A Santanello, M Toohey, O Van Orden, M Soll, S Batten
Background: Anger is often indirectly addressed in posttraumatic stress disorder (PTSD) treatment despite reports that traumafocussed therapy may not be suf cient in improving anger responses1 Thus, it is possible that anger in the context of PTSD might be related to additional factors such as personality traits and coping styles.
Purpose: We sought to identify additional elements influencing anger. We hypothesised that trait anger (a personality trait) and experiential avoidance (a coping style) would account for significant variance in state anger when controlling for PTSD symptom severity in a sample of treatment-seeking veterans with a history of PTSD.
Material and Methods: One-hundred and one veterans completed pencil-and-paper surveys including the Posttraumatic Stress Disorder Checklist, Acceptance and Action Questionnaire-II, and State-Trait Anger Expression Inventory-2, and a subsample of 79 veterans meeting screening criteria for PTSD were included in these analyses.
Results: Experiential avoidance, PTSD symptom severity, and trait anger had significant zero-order correlations with state anger while only PTSD symptom severity and trait anger had a significant partial effect in the full model (p< 0.01). Stepwise multiple regression produced a three-predictor model accounting for 27% of the variance in state anger, explained partially by PTSD symptom severity and trait anger (F(3, 75) = 12.86, p < 0 .001).
Conclusion: Implications include addressing anger directly in treatment and not as a symptom that will be indirectly reduced through the treatment of PTSD.
Key Words/Phrases: State Anger, Trait Anger, PTSD, Experiential Avoidance, Veterans
Conflicts of interest: none declared.
Anger is among the most common symptoms reported by veterans who have been diagnosed with PTSD2. Veterans diagnosed with PTSD who served in Vietnam3-4 and who served in Operations Iraqi and Enduring Freedom5-6 tend to report a significantly higher degree of anger than veterans not diagnosed with PTSD. Congruently, partners of veterans diagnosed with PTSD often report observing higher levels of anger in their spouses when compared to partners of combat veterans without PTSD diagnoses7.
Anger is associated with the development and maintenance of PTSD. The presence of anger at initial assessment has been found to predict symptoms of PTSD four weeks post-assault in survivors of sexual and nonsexual assault8 as well as chronic PTSD fourteen years post-assault in a sample of Vietnam veterans9. Anger can interfere with motivation to engage in treatment1 and is also associated with attenuated effectiveness of treatment in both combat10 and non-combat11 veterans presenting with symptoms of PTSD. Among veterans and civilians diagnosed with PTSD, anger is associated with a variety of family, vocational, medical, and functional impairments2,3,12-15.Despite the prevalence and salient consequences of PTSD-related anger, the lack of literature on the topic makes it difficult to conceptualise and treat effectively1. As a result, available interventions for PTSD-related anger are limited in their effectiveness16.
Spielberger differentiated between two types of anger: trait anger and state anger17. While state anger is conceptualised as the episodic and momentary experience of angry affect, trait anger is defined as the general, dispositional tendency to experience anger. State and trait anger have been shown to have a strong, direct relationship with each other in police officers diagnosed with PTSD18 and in male college students who report exposure to a traumatic event19. Iraq and Afghanistan war veterans who screened positive for PTSD symptoms were observed to have significantly higher trait anger than veterans who did not screen positive for PTSD symptoms5. In addition, trait anger appears to have a significant relationship with emotional numbing, hyperarousal, and aggressive behaviour in veterans of the recent conflicts in Iraq and Afganistan20.
Previous research suggests that experiential avoidance may also play an important role in the development and maintenance of a variety of psychological disorders including PTSD21-23. Experiential avoidance is de ned as a general coping style in which individuals engage in strategies intended to alter the frequency or experience of private events such as thoughts, feelings, memories or bodily sensations or the contexts that occasion them22. Studies have shown that experiential avoidance predicts PTSD in veterans24-25.
Experiential avoidance, that is avoidance of both trauma and non-trauma related internal experiences such as unwanted thoughts and feelings, might also contribute to maladaptive behavioural patterns related to anger26-27. For example, individuals might try to avoid feeling angry in the moment when it seems too intense, lasts too long, or occurs too frequently. Gardner and Moore proposed an “anger avoidance model” which suggests that strong motivation to avoid or terminate the experience of anger may play a central role in anger dysregulation27. Specifically, it is proposed that an early-life history of abuse and neglect may contribute to the development of cognitive biases toward threat and an exaggerated sense of vulnerability that fuels chronic feelings of anger. Anger is experienced as intolerable, and difficulties in processing angry affect contribute to overgeneralised internal (e.g., cognitive rumination) and external (e.g., aggressive behaviour) coping strategies that function to decrease or prevent the experience of anger in the short term. Paradoxically, these strategies may simultaneously maintain and even create situations that are likely to occasion feelings of anger and vulnerability in the future. Similarly, Eifert and Forsyth proposed that anger- related behaviours aimed at reducing emotional discomfort (i.e., experiential avoidance), such as the experience of anger, are often ineffective in the long-term and are likely to interfere with life satisfaction and functioning in various domains of living26. Treatment specifically targeting experiential avoidance has been shown to be a promising approach to reduce the impact of anger on functional impairment28.
The relationship between anger and avoidance coping may also have significant implications for individuals who have been diagnosed with PTSD. Anger in the context of PTSD may serve as an “active” form of avoidance which may afford the trauma survivor a feeling of agency that distracts them from feeling helpless or vulnerable29. This may interfere with the processing of more “vulnerable” emotions such as fear, which is often a key in successful recovery from traumatic experiences8,30. Many trauma survivors may begin to fear the experience of anger and the consequences of angry behaviour. Fear of (and perhaps subsequent avoidance of) anger and angry reactions that may be occasioned by trauma- focussed therapy have been shown to partially account for suboptimal responses to treatment31. Unfortunately, few studies have examined the possible relationships between anger and avoidance coping directly, and these relationships are not well understood. Additional research in this area is needed as it may help to improve clinicians’ ability to treat individuals presenting for PTSD treatment with high levels of anger32.
In the current study, the authors sought to investigate the impact of experiential avoidance (a coping style), trait anger (a personality trait), and PTSD symptom severity on state anger in veterans with PTSD. The study included the following hypotheses: a) PTSD symptoms would predict state anger, b) experiential avoidance would predict state anger, c) trait anger would predict state anger, d) experiential avoidance would predict state anger independent of trait anger, and e) trait anger would predict state anger independent of experiential avoidance and PTSD symptom severity.
Participants were drawn from a sample of 101 veterans receiving outpatient or residential treatment for PTSD at the Baltimore Veterans Affairs Medical Center (VAMC). Specific details regarding treatment setting (e.g., outpatient, residential) were not collected. Recruitment efforts were mostly directed toward veterans participating in outpatient group therapy with some participation by veterans receiving residential treatment. Therefore, at the time of their participation we estimate that approximately two-thirds of the sample was receiving outpatient treatment, and one-third was receiving residential treatment. All veterans seeking treatment for PTSD at the Baltimore VAMC met full criteria for the disorder during a semi-structured intake prior to receiving a referral to the Trauma Recovery Program; the diagnosis of PTSD could have been given related to either a military or civilian trauma. No formal data were recorded for individual participants about the process of PTSD diagnosis, co-occurring psychiatric disorders, or substance use problems. However, the majority of the initial sample reported service in a war zone (75.2%) and exposure to potentially traumatising events such as receiving either friendly or hostile re (76.2%), and a minority reported witnessing or participating in atrocities (48.5%), experiencing military sexual assault (16.8%), and being threatened with sexual assault in the military (10.9%). Only those participants completing the survey whose responses on the PTSD Checklist indicated that their past- month symptoms were suggestive of PTSD, based on the cutoff score of 50, as recommended by Weathers et al. 33 , were included in the analyses for this study. Using this criterion, the study sample was predominantly male (n = 73) with six female participants. The average age of participants was 52.1 years (SD = 9.99), with a range from 25 to 74 years. The sample was 50.6% African American, 39.2% Caucasian, 2.5% Latino, and 1% Native American, with 6.4% from other, unspecified racial/ethnic groups.The majority of the sub-sample reported service in a war zone (70.9%) and exposure to potentially traumatising events such as receiving either friendly or hostile re (70.9%), and again a minority reported witnessing or participating in atrocities (30.4%), experiencing military sexual assault (20.3%), and being threatened with sexual assault in the military (12.7%).Table 1 shows additional demographic and military service characteristics of the study sample.
[See PDF for table 1].
Posttraumatic Stress Disorder Checklist33. The PTSD Checklist (PCL) is a 17-item, self-report measure of the frequency of posttraumatic stress symptoms (DSM-IV TR) in the past month. The PCL appears to have adequate sensitivity to the presence of PTSD symptoms in veterans34. Cronbach’s Alpha coefficient for the current sample was 0.798.
Acceptance and Action Questionnaire-II35. The Acceptance and Action Questionnaire-II (AAQ-II) is a ten-item, self-report measure of experiential avoidance. The measure includes a list of ten statements (e.g., “It’s OK if I remember something unpleasant”) rated on a 7-point scale from Never True to Always True. Lower total scores re ect greater experiential avoidance, and higher scores indicate greater psychological flexibility. The AAQ-II appears to have good internal consistency (average Cronbach’s Alpha coefficient of 0.83) and validity in preliminary validation research35. Cronbach’s Alpha coefficient for the current sample was 0.622.
State-Trait Anger Expression Inventory-2 17. The State-Trait Anger Expression Inventory (STAXI-2; Spielberger,17) consists of 57 items representing dispositional and momentary anger rated from 1 (Not at all) to 6 (Very much so). The STAXI-2 internal reliability estimates range from 0.73 to 0.95 for the total scale and from 0.73 to 0.93 for the subscales. Spielberger17 also reported construct-related validity for the scales and subscales. Concurrent validity of the original STAXI was demonstrated by comparing it to several scales, including the Minnesota Multiphasic Personality Inventory (Hostility and Overt Hostility scales), Buss-Durkee Hostility Inventory, and the Eysenck Personality Questionnaire (Psychoticism and Neuroticism)� Chronbach’s Alpha coefficient for the total scale in the current sample was 0.906.
The study was approved by the medical cenrer’s designated Institutional Review Board to recruit participants from the population of veterans who were seeking treatment in the Trauma Recovery Program of the Baltimore VAMC. Veterans receiving outpatient or residential treatment for PTSD were approached by members of the research team and asked to participate in this study. Study personnel explained the general rationale for the study, described the procedures, and obtained informed consent. After completing the consent process, participants were provided with a demographic questionnaire, the AAQ-II, PCL, STAXI-2, and several other paper-and-pencil measures not included in the present analyses. Participants completed all study measures at the time of consent in the presence of a member of the research staff.
Data Analyses & Results
Multiple linear regression analysis was used to examine a model predicting state anger (STAXI-2: State) from past-month PTSD symptom severity (PCL), experiential avoidance (AAQ-II; with higher scores reflecting lower experiential avoidance), and trait anger (STAXI-2:Trait).Table 2 shows descriptive statistics as well as full-model and semi-partial regression coefficients for each variable in the model; raw scores were transformed into z-scores for ease of interpretation of coefficients.
[See PDF for table 2].
[See PDF for table 3].
Each of the variables had a significant zero- order correlation with state anger, and both PTSD symptoms severity (p < 0.05) and trait anger (p < 0.01) had significant partial effects in the full model. Holding the other variables constant, for every one- standard deviation increase in PTSD severity, there was a 0.20-point increase in state anger and for every one-standard deviation increase in trait anger, there was a 0.36-point increase in state anger.
Next, hierarchical regression analysis was employed to predict state anger, with each variable listed in Table 3 entered in a sequential step.
PTSD symptom severity significantly predicted state anger (p < 0.01). The addition of experiential avoidance in step two did not improve the model. The addition of trait anger in step three demonstrated a significant effect in predicting state anger (p < 0.01), and the three-predictor model accounted for 27% of the variance in state anger, explained largely by trait anger, F(3, 75) = 12.86, p < 0.001.
The purpose of the current study was to explore the impact of experiential avoidance (a coping style), trait anger (a personality trait), and PTSD symptom severity on state anger in veterans with PTSD. Experiential avoidance, PTSD symptom severity, and trait anger had significant zero-order correlations with state anger. As hypothesised, trait anger was observed to predict state anger independent of experiential avoidance and PTSD symptom severity. Although PTSD symptom severity was also a significant predictor of state anger in the final model, it is noteworthy that the trait anger was the strongest predictor. The relationship between state and trait anger observed in this study is consistent with previous findings demonstrating the strong relationship between these constructs18-19. Anger is a common residual symptom even when other symptoms of PTSD are treated effectively and appears to be more closely related to symptoms of “dysphoric arousal” rather than “anxious arousal” symptoms that are most directly targeted by trauma focused treatment36-37. The unique relationship between trait anger and state anger, independent of PTSD symptom severity, observed in this study lends support to recent recommendations suggesting that anger- focussed assessment and treatment be included for veterans and service members presenting with co- morbid PTSD and anger dysregulation38.
Consistent with the fear avoidance models proposed by Foa et al 30. and Kulkarni et al. 29, we hypothesised that experiential avoidance would predict state anger independently of PTSD symptom severity and trait anger. Although experiential avoidance was not found to predict anger independently, a significant, zero-order correlation between experiential avoidance and state anger was observed. This suggests that avoidance-based coping may have a subtle relationship with anger in the context of PTSD that could not be adequately measured due to the limitations of this study outlined below. Another explanation for these results could be related to the constructs of state anger and experiential avoidance as measured by the STAXI-2 and AAQ-II, respectively. Experiential avoidance has been conceptualised as the tendency to cope with unwanted private events by avoiding or altering their form, frequency, or intensity or the contexts that occasion them22. State anger is conceptualised as the episodic experience of angry affect17, and may be a slightly different construct than anger as a form of active avoidance such as proposed by the fear avoidance model. State anger may be better conceptualised as the topography of the emotional state to which it refers, whereas “anger-as-avoidance” may be a new concept, reflecting the function of this affective state in the context of PTSD symptoms. There may be a need to develop a specific measure of “anger-as-avoidance” that more directly reflects this construct.
There are several limitations to the current study. First, given the small sample size, there may have been limited power to detect influences of experiential avoidance on state anger. Another possible limitation is the potential range restriction in anger scores due to the composition of the current sample. Because all of the study participants were participants in a PTSD treatment program, and because anger is a frequent component of the presentation of PTSD, it is possible that the relatively higher levels of anger and concomitant PTSD symptoms introduced some issues with multi-collinearity.However, the current study was designed specifically to look at the relationship of PTSD and anger within a PTSD treatment-seeking sample. Future studies will be needed to determine whether these relationships apply in samples with a wider range of anger levels and posttraumatic symptom scores that do not all rise to the level of requiring mental health treatment. Third, all observations occurred at one assessment point, precluding the use of more sophisticated statistical analyses and inferences regarding the relationships between factors over the course of time. This cross-sectional design provides only retrospective information on psychological factors that may influence momentary anger. Use of a longitudinal design (e.g., collection of study measures prior to initiating treatment for PTSD and following participation in a course of therapy) would afford the opportunity to investigate more complex hypotheses.However, we believe that a causal relationship between variables can be inferred due to the differences in the temporal nature of each measure (comparing a disposition, coping style, and disorder to a momentary construct). Fourth, methods used to determine inclusion in the current study sample, particularly those used to establish the presence of PTSD, were suboptimal given that all veterans included in this sample were judged to meet DSM-IV criteria for PTSD based on an unstructured intake interview. Thus, only participants reporting significant symptoms of PTSD (PCL scores of 50 or higher) were included in the study sample.Utilisation of a structured, diagnostic interview such as the Clinician Administered PTSD Scale would have been a more reliable and accurate method of determining inclusion in the study sample.
Despite the limitations of the current study, trait anger and experiential avoidance might be useful factors to consider in future research with veterans presenting with symptoms of PTSD. The findings of this cross-sectional study are consistent with emotion-processing and cognitive theories of PTSD and might also suggest the importance of addressing a broader range of avoidance tendencies (emotional, experiential) that may uniquely influence the persistence of anger-related problems in veterans with PTSD. In treatment outcome research and prospective studies, inclusion of measures that examine these often excluded constructs of these variables would further elucidate the relationship between problematic anger and PTSD and may suggest novel approaches for intervention.
Corresponding author: Michael Toohey [email protected]
Authors: M Toohey1, A Santanello2, O Van Orden2, M Soll3, S Batten4
Author Af liations:
1 Eastern Washington University
2 VA Maryland Health Care System
3 Portland, Oregon
4 Booz Allen Hamilton, McLean, Virginia