Introduction

Traumatic events such as serious accidents, physical and sexual abuse, and sudden deaths of loved ones are common occurrences globally [1–3]. The World Mental Health Surveys recently assessed the prevalence of 29 types of traumatic experiences across 24 countries and found that 70.4% of respondents experienced at least one traumatic event in their lifetime [2, 3]. In high-income countries like Norway, at least one serious lifetime event was reported by 85% of men and 86% of women [4]. Studies focusing on trauma exposure in sub-Saharan Africa have yielded variable findings. For example, a study among urban schools in Kenya and South Africa found that more than 80% of participants reported either directly experiencing or witnessing at least one traumatic event [5]. Another study in Kenya reported that 64.4% of participants had experienced at least one potentially traumatic event [6]. A study in Ethiopia noted that the prevalence of experiencing a potentially traumatic event was 48.7% [7]. In Uganda, most of the studies on traumatic events have focused on conflict areas such as northern Uganda. For example, the Wayo-Nero study in northern Uganda found that 80.8% of participants experienced at least one war-related traumatic event (e.g., war-related physical trauma, war-related psychological trauma, and war-related sexual trauma), and 47.2% experienced at least three war-related traumatic events [8]. However, there is limited information on how measures perform for assessing traumas in Uganda.

Potentially traumatic events (PTEs) are associated with many serious adverse mental health effects. It is estimated that about 15–25% of those who experience a traumatic event will go on to develop post-traumatic stress disorder [3] and other psychiatric conditions, including depressive disorders, substance use disorders, anxiety disorders, and psychotic disorders, among others [9, 10]. Studies have also shown that people with psychiatric disorders are more likely to have experienced PTEs [11, 12]. Psychiatric conditions impair functioning and add to the global burden of disease [13]. Therefore, screening for PTEs is important to prevent the development of psychiatric disorders and provide timely treatment.

Given the high prevalence and impact of traumatic events, it is imperative to have trauma assessment tools that can be utilized in clinical settings as well as within the general population. Individuals with psychotic disorders, in particular, are at an increased risk of experiencing traumatic events, and up to 25–50% of patients with psychotic spectrum disorders in high-income countries have co-occurring PTSD [14–16]. Additionally, high rates of traumatic events and PTSD, particularly interpersonal traumas, have been reported for individuals with psychotic disorders in some low- and middle-income countries, including India [17, 18]. Therefore, understanding traumatic events relevant to patients with psychotic disorders may be especially important for screening and treating these patients with the proper tools. Certain PTEs may group together and have a higher chance of leading to certain symptoms (e.g., interpersonal PTEs and psychotic disorders). The relationship between these clusters may be helpful to explore their relationship to psychopathology [19, 20]. Measuring and modeling trauma exposure as composite variables may be helpful to this end by aggregating two or more highly related traumatic events [21].

Multiple assessment tools are available for use in evaluating PTEs. The Life Events Checklist (LEC-5; Weathers et al., 2013) for the Diagnostic and Statistical Manual of Mental Disorders, 5^th edition (DSM-5; [22] is a widely used measure. The LEC has previously been applied in many cultural settings, including Africa [23, 24]. However, to our knowledge, it has not been used in the Ugandan setting before. While the LEC is one of the most widely used instruments to assess exposure to PTEs [25], only a handful of studies have evaluated the factor structure of the scale [6, 7, 26, 27] including our colleagues’ findings from the larger parent study for Ethiopia, Kenya, and South Africa [6, 7, 27]. These studies showed good factorial validity for the LEC-5 [6, 7, 26, 27]. The factor structure of most tools for assessing traumatic events, including LEC-5, have not been assessed for use in the Ugandan setting.

This study aimed to determine the prevalence of PTEs for Ugandan adults, including specifically for cases (i.e., individuals with a psychosis-related diagnosis) and controls (i.e., individuals in a general medical setting) and to evaluate the factor structure of the LEC-5 in clinical settings in Uganda with data from a large neuropsychiatric study. Understanding the prevalence of PTEs between cases and controls is essential for clinical relevance and can guide interventions, treatments, and strategies aimed at addressing the specific needs of each group.

Materials and methods

Results

This study involved a total of 4,479 participants (2,375 cases and 2,104 controls), with a mean age of 35.9 years (SD = 12.2), the majority (76.6%) of whom were aged below 45 years. There were slightly more female (54.1%) than male participants, and the majority of the sample (74.5%) had either some primary or some secondary education (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Ugandan sample demographics (N = 4479) ^*.

https://doi.org/10.1371/journal.pone.0298385.t001

Overall, 60.5% of participants reported experiencing one or more PTEs, while 19.3% endorsed three or more PTEs (Table 2). The most frequently endorsed PTEs were physical assault (22.8%) and exposure to combat/war-zone (15.1%). Exposure to toxic substances was the least endorsed item (1.7%). The frequency of traumatic events stratified by case-control status is also presented in Table 2. Cases (64.2%) were more likely to have experienced one or more PTEs compared to controls (55.4%, p < 0.001). Cases were also more likely than controls to report experiencing serious accidents (8.9% vs. 6.6%; p = 0.005), physical assault (28.6% vs 16.2%; p < 0.001), weapon assault (12.8% vs. 10.1%; p = 0.006), sexual assault (11.5% vs. 5.0%; p < 0.001), and unwanted sexual experiences (7.3% vs. 5.1%; p = 0.004). Controls were more likely than cases to have experienced illness/serious injury (11.4% vs. 8.1%; p < 0.001) and sudden violent or accidental death of a close loved one (12.9% vs. 11.3%; p = 0.039).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Frequency of traumatic events according to case control status.

https://doi.org/10.1371/journal.pone.0298385.t002

The EFA analyses conducted on a random split-half of our sample resulted in a five-factor solution with item loadings varying from 0.30 to 0.82 (Table 3). The five factors were labeled as follows based on item grouping: sexual violence, war and death, physical violence, environment and war-related trauma, and accidents and injury. Three items (combat, severe suffering, and fire/explosion) cross-loaded on more than one factor but were assigned to the factor with the higher loading. Lastly, the item serious accident had a factor loading below the 0.3 cut-off and was not assigned to any of the five factors from the EFA.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Principal component analysis with varimax rotation of Life Events Checklist for DSM-5 (LEC-5).

https://doi.org/10.1371/journal.pone.0298385.t003

Next, we examined the factor structure of the LEC-5 by testing and comparing all three models for fit. The first model was the EFA from this study in the other split-half of our sample. The second model was based on a prior six-factor model and the WHO—World Mental Health Survey findings [2, 26], and the third was a theoretical seven-factor model [33, 34]. CFA models were analyzed with a sample variance-covariance matrix and a maximum likelihood minimization function. Overall, the goodness-of-fit indices showed that the theoretical seven-factor model provided the best fit comparatively (Table 4; Chi-square (85) = 337.08, p < 0.001; SRMR = 0.022, RMSEA = 0.022, TLI = 0.908, CFI = 0.935). In addition, completely standardized parameter estimates from each model solution are presented in Figs 1 and 2. Factor loading estimates showed that items from LEC-5 were strongly related to their specified latent factors in the seven-factor model, except natural disasters (0.18), toxic substance exposure (0.16), and serious illness (0.23), which fell below the cut-off of 0.3 for factor loadings (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Final seven-factor model selected for Life Events Checklist for DSM-5 (LEC-%) in Uganda.

https://doi.org/10.1371/journal.pone.0298385.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Additional models for Life Events Checklist for DSM-5 (LEC-5) in Uganda.

a) EFA-based five-factor model for the LEC-5. b) Prior six-factor mode; for the LEC-5.

https://doi.org/10.1371/journal.pone.0298385.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Fit indices for comparison of confirmatory factor analysis models.

https://doi.org/10.1371/journal.pone.0298385.t004

Discussion

The prevalence of experiencing at least one traumatic event among all study participants was 60.5%. We conducted an EFA for a random split-half of the sample, which showed a five-factor solution with three items (combat, severe suffering, and fire/explosion) cross-loading on more than one factor and one item, serious accidents, falling below the 0.3 cut-off for all five factors. A possible explanation for this may be in the way that “serious accidents” are interpreted by the participants which may be different from the construct being assessed. In psychometric analysis, theoretical constructs vary in complexity and some are inherently more intricate and multifaceted than others. Differences in the understanding and interpretation of these constructs may yield lower factor loadings. We then conducted a confirmatory factor analysis to establish an adequate measurement model. These metrics were measured against the six-factor model based on the South Korean EFA of the LEC combined with the WHO—World Mental Health Survey findings [2, 26] and the theoretical seven-factor model used in the South African Stress and Health Survey [33, 34]. The EFA for the LEC-5 in the Ugandan setting showed the poorest model fit for the Ugandan adult population of the three models and indices that did not meet criteria for most cut-offs. These five categories were: sexual violence, war and death, physical violence, environment and war-related events, and accidents and injury. These categories are similar to the six-factor model based on the South Korean EFA of the LEC combined with the WHO—World Mental Health Survey findings [2, 26]. The results of the CFA found that the theoretical seven-factor model was the best fitting model for the present study.

The prevalence of experiencing any traumatic event in our study is lower than in a prior study conducted in northern Uganda, where 80.8% of participants reported experiencing at least one PTE [8]. Mugisha et al. used a community sample and the Childhood Trauma Questionnaire (CTQ)–Short Form [36] to assess exposure to traumatic events before the age 18 years old. Their reported prevalence of traumatic events was higher than in our study (even though we include a lifetime timeframe and half of our sample is a clinical population). One possible explanation for the difference in prevalence may be the number of items in each measure with the CTQ which includes 28 items, meaning more opportunities for affirmative answers. Another factor accounting for the prevalence difference may be study’s geographical area, with their study solely taking place in a post-conflict region in northern Uganda [8], and our study taking place across several regions in Uganda. Our study results can also be compared with prior studies conducted in similar settings. For instance, in Ethiopia, researchers examining data from our parent study found a much lower percentage of 48.7% reported exposure to at least one traumatic event [27], whereas in South Africa 92% of participants had one or more traumatic events in their lifetime [27]. Also in South Africa, Atwoli et al. found 73.8% using the World Mental Health Composite International Diagnostic Interview (WMH-CIDI; [33, 37]. In Kenya, Kwobah et al also using the LEC-5 as part of the NeuroGAP study found that 64.4% of cases reported having experienced at least one PTE in their lifetime [6], which was similar to our findings. Also, in Kenya, Jenkins and colleagues using a structured epidemiological assessment, found 48% of study participants reported traumatic events [38], or lower than in the present study. In high-income countries such as Norway, 86% of women and 85% of men reported experiencing one significant lifetime event [4]. Such variability between and within countries is likely due to differences in the populations studied and measurement instruments used. For example, Jenkins et al used open-ended questions with seven examples, which may have elicited fewer endorsements of traumatic events than the LEC-5 used in the Kenyan sample or our sample as part of NeuroGAP-Psychosis. Other possible reasons for the observed heterogeneity in prevalence estimates than differences in assessment scales may be attributed to differences in risk of exposure to PTEs as well as types of PTE (e.g., accidents versus war exposure) [39], definitions of PTE, and time frame references for assessing PTEs [40, 41].

The prevalence of traumatic events in this sample was higher among cases (64.2%) than in controls (55.4%). This finding is similar to other studies across varied settings [42–45]. Also, cases reported more physical and sexual assault than controls, among other traumas. This finding is reflected in studies in Uganda that report a high prevalence of sexual assault and abuse [46, 47], which is consistent with research that has highlighted violent victimization of persons with severe mental illness such as psychosis [48–51].

A theoretical seven-factor model fits our data comparatively well in the present study. This model included the following latent categories: war, physical violence, sexual violence, accidents, network trauma, severe suffering, and witnessing. These groupings of traumatic events from prior research may fit our data well since traumatic events serve less as latent factors and more as composite phenomena [21, 52]. Our findings will probably facilitate the comparability of our results with other studies that utilize similar scales and may help identify and bridge the gap between theoretical constructs and clinical observations specific to the Ugandan setting. Based on the theoretical model, severe injury and severe suffering had the strongest loadings, while items aggregating under the accident latent factor were least strongly associated with it. Sexual assault and other sexual abuse were also strongly related to the sexual violence latent trait. Most Ugandan cultures engage in early marriages, which sets the premise for sexual assault and abuse, especially among women [53]. This may explain the high frequency of sexual assault reported among the participants. Early marriage refers to formal or informal unions in which the girl is less than 18 years [54]. In Uganda, 10% of girls are married before age 15, and 40% will be married before their 18^th birthday [55]. Marriage and cohabiting are also important aspects related to sexual violence in Uganda, primarily due to the significance that sexual relationships are given and the roles of spouses in fulfilling sexual obligations [56, 57].

Strengths and limitations

The strengths of this study include that it is the first study to assess the psychometric properties of the LEC-5 in Uganda in a large sample from different regions and using a standard administration of the LEC-5. However, some limitations should be considered when interpreting the results. First, our sample was comprised of both participants with psychosis and controls, and as such, there may have been an overrepresentation of the types of PTEs among the cases that do not align with the general population, limiting the generalizability of our findings. We do not believe this to be the case, given that the theoretical model fits better than the data driven EFA model, suggesting the PTEs are grouping together in a similar manner as other populations and not simply determined by the makeup of our study sample.

In addition, retrospective reports of PTEs could be biased due to forgetting over time and reality distortion experienced by many patients with psychosis [58, 59]. Moreover, we did not consider cultural variations in perceptions of trauma (e.g., [60] and whether participants perceived their experiences as traumatic. Furthermore, our findings on sexual assault may be underreported as sexual assault and abuse are considered taboo in Ugandan culture, and survivors of abuse who come forward are likely to face social stigma [61, 62].

Conclusion

The findings from our study support the LEC-5 as a reasonable tool for assessing PTEs in the Ugandan context. This is based on the comparability of our EFA and CFA metrics to the theoretical seven-factor model used in the South African Stress and Health Survey. Therefore, the LEC-5 can be utilized as a screening tool for PTEs in clinical populations. This may aid clinicians in carrying out trauma-related assessments and subsequently providing patient-focused treatment. Further studies should assess the criterion validity and reliability of the LEC-5 in Uganda.

Acknowledgments

We would like to acknowledge the data managers, research assistants, and project managers who have worked on this study: Adiru Tamali, Apio Racheal, Clare Samba Nalwoga, Francis Ojara, Julius Okura, Naome Nyinomugisha, Samalie Nsangi, Stanley Baniyo, and Stella Anena. We would also like to thank the participants who shared their time and their experiences with us. Without them, this work would not be possible.