Distress and disorders

We think an important topic is how terminology to describe impacts is still evolving to reflect broad constructs and developments in bringing together social, psychological and psychiatric considerations. An important question is how we construe the psychosocial and mental health impacts of violence and terrorism. When one of us, R.W., was first engaged in practice in this area more than 40 years ago, there were uncertainties about how we should apply diagnoses to people’s reactions to terrorism and post-traumatic stress disorder (PTSD) was a new diagnostic possibility. Subsequently, acute stress disorder (ASD) was conceived to describe the reactions of people that resembled PTSD in the first 4 weeks after incidents as a way of identifying people at higher risk of developing PTSD.

However, we note that the authors of many of the papers in the thematic series describe people’s early reactions more simply as their suffering ‘distress’. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3–Reference Stene, Glad, Stensland, Nilsen and Dyb5,Reference Dean, Eggleston, Ali, Thaufeeg, Wells and Zarifeh7–Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9} Thus, a number of the papers in this series describe participants as distressed rather than applying diagnoses in the short-term. By contrast, Karam et al espouse the importance of practitioners recognising the presence of symptoms of ASD, and particularly their intensity. ^{Reference Karam, Al Barathie, Dimassi, Mascayano, Slim and Karam13}

This takes us to a related matter, which is raised by one of the papers in the series, concerning how we should measure the impacts on people of incidents that include terrorism. ^{Reference Wesemann, Applewhite and Himmerich10} Some researchers report epidemiological findings from validated and commonly used measures of people’s symptoms. Often, they employ self-report methods and use cutoffs that researchers have calculated in previous research in other circumstances to offer estimates of caseness. However, what is becoming clearer is that using self-report survey techniques with convenience samples that are poorly defined tends to provide estimates of the prevalence of common mental disorders and PTSD that are notably higher than findings from standardised clinical interviews. This matter was brought to prominence by surveys conducted during the COVID-19 pandemic. ^{Reference Scott, Stevelink, Gafoor, Lamb, Carr and Bakolis14} In the series, the authors of the systematic review of 33 papers draw attention to this matter in relation to terrorism and the authors opine that when ‘questionnaire results … [are] utilised in future research, this should always be described in the limitations’.[ ^{Reference Wesemann, Applewhite and Himmerich10} p.10] Our opinion is that this is a reason for being cautious about ascribing diagnoses in the early days and weeks after incidents to people affected by terrorism and collective violence when self-report methods are used in surveys. We qualify the recommendation from Karam et al ^{Reference Karam, Al Barathie, Dimassi, Mascayano, Slim and Karam13} by suggesting that researchers should take into account how the data they interpret were collected when deciding on whether they are confident to offer diagnoses for people whom they have not examined. Certainly, these observations indicate the need for further research into how best to compare and integrate the results of self-report questionnaire and clinical interview studies.

Similar challenges of terminology and measurement apply in the medium- and longer-terms as well as the short-term. In the past, we anticipated that distress would ameliorate relatively quickly in the early aftermath before diagnoses of common mental disorders and PTSD became appropriate, but the findings from the papers about the Manchester Arena bombing and the Utøya shooting imply that a significant proportion of uninjured people may continue to be distressed over many months. Whether they do or do not appears to turn, to an extent at least, on the intensity of their early suffering, their previous childhood experiences, how they are assisted, their current relationships and their levels of affluence. Consequently, the notion of active monitoring over time has gained sway. ¹⁵ The paper regarding psychosocial follow-up after the Utøya shootings evaluates a model of active monitoring of people involved in incidents. While its authors were not able to reach a firm conclusion about proactive follow-up, there were clear indications from the Manchester Arena studies that relationships shared with other people affected by the same incident may prove mutually beneficial in line with science about the social cure. ^{Reference Drury, Stancombe, Williams, Collins, Lagan and Barrett8}

All these circumstances are reminiscent of the literature on the impacts of COVID-19 and other disasters. ^{Reference Murray, Kaufman and Williams16,Reference Bryant, Gibbs, Gallagher, Pattison, Lusher and MacDougall17} Immediate surveys may have some merit, but our opinion is that using mixed-method, longitudinal studies is key to understanding symptom progression, implicated factors and causality. The papers from Norway, New Zealand and the UK illustrate this matter.

Impacts on families

The impacts on families affected by terror attacks are consistently noted in many papers. A number of the studies included young people among the participants. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3–Reference Stene, Glad, Stensland, Nilsen and Dyb5,Reference Dean, Eggleston, Ali, Thaufeeg, Wells and Zarifeh7–Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9} The research on the Manchester Arena bombing showed that the quality of close relationships is pivotal to long-term outcomes. ^{Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9}

As previously reported, deaths of close relatives in terror attacks are associated with mental disorders, particularly PTSD and depression. ^{Reference Bell, Sulaiman-Hill, Tanveer, Porter, Dean and Schluter6} But, the families of survivors also report the toll on them from witnessing the suffering and enduring distress experienced by their loved ones. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3,Reference Dean, Eggleston, Ali, Thaufeeg, Wells and Zarifeh7} Stancombe et al report that some parents described high levels of stress and feelings of guilt in exposing their children to the Manchester bombing, feeling unable to provide adequate care themselves, and stress associated with not being able to access support they thought appropriate. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3}

What is not addressed in the first 11 papers is what occurs when parents bring children to riots or coach them to participate. The 12th paper in our series considers that

Conversely, several papers in our series highlight the important roles families play in providing support after terror attacks, echoing previous findings that they are very often the preferred avenue of care. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3,Reference Bell, Sulaiman-Hill, Tanveer, Porter, Dean and Schluter6–Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9} However, we should not assume that helpful support occurs naturally, and survivors can be constrained from sharing feelings with their families if the latter are perceived as being unable or less inclined to understand their experiences. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3,Reference Drury, Stancombe, Williams, Collins, Lagan and Barrett8} This could potentially exacerbate distress and prolong recovery. These authors suggest services should facilitate processes that ensure that this support is constructive by increasing families’ understanding of survivors’ experiences and intra-familial communication. ^{Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9}

Longer-term impacts

Disasters and terrorism have negative impacts on mental health with increased rates of symptoms of mental disorders. However, relatively less is known about longer-term as compared with short-term impacts. Several papers in this thematic series address this gap, examining outcomes 2 to 13 years post-incident. ^{Reference Stene, Glad, Stensland, Nilsen and Dyb5–Reference Dean, Eggleston, Ali, Thaufeeg, Wells and Zarifeh7,Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9–Reference Wesemann, Mahnke, Polk and Willmund11}

Long-lasting impacts on survivors and emergency responders have been reported consistently and three papers in this series make this point. ^{Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9–Reference Wesemann, Mahnke, Polk and Willmund11} Oppo et al report a qualitative systematic review and meta-analyses of 10 papers on studies of adult participants. They identified four symptom trajectories in 64% of studies by latent growth modelling: low-stable (the modal response with pooled prevalence of 58%), high-stable (7%), decreasing (13%) and delayed-worsening (8%); they also found a fifth trajectory of moderately stable (19%) in 50% of the studies. ^{Reference Oppo, Forresi, Barbieri and Koenen18} A paper in our series explores the trajectories of distress and recovery; it found three broad modal patterns of responding: low symptom levels and short-term distress (the largest group of people), more persistent distress and slower recovery (a substantial minority) and high stress and deterioration (in a smaller but significant minority). ^{Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9} Thus, despite variations in the detail, terms in which the trajectories are described and measured, and number of trajectories, there appears to be broad consistency between the reports from Stancombe et al, ^{Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9} Oppo et al ^{Reference Oppo, Forresi, Barbieri and Koenen18} and Bryant et al ^{Reference Bryant, Nickerson, Creamer, O’Donnell, Forbes and Galatzer-Levy19} about how people respond over time to disasters and collective violence in a variety of different contexts.

Several papers in this issue explore some of the potential drivers of these patterns of responding. Confirming previous findings, ^{Reference Bugge, Dyb, Stensland, Ekeberg, Wentzel-Larsen and Diseth20,Reference Stensland, Thoresen, Jensen, Wentzel-Larsen and Dyb21} the presence of physical symptoms such as changes in sleep and appetite, somatic concerns and inability to perform everyday activities were reported by people with more enduring functional impairment. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3} The opinion of Stensland et al is that their ‘findings indicate that survivors’ early pain and related somatic symptoms strongly and consistently predict later psychopathology.’ ^{Reference Stensland, Thoresen, Jensen, Wentzel-Larsen and Dyb21} As identified, shame following a terror attack was explored by two of the studies we highlight. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3,Reference Glad, Aakvaag, Wentzel-Larsen, Dyb and Thoresen4} Both report that experiences of shame are prevalent years after the event with significant associations between shame and mental health outcomes. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3,Reference Glad, Aakvaag, Wentzel-Larsen, Dyb and Thoresen4} These findings suggest that somatic reactions and shame might serve as markers of risk or enduring problems. They may be important symptoms to include in research and are potential targets for intervention. They may also help us to differentiate people who are distressed from those who may progress to developing disorders.

Secondary stressors ^{Reference Williams, Ntontis, Alfadhli, Drury and Amlôt22} are discernable in a number of the papers whether or not they are described using that terminology. Their deleterious impacts on recovery are explored in several papers. ^{Reference Gottschick, Diexer, Massag, Klee, Broda and Purschke1–Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3,Reference Dean, Eggleston, Ali, Thaufeeg, Wells and Zarifeh7–Reference Stancombe, Williams, Drury, Hussey, Gittins and Barrett9,Reference Bhui, Roberts, Lashley, Jones and Kaufman12} They included family concerns, work stress, media involvement and barriers to or lack of healthcare and support. Interestingly, while social support is typically protective, one paper also reports that inappropriate or invalidating responses from families, friends and service providers can themselves act as secondary stressors. ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3} Thus, there is a component of risk concerning the nature and quality of social support that people are offered.

These findings argue for creating a clear but flexible framework for interventions to support the likely different levels of need after a terror attack (i.e. supporting the well-being of everyone affected, providing psychosocial interventions for distressed people who are struggling, and assessing and treating people who have mental disorders) [ ^{Reference Murray, Kaufman and Williams16} identifies this framework]. Fig. 1 offers a diagramatic representation of this approach. A consistent finding is that the impacts are often long-lasting, and it is now clear that support services should be available for a number of years after incidents.

Fig. 1 A strategic approach to responding to the needs of people affected by terrorist incidents (reproduced with permission from R Williams and V Kemp, 2020 – all rights reserved).

Responders

Two papers in this thematic series explore the impacts of terrorist attacks on responders. ^{Reference Wesemann, Applewhite and Himmerich10,Reference Wesemann, Mahnke, Polk and Willmund11} The main findings of the first showed that prevalence rates of PTSD, major depression and specific anxiety disorders were inconsistent and that the spectrum of symptoms is much broader than those of PTSD and other diagnoses. ^{Reference Wesemann, Applewhite and Himmerich10} As noted earlier, this raises possibilities that self-report surveys might risk confusing distress with diagnosable disorders; this might be one reason why prevalence rates reported in studies using self-reported questionnaires are so much higher than when people are assessed in clinical interviews. ^{Reference Scott, Stevelink, Gafoor, Lamb, Carr and Bakolis14}

We make two further points. First, our experiences of the COVID-19 pandemic drew our attention to the needs of healthcare staff and, indeed, all responders, but it is also clear that the needs of staff had gradually been becoming more apparent and better recognised over a much longer time frame. ^{Reference Williams and Kemp23,Reference Oeppen, Melville and Brennan24} The needs of carers and responders have not been resolved as the prevalence of COVID-19 has slowly but sporadically reduced across many countries’ populations. Thus, it is not solely emergency responders who require improved care and support, but all healthcarers. ^{Reference Oeppen, Melville and Brennan24}

Second, what form should responses to the needs of staff take? We are not drawn to specific mental health treatments as the foundation of society’s initial responses to the experiences and needs of responders, though some of them may require mental health care as time progresses. The discussion in the paper by Wesemann et al is helpful. ^{Reference Wesemann, Applewhite and Himmerich10} It includes the statement that institutional peer support was found to be effective after the 9/11 World Trade Center attacks and is probably the most appropriate way to deal with such huge incidents. ^{Reference Dowling, Moynihan, Genet and Lewis25} Several papers in our series suggest that survivors are helped by social support though the papers by Stene et al, ^{Reference Stene, Glad, Stensland, Nilsen and Dyb5} and Stancome et al ^{Reference Stancombe, Williams, Drury, Collins, Lagan and Barrett3} indicate that the researched position on effectiveness may be complicated and turn on other interacting factors (such as past experiences and affluence). ^{Reference Stene, Glad, Stensland, Nilsen and Dyb5} Others papers, including many that are published elsewhere, indicate that staff of healthcare services benefit from peer support as incidents unfold, in the immediate aftermath and also in the longer-term. ^{Reference Wesemann, Applewhite and Himmerich10,Reference Oeppen, Melville and Brennan24–Reference Williams, Kemp, Burgess, Murray, Stokes and Wood26} Thus, ‘Psychosocial Care’ in Fig. 1 includes social support for survivors and peer support for responders and other staff.

Lessons for designing and delivering services in the future

The papers included in this series illustrate knowledge that should influence how we design, deliver and sustain services and supports for people affected by collective violence. Some papers explore events and people’s perceptions of them through qualitative approaches. Others use quantitative, epidemiological research techniques and/or analysis of large data sets. Many show the importance of statistical analysis and the growing influence of longitudinal statistical approaches. Some of the papers included in the series reflect the themes of prevention through, for example, social support systems and minimising secondary stressors. Bhui et al recommend a range of other approaches including the importance of addressing the powerful influences of social media. ^{Reference Bhui, Roberts, Lashley, Jones and Kaufman12}

In conclusion, we think that further development is needed on many fronts including how best to define, prevent, determine, monitor and respond to the impacts of violent incidents on survivors so that our interventions become synergistic and promote effective person-centred, culturally appropriate care through alliances between survivors of past incidents, people at risk, practitioners, researchers, policymakers and the responsible authorities.