table 3). Most differences between the two groups were small to medium based on effect size estimates (see Tables 1 and 3).
Clinical measures of injured workers admitted to a traumatic psychological injury vocational rehabilitation program between the years 2017–2019
Clinical measures of injured workers admitted to a traumatic psychological injury vocational rehabilitation program between the years 2017–2019
Note. *p < 0.001. RTW = Return-to-work; FTW = Fit-to-work; SF-36 = 36-item Short Form Health Survey; TSI-2 = Trauma Symptom Inventory-2; TR = TSI-2 Trauma Factor Scale; SE = TSI-2 Self-Disturbance Factor Scale; EX = TSI-2 Externalization Factor Scale; SO = TSI-2 Somatization Factor Scale. 1Psychology Log subscale scores are expressed as an integer ranging between 0 and 10, except for ‘Sleep,’ which represents average hours of sleep per night, and ‘Readiness to RTW,’ which is displayed as a percentage from 0 to 100. cCohen’s d effect size measurement (small d = 0.20, medium d = 0.50, and large d = 0.80). vCramer’s V effect size measurement (weak φc > 0.05, moderate φc > 0.10, strong φc > 0.15, and very strong φc > 0.25).
Univariate and adjusted odds ratios (AOR) for each predictor variable’s association with the RTW outcome are shown in Table 4. The final multivariable logistic regression model indicated that workers had higher odds of successful RTW if they had a single psychiatric diagnosis (AOR 2.20, 95% CI 1.37 –3.51, p < 0.001). Individuals were also found to have higher odds of RTW if they had non-elevated scores on the TSI-2 trauma factor (AOR 1.85, 95% CI 1.27 –2.69, p < 0.001) and if they reported a higher readiness for RTW at program intake (AOR 1.24, 95% CI 1.14 –1.34, p < 0.001). Lower odds of RTW were found for individuals in education, law and social, community and government occupations (AOR 0.50, 95% CI 0.32 –0.79, p = 0.003). Finally, trauma history (AOR 0.68, 95% CI 0.45 –1.04, p = 0.07), working in trades occupations (AOR 0.72, 95% CI 0.47 –1.09, p = 0.12), being male (AOR 1.14, 95% CI 0.79 –1.64, p = 0.49), and age (AOR 1.01, 95% CI 0.99 –1.03, p = 0.30) were not statistically significant in the final regression model.
Logistic regression analysis predicting return-to-work at time of discharge from traumatic psychological injury vocational rehabilitation programs (n = 562)
Logistic regression analysis predicting return-to-work at time of discharge from traumatic psychological injury vocational rehabilitation programs (n = 562)
Note. *p < 0.05; **p < 0.01; ***p < 0.001. TSI-2 = Trauma Symptom Inventory-2.
Statistically significant interactions between claimant gender and factors predictive of RTW were also identified. In the final model, females were less likely to RTW than males if they had multiple psychiatric diagnoses (AOR 0.35, 95% C 0.13 –0.94, p = 0.04). Further, females in trades industries (AOR 0.41, 95% CI 0.17 –0.96, p = 0.04) and in education, law and social, community and government occupations (AOR 0.34, 95% CI 0.14 –0.85, p = 0.02) were also less likely to RTW.
We examined multiple simultaneous psychiatric diagnoses in the context of factors predictive of RTW outcomes for individuals admitted into a multidisciplinary rehabilitation program. Numerous factors were found to be predictive of RTW outcome at program discharge, such as occupational category, trauma-related symptomology, and self-reported readiness to RTW. Primarily, answering our first research question, we found that individuals with more than one psychiatric diagnosis were less likely to RTW following treatment in a PTSI rehabilitation program. These results appear to be among the first demonstrating worse RTW outcomes in individuals with multiple psychiatric diagnoses following rehabilitation for PTSI. Understanding the adverse impact of multiple psychiatric diagnoses on the outcomes of RTW rehabilitation can help inform rehabilitation providers as to which individuals are at an increased risk of failure to RTW and possibly in need of more intensive intervention.
Many PTSI rehabilitation programs, such as the one examined in the current study, do not provide differing levels of services or specifically designed interventions for individuals with multiple psychiatric diagnoses. In fact, in the current study, there were no significant differences between the single and multiple psychiatric diagnoses groups in terms of the type of program (i.e., level 1, 2, or 3) to which they were assigned (see Table 1). This suggests that the presence of multiple psychiatric diagnoses did not play a role in determining what treatment program individuals were triaged into. Despite this, our findings indicate numerous differences in demographic, occupational, injury-related, and clinical variables between these groups, suggesting that a different level of care may better meet the needs of individuals with multiple psychiatric diagnoses. Therefore, our findings could inform the development of PTSI rehabilitation programs specifically designed for individuals with multiple psychiatric diagnoses.
Although the impact of multiple psychiatric diagnoses on RTW outcomes has not been previously examined, other researchers have demonstrated, in different settings, that various combinations of psychiatric diagnoses are associated with a range of negative consequences. For example, researchers have suggested that various combinations of psychiatric conditions are associated with greater distress, impairment, and health care utilization (Rytwinski et al., 2013), lower levels of physical, psychological, and social functioning (McMillan et al., 2017), higher levels of occupational disability (Momartin et al., 2004), and a worse treatment prognosis (McCauley et al., 2012; Roberts et al., 2015). As such, our observation that RTW outcomes are worse for individuals in the multiple psychiatric diagnoses group appears consistent with the difficulties experienced by individuals with multiple psychiatric conditions in other settings, such as veterans or general clinical populations. However, our data did not allow us to examine whether certain combinations of diagnoses limit RTW more than others or the influence of pre-existing psychiatric disorders.
Interaction effects with gender and occupation were also assessed. It is worth noting that gender was only collected by WCB-Alberta on a binary scale and, as such, may be unrepresentative of the range of gender identities present in the population. Our results indicated that gender interacted with the psychiatric diagnosis variable, resulting in more favourable RTW outcomes for men with multiple psychiatric diagnoses. This finding suggests that the presence of two or more simultaneous psychiatric conditions could adversely impact females’ RTW outcomes more than males’. Further, our results indicated that gender and occupation also interacted, resulting in women in trades, government, education, law and social and community occupations having worse RTW outcomes. One possible explanation for the discrepancy between women and men may be related to certain types of psychiatric diagnoses differing in their prevalence based on gender. Another potential explanation is that the PTSI rehabilitation program itself has difficulties adequately addressing the needs of certain populations, such as women diagnosed with multiple psychiatric diagnoses. To our knowledge, previous researchers have not examined the influence of gender on RTW in individuals with multiple psychiatric diagnoses. Future research should focus on examining why these gendered interaction effects exist so that PTSI rehabilitation programs can be improved.
The demographic, occupational, injury-related and clinical variables were also compared between workers with single and multiple psychiatric diagnoses. Previous researchers have suggested that individuals with multiple psychiatric diagnoses report a higher quantity and severity of mental health symptoms (Momartin et al., 2004; Knowles et al., 2019). Our findings supported this; particularly, we found that psychiatric symptom severity (i.e., depression, trauma, hyperarousal, suicidal ideation, defensive avoidance) was higher in the multiple diagnoses group. Additionally, we found that individuals in the multiple diagnoses group had longer program durations and were more likely to self-report problematic substance use.
Individuals in the multiple diagnoses group also had a more extended period between the date of accident and assessment. Although the association between multiple psychiatric conditions and longer accident to assessment periods has not been previously demonstrated, it is theoretically supported by the network model of comorbidity, which suggests that mental health symptoms, if left untreated, may cause the development of further symptoms and subsequent diagnoses over time (Cramer et al., 2010). This finding may suggest that leaving psychological injuries untreated over time can increase the risk of multiple psychiatric conditions. However, this finding could also reflect the idea that individuals with more than one psychiatric condition experience an increased number of administrative difficulties when filing WCB-Alberta claims. Future research is needed to understand why the association between multiple psychiatric diagnoses and a longer accident to assessment time exists. The differences we observed in demographic, occupational, injury-related and clinical variables between workers with single and multiple psychiatric diagnoses highlight the need to consider these groups separately, which could be used to inform the development of PTSI rehabilitation programs for individuals with multiple psychiatric diagnoses.
Limitations
There are some limitations in the present study, which may indicate possibilities for future research. First, the current study utilized secondary analysis of archived WCB-Alberta data. This may restrict the current results when applied to treatment contexts outside Alberta. Second, because we used archived data, certain variables such as race and ethnicity could not be collected and reported upon. The impact of race and ethnicity would have been beneficial to examine, as our findings indicated significant interactions with demographic variables, including gender. Other demographic variables that were not available, such as race and ethnicity, could also impact RTW outcomes. Third, only a small number of workers had a full RTW without restrictions; therefore, we could not further analyze our results based on the degree of RTW achieved at program discharge. However, the data represent real-world observations from all workers with PTSI treated within the jurisdiction during the study period.
Fourth, we could not run internal consistency measures on our clinical measures (i.e., BDI-II, BAI, TSI-2) because the raw score data were unavailable. However, comparing these variables between the psychiatric diagnoses groups was primarily exploratory, so the lack of internal consistency measures does not greatly diminish our overall findings. Fifth, there was a considerable amount of missing data for the patient-reported outcome measures. Despite this, the sample size for the final regression model (n = 562) was still ample for logistic regression modelling. Data were also found not to be missing at random across numerous variables. As such, results should be interpreted with caution due to the differences between individuals with and without missing data. Finally, a comparison between different combinations of psychiatric diagnoses was not possible due to insufficient sample size.
Conclusion
Being diagnosed with multiple psychiatric conditions appears to be associated with more negative RTW outcomes following rehabilitation for PTSI. Workers with multiple compared to single psychiatric diagnoses also take longer to be assessed by treatment providers following their injury, report a lower readiness to RTW, have higher depression and traumatic stress severity scores, and are more likely to have self-reported substance use problems. These findings are important as the impact of multiple psychiatric diagnoses has primarily been examined in other treatment contexts and has not focused on examining RTW as a treatment outcome. Future research is needed to examine the influence of pre-existing psychiatric disorders and identify if certain combinations of diagnoses limit RTW more than others. Future research should also examine why PTSI rehabilitation programs may be less effective for certain populations (i.e., women with multiple psychiatric diagnoses), as these differences in effectiveness could be due to inadequately meeting the needs of specific populations. Finally, we suggest that our results inform future research focused on developing RTW treatment programs for individuals with multiple psychiatric diagnoses following a PTSI.
Footnotes
Acknowledgments
The authors would like to thank the Workers’ Compensation Board of Alberta for providing the data used in this research.
Conflict of interest
The authors report no conflict of interest.
Ethics statement
The University of Alberta Research Ethics Board approved this research on January 21st, 2022 (Pro00117127).
Funding
The results reported herein correspond to specific aims of grant 20SPHIFR33-2 awarded to Dr. Douglas Gross from Alberta Labour and Immigration. This work was also supported by a Research Grant from the Workers’ Compensation Board of Alberta. Research support to Professor Straube was provided by a grant from the Workers’ Compensation Board of Alberta, ‘Program of Research and Training in Occupational Medicine’, which was used to employ Dr. Jackson.
Informed consent
This study did not require informed consent.
