Critical Incident Stress Debriefing (CISD) has come under considerable debate over the last few years. Several studies have failed to document any effect of the intervention, while other studies document a clear effect. Most studies, be that in favor of debriefing or not, have serious methodological flaws. These concern aspects such as researching interventions not qualifying as a PD; using self-selection to the groups; inadequate timing of the intervention; and interventions of dubious clinical value. It is concluded that the quality of the studies to date does not justify a discontinuation of its use.
Psychological Debriefing – An Effective Method?
The Critique of PD and it’s Setting
Psychological debriefing (PD), also termed critical incident stress debriefing (CISD) was originally described by Mitchell (1983) as “either an individual or group meeting between the rescue worker and the caring individual (facilitator) who is able to help the person talk about his feelings and reactions to the critical incident“(p. 37). Some of the conceptual confusions originated in Mitchell using the term debrief for individual contacts although he later refer to CISD as “a group meeting for discussion about a distressing critical incident” (Mitchell & Everly, 1996, p. 8). Dyregrov (1989) presented the following definition: “A psychological debriefing is a group meeting arranged for the purpose of integrating profound personal experiences both on the cognitive, emotional and group level, and thus preventing the development of adverse reaction” (p.25). A detailed review of facts, thoughts, impressions and reactions can obviously be used in conversations with individuals as part of normal crisis intervention, while psychological debriefing as a method was designed for groups. For group meetings to achieve their aims they should be instigated within a brief time after the traumatic event, those who lead the group must be trained and experienced in leading the debrief process, the group must have experienced a common stressor, time must allow a thorough review of the different “phases”, and the meetings used to screen those who need extra help. The many factors influencing the debrief meeting and its ability to achieve its purpose is described in Dyregrov (1997). From Mitchell’s (1983) first description of the method until today’s practice there has been significant changes, and presently the method usually is used as one part of a more integrated system of interventions referred to as critical incident stress management or CISM (Mitchell & Everly, 1996).
During the last years several critical reports have been published regarding the use of PD or CISD. A heated debate has been going on both in Australia, the States and in Great Britain, even leading to a suggested discontinuation of the use of CISD protocols (see Avery & Orner, 1998). Since the debate continues and has led to premature suggestions for discontinuation of the use of PD it is necessary to look more closely at some of these studies.
The above mentioned debate grew rapidly after Beverly Raphael, Leonore Meldrum and Alexander C. McFarlane wrote a letter to the editor of the British Medical Journal in 1995, asking for more randomized, controlled studies of the method. Furthermore, they stated that several studies reported a negative effect of the method. In addition they wrote that the method actually could aggravate the traumatic process, and that it has an ideological and symbolic more than a helping value. Several studies reviewed later in this article were also taken as “proof” that PD has no effect. Raphael, Meldrum and McFarlane are respected experts in the traumatology field, and their reservations against debriefing naturally have left many professionals in doubt about the necessity and effectiveness of debriefings. In this article it is proposed that their critique was based on studies that did not warrant the negative presentation they gave of debriefing.
Studies Reporting No Effect of PD
In the following an evaluation will be presented of the studies that purport that PD or CISD does not have the desired effect. Following this a description of studies supporting PD is given.
Deahl, Gillham, Thomas, Searle and Srinivasan (1994) investigated the proneness for disease in soldiers from the Gulf War, finding that debriefing did not reduce later psychiatric morbidity. They do not make clear what the debriefing consisted of since the intervention is only briefly mentioned in the article. They furthermore inform the reader that they have used Dyregrov’s model (1989). This is rather strange since neither has this author developed such a model, nor have this group received training from me. In addition the timing of the debriefing is highly variable. The most serious methodological objection is, however, the self-selection that has taken place to the debriefing-group. This means that the participants in the debrief group personally wanted to take part in the debrief, most likely as a result of a greater need to talk about the event than the individuals who automatically became part of a control group. This of course becomes a serious source of error that may explain the possible differences between the groups. The authors themselves are aware of these methodological limitations, and conclude, despite their negative results, that: “..despite our findings we remain committed to the principle of debriefing…” (p.64). They emphasize the importance of a rapid and locally held debriefing, and that the soldiers were suspicious towards “outsiders”, included mental health personnel.
In Australia, Justin Kenardy and colleagues (1996) have conducted research on rescue personnel after an earthquake. They found no effect of the debriefing during the two years following the disaster (measured with GHQ-12 and IES on four different occasions). The groups were established through self-selection with the sources of error this entails and that already are commented upon. The participants taking part in the debriefing group were significantly different from the control group regarding level of education, their self-report on being helpful in non-threatening situations, by having higher professional prestige, and by being more females (females usually report more distress on most measures, cf. Breslau, Davis, Andreski, Peterson., & Schultz, 1997). The authors had no control over the debriefing procedures, thus it is not known what they consisted of. Furthermore, background and training of the debriefing leaders is not documented, and the same goes for the timing of the debriefing. The authors report that 80% of the participants felt that the debriefing was of help.
In Great Britain, Lee, Slade and Lygo (1996) have offered what they call “psychological debriefing” to women having had a miscarriage. In this study the women were randomized to two different groups and they were offered a one-hour consultation in their own home two weeks following the miscarriage. It is obvious that they had emphasized a schematic procedure in the various phases of the debriefing. One week and four months following this consultation the women were screened for anxiety, depression, intrusive memories and avoidance reactions. This screening showed no differences between the groups. This study is peculiar in several ways. First of all, it is not a study of debriefing. It does not describe how a group of subjects, being gathered for intervention following a disaster they all have been involved in, is coping. It could have been a description of a crisis intervention if the help had arrived earlier. However, the most serious aspect of this study is that the women only were offered a one-hour consultation following their miscarriage. The author of this article has for many years worked with families who had lost a child. The follow-up necessary in such situations demands a much more intensive approach (Dyregrov, 1989, 1990). It is indeed very doubtful whether the PD format is adequate in a one-hour follow-up in these situations. A short conversation around a very emotionally demanding situation may open up emotional channels without adequate time to talk through the event
Hobbs, Mayou, Harrison and Worlock (1996) randomized a group of victims after traffic accident to an intervention group and a control group. The intervention consisted of what the authors called psychological debriefing which lasted for one hour, and it was usually carried out between 24 and 48 hours following the accident. While the groups were not different regarding symptoms preceding the intervention, the intervention group had experienced more serious physical injuries following the accident and they stayed longer in hospital than the controls. Four months following the intervention the researchers found no significant decline in different symptoms in any of the two groups. In two sub-scales of the “Brief Symptom Inventory” the intervention group had higher scores (more problems). The intervention is carried out individually and not in a group, and the session lasts for one hour only, without any follow-up. This is more a study of crisis intervention of dubious quality more than it is a study of debriefing. Clinically it is also questionable whether the use of an intervention following the “debriefing model” is correct at this point in time following the event. This authors clinical experience has been that the physical healing must take place before the psychological healing processes can continue. That they in this study try to pressure the wounded person into cognitive and emotional processing of the accident is a questionable clinical procedure in my opinion. It seems quite clear that this and the previously mentioned studies look at the effect of one hour of individual consultation, more than study the effect of PD.
Bisson, Jenkins, Alexander and Bannister (1997) randomized patients wounded in a fire to what they called a debrief-group and a control group. The intervention was given to each single patients or couples, and it lasted on average 44 minutes (30-120 minutes), and was carried out by a nurse or a research psychiatrist who were tutored by the first author (psychiatrist). The results showed that sixteen (26 %) of the PD group was found to have PTSD following thirteen months, while in the control group 9 % were diagnosed with PTSD. Even before the intervention, the PD group was described as having experienced twice as many important past traumas, and in addition the PD group had experienced more serious fire traumas than the control group. Both these aspects can explain why the PD group’s PTSD diagnoses were higher in number than for the control group. Turnbull, Busuttil and Pittman (1997) and Reiss and Leese (1997) have raised several other methodological objections against this study. Bisson and co-workers also reported that the earlier the intervention was carried out following the accident, the worse they were doing later on. This pertain to the criticism raised against Hobbs and colleagues – it is clinically unsound to intervene following PD principles while physical healing is taking place. In a letter to British Journal of Psychiatry following this study, a doctor who himself have been wounded in a fire stated that the timing of PD was incorrect (Kraus, 1997).
A question should also be raised regarding the type of quick intervention that took place in the study by Bisson et al. Paramount when practicing PD is to spend the amount of time required. In this study 44 minutes on average was spent with the patients. Within this short amount of time the author of this article would not even have finished listening to the participants telling the facts and thoughts regarding the event. If anything is measured in Bisson et al’s study, it must be the effect of a badly timed, rapid conversation, and not a sound clinical intervention.. In events of this nature, studies need to investigate the effect of a timely crisis intervention that includes several conversations with the patients. Bisson is one of the professionals who is very critical to PD (see Bisson, 1997; Bisson & Deahl, 1994). When the critic is based on intervention of an insufficient quality, it does not help the cause to have a good research design.
In addition to these studies Hytten and Hasle (1989) did not find any differences in Impact of Event scores between fire-personnel that participated in debriefing and those who did not following a hotel fire, even though the participants in the debriefing viewed it favorable. Again self-selection determined the group composition, something also present in a study by Matthews (1998). She studied a group of 63 health care workers who experienced violence or other trauma in their work at psychiatric institutions. One week after the event she compared 14 workers who wanted and got debriefing with 18 who were offered but refused, all within the same health district. In addition she used 31 persons who experienced the same kind of events in another health district as a comparison group. She found the lowest stress level in the district where debriefing was available. The group that chose not to be debriefed had the lowest level of posttraumatic stress after one week. One may easily think that those who participated in debriefing were the worst off, since, before the intervention they were significantly more distressed by what had happened than those who chose not to be debriefed. The author correctly points out that those who experienced the event as most stressful sought the most appropriate way of handling the event; to participate in debriefing. The fact that the stress level was lowest in the area where debriefing was available lead the author to the conclusion that debriefing have a part to play following work related trauma. It must be assumed that debriefing first lead to an increase in distress because it activates emotional networks, as when participants write about a traumatic event. Here research shows a temporary increase in distress but a long-term improvement on different health measures (see Smyth, 1998 for a summary). The sole measurement after one week in Matthews’s study will not uncover such effects. The study has other obvious limitations, such as self-selection, and shows the complexity of conducting research in this area.
In conclusion, it seems that studies which report no effect of debriefing (or a negative effect, see Bisson et al, 1997) have several methodological weaknesses: a) they analyze interventions that only to a limited degree can be called psychological debriefing, b) several studies use self-selection to intervention group and control group, c) it is not clearly defined what the debrief consisted of, d) the timing of the intervention is variable and partly outside the time period recommended for PD, e) the intervention used seems to be clinically insufficient regarding the traumatic event experienced, f) the background and training of the persons who have carried out the interventions is unclear or possibly inadequate, g) the groups in the studies are not comparable, and h) debriefing is investigated in isolation, and not as part of an integrated chain of assistance (CISM).
Particularly the self-selection is a problem in these studies, because it must be presumed that persons who are characterized by avoidance and repression will avoid meetings where they are expected to talk about the event. First of all those who do not feel the need for debriefing because they were peripheral to the event or felt that the event was of little consequence to them will be part of the control group. Secondly, people who use avoidance and denial as a coping strategy will tend to stay away from such meetings. If this “control” group is compared with a group that through debriefing meetings are encouraged to and “learn” to put their thoughts and reactions into words, then one would expect the debriefed group to score higher on self-reported reactions (normally being studied).
It is also alarming that most reports study the effect of short-term individual intervention, while PD has been developed just to profit from being in a group following the event.
Additional Studies of PD
A number of studies have concluded that PD or CISD is followed by a positive effect for the participants (Bohl, 1991; Ford et al., 1993; Jenkins, 1996; Robinson & Mitchell, 1993; Stallard & Law, 1993; Yule & Udwin, 1991). Everly, Flannery & Mitchell (1998) have in addition revised a number of published and unpublished reports and case studies showing positive effects of debriefing. In almost all of the reports (also the negative studies previously described) the participants of the debriefing groups (or individual meetings) when asked to rate their satisfaction or helpfulness experience the meetings as being helpful.
Everly, Boyle and Lating (1998) conducted a metaanalysis based on debriefing studies found in medical and psychological databases. They identified 14 empirical investigations of which 10 were utilized for the analysis. Three were excluded as they failed to use group debriefing interventions and one as it failed to yield data that meaningfully could be used in the analysis. They found a significantly positive effect size (mean Cohen’s d = .54, p < .01) resulting from the CISD intervention. The authors comment that this beneficial effect was revealed despite the wide variety of subject groups, the wide range of traumatic events, and the diversity of outcome measures.
However, many of the methodological objections raised in relation to the critical studies also goes for the studies where participants report positive results. A number of very different interventions are being called debriefing, and the extent and the timing of these interventions vary. In addition the training and background of the debriefers are variable, and a lack of control group or self-selection procedure to intervention and control group has taken place. Instead of going through all these studies, a few of the studies will be discussed more thoroughly.
Chemtob, Tomas, Law and Cremniter (1997) carried out a thorough study regarding “the influence of debriefing on psychological distress”. In this study they describe how victims of a hurricane had their problems reduced compared to a group who only later received the same type of intervention and who then, after debriefing, report the same reduction in problems. The effectiveness of the intervention was evaluated by the use of the Impact of Event scale used before and following the intervention. There are several objections to this type of design. In addition to lack of data regarding the participants ahead of the debriefing, the participating group was very heterogeneous. Furthermore, the intervention, consisting of PD plus a two hour long lecture on “post disaster recovery”, was carried out six to nine months following the disaster. This study confirm that PD can be effective a long time after the time period recommended for debriefing, a finding similar to what was reported by Stallard and Law (1993) in their study of adolescents who survived a mini-bus traffic accident.
Usually PD is practiced as one of several interventions following a critical event, often called Critical Incident Stress Management (CISM). Leeman-Conley (1990) documented that an Australian bank introducing CISM following a bank robbery experienced a decline in the number of sick leaves and “worker’s compensation claims” with more than 60% compared with the year ahead of the introduction of the program, even though the assaults became increasingly more brutal. Flannery and colleagues (reported in Everly, Flannery & Mitchell, 1998) report that following the implementation of a CISM program violence and attacks from patients within a psychiatric setting was reduced with 63% over a two-year period. In addition a reduction in personnel turnover and sick-leaves took place together with a decline in “workers compensation claims”. Medical and juridical expenses were also reduced. Because of these results, similar programs have been started other places.
In Canada, Western Management Consultants (reported in Everly, Flannery & Mitchell, 1988) evaluated a comprehensive CISM program consisting of among other things preparatory training, individual counseling, and CISD (based on Mitchell) for nurses. Of the persons who participated in a debriefing in this study, 24% reported a decline in personnel turnover, and 99% reported a decline in sick leave days.
The author of this article have not carried out studies on the effect of debriefing. Unpublished data seem to indicate, however, that the participants’ perception of the value of PD depends on the amount of experience of the debrief leader (Dyregrov, 1996). Where the leader had much experience with the method, almost everyone participating in the debriefing reported that they found the PD useful. In the cases where the leader had less experience, however, a much lower number of participants found the session useful (the majority said that the PD was of some use). Even though the following example has no scientific value, it is still worth mentioning that the person in charge of the follow-up of post office personnel following armed robberies in Bergen, Norway, stated that it was a huge difference between the period before and after PD was routinely installed as part of a CISM program. Ten years ago, before the psychological follow-up and PD was introduced, a high number of and long sick leaves were common. Today, however, this is no longer a problem.
The many variables effecting the debriefing process and it’s outcome has been described in more detail elsewhere (Dyregrov, 1997). Particularly important is the background of the debriefing leader, and his or her competence in leading such meetings. With the possibility that debriefing can have a negative effect on the participants (Dyregrov, 1997), the group process should be analyzed more thoroughly, together with group design, leadership competence, timing, areas of use, individual suitability, etc.
Conclusion
Several studies have been published over the last years concluding that debriefing does not have a positive effect on mental health measures following critical events. These studies are founded on weak methodological designs, and it would be wrong to draw firm conclusions regarding the usefulness of the technique. The majority of the studies have described one single intervention with individual patients, and not the group intervention that PD really is intended to be. When individuals are receiving help during a crisis, it is not possible to apply the same structure as in PD, without adjusting the process. The group cannot be used in order to normalize reactions, and it cannot be a source of help for the individual members of the group (see Dyregrov, 1997, for a description of this group support). Intervention with individuals assumes a different procedure, where the same areas being covered in PD is being processed (facts, thoughts, impressions and reactions). However, the normalization of reactions depend on the therapist’s experience, and there is more freedom to go back and forth between “phases” talking about the facts, thoughts, impressions and reactions related to the event. This usually demand longer individual follow-up sessions than what has been reported in the studies described herein.
In my opinion the debate on debriefing is not only a scientific but also a political debate. It entails power and positions in the therapeutic world. As a technique presented by Jeffrey T. Mitchell in 1983, PD represented a threat to the psychiatric professional elite. Throughout his teachings Mitchell has argued that the traditional psychiatric way of thinking was not appropriate for the population (emergency personnel) for whom the method was developed for, and that health personnel with a psychiatric background often would need to relearn some of their way of thinking and their work methods in order to become good practitioners of the method. In addition, many of the people being trained in the technique both in Australia and North America were peer support personnel and mental health workers working outside psychiatric institutions. PD thus has been partly self-help and consumer driven where the recipients of services have had more control than in traditional academic or medical approaches based on a more psychiatric disease model. Mitchell have also strongly emphasized that debriefing is not a form of or substitute for psychotherapy (see Mitchell & Everly, 1996, p. 211). Using PD or CISD as part of crisis intervention has thus been part of a non-psychiatric approach, and therefore it was only natural that there would be a reaction from the “psychiatric establishment”. The critique in Australia was raised by some of the best known psychiatrists within the trauma field, but it was based on studies that either lacked the methodological quality necessary to support the critique, or studies that investigated the effect of individual follow-up. This, in my point of view, indicates that the debate not only entailed the question of whether debriefing worked or not, but also a more political stance.
Participants of PD report in most cases that they find the groups helpful. Depending on what the group goals are, the groups will be regarded as more or less successful. It is obvious that the groups have a symbolic function, for instance where a group of colleagues or friends have experienced a critical event. Moreover, the group can become a sign that the employer or community cares. So far no good documentation of the preventive effect regarding post traumatic stress syndrome has been produced, however, several studies indicate a marked reduction in costs due to reduction in sick leaves. It must be presumed that PD on its own, without being followed by support and care from leaders and colleagues, or without the possibility for individual follow-up when necessary, only has more limited value. PD should be practiced in a caring environment and as part of a strategy where the building of a social fellowship within the company or in the local community is central.
One of the positive outcomes of the debate regarding debriefing is the highlighting of several factors that we do not have sufficient knowledge of. Future studies will presumably improve the quality control of PD.
References:
Avery, A., & Orner A. A., (1998). First report of psychological debriefing abandoned – the end of an era? Traumatic Stress Points, 12, 3-4.
Bisson, J. I. (1997). Is post-traumatic stress disorder preventable? Journal of Mental Health. 6, 109-111
Bisson, J. I., & Deahl, M. P. (1994). Psychological debriefing and prevention of post-traumatic stress. British Journal of Psychiatry, 165, 717-720.
Bisson, J. I., Jenkins, P. L., Alexander, J., & Bannister, C. (1997). Randomized controlled trial of psychological debriefing for victims of acute burn trauma. British Journal of Psychiatry, 171, 78-81.
Bohl, N. (1991). The effectiveness of brief psychological interventions in police officers after critical incidents. In J. Reese, J. Horn, & C. Dunning (Eds.). Critical incidents in policing. Revised (pp. 31-38). Washington, D. C.: U.S. Government Printing Office.
Breslau, N., Davis, G. C., Andreski, P., Peterson. E., L., & Schultz, L. R. (1997). Sex differences in posttraumatic stress disorder. Archives of General Psychiatry. 54, 1044-1048
Chemtob, C. M., Tomas, S., Law, W., & Cremniter, D. (1997). Postdisaster psychosocial intervention: a field study of the impact of debriefing on psychological distress. American Journal of Psychiatry, 154, 415-417.
Deahl, M., Gillham, A. B., Thomas, J., Searle, M. M., & Srinivasan, M. (1994). Psychological sequelae following the Gulf war. Factors associated with subsequent morbidity and the effectiveness of psychological debriefing. British Journal of Psychiatry, 165, 60-65.
Dyregrov, A. (1989a). Caring for helpers in disaster situations: Psychological debriefing. Disaster Management, 2, 25-30.
Dyregrov, A. (1989b). Retningslinjer for hjelp til familier etter barns død. Tidsskrift for Den norske lægeforening, 109, 3408-3411.
Dyregrov, A. (1990). Crisis intervention following the loss of an infant child. Bereavement Care, 9, 32-35.
Dyregrov, A. (1996). The process in critical incident stress debriefings. Paper presented at the European Conference on Traumatic Stress in Emergency Services Peacekeeping Operations & Humanitarian Aid Organizations. University of Sheffield, England, 17th – 20th March.
Dyregrov, A. (1997). The process in critical incident stress debriefings. Journal of Traumatic Stress, 10, 589-605.
Everly, G. S., Boyle, S. H., & Lating, J. M. (1998). Psychological debriefing: A meta-analysis. Paper presented at the First Nation’s Emergency Services Society Critical Incident Stress Conference, May 10-14, Vancover, British Colombia, Canada.
Everly, G. S., Flannery, R. B., & Mitchell, J. T. (1998). Critical Incident Stress Management (CISM): A review of the literature. Manuscript submitted for publication.
Ford, J. D., Shaw, D., Sennhauser, S., Greaves, D., Thacker, B., Chandler, P., Scwartz, L., & McClain, V. (1993). Psychosocial debriefing after operation desert storm: marital and family assessment and intervention. Journal
of Social Issues, 49, 73-102.Hobbs, M., Mayou, R., Harrison, B., & Worlock, P. (1996). A randomised controlled trial of psychological debriefing for victims of road traffic accidents. British Medical Journal, 313, 1438-1439.
Hytten, K., & Hasle, A. (1989). Fire fighters: A study of stress and coping. Acta Psychiatrica Scandinavica, 80, Supplementum, 50-55.
Jenkins, S. R. (1996). Social support and debriefing efficacy among emergency medical workers after a mass shooting incident. Journal of Social Behavior and Personality, 11, 477-492.
Kenardy, J.A., Webster, R.A., Lewing, T.J., Carr, V.J., Hazell, P.L., & Carter, G.L. (1996). Stress debriefing and patterns of recovery following a natural disaster. Journal of Traumatic Stress, 9, 1, 37-49.
Kraus, R. P. (1997). Psychological debriefing for victims of acute burn trauma (Letter to the editor). British Journal of Psychiatry, 171, 583.
Lee, C., Slade, P., & Lygo, V. (1996). The influence of psychological debriefing on emotional adaptation in women following early miscarriage: a preliminary study. British Journal of Medical Psychology, 69, 47-58.
Leeman-Conley, M. M. (1990). After a violent robbery... Criminology Australia. April/May, 4-6.
Matthews, L. R. (1998). Effect of staff debrieing on posttraumatic stress symptoms after assaults by community housing residents. Psychiatric Services, 49, 207-212.
Mitchell, J. T. (1983). When disaster strikes.... The Critical Incident Stress Debriefing. Journal of Emergency Medical Services, 8, 36-39.
Raphael, B., Meldrum, L., & McFarlane, A.C. (1995). Does debriefing after psychological trauma work? British Medical Journal , 310, 1479-1481.
Reiss, D., & Leese, M. (1997). Psychological debriefing for victims of acute burn trauma (Letter to the editor). British Journal of Psychiatry, 171, 584.
Robinson, R. C., & Mitchell, J. T. (1993). Evaluations of psychological debriefings. Journal of Traumatic Stress, 6, 367-382.
Smyth, J. M. (1998). Written emotional expression: Effect sizes, outcome types, and moderating variables. Journal of Consulting and Clinical Psychology, 66, 174-184.
Stallard, P., & Law, F. (1993). Screening and psychological debriefing of adolescent survivors of life-threatening events. British Journal of Psychiatry, 163, 660-665.
Turnbull, G., Busuttil, W., & Pittman, S. (1997). Psychological debriefing for victims of acute burn trauma (Letter to the editor). British Journal of Psychiatry, 171, 582.
Yule, W., & Udwin, O. (1991). Screening child survivors for post-traumatic stress disorders: Experiences from the «Jupiter» sinking. British Journal of Clinical Psychology, 30, 131-138.
* German version, translated by Angelika Schrodt