Discussion

This study investigated the temporal association between moral injury appraisals and psychological outcomes in a community sample of refugees. The finding that moral injury appraisals predicted subsequent changes in psychological symptoms (and not, for the most part, vice versa) indicates that the way in which an individual thinks about their experiences, and specifically whether these experiences represent moral transgressions, is important in influencing their subsequent psychological functioning. This is consistent with theoretical models and empirical evidence that highlight the central role of trauma appraisals in impacting on post-trauma responses (Dunmore et al., Reference Dunmore, Clark and Ehlers1999; Ehlers & Clark, Reference Ehlers and Clark2000; LoSavio, Dillon, & Resick, Reference LoSavio, Dillon and Resick2017; O'Donnell et al., Reference O'Donnell, Elliott, Wolfgang and Creamer2007; Resick & Schnicke, Reference Resick and Schnicke1992). Furthermore, these findings have important implications for models of moral injury. Theorists vary in their conceptualization of moral injury, with some proposing that specific events or transgressions are, by definition, morally injurious (Litz & Kerig, Reference Litz and Kerig2019; Nash & Litz, Reference Nash and Litz2013), while others have considered a moral injury to represent a constellation of outcomes including fear, anger, shame, guilt etc. (Jinkerson, Reference Jinkerson2016). The findings from this study provide empirical support for recent theoretical frameworks that propose that it is the appraisal of an event as morally injurious that is critical in influencing subsequent outcomes rather than the nature of the event itself (e.g. Farnsworth, Reference Farnsworth2019; Kopacz et al., Reference Kopacz, Connery, Bishop, Bryan, Drescher, Currier and Pigeon2016).

In this study, we found that MI-Other and MI-Self appraisals were differentially related to psychological symptoms over time. Consistent with our hypotheses, MI-Other appraisals were associated with subsequent increases in re-experiencing and arousal symptoms, feelings of sadness, and feelings of anger (although the association with anger was marginally significant, p = 0.066). These findings are in accordance with studies that have observed associations between MI-Other appraisals and anger responses, depression and PTSD in both military (Bryan et al., Reference Bryan, Bryan, Anestis, Anestis, Green, Etienne and Ray-Sannerud2015; Jordan et al., Reference Jordan, Eisen, Bolton, Nash and Litz2017; Lancaster, Reference Lancaster2018; Stein et al., Reference Stein, Mills, Arditte, Mendoza, Borah, Resick and Consortium2012) and refugee (Hoffman et al., Reference Hoffman, Liddell, Bryant and Nickerson2018) samples. In contrast, these findings were inconsistent with previous longitudinal studies that found no association between MI-Other appraisals and PTSD, depression and/or anger in a treatment-seeking refugee sample (Nickerson et al., Reference Nickerson, Hoffman, Schick, Schnyder, Bryant and Morina2018) and military veterans (Currier et al., Reference Currier, McDermott, Farnsworth and Borges2019). It is important to note, however, that there were significant methodological differences between these other longitudinal studies and our study that make it difficult to compare the results. Specifically, the Nickerson et al. (Reference Nickerson, Hoffman, Schick, Schnyder, Bryant and Morina2018) study was limited by small sample size, restricted measurement of MI appraisals and focus on treatment-seeking refugees. Furthermore, the Currier et al., study (Currier et al., Reference Currier, McDermott, Farnsworth and Borges2019) conceptualized of MI-self and MI-other as a constellation of outcomes (i.e. shame, anger, resentment) instead of appraisals, and controlled for depression symptoms rather than investigating their association with MI over time. The results from the current study can be understood in the context of theoretical models that implicate perceptions of control in posttraumatic mental health (Ehlers & Clark, Reference Ehlers and Clark2000; Foa, Zinbarg, & Rothbaum, Reference Foa, Zinbarg and Rothbaum1992). Specifically, the appraisal that someone has transgressed important moral rules suggests that the individual did not have personal control over the event, but instead observed morally transgressive acts. Empirical research has indicated that perceived lack of control is associated with the development of heightened fear reactions and intrusive memories in trauma survivors (Frazier, Reference Frazier2003; Frazier, Bergman, & Steward, Reference Frazier, Bergman and Steward2002; Kushner, Riggs, Foa, & Miller, Reference Kushner, Riggs, Foa and Miller1993), including those exposed to torture (Basoglu et al., Reference Basoglu, Livanou, Crnobaric, Franciskovic, Suljic, Duric and Vranesic2005; Basoglu & Mineka, Reference Basoglu, Mineka and Basoglu1992; Le et al., Reference Le, Morina, Schnyder, Schick, Bryant and Nickerson2018). Taken together, these findings suggest that exposure to events that give rise to MI-other appraisals may lead to increases in arousal, resulting in greater fear (i.e. re-experiencing and arousal symptoms) and anger responses. Further research is needed, however, to determine whether perceptions of control acts as a mechanism that underlies the association between exposure to a PMIE, MI-other appraisals, and subsequent psychological responses.

The pattern of association between MI-self appraisals and subsequent psychological responses was markedly different from that of MI-Other appraisals. Notably, we found that MI-Self appraisals at baseline were associated with decreases in re-experiencing symptoms 6 months later. This is consistent with a previously-found negative cross-sectional association between MI-self appraisals and re-experiencing symptoms in a refugee sample (Hoffman et al., Reference Hoffman, Liddell, Bryant and Nickerson2018). One possible explanation for this finding relates to the potential for differential processing of adverse experiences in the presence and absence of significant fear responses. Models of posttraumatic mental health have proposed that elevated arousal reactions at the time of trauma exposure disrupts memory encoding, giving rise to re-experiencing symptoms such as intrusive memories, nightmares, and distress to reminders of the event (Brewin, Dagleish, & Joseph, Reference Brewin, Dagleish and Joseph1996; Ehlers & Clark, Reference Ehlers and Clark2000; Foa et al., Reference Foa, Zinbarg and Rothbaum1992; Foa & Rothbaum, Reference Foa and Rothbaum1998; Resick & Schnicke, Reference Resick and Schnicke1992). In contrast to MI-other appraisals, events that give rise to MI-self appraisals are likely to involve a sense of agency, as the individual believes they have acted in a way that transgresses their own moral rules. This agency may lead to lower levels of fear as a function of heightened perceptions of control (Basoglu & Mineka, Reference Basoglu, Mineka and Basoglu1992; Ehlers & Clark, Reference Ehlers and Clark2000; Foa et al., Reference Foa, Zinbarg and Rothbaum1992; Frazier et al., Reference Frazier, Bergman and Steward2002; Kushner et al., Reference Kushner, Riggs, Foa and Miller1993), with this absence of fear potentially facilitating more complete memory processing. This may then lead to decreased intrusive memories over time as the individual continues to process the event. This hypothesis is supported by the finding in the current study that MI-self appraisals were not associated with subsequent increases in arousal symptoms. It is important to note, however, that findings regarding the relationship between MI-self and re-experiencing symptoms in the moral injury literature have been mixed, with some studies finding a positive relationship between MI-self and re-experiencing symptoms (Litz et al., Reference Litz, Contractor, Rhodes, Dondanville, Jordan, Resick and Consortium2018; Stein et al., Reference Stein, Mills, Arditte, Mendoza, Borah, Resick and Consortium2012), while others have failed to find a relationship between these two constructs (Bryan et al., Reference Bryan, Bryan, Anestis, Anestis, Green, Etienne and Ray-Sannerud2015). Methodological differences may account for these discrepancies. Studies that have found positive associations between MI-self and re-experiencing symptoms have generally considered the association between psychological outcomes and specific events that have pre-categorized by the researchers as either representing MI-self or MI-other events (Litz et al., Reference Litz, Contractor, Rhodes, Dondanville, Jordan, Resick and Consortium2018; Stein et al., Reference Stein, Mills, Arditte, Mendoza, Borah, Resick and Consortium2012). Given that there is evidence from both refugee and military populations that (i) individuals can hold both MI-self and MI-other appraisals concurrently and (ii) different individuals can interpret the same event as representing another- or self-transgression (Hoffman, Liddell, Bryant, & Nickerson, Reference Hoffman, Liddell, Bryant and Nickerson2019; Schorr et al., Reference Schorr, Stein, Maguen, Barnes, Bosch and Litz2018), it may be that the focus on pre-categorized event type rather than subjective appraisals led to inconsistent results across studies. In contrast, in this study, we investigated the association between MI-Self and MI-Other appraisals (which were not linked to specific events) and changes in psychological outcomes. Taken together, these findings suggest that it is the appraisal of an event as representing a self-transgression that is associated with reduced re-experiencing symptoms over time, irrespective of the event type. Nevertheless, further research should be conducted to elucidate the temporal relationship between MI appraisals and re-experiencing symptoms across a variety of contexts. For example, the use of experimental paradigms to investigate these causal relationships between MI-self appraisals and intrusive memories represents a promising future avenue of enquiry.

One surprising finding in this study was the association between MI appraisals and shame. Consistent with research that has linked MI-Self with shame and guilt in US military veterans and active-duty Marines, we hypothesized that MI-Self appraisals at T1 would be associated with increased shame over the subsequent 6 months. We did not predict an association between MI-Other appraisals and shame. Contrary to our hypotheses, we found that MI-Self appraisals at T1 predicted decreases in shame over the following 6 months, while MI-Other appraisals at T1 predicted subsequent increases in shame. To our knowledge, this is the first longitudinal study to investigate the temporal associations between specific MI appraisals and shame over time. There are several potential explanations for these findings, which highlight the complex interplay of MI appraisals and emotional responses in the refugee context. The finding that higher levels of MI-Self appraisals predicted subsequent decreases in shame may indicate that higher MI-Self appraisals at baseline were associated with greater awareness of one's own role in a moral transgression, potentially allowing for greater cognitive processing of the experience and subsequent decreases in feelings of personal responsibility. This is consistent with our finding that greater MI-Self appraisals at baseline were associated with subsequent decreases in re-experiencing symptoms which points to the continued processing of PMIE-related memories over time. This hypothesis is highly speculative, however, and further investigation of the temporal associations between specific emotional responses, MI appraisals and other cognitive factors (e.g. perceptions of control, self-blame, personal responsibility) is necessary to draw firm conclusions. In contrast, our finding that greater MI-Other appraisals at T1 were associated with subsequent increases in shame is consistent with the association between MI-Other appraisals and other negative emotional responses measured in this study (i.e. fear, anger, sadness). This suggests that MI-Other appraisals are associated with pervasive, rather than specific, negative emotional responses. Given that ‘negative alterations in cognition and mood’ is one of the key DSM-5 criteria for PTSD (APA, Reference American Psychiatric Association2013), this finding also further underscores the robust relationship between MI-Other appraisals and PTSD symptoms.

This study has important implications for the conceptualization and measurement of moral injury. The approach of conceptualizing moral injury in terms of appraisals [while increasingly common (Farnsworth, Reference Farnsworth2019; Kopacz et al., Reference Kopacz, Connery, Bishop, Bryan, Drescher, Currier and Pigeon2016; Lancaster & Erbes, Reference Lancaster and Erbes2017)] differs substantially from other approaches which have examined the direct association between PMIEs and psychological outcomes (Currier, Holland, Drescher, & Foy, Reference Currier, Holland, Drescher and Foy2015; Nazarov, Fikretoglu, Liu, Thompson, & Zamorski, Reference Nazarov, Fikretoglu, Liu, Thompson and Zamorski2018). While the latter approach has the advantage of separating event exposure and associated distress, this has key disadvantages as it requires a presupposition on the part of the researcher regarding whether specific events are potentially morally injurious. As noted above, there is growing evidence that individuals differ substantially in (a) whether they consider particular event types to be morally injurious and (b) whether these represent moral transgressions related to the self, the other, or both (Hoffman et al., Reference Hoffman, Liddell, Bryant and Nickerson2019; Schorr et al., Reference Schorr, Stein, Maguen, Barnes, Bosch and Litz2018). Accordingly, measuring moral injury in terms of appraisals allows the individual to determine for themselves whether their experiences represent moral transgressions and give rise to associated distress, and whether these relate to others or the self. This is particularly important when working in a cross-cultural context as moral frameworks may vary markedly across (and within) cultural groups (Graham, Meindl, Beall, Johnson, & Zhang, Reference Graham, Meindl, Beall, Johnson and Zhang2016). Nevertheless, it may be important to consider whether particular types of events are more likely to give rise to moral injury appraisals; accordingly, future research could investigate the association between specific PMIEs or traumatic events, and MI-other and MI-self appraisals.

A second measurement issue arising from this study relates to the incorporation of distress in the MIAS items (e.g. ‘I am troubled because I did things that were morally wrong), and the use of these items to predict psychological outcomes. To a certain extent, this approach involves predicting distress from distress, which warrants further consideration from a conceptual perspective. This approach was implemented in the MIAS scale to capture both (i) whether an individual perceived that he/she had experienced a moral transgression, and (ii) whether this transgression continued to trouble the individual. The finding that MI-other and MI-self appraisals differentially predicted specific psychological symptoms suggests that this approach was useful in linking PMIEs, appraisals and psychological outcomes. Nevertheless, it is important to acknowledge that there is a level of conceptual overlap in simultaneously assessing perceived moral transgressions and distress, and to consider ways in which this could be disentangled in the future. One possible approach would be to remove reference to distress in the MIAS items. In fact, the MIAs included three items that did not directly reference distress (e.g. ‘I violated my own morals by failing to do something I should have done’). The two factor-higher order CFA model presented in this paper fit the data well, and these items loaded onto the relevant factors (online Supplementary Table A). This is perhaps not surprising as it could be argued that endorsing an item relating to violating one's morals or having one's morals violated implies a level of distress. It should be noted, however, that these three items evidenced the lowest standardized factor loadings (0.88, 0.79, and 0.73), although these were still relatively strong. Accordingly, it may be useful to investigate whether altering the MIAS items to remove reference to distress results in a ‘purer’ measure of appraisals, and yields different results when testing the association between moral injury appraisals and psychological outcomes. This approach was recently taken in a study conducted by Chaplo, Kerig, and Wainryb (Reference Chaplo, Kerig and Wainryb2019) who found that where a measure indexing perceived moral violations (without associated distress) in youth yielded significant correlations with the Moral Injury Events Scale, which assesses both violations and distress (Chaplo et al., Reference Chaplo, Kerig and Wainryb2019). Further investigation of the association between perceived violations and distress would elucidate the association between these constructs, PMIEs and psychological outcomes.

This study had a number of limitations. First, sampling was undertaken using convenience and snowball methods. While this is an effective way to recruit difficult-to-access populations (Sadler et al., Reference Sadler, Lee, Lim and Fullerton2010), this may mean that findings are not representative across refugees in Australia. Second, participants were limited to those who were literate in Arabic, Farsi, Tamil, or English. The online nature of this study afforded high levels of confidentiality and privacy, possibly enhancing accurate reporting of sensitive experiences (i.e. moral injury-self appraisals). However, this may have resulted in a cohort with relatively high education levels, again limiting generalizability of findings. Third, we did not ask participants to identify specific events that may have given rise to moral injury appraisals, and thus it is difficult to ascertain exactly which types of experiences are linked to specific moral injury appraisals; future research should investigate this. Related to this, as noted above, the measure employed in this study (the MIAS) focused on examining the association between appraisals of PMIEs and psychological outcomes. We note that there is likely to be conceptual overlap between these constructs, and future research should disentangle the association between these variables. Fourth, there are a number of key cognitive constructs implicated in posttraumatic mental health, such as posttraumatic cognitions, perceptions of control and attributions of responsibility (Delahanty et al., Reference Delahanty, Herberman, Craig, Hayward, Fullerton, Ursano and Baum1997; Foa, Ehlers, Clark, Tolin, & Orsillo, Reference Foa, Ehlers, Clark, Tolin and Orsillo1999; Hassija & Gray, Reference Hassija and Gray2013; Hickling, Blanchard, Buckley, & Taylor, Reference Hickling, Blanchard, Buckley and Taylor1999), that are likely to be associated with moral injury appraisals, and were not examined in this study. Investigation of whether these represent distinct or overlapping concepts represents an important future line of enquiry. Fifth, this study focused on investigating the association between moral injury appraisals and specific state emotional responses such as anger and shame. There are likely to be other important state emotional responses that are associated with moral injury appraisals, such as guilt and disgust, that were not examined in this study. Future research could investigate how moral injury appraisals are differentially associated with a wider range of emotional responses. Furthermore, theorists have highlighted the potential relationship between moral injury and trait emotions, such as shame proneness (Zalta & Held, Reference Zalta and Held2020). Shame proneness refers to the tendency an individual might have to attribute events to their own negative traits (Zalta & Held, Reference Zalta and Held2020). Accordingly, it may be the case that such stable characterological variables might influence the way in which an individual appraises a PMIE, and the extent to which this then leads to subsequent psychological distress. Further research should investigate the role that trait emotions and attributional styles play in this association. A final limitation of this study is that it examined perceptions of moral transgressions and how these related to psychological symptoms longitudinally without considering the stability of an individual's framework over time. There is evidence from the field of moral psychology that morals and values can change over time, with these changes being influenced by a number of factors (Lindstrom, Jangard, Selbing, & Olsson, Reference Lindstrom, Jangard, Selbing and Olsson2017; Myyry, Juujarvi, & Pesso, Reference Myyry, Juujarvi and Pesso2013; Young, Reference Young2015). The DSM-5 has recognized the important impact that exposure to traumatic events can have on beliefs about the self, the world and others (APA, Reference American Psychiatric Association2013). Furthermore, in the moral injury literature, Farnsworth, Drescher, Nieuwsma, Walser, and Currier (Reference Farnsworth, Drescher, Nieuwsma, Walser and Currier2014) assert that the function of moral emotions can change over time with changes in context; such that emotions that may be considered appropriate in one context (e.g. aggression in situations of danger) may not be appropriate in an alternative context (e.g. daily life in the resettlement country). Accordingly, the change in context can make the functional impairment associated with the emotion more salient. It may well be the case that exposure to PMIEs fundamentally alter the moral framework of the individual, which may have important consequences for subsequent distress and functioning. Accordingly, future research could investigate the way in which moral injury appraisals map onto an individual's moral framework, whether these change over time, and how this is associated with emotional and psychological responses.

This study has important potential clinical implications. First, the finding that moral injury appraisals influence subsequent psychological responses highlight them as a promising potential intervention target. This raises the question as to the appropriateness of current best-practice interventions for trauma-affected populations for moral injury appraisals. There is evidence that changes in trauma-related appraisals mediate reductions in PTSD symptoms in intervention studies (Holliday, Holder, & Suris, Reference Holliday, Holder and Suris2018; McLean et al., Reference McLean, Zang, Gallagher, Suzuki, Yarvis, Litz and Consortium2019; Schumm, Dickstein, Walter, Owens, & Chard, Reference Schumm, Dickstein, Walter, Owens and Chard2015; Wiedemann et al., Reference Wiedemann, Stott, Nickless, Beierl, Wild, Warnock-Parkes and Ehlers2020). However, Farnsworth (Reference Farnsworth2019); Farnsworth, Drescher, Evans, & Walser (Reference Farnsworth, Drescher, Evans and Walser2017) notes that descriptive appraisals (which often involve factual inaccuracies; e.g. If I leave my house, I am certain to be assaulted) may be more amenable to traditional cognitive interventions than moral injury cognitions which are often prescriptive in nature (i.e. I shouldn't have left family behind). Nevertheless, evidence-based trauma-focused interventions that specifically target cognitions that give rise to feelings of anger and shame (Ehlers & Clark, Reference Ehlers and Clark2000; Resick & Schnicke, Reference Resick and Schnicke1992) and specific interventions developed to target moral injury (Gray et al., Reference Gray, Schorr, Nash, Lebowitz, Amidon, Lansing and Litz2012; Litz, Lebowitz, Gray, & Nash, Reference Litz, Lebowitz, Gray and Nash2017) may be beneficial in reducing distress related to MI-other and MI-self appraisals in refugee groups. Taken together, these findings point to a potentially important treatment direction for moral injury-related psychological distress, namely the targeting of moral injury appraisals.

Overall, findings from this study highlight the temporal impact of moral injury appraisals on emotional and psychological responses in individuals from a refugee background. These results underscore the importance of considering cognitive responses to adverse experiences to best understand the nature and extent of subsequent psychopathology. Furthermore, findings suggest that clinical interventions should take into account moral injury appraisals when seeking to alleviate the mental health burden of experiences of trauma and displacement in refugees.