Deconstructing Psychological Debriefing and the Emergence of Calls for Evidence-Based Practice

Anna Avery, Steve King, Roger Bretherton and Roderick Ørner

Recent issues of ISTSS and European Society for Traumatic Stress Studies newsletters have provided a welcome opportunity for protagonists in the debriefing outcome debate to state their positions about practice and policy. It is to the credit of the field of psychotraumatology that this debate has progressed from its source, in clinical and academic disciplines, to the public domain. In this context, new and complex considerations, such as those arising from legal accountability, exert an increasing influence on the direction of the debate (Freckelton, 1998).

Under pressure to defend original theses and give account of the rationales for critical antitheses, we trust the current dialectic will prove a moderating and modernizing influence in the long term. An opportunity is being created for doctrinaire defensiveness to be replaced by tolerance of sensible pluralism and a search for common ground in published evidence.

On entering public discourse about past claims and current reappraisals, all parties in the debate share a common predicament of risking professional or personal embarrassment. The specter of costly litigation looms large.

We wonder if, at least in part, these considerations of public accountability have contributed to the shift toward more qualified claims for psychological debriefing in the fall issue of Traumatic StressPoints (Mitchell and Everly, 1998). When compared with earlier pronouncements about its apparently realizable objectives and aims (Mitchell and Everly, 1995), recent clarifications recommend themselves for recognizing the existence of an evidence base pointing to the limitations of Critical Incident Stress Debriefing. Our view is that Mitchell and Everly have presented important and timely positional statements.

As noted with surprise and interest by psychotraumatologists worldwide, it now appears Mitchell and Everly's pronouncements about the objectives and aims of CISD, made during the 1980s and as recently as 1995, should have been interpreted as pertaining exclusively to outcomes achievable by a comprehensive package of critical incident stress management services. By way of a supportive example, the authors refer to flexible, need-driven psychosocial interventions developed in a community many months after it was afflicted by natural disaster (Chemtob, Tomas, Law and Cremniter, 1997).

We wish to draw attention to a number of logical flaws in this case study approach to evidence of CISMS efficacy. For instance, claims that this systematic study confers credence on the historical, or even current, position taken by Mitchell and Everly (1998) is challenged by an examination of the list of services actually provided by Chemtob and colleagues (1997). They coordinated, to good effect, a comprehensive therapeutic program of clinical and psychosocial interventions that have but a tenuous link to practice guidelines for CISMS. Furthermore, their services hardly qualify as early interventions (Chemtob et al., 1997).

To foster a more informed view of the implications of outcome studies published to date, we would like to make the point that research reports subjected to a meta-analysis by Everly, Boyle and Lating (1998) and Everly and Boyle (1997) appear not to be a representative sample of published outcome studies. Because results derived from meta-analysis are a function of the studies included, entirely different conclusions are possible had alternative information sources been used (Davies and Crombie, 1998).

This is exactly what happened when Rose and Bisson (1999) used the Cochrane database for a different meta-analysis of other outcome studies. Any impression of bias in inclusion and exclusion criteria for meta-analyses does scant justice to the standard of scientific rigor now expected in the field of psychotraumatology. The same is true for conclusions of any one review that fails to acknowledge different outcomes reported in other publications.

Notwithstanding these obvious methodological limitations, Mitchell and Everly (1998) refer at some length to these meta-analyses that purportedly demonstrate a large positive effect attributable to the CISD intervention and the probability that such a finding might be derived by chance alone approaches zero. Not only does a study by Carlier, Lamberts, van Uchelen and Gersons (1998), which should have been acknowledged by Mitchell and Everly but was not, expose their conclusions to be statements of faith, but their declarations also belie a bizarre leap of logic. Having, with some justification, warned of methodological flaws inherent in debriefing outcome studies, these authors proceed to develop a line of reasoning premised on results of meta-analyses of exactly the type of studies they themselves consider suspect.

A further logical inconsistency warrants comment. The effect sizes purportedly revealed by the CISD meta-analyses are of a magnitude that would warrant great confidence in this particular and specific type of intervention. This leads us to wonder why Mitchell and Everly consider it necessary to argue their previous claims about achievable outcomes of early interventions pertain to a range of psychosocial interventions (CISMS) and not to any single intervention protocol. Conclusions derived from the quoted meta-analyses of group CISD, would surely justify considerable confidence in this particular intervention protocol achieving outcomes listed in their 1995 publication (Mitchell and Everly, 1995).

We see a further problem. What advice has Mitchell and Everly got for those who, in a court of law, seek a coherent rationale for the confidentiality rule applying to CISMS? Their disclaimer of specific therapeutic effect or outcome for any given intervention makes it difficult to uphold the principle of confidentiality using rationales from clinical, legal or pastoral practice. It would appear ill advised ever to have proffered this guarantee to service users.

The ESTSS has since September 1998 convened a series of European Regional Conferences On Psychological Debriefing and Early Interventions. These are part of a continent-wide consultation process involving practitioners and researchers with experience in this specialist field.

So far, evidence systematically collated in a variety of settings within the United Kingdom fails to demonstrate a distinct beneficial impact of one-off early interventions provided on a one-to-one basis (Hobbs, 1998; Brewin and Rose, 1998; Bisson, 1998). The impression formed is that the course, development and resolution of reactions to exceptionally stressful events is determined by factors, the control of which is beyond the realistic scope of early interventions.

Yet presenting opinions at a European Regional Conference is the result of a published study involving police officers in the Netherlands, some of whom underwent a CISD and some of whom did not. Evidence from follow up warns of adverse reactions that can, under certain circumstances, accrue to emergency services personnel taken through the CISD protocol by trained and experienced debriefers. (Carlier, et al., 1998)

Other speakers will have their opportunities to present original findings in a variety of European venues before the 6th European Conference on Traumatic Stress in Istanbul, Turkey, in June 1999. For this conference, a number of symposia are being convened to assess the status of evidence about early interventions and to draft guidelines on evidence-based practice.

The European Regional Conferences have so far been instrumental in engendering an appreciation that studies of CISD symptom outcome furnish, but one of a number of possible strands of evidence that should inform future practice and policy. We consider the debriefing debate remarkable for not having sought consensus about the nature of these alternative sources of evidence and the appropriate methodologies for their investigation.

In our estimation, a mistake was made in championing psychological debriefing without first generating evidence of its efficacy in terms appropriate for this type of support. The consequence of this weak evidence base is that the notion of early intervention has not been critically scrutinized. For instance, scant consideration has been given to the contradiction inherent in first stating that CISD is not a therapeutic intervention and then proffering unsubstantiated claims of its efficacy in terms drawn from clinical practice.

With such an inauspiscious start, it should come as no surprise to anyone that, true to the traditions of clinical disciplines that seek rationales for therapeutic practice through symptom outcome evaluations, priority has been given to considerations of clinical efficacy. Evidence gathered this way has not substantiated claims made for CISD. The counter-response has been to criticize systematic studies for inadequate methodological rigor. Mitchell and Every (1998) seem impervious to the monumental irony of stipulating unrealizable methodological improvements as a precondition for decisive evidence when it was the manner in which they promoted CISD in the first place that led to the adoption of these particular research methodologies.

We believe evidence in this specialist field points to a need to abandon the notion of early intervention on at least two counts. "Early" implies a time limited focus when evidence suggests informed clinical practice calls for intermediate and long-term follow up to take account of the diversity of traumatic stress reaction that come to light over time (Carlier et al. , 1998).

More significantly, we venture a view the notion of intervention shortly after exposure to traumatic stressors may be misplaced. It makes more sense to aspire to establish flexible, evidence-based support strategies developed on the basis of what victims themselves would welcome as well as the types of psychosocial support they can demonstrably put to good effect in the various stages of adjustment occurring in the wake of trauma. Outcome may more usefully be formulated in terms of facilitating improvements in the quality of the recovery environment, rather than postulates about symptomatic change, homeostatic equilibrium and identifying individuals at particular risk (Ørner, Avery and King, 1999

And finally, we have, without success, set out in search of a most elementary evidence base that confirms that emergency responders and other user groups actually want to be debriefed within the recommended 48 to 72 hours of a critical event. If this evidence base is as weak as we suspect, it would appear there is no valid rationale for advocating this specific type of early intervention. Given that Mitchell and Everly took the lead in recommending CISD during the 1980s, it may help raise the standard of the debriefing debate if they were to explain the evidence-based rationales for publicly advocating early intervention in general, and then give account of how this evidence base in turn gave rise to designated protocols for delivering CISD and other CISMS.

References
Bisson, J.I. (1997) Is post-traumatic stress disorder preventable? Journal of Mental Health 6, 2, 109-111
Brewin, C.R., Andrews, B. & Rose, S. (1999) A Preventative Programme for Victims of Violent Crime. In press
Carlier, I.V.E., Lamberts, R.G., van Uchelen, A.J. & Gersons, B.P.R. (1998). Disaster related post-traumatic stress in Police Officers: A field study of the impact of debriefing. Stress Medicine, 14, 143-148
Chemtob, C., Tomas, S., Law, W. & Cremniter, D. (1997) Postdisaster psychological intervention: A field study of debriefing on psychological distress. American J Psychiatry 154, 415-417
Davies, H.T.O. & Crombie, I.K. (1998) Getting to grips with systematic reviews and meta-analyses. Hospital Medicine 59, 12, 955-958
Everly, G.S. & Boyle, S. (1997) Critical Incident Stress Debriefing (CISD): A meta analysis. Paper Presented at the Fourth World Congress on Stress, Trauma and Coping in the Emergency Services Professions, Baltimore, MD, USA.
Everly, G.S., Boyle, S., & Lating, J. (1998) Psychological debriefing: A Meta Analysis. Paper presented at First Nations' Emergency Services Society Critical Incident Stress Conference, Vancouver, BC, Canada.
Freckleton, I. (1998) Critical Incident Stress Intervention and the Law. Journal of Law and Medicine. November 1998
Hobbs, M., Mayou, R., Harrison, B. & Worlock, P. (1996) Randomized controlled trial of psychological debriefing for victims of road traffic accidents. British Medical Journal 313, 1438-1439
Mitchell, J.T. & Everly, G.S. (1995) Critical Incident Stress Debriefing (CISD); and the prevention of work related traumatic stress among high risk occupational groups. In G.S. Everly and J.M. Lating (Eds.), Psychotraumatology: Key Papers and Core Concepts in Post Traumatic Stress (PP 267-280. New York: Plenum Press.
Mitchell, J.T. & Everly, G.S. (1998) Critical Incident Stress Management: A New Era in Crisis Intervention. Traumatic StressPoints, Fall 1998, 6-11
Ørner, R.J., Avery, A. & King, S. (1999). An evidence based rationale for early intervention after critical incidents. Paper presented at European Regional Conferences on Early Interventions and Psychological Debriefing, Aberdeen, Scotland.Raphael, B., Meldrum, L. & McFarlane, A.C. (1995) Does debriefing after psychological trauma work ? Time for randomised controlled trials. British Medical Journal, 310, 1479-1480
Rose, S. & Bisson J.I. (1999) One-off psychological interventions following trauma: A systematic review of the literature. Journal of Traumatic Stress (in press)