Introduction
United Nations (UN) personnel seek to address a wide range of political, social, environmental, and cultural crises throughout the world (e.g. UN Department of Economic and Social Affairs (n.d.)). Although the focus of UN work varies widely across entities and job type, UN personnel may be at high risk for psychiatric disorders as their work may involve exposure to high levels of stress and trauma. For example, several UN entities are directly involved in providing humanitarian relief in response to war, environmental disasters, and other emergencies (United Nations Disaster Assessment and Coordination. UNDAC Field Handbook, 2013). UN personnel may also be exposed to considerable stress and trauma through human rights policy and advocacy work, which often entails the documentation human rights violations through interviewing victims, reviewing evidence, and visiting physical sites of human rights violations (Gallagher, Reference Gallagher, Burkey, Skåre and Mørk2008). Moreover, the UN also employs sanctioned peacekeepers – military personnel under the UN command – involved in monitoring and observing peace processes in post-conflict settings and providing security in regions where there are current conflicts (Ratner, Reference Ratner1995). Although the primary objectives of UN personnel may vary considerably, the stressors and challenges they face may overlap considerably as they often work in the similar contexts, engage in each other's roles (e.g. advocates and peacekeepers may help deliver aid), and influence one another's work.
Although no studies to our knowledge have been conducted with UN personnel, a growing body of research suggests that individuals engaged in humanitarian aid, human rights advocacy, and peacekeeping report high levels of direct and secondary trauma. The International Committee of the Red Cross (ICRC) recently reported that over the past 2 years there have been approximately 1200 incidents of violence against health-care workers in 16 countries (ICRC, 2018). There is also a growing body of evidence documenting risks associated with human rights advocacy and activism. Amnesty International recently launched a campaign, ‘Brave. Human Rights Defenders Under Threat: A Shrinking Space for Civil Society’ documenting an increase in death threats against human rights advocates6 and advocates report high levels of direct and secondary trauma (Amnesty International, 2017). Peacekeepers are also exposed to direct threat and violence. A recent study found that a majority of Australian peacekeepers reported being in a situation in which there was a threat of injury (83%) or death [73%, (Forbes et al., Reference Forbes, O'Donnell, Brand, Korn, Creamer, McFarlane, Sim, Forbes and Hawthrone2016))]. In a 2015 report, the High-Level Independent Panel on UN Peace Operations emphasized that UN personnel operate in ‘increasingly dangerous environments’ (Willmot et al., Reference Willmot, Sheeran and Sharland2015).
Compared to other populations routinely exposed to traumatic stress in the context of their work, such as combat veterans, there remains a paucity of research on psychiatric morbidity associated with individuals involved in aid work, peacekeeping, and human rights advocacy. To date, studies investigating rates of psychiatric disorders in humanitarian aid workers vary considerably from 8 to 43% for PTSD and 8–20% for depression (Connorton et al., Reference Connorton, Perry, Hemenway and Miller2012). Only two studies have been conducted with human rights advocates in which rates of depression ranged from 8.6 to 16% and PTSD from 7.1 to 20% (Holtz et al., Reference Holtz, Salama, Lopes Cardozo and Gotway2002; Joscelyne et al., Reference Joscelyne, Knuckey, Satterthwaite, Bryant, Li, Qian and Brown2015). A meta-analysis of peacekeeper studies reported that rates of PTSD ranged from 0.05 to 2.8% [pooled prevalence rates = 5.3% (Souza et al., Reference Souza, Figueira, Mendlowicz, Volchan, Portella, Mendonça-de-Souza and Coutinho2011)].
The extant data are based primarily on studies with small samples, in specific contexts, employing different methods of assessment and as such, it is difficult to determine the levels of psychiatric disorders in this population. Furthermore, even less is known about the occupational and organizational factors that might be associated with psychiatric vulnerability. Clarifying those factors associated with risk for psychiatric disorders could serve as an important guide for policy-makers to prevent and reduce the psychiatric burden on this global workforce. Furthermore, little is known about the potential barriers to accessing mental healthcare in this population. To fill these gaps, UN personnel were asked to complete an online survey screening for psychiatric disorders, socio-demographic questions, and utilization of mental healthcare services.
Methods
Participants
Organizations which had a direct service–delivery relationship with the Medical Services Division of the UN were invited to participate. This included the New York-based agencies, funds and programs, and the various UN secretariat entities around the world, including peacekeeping missions, Economic Commissions and the UN offices in Geneva, Nairobi, and Vienna. Specifically, the participating entities included Department of Field Support (DFS), United Nations Development Programme (UNDP), United Nations Office at Nairobi (UNON), the United Nations Children's Fund (UNICEF), Secretariat Headquarters in NY, United Nations Department of Safety and Security (UNDSS), United Nations Office at Geneva (UNOG), Economic Commission for Latin America and the Caribbean (ECLAC), Economic and Social Commission for Asia and the Pacific (ESCAP), United Nations Office at Vienna (UNOV), and Economic and Social Commission for Western Asia (ESCWA).
A total of 45 544 participants from 11 UN entities were invited to participate in the study.Footnote 1Footnote † In total, 17 363 UN personnel participated in the survey, representing a convenience sample of 32.8% personnel employed by participating agencies and 17.7% of all UN personnel (see Table 1 for demographic and occupation-related sample characteristics of survey respondents). Although our sample differed from available UN-wide demographic data on Gender, Age, Parental Status, Contract Type, and Duty Station Type (p < 0.0001), Cramer's V showed that the effect sizes were small (φc = 0.07–0.14).
Demographic and occupational characteristics of UN staff members (n = 17 363)
Note: count (percentage).
Among those who participated in the survey, 79.27% (N = 13 763) completed all of the questions. A series of χ2 tests were conducted comparing the demographic variables among survey responders that completed all three clinical measures, partially completed the clinical measures, and did not complete any of the clinical measures (see supplemental materials for complete analyses). No differences were found between those who completed v. those who partially completed the clinical self-report measures on Gender, Age, Relationship Status, Duty Station Type, and Duration of Employment for the UN. A marginally significant difference was found for Contract Type. Those who completed the clinical measures were more likely to be on fixed-term appointments (χ2 = 3.84, p = 0.05) whereas those who partially completed the clinical measures were more likely to be on temporary appointment (χ2 = 4.00, p = 0.05). When comparing those who completed the clinical measures, compared to those that did not complete any of the clinical measures, there were no differences found for gender or duty station type. Individuals that did not complete the surveys were more likely to be below 34 and younger (χ2 = 72.95, p < 0.0001), not in a relationship (χ2 = 17.00, p < 0.0001), worked for the UN for <1 year (χ2 = 18.48, p < 0.0001) or between 1 and 3 years (χ2 = 4.43, p < 0.05), and were on temporary (χ2 = 16.25, p < 0.05) or consultancy appointments (χ2 = 3.87, p < 0.05).
All UN personnel from these 11 UN entities were eligible for inclusion in the study. Missing data were accommodated using case-wise deletion. The data for this cross-sectional prevalence study was originally collected as part of an internal assessment carried out by the Medical Services Division of the UN. The de-identified data were shared with the NYU School of Medicine for analysis. The NYU School of Medicine Institutional Review Board determined that the analysis of these de-identified archival data was exempt from IRB review per 45 CFR 46.101(b) #4.
Procedures
The survey was distributed by the UN Medical Service Division to UN personnel at the 11 UN agencies over an 8 month period between 2015 and 2016. Data were collected via an on-line questionnaire using SurveyMonkey. An e-mail with information on the purpose of the survey, confidentiality clauses, the approximate time needed to complete the survey, and a link to the survey was sent to the personnel of all participating organizations. No personally identifying information was collected.
Measures
Demographics
Individuals were asked to complete a series of demographic questions about gender, age, relationship status, parental status, and duration of UN employment. Participants were also asked about the type of appointment (e.g. permanent, temporary, consultant), recruitment type (local v. international), and duty station type (family v. non-family). Participation in the survey was voluntary. The survey was available primarily in English for the majority of the organizations.Footnote 2
Trauma exposure
Participants were asked if they had experienced or witnessed an event in the past 12 months that involved actual or threatened death or serious injury either during work or while they were off-duty.
Generalized anxiety disorder (GAD)
Gad was measured with the 7-item self-report instrument GAD-7 (Spitzer et al., Reference Spitzer, Kroenke, Williams and Löwe2006). The GAD-7 assesses symptoms of GAD within the past 2 weeks. A cut-off score of ⩾10 is defined as a positive screen for GAD (Spitzer et al., Reference Spitzer, Kroenke, Williams and Löwe2006).
Major depressive disorder (MDD)
MDD was screened with the Patient Health Questionnaire-9 (PHQ-9), which is a validated 9-item self-report screening instrument for depression (Kroenke et al., Reference Kroenke, Spitzer and Williams2001). The PHQ-9 assesses symptoms of depression over the past 2 weeks. A cut-off score of ⩾10 defines a positive screen for MDD (Kroenke et al., Reference Kroenke, Spitzer and Williams2001).
Posttraumatic stress disorder (PTSD)
To screen for PTSD, the 6-item PTSD Checklist-6 (PCL-6) was used (Lang et al., Reference Lang, Wilkins, Roy-Byrne, Golinelli, Chavira, Sherbourne, Rose, Bystritsky, Sullivan, Craske and Stein2012). The PCL assesses symptoms of PTSD within the past month. A cut-off score of ⩾14 is considered a positive screen for PTSD in the PCL-6 (Lang et al., Reference Lang, Wilkins, Roy-Byrne, Golinelli, Chavira, Sherbourne, Rose, Bystritsky, Sullivan, Craske and Stein2012).
Mental healthcare utilization
Participants were asked if they had felt the need to see a mental health professional (e.g. staff counselor) within the past 12 months and if they were currently receiving mental health services within or outside the UN.
Statistical analysis
Estimated point prevalence rates of participants who screened positive for GAD, MDD, or PTSD were ascertained as were the rates of reported mental healthcare utilization in the past 12-months. Multivariable logistic regression analyses were used to examine the associations among demographic as well as occupational factors and GAD, MDD, or PTSD. The binary outcome variable was case status for each participant with separate models run for each of the three disorders. The statistical significance level employed in the models was α = 0.05. The preparation of the paper followed the statement for strengthening the reporting of observational studies in epidemiology [STROBE (Vandenbrouckel et al., Reference Vandenbrouckel, von Elm, Altman, Gøtzsche, Mulrow, Pocock, Schlesselman and Egger2007; von Elm et al., Reference von Elm, Altman, Egger, Pocock, Gotzsche and Vandenbroucke2007) (see Figure 1)].
Flow diagram: United Nations Mental Well-Being Survey.
Results
Estimated point prevalence of screen positives for GAD, MDD, and PTSD
Among survey respondents, 9118 individuals (65.32%) did not meet the screening criteria for any for the measured mental health disorders. However, 17.9% (N = 2759) reported symptoms that met the threshold for GAD, 22.8% (N = 3417) for MDD, and 19.9% (n = 2823) for PTSD (see Fig. 2). In addition, 34.67% (n = 4840) of UN members screened positive for at least one mental health disorder, 18.01% (n = 2523) screened positive for at least two mental health disorders, and 7.62% (n = 1064) screened positive for all three mental health disorders.
Percentages of UN employees who screen positive for mental health disorders.
Predictors of mental health disorders
To identify predictors of GAD, MDD, and PTSD a series of multivariable logistic regressions were conducted (see Tables 1–4).
Generalized anxiety disorder
The parameter estimates of multivariable logistic regressions are reported in Table 2. When accounting for all predictors in the multivariable logistic regression, several factors remained statistically significant as predictors of GAD (see Table 2). GAD was significantly associated with low job satisfaction (p < 0.0001), followed by longer duration of UN employment (p < 0.0001), younger age (p < 0.001), and trauma exposure on (p < 0.001) or off-duty (p < 0.0001).
Parameter estimates for the multivariable logistic regression including covariates to predict the presence of generalized anxiety disorder in UN staff members (n = 12 732)
Note: AUC = 0.73. Values in bold text connote statistical significance, p < 0.05, p < 0.01, p < 0.001, p < 0.0001.
Major depressive disorder
Similarly, multivariable logistic regression revealed that MDD associated with low job satisfaction (p < 0.0001), followed by longer duration of UN employment (p < 0.0001), younger age (p < 0.01), and trauma exposure on (p < 0.01) or off duty exposure (p < 0.01), (see Table 3).
Parameter estimates for the multivariable logistic regression including covariates to predict the presence of major depressive disorder in UN staff members (n = 12 732)
Note: AUC = 0.74. Values in bold text connote statistical significance, p < 0.05, p < 0.01, p < 0.001, p < 0.0001.
Posttraumatic stress disorder
When accounting for all covariates in the multivariable logistic regression analysis, six factors remained statistically significant as predictors of PTSD (see Table 4). These factors were younger age (p < 0.0001), not being in a relationship (p < 0.05), parents with a dependent child (p < 0.05), longer duration of UN employment (p < 0.0001), lower job satisfaction (p < 0.0001), and trauma exposure during work (p < 0.0001) and off-duty (p < 0.0001).
Parameter estimates for the multivariable logistic regression including covariates to predict the presence of posttraumatic stress disorder in UN staff members (n = 12 732)
Note: AUC = 0.79. Values in bold text connote statistical significance, p < 0.05, p < 0.01, p < 0.001, p < 0.0001.
Predictors of comorbidity: combination of GAD, MDD, and PTSD
In an attempt to identify the most distressed population of UN employees, multivariable logistic regression was conducted to identify predictors of triple comorbidity (GAD, MDD, and PTSD). After accounting for all covariates in the model, younger age (p < 0.0001), female gender (p < 0.01), partnership status (single) (p < 0.0001), parents with a dependent child (p < 0.0001), longer duration of employment with the UN (p < 0.0001), lower job satisfaction (p < 0.0001), and trauma exposure on (p < 0.0001) and off duty (p < 0.0001) predicted the group that met current screening criteria for all three disorders (see Table 5).
Parameter estimates for the multivariable logistic regression including covariates to predict comorbidity of generalized anxiety disorder, major depressive disorder and posttraumatic stress disorder UN staff members (n = 12 732)
Note: AUC = 0.75. Values in bold text connote statistical significance, p < 0.05, p < 0.01, p < 0.001, p < 0.0001.
Mental healthcare services
Next, the self-reported frequency of mental health services utilization was examined. In the past year 1.68% (n = 197) of the total UN personnel in the survey reported receiving mental health services within the UN and 4.58% (n = 538) sought mental healthcare services outside the UN. Among those who screened positive for one or more of GAD, MDD, or PTSD, 10.01% reported using current mental health treatment outside of the UN and 2.08% within the UN system. The mental healthcare rate of utilization for individuals that screen positive for GAD, MDD, and PTSD was 7.6%.
Within the past year, 93.75% (n = 11 017) of the UN personnel in the total sample reported that they did not receive mental healthcare during the previous year. Among this group, 15.1% (n = 858) responded that they did not do so because they were ‘not comfortable’ and 10.2% (n = 582) stated that they did not know if such services were available.
Discussion
This survey found that a high number of UN personnel in this study screened positive for mental health disorders. The percentage of those screening positive for GAD, MDD, or PTSD is considerably higher than the prevalence in the world general population (e.g. Kessler et al., Reference Kessler, Berglund, Demler, Jin, Merikangas and Walters2005). Remarkably, about one-third of all UN personnel surveyed reported symptoms that qualify for screening positive for at least one mental health disorder. This rate is in the upper range of individuals screening positive for anxiety disorders, MDD or PTSD found in personnel that are recurrently exposed to occupational and potentially traumatic stress, such as combat veterans and police (Hoge et al., Reference Hoge, Castro, Messer, McGurk, Cotting and Koffman2004; Marmar et al., Reference Marmar, McCaslin, Metzler, Best, Weiss, Fagan, Liberman, Pole, Otte, Yehuda, Mohr and Neylan2006).
Relative to the high percentage of those screening positive for mental health disorders in this study, UN personnel reported low levels of mental healthcare utilization. These findings suggest that there is a large treatment gap for UN personnel and further underscore the large treatment gaps that have been identified globally (Patel and Prince, Reference Patel and Prince2010). These data are also consistent with unpublished data from UNHCR (Dubravka et al., Reference Dubravka, Thomas, Jachens and Mihalca2016) in which staff reported ‘not feeling comfortable’ with mental health services or not knowing if services were ‘available’. The treatment gap for this population is probably even larger given that all of the individuals in this survey were employed by the UN, an intergovernmental organization that has greater resources than many smaller non-governmental organizations (NGOs) and local agencies throughout the world. Our findings suggest that stigma about seeking mental health services may be a key barrier to care and that organizations in this field would benefit from studying negative attitudes towards help seeking within their staff and incorporating established methods for reducing mental health gaps in occupational settings (Joyce et al., Reference Joyce, Modini, Christensen, Mykletun, Bryant, Mitchell and Harvey2016).
In addition, a number of factors were significantly associated with GAD, MDD, and PTSD in UN personnel. A consistent factor across all three psychiatric symptom categories was low levels of job satisfaction. Although it cannot be determined from this survey whether low job satisfaction is a risk factor or an outcome of mental health disorders, other studies have found that job satisfaction is a robust predictor of mental health outcomes (Faragher et al., Reference Faragher, Cass and Cooper2005) and the strong association between job satisfaction and mental health underscores the need for aid and advocacy groups to clarify which factors impact job satisfaction within their organizations.
In addition, after working for 5 years for the UN, the likelihood of screening positive for a mental health disorder is two times higher than working for <1 year and after 10 years, approximately three times higher than working for <1 year.
The link between duration of employment and risk for psychiatric disorders may be due to cumulative exposure to stress and trauma during and outside of work, which previous research has found to be associated with negative mental health outcomes, such as PTSD and depression (Kraaij and de Wilde, Reference Kraaij and de Wilde2001; Breslau et al., Reference Breslau, Peterson and Schultz2008). Trainings and the identification of mental health vulnerabilities within the first year of work may represent a critical window for preventing and reducing mental health issues within aid and advocacy organizations. However, given the cross-sectional nature of this study, prospective longitudinal studies are warranted to further shed light on these findings.
Another significant predictor for psychiatric disorders was the exposure to a traumatic event within the past 12 months as reported by UN personnel, either during work or while off-duty. These findings suggest that given the high rates of trauma exposure among UN personnel and its relation to negative mental health outcomes, mental health programs within the UN can target those staff whose jobs will most likely involve direct or indirect exposure. Future research would benefit from testing whether resilience training programs or mental health screening initiatives employed by other fields in which personnel are often exposed to trauma could be adapted for this population (Adler et al., Reference Adler, Bliese, McGurk, Hoge and Castro2009; McCraty and Atkinson, Reference McCraty and Atkinson2012).
Although it might be expected that UN staff would be exposed to high levels of trauma during work, rates of exposure were also high when UN staff were off-duty. Additional follow-up work warrants a better understanding of the nature of this off-duty trauma.
Limitations of this survey are the inability to determine causality due to the cross-sectional nature of the study. A second limitation is that these data were obtained through convenience sampling and cannot be determined if these findings are representative of the UN as a whole. Given the large sample size, it is possible that the high rates of individuals screening positive psychiatric disorders and low rates of mental healthcare utilization observed in this study may be found across the UN. Therefore, future research with a representative sample of personnel will be critical for understanding the extent to which these findings are generalizable across the entire organization. Along those lines, there may be differences in the screening rates and factors associated with psychiatric disorders and mental healthcare utilizations across the UN agencies that participated in the survey. However, the data-sharing agreement between the UN and NYU stipulated that the data would not be analyzed or reported at the individual agency level. Another limitation was the total of potential survey participants was estimated, as the exact number of consultants employed at the time of the survey is not included in the published figures on staff totals. It is believed that the presence of consultants did not have a major impact on the findings given that only 1% of the total sample comprised consultants. Additionally, the data in this study may represent non-response bias as it cannot be excluded that missing values were not missing at random. For example, individuals that were younger, not in a relationship, and with less permanent appointments were less likely to answer the clinical measures. Follow up research is needed to better understand whether these patterns are clinically relevant. An alternative explanation may be that those who were newer to the UN with temporary appointments may have been less motivated to complete all of the measures or they may have been more concerned about disclosing information related to psychiatric symptoms. Furthermore, the presence of psychiatric disorders was determined with self-report measures. Previous work has suggested that self-report measures, compared with screening based on clinical interviews, may yield higher rates of those screening positive for mental health disorders due to memory or response biases, and the interpretation of the instructions or items on each self-report measure (e.g. Engelhard et al., Reference Engelhard, Arntz and van den Hout2007).
Taken together, the findings of this study strongly suggest that the implementation of a comprehensive mental health strategy is urgently needed throughout the UN. Proactive measures should be taken to actively target those factors at the organizational, environmental, and individual level associated with mental health disorders in this population. UN work is quite diverse in terms of its occupational demands and its environmental and geographic contexts, and interventions may need to be tailored to specific occupational or demographic subgroups. Furthermore, these findings point to the need to better understand barriers to care given to the large mental health services gap observed in this study. Humanitarian aid workers, human rights advocates, and peacekeepers are charged with work such as carrying out research and policy development to the direct provision of basic aid for the world's most vulnerable communities. We recommend that aid and advocacy organizations and agencies partner with mental health researchers to develop policies and programs for addressing psychiatric disorders among people working on many of the world's most urgent issues.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/gmh.2019.29.
Acknowledgements
Katharina Schultebraucks was supported by the German Research Foundation (SCHU 3259/1-1). We also appreciate contributions from Webster University and UNHCR in the development of the survey. JFF, YB, and AM, are employed by the United Nations and views expressed herein are those of the author(s) and do not necessarily reflect the views of the United Nations.
Open access
