Anda di halaman 1dari 12

British Journal of Medical Psychology (2000), 73, 8798 q 2000 The British Psychological Society

Printed in Great Britain

87

The in uence of occupational debrie ng on post-traumatic stress symptomatology in traumatized police of cers
I. V. E. Carlier,* A. E. Voerman and B. P. R. Gersons
Department of Psychiatry, Academic Medical Centre, University of Amsterdam, The Netherlands Certain individuals, such as police of cers, are exposed to traumatic events as part of their occupational roles. In an effort to prevent psychological illnesses, notably the post-traumatic stress disorder, from arising out of work-related traumatic incidents, psychological interventions have been developed such as Critical Incident Stress Debrie ng (Mitchell, 1983; Mitchell & Everly, 1996). The present study tests the hypothesis that debrie ng reduces the psychological morbidity caused by work-related incidents. Because debrie ng techniques were not designed for application on a one-off basis (Robinson & Mitchell, 1993), the procedure studied here consisted of three successive debrie ng sessions (at 24 hours, 1 month and 3 months post-trauma), which included traumatic stress education. In a sample of 243 traumatized police of cers, a subgroup of debriefed of cers (N = 86) was compared with non-debriefed internal (N = 82) and external (N = 75) control groups. No differences in psychological morbidity were found between the groups at pre-test, at 24 hours or at 6 months post-trauma. One week post-trauma, debriefed subjects exhibited signi cantly more post-traumatic stress disorder symptomatology than non-debriefed subjects. High levels of satisfaction with debrie ng were not re ected in positive outcomes. The ndings are translated into recommendations for the future use of debrie ng in police practice.

Critical incident stress is the emotional stress that individuals experience after being exposed to a speci c incident or incidents (Blacklock, 1997, p. 23). Certain people are exposed to critical or traumatic events as part of their occupational roles, for instance police of cers (Carlier, Fouwels, Lamberts, & Gersons, 1996; Carlier & Gersons, 1992, 1994, 1995; Carlier, Lamberts, & Gersons, 1997; Gersons & Carlier, 1994; Wagner, 1986). Traumatic experiences on the job can cause psychological damage, commonly in the form of post-traumatic stress disorder (PTSD). In an effort to prevent or minimize psychological morbidity following work-related traumatic events, calls have been made for the routine provision of early psychological intervention. An example of such early intervention is critical incident stress debrie ng (CISD; Dyregrov, 1989; Mitchell, 1983) advocated for the prevention of PTSD. Debrie ng is founded on the belief that promptly talking through traumatic experiences will aid people in recovering from psychological damage (Bolwig, 1998, p. 169); it is
*Requests for reprints should be addressed to Dr Ingrid V. E. Carlier, Department of Psychiatry, University of Amsterdam, Tafelbergweg 25, 1105 BC Amsterdam, The Netherlands (e-mail: I. V. Carlier@AMC.UVA.NL).

88

I. V. E. Carlier et al.

based on the principle of early intervention (Mitchell, 1983). A typical debrie ng takes place in a session 23 days after the trauma. Although initially designed for use in groups, debrie ng has also been applied to individuals, couples and families. To date, three randomized controlled trials have been published on debrie ng, mainly performed in a single-session format (Bisson, Jenkins, Alexander, & Bannister, 1997; Hobbs, Mayou, Harrison, & Worlock, 1996; Lee, Slade, & Lygo, 1996). In the Lee study, individual debrie ng was provided 2 weeks post-trauma. At a 4-month follow-up, no signi cant differences in symptomatology were evident between debriefed and nondebriefed respondents. The Hobbs study, in which individual debrie ng occurred 2448 hours post-trauma, found signi cantly greater symptomatology in debriefed respondents at a 4-month follow-up than in non-debriefed respondents. In the Bisson study, debrie ng was carried out, either individually or in pairs, 219 days post-trauma. Although 3 months later no signi cant differences had emerged between respondents, a 13-month follow-up recorded signi cantly greater symptomatology (PTSD, depression, anxiety) in the debriefed group than in the non-debriefed group. We may conclude that these randomized trials offer no evidence for a positive effect of individual debrie ng (see also Rose & Bisson, 1998; Wessely, Rose, & Bisson, 1997). Busuttil and Busuttil (1995) have suggested that the lack of positive effects from individual debrie ng may be due to the absence of therapeutic group factors such as support and recognition. This explanation does not seem very plausible, however, given that several controlled non-randomized studies of group debrie ng have replicated the results of the three randomized trials (Carlier, van Uchelen, Lamberts, & Gersons, 1998a; Deahl, Gillham, Thomas, Searle, & Srinivasan, 1994; Grif th & Watts, 1994; Hytten & Hasle, 1989; Kenardy et al., 1996; Matthews, 1998; McFarlane, 1988). Other authors (Turnbull, Busuttil, & Pittman, 1997; Watts, 1994) have blamed the absence of positive effects on the practice of providing only one session of debrie ng. The assumption is that multiple debrie ng sessions may deliver a more powerful effect on post-traumatic stress symptomatology. This is the hypothesis that we investigate in the present study on traumatized police of cers. In the Netherlands, it is standard practice to offer a series of three debrie ng sessions to every operational police of cer who experiences a critical incident. The types of critical incident that can prompt an offer of debrie ng are precisely delineated, and they conform to the stressor A(1) criterion for PTSD diagnosis (APA, 1994; Carlier & Gersons, 1992). This distinguishes them from other stressful life events, such as relationship problems and work con icts, which do not qualify subjects for the debrie ng programme. Methods Design and participants
Our study employed a pre-testpost-test control group design. Randomization was not feasible, because police regulations require that all traumatized of cers receive an offer of debrie ng and that no one is to be excluded. A partial solution to the randomization problem was to form a control group of of cers who had undergone traumas before occupational debrie ng was introduced in 1992. In our study this group is designated as the external control group; it was drawn from a research project that the authors (Carlier et al., 1997) had conducted previously amongst 75 police of cers. These existing data, which involved the same types of police traumas and assessment methods but no intervention, were subjected to secondary analysis.

Post-traumatic stress in police of cers

89

The external control group was compared with a sample of 86 traumatized of cers the experimental group who had assented to and received debrie ng. We also recruited an additional group of 82 traumatized of cers referred to here as the internal control group who had declined the offer of debrie ng. Their reasons for doing so were that they did not perceive the event as shocking or had no time for a debrie ng interview. More detailed information on the research samples can be found in the results section, where we compare the three groups in terms of their pre-test data.

Measures
The total period of assessment was 6 months, measured from the traumatic event itself. Four assessments were made: pre-test (shortly before the start of debrie ng), 24 hours post-trauma (shortly after the rst debrie ng session), 1 week post-trauma, and 6 months post-trauma (after the second and third debrie ng sessions). For the pre-testing and the retesting 24 hours post-trauma, the SpielbergerStateTrait Anxiety Inventory (STAI, State Version; Spielberger, Gorsuch, & Lushere, 1970) was administered, taking about 5 minutes. For the external control group, no STAI data were available. At the third assessment point, 1 week post-trauma, the Self-Rating Scale for PTSD (SRS-PTSD; Carlier, van Uchelen, Lamberts, & Gersons, 1998b; Davidson et al., 1997) was administered. It recorded the 17 PTSD symptoms (DSMIV; APA, 1994) and took about 10 minutes. The SRS-PTSD has been shown to have satisfactory sensitivity and speci city (Carlier et al., 1998b). In the external control group, the Impact of Events Scale was used (IES; Horowitz, Wilner, & Alvarez, 1979). As the IES does not include hyperarousal, only re-experiencing and avoidance were scored for the external control group at this assessment point. Both instruments, SRS-PTSD and IES, record both the frequency and the intensity of the PTSD symptoms. The two instruments correlated strongly in the mean scores for both reexperiencing (Pearsons r = 0.88, p < .01) and avoidance (Pearsons r = 0.73, p < .01). Also at this assessment point, the Peritraumatic Dissociative Experiences Questionnaire (PDEQ-R; Marmar et al., 1994) was administered. Certain general data were also solicited, such as trauma-related variables, background variables, satisfaction with debrie ng, and (in the internal control group) reasons for refusing debrie ng. At the nal assessment point 6 months post-trauma, the Structured Interview for PTSD (SI-PTSD; Carlier et al., 1998b; Davidson et al., 1989) was administered. The SI-PTSD has shown good reliability and validity, and it correlates well with the diagnoses for both current and past PTSD made with the Structured Clinical Interview for DSMIV Axis I Disorders (SCID; Blake et al., 1995; Davidson et al., 1989). In addition, the Anxiety Disorders Schedule-Revised (ADIS-R) was administered to detect any other psychopathology (DiNardo & Barlow, 1988). It assessed the present state of psychological health (current diagnosis) as well as that in the past (lifetime diagnosis). The duration of the entire interview was between 30 minutes and 1 hour. Finally, the following relevant issues were recorded: perceived support at work and occupational stress (Caplan, Cobb, French, van Harrison, & Pinneau, 1980; Carlier et al., 1997), perceived private support (Cohen & Hoberman, 1983), neuroticism and introversion (Eysenck, 1957; Jensen, 1958); dissociation (Bernstein & Putnam, 1986; Draaijer & Boon, 1993), personality disorder (Hyler et al., 1988), past or present illnesses, and other traumatic experiences (Green, 1993; Mollica et al., 1992). All assessors were trained research psychologists (A.V., R.H., G.W., J.K.) who were blind to the debrie ng status of the subjects. Non-parametric (chi-square) tests were run to detect any researcher effects, and none were found. The interviews were conducted in small, comfortable rooms at the police departments. The assessors took extensive notes during the interviews, and these were scored by an independent rater (I.C.). Inter-assessor agreement with regard to post-traumatic stress symptoms was .98.

The intervention
The debrie ng adhered to the structure rst described by Mitchell (1983; see also Mitchell & Everly, 1996), adapted for use with individual trauma victims. It goes without saying that debrie ng in the police force takes place on a voluntary basis. When someone has experienced a critical incident, they are provided three

90

I. V. E. Carlier et al.

successive debrie ngs approximately 24 hours, 1 month and 3 months after the incident. The debriefer applies a seven-stage, semi-structured procedure, comprising: an introduction; facts; thoughts and impressions; emotional reactions; normalization and traumatic stress education; planning for the future; and disengagement. The last two interviews also re ect an important auxiliary role of debriefinga screening function to identify people who may need additional intervention. The debrie ng itself focuses solely on the critical incident and its effects, and it is supplemented by written documentation that explains common reactions to traumatic events, suggests strategies to deal with such reactions, and gives telephone numbers to contact if further immediate help is required. Initial debrie ngs averaged 41.4 minutes (SD = 24.9; range 5120 minutes); second debrie ngs 17.4 minutes (SD = 11.1; range 560 minutes); and third debrie ngs 15.9 minutes (SD = 8.4; range 130 minutes). All debrie ngs were conducted in the same fashion. They were carried out by 43 police of cers, trained in the technique and supervised by one of three police social workers (E.M., P.S., H.G.), all of whom had considerable experience in the comprehensive Critical Incident Stress Management (CISM) programme. Debrie ngs took place in the privacy of a separate room at the police station.

Results Pre-test data on background variables and traumatic events Table 1 shows the most important background variables of the three subsamples. Summarizing Table 1, we may conclude that the three subsamples are reasonably comparable in terms of their background variables, and that the experimental group and the internal control group exhibited similar anxiety levels (the external control group was not tested). Although the average age in the experimental group was signi cantly lower than that in control groups (F(2) = 7.4, p < .01), the absolute difference is barely 3 years, making it improbable that age would bias the results. A signi cantly smaller proportion of the respondents in the experimental group were married or cohabiting than in the control groups (x 2 (1) = 6.6, p < .05, odds ratio (OR) + 2.7, 95% con dence interval (CI), 1.25.7; x 2 = 3.9(1), p < .05, OR = 2.2, 95% CI, 1.04.7), but the divergence was not extreme. The experimental group and the external control group contained signi cantly more patrol of cers than the internal control group (x 2 (1) = 6.2, p < .05, OR = 0.3, 95% CI, 0.090.8; x 2 (1) = 6.1, p < .05, OR = 3.9, 95% CI, 1.212.6). Conversely, the internal control group contained more detectives than the experimental group and the external control group (x 2 (1) = 7.2, p < .01, OR = 10.4, 95% CI, 1.3 84.6; x 2 (1) = 4.1, p < .05, OR = 0.2, 95% CI, 0.041.06). Respondents in the experimental group had signi cantly fewer years of police experience than those in the control groups (F(2) = 10.2, p < .01), but here too the absolute differences were small. In the other characteristics we recorded, which included occupational stress, perceived private support, past or present illnesses, neuroticism, introversion, dissociation and peritraumatic dissociation (not shown in Table 1), no signi cant differences emerged between the three subsamples on the pre-test. Table 2 gives an overview of the traumatic police events experienced by the respondents. In relation to the critical police incidents that prompted debrie ng, few statistically signi cant differences appeared between the three subsamples. Event 9, an accident involving death or serious injury, was slightly less common in the experimental group than in the internal control group (x 2 (1) = 4.7, p < .05, OR = 0.3, 95% CI, 0.12 0.93), where it was again less common than in the external group (x 2 (1) = 12.6, p < .01,

Post-traumatic stress in police of cers


Table 1. Pretest data on background variables for the three comparison groups (N = 243)
Comparison groups Experimental group (N = 86) Variables Sex Male Female Mean age (years) Civil status Married/cohabiting Task Patrol of cer Supervisory Detective Years of police experience National origin No information Netherlands Foreign country Sick leave in previous year (days) Number of previous police traumas Number of previous private traumas Previous psychotherapy from psychologist/psychiatrist Pre-test state anxiety (STAId)
a b c

91

Internal control group (N = 82) N 53 29 (%) (65) (35) (SD = 7.1) (71) (82) (7) (11) (SD = 7.9) (2) (89) (9) (SD = 24.7) (SD = 6.3)

External control group (N = 75) N 60 15 32.4 63 71 2 2 9.4 73 2 7.3 13.4 (%) (80) (20) (SD = 5.9) (84) (94) (3) (3) (SD = 5.7) n.s. n.s. <.01a,c <.05b,c <.01a,b,c n.s. <.01a,c <.01a,c n.s. n.s. n.s. p

N 60 26 28.9 57 81 4 1 5.1 1 79 6 3.1 10.3

(%) (70) (30)

(SD = 5.6) 31.7 (66) (94) (5) (1) (SD = 6.0) (1) (92) (7) (SD = 6.3) 58 67 6 9 8.7 2 73 7 7.2

(97) (3)

(SD = 1.57) n.s. (SD = 6.1) <.01a,b n.s. n.s. n.s.

(SD = 5.5) 14.4

0.08 (SD = 0.3) 5 31.4 (6)

0.08 (SD = 0.3) 8 (10) (SD = 7.4)

0.03 (SD = 0.2) 6 (8)

(SD = 9.9) 29.3

Signi cant difference between experimental group and internal control group. Signi cant difference between experimental group and external control group. Signi cant difference between internal and external control group. d The mean STAI score in a representative norm group is 38.9 (SD = 8.7).

OR = 5.3, 95% CI, 2.013.8). Failed resuscitation attempt, event 10, was signi cantly less common in the external control group than in the experimental group or the internal group (x 2 (1) = 7.3, p < .01, OR = 0.15, 95% CI, 0.030.71; x 2 (1) = 9.9, p < .01, OR = 0.12, 95% CI, 0.020.55). No signi cant differences emerged between groups with regard to speci c traumarelated variables such as duration of the incident, mortal danger, injury to self or colleague, death of colleague, weapon use, injury or death to a member of the public, seeing a dead or injured person, emotional exhaustion at the time of the incident, or a child as victim of the incident (a table of trauma-related variables is available from the rst author).

92

I. V. E. Carlier et al.
Comparison groups Experimental group (N = 86) Traumatic police incidents N 6 6 0 2 3 4 3 11 17 13 15 4 1 1 0 0 (%) (7) (7) (2) (4) (5) (4) (13) (19) (15) (17) (5) (1) (1) Internal control group (N = 82) N 1 7 2 1 2 4 1 13 8 15 16 6 2 1 3 0 (%) (1) (8) (2) (1) (2) (5) (1) (16) (10) (19) (20) (7) (3) (1) (4) External control group (N = 75) N 2 3 0 5 5 4 4 9 22 2 12 0 4 0 0 3 (%) (3) (4) (7) (7) (5) (5) (12) (29) (3) (16) (5) (4) n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. <.01a,c <.01b,c n.s. n.s. n.s. n.s. n.s. n.s. p

Table 2. Pre-test data on traumatic incidents reported by the three comparison groups (N = 243)

1. Colleague was killed or injured at work, or committed suicide 2. Seriously injured or dead children 3. Confrontation with aggressive mob 4. Being attacked 5. Having to use violence 6. Accident involving injury to self 7. Being threatened 8. Finding a corpse (natural death) 9. Bad accident with deaths or serious injuries 10. Failed resuscitation 11. Finding a corpse (suicide) 12. Suicide attempt with serious injury 13. Murder (adult) 14. Murder attempt with serious injury 15. Murder (child) 16. Assistance at large re
a

Signi cant difference between experimental group and internal control group. b Signi cant difference between experimental group and external control group. c Signi cant difference between internal and external control groups.

Effects of debrie ng on anxiety and post-traumatic stress Table 3 indicates the levels of state anxiety (STAI) and the post-traumatic stress symptoms recorded at the interim assessments 24 hours and 1 week post-trauma. It shows the numbers and percentages of respondents who were exhibiting each PTSD symptom. The STAI detected no signi cant differences between the experimental and internal control groups 24 hours post-trauma. One week post-trauma, the experimental group reported a signi cantly higher rate of re-experiencing than the control groups. Signicantly more respondents in the experimental group also reported the avoidance symptom loss of recall. At the post-test carried out 6 months post-trauma, the effect assessment for posttraumatic stress symptoms (not shown in table) yielded the following picture. In contrast to 1 week post-trauma, there were no longer any signi cant differences between the comparison groups. All symptoms had strongly receded with the passage of time. No one in the sample quali ed for the PTSD diagnosis at post-test. Effects of debrie ng on other trauma-related psychopathology and on sick leave and work resumption The ADIS-R interview administered 6 months post-trauma showed very little evidence of other types of psychopathology in the sample. Trauma-related psychopathology was

Post-traumatic stress in police of cers

93

Table 3. Effect assessments 24 hours post-trauma (STAI) and 1 week post-trauma (PTSD symptoms) (N = 243)
Comparison groups Experimental group (N = 86) Internal control group (N = 82) External control group (N = 75)

Symptomatology
STAI immediately after debrie ng (24 hours post-trauma) PTSD symptom clusters (1 week post-trauma) 1. Recurrent recollections 2. Dreams 3. Acting/feeling as if 4. Intense distress 5. Physiological reactivity Re-experiencing (one or more symptoms) Mean re-experiencing score b 6. 7. 8. 9. 10. 11. 12. Avoiding thoughts/ feelings Avoiding activities Loss of recall Diminished interest Detachment Restricted affect Foreshortened future
a

N
28.2

(%)

(%)
(SD = 7.8)

(%)

p
n.s.

(SD = 6.7) 29.2

21 9 17 17 3 38 4.3 3 0 10 2 0 0 0 0 0.18 2 2 2 0 0 1 0.27 0

(24) (11) (19) (20) (4) (44) (SD = 3.2) (4) (12) (2)

17 5 6 10 0 21 3.4 0 0 2 1 0 0 0 0

(21) (6) (7) (12)

(SD = 3.6)

n.s. n.s. <.05c n.s. n.s. <.01c <.01c,d n.s. <.05c n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.

(26) (SD = 2.9) 2.7

(2) (1)

Avoidance (three or more symptoms) Mean avoidance score 13. 14. 15. 16. 17.
b

(SD = 0.4) (2) (2) (2)

0.12 4 0 1 0 0 0 0.15 0

(SD = 0.4) 0.44 (SD = 1.3) (5) (1) (SD = 0.5)

Dif culty sleeping Irritability Dif culty concentrating Hypervigilance Startle responses

Hyperarousal (two or more symptoms) Mean hyperarousal score b PTSD diagnosis


a b

(1) (SD = 0.9)

PTSD scores are based on self-scoring. Mean scores are based on a combined frequency intensity score. For re-experiencing the scores ranged from 0 to 21, for avoidance from 0 to 24 and for hyperarousal from 0 to 18. c Signi cant difference between experimental group and internal control group. d Signi cant difference between experimental group and external control group.

94

I. V. E. Carlier et al.

diagnosed in 1% of the respondents in the experimental group and 3% of those in the control groups. It involved mainly major depressive episode and agoraphobia without history of panic disorder. The differences between the three comparison groups were not signi cant. The results on the outcome measures sick leave and work resumption showed that the largest proportion of the of cers in all groups resumed work immediately following the incident. No signi cant differences between groups could be established at any of the assessment points with regard to rates of sick leave or work resumption (a table of the outcome measures sick leave and work resumption is available from the rst author). Satisfaction with debrie ng Respondents who received debrie ng generally expressed great satisfaction with it: 98% were satis ed with the rst and second debrie ng sessions (24 hours and 1 month posttrauma), and 2% were satis ed to a degree; 88% were satis ed with the third session (3 months post-trauma) and 12% were satis ed to a degree. There was, however, no statistical correlation between the degree of satisfaction with debrie ng and the number of psychological symptoms reported (Pearsons r = .07, p > .05), nor was satisfaction associated with the numbers of sick days taken (t(72) = 0.02, p > .05) or with work resumption rates (x 2 (1) = 0.80, p > .05). In uencing factors We included in the analysis the following factors which could potentially in uence the effects of debrie ng: background characteristics of the respondents, type of traumatic event and other trauma-related variables, dissociation and peritraumatic dissociation, neuroticism and introversion, personality disorder, perceived support at work and occupational stress, perceived private support, past and present illnesses, previous traumas at work and in private life, and debrie ng-related variables (gender of debriefer, duration of debrie ng, amount of experience of debriefers, satisfaction). None of these variables was found to have in uenced the effects of debrie ng at any of the assessment points (outcome measures were symptomatology, sick leave and work resumption). Discussion How consistent are these ndings with those in other studies? First, our ndings on satisfaction are in agreement with evidence that participants generally appreciate the intervention (Carlier et al., 1998a; Jenkins, 1996; Robinson & Mitchell, 1993; Turner, Thompson, & Rosser, 1993). This seems to re ect a natural tendency in people who have recently been traumatized to seek emotional support, recognition, understanding and endorsement. Nevertheless, this high degree of satisfaction with debrie ng was in no way re ected in fewer stress symptoms, lower rates of sick leave or a more rapid resumption of work. This too, corresponds with the ndings of other studies (Bisson et al., 1997; Doctor, Curtis, & Isaacs, 1994; Lee et al., 1996). Secondly, we can compare our ndings on the ef cacy of debrie ng with those of the three randomized controlled trials that have been published on this issue. Our

Post-traumatic stress in police of cers

95

observation that debrie ng, contrary to expectations, did not lead to a reduction in stress-related symptomatology is in line with ndings by Lee et al. (1996), who likewise found no signi cant differences between debriefed and non-debriefed subjects. Our longer-term ndings are at variance with those by Hobbs et al. (1996) and Bisson et al. (1997), who found signi cantly greater symptomatology after debrie ng. In terms of research design, the present study can best be compared to that by Matthews (1998), who also used two comparison groups. Alongside a group of debriefed direct care workers, Matthews placed one group of workers who had access to debrie ng but chose not to attend, and a second group who worked in an area where the intervention was not available. One week post-trauma, he found the rst nondebriefed group to have consistently lower ratings on individual stress symptom measures than the debriefed group, fully in line with our observations at this shortterm assessment point. As Matthews points out, it could be argued that this nding con rms the validity of allowing workers to choose whether to take part in debrie ng. Alternatively, Matthews draws attention to the danger that involvement in debrie ng may intensify the workers stress response or their perception of stress. One reason for this might be that the psychoeducational component of debrie ng and the scrutiny given to individual reactions to the event serve to heighten the participants awareness of their own reactions, thus eliciting higher levels of stress (Matthews, 1998). In our view, however, such an unexpected nding 1 week after the traumatic event cannot reasonably be regarded as a negative effect of debrie ng, as no PTSD diagnosis can be made at such an early stage. Most studies carried out so far have focused on one-session critical incident stress debrie ng. The debrie ng procedure we studied here consisted of three sessions. In their original form, debrie ng techniques were not designed for one-off or stand-alone interventions, but as part of the CISM procedure (Robinson & Mitchell, 1993; Turnbull et al., 1997). A randomized control trial by Brom, Kleber, and Hofman (1993) is relevant in this connection, because it, too, evaluated a procedure that conformed to the threestage debrie ng model applied in our study. The Brom assessments were carried out 1 month and 6 months post-trauma, and the ndings were consistent with those in our study (see also Lilford, Stratton, Godsil, & Prasad, 1994). Even though our sample was much larger than those in most other debrie ng studies hitherto, the rates of post-traumatic stress symptomatology were still very low. This could perhaps be attributed to the fact that debrie ng is offered by Dutch police forces after every critical police incident, regardless of whether it has triggered intense emotions in the of cer in question. For PTSD to actually develop, however, a person must not only have been exposed to the distressing event, but must also have responded to it with intense fear, helplessness, or horror and have suffered longer-term distress or impairment in social, occupational, or other important areas of functioning (APA, 1994). This may also explain the brief duration of many of the debrie ngs, especially at the second and third sessions. The briefness of a session is not necessarily a bad sign. The study by Bisson et al. (1997) showed that longer individual debrie ngs were more likely to be associated with poor outcomes. Also Bunn and Clarke (1979) evaluated the use of a 20-minute supportive interview comparable to debrie ng. The subjects of the intervention group showed less anxiety immediately after the interview than those in the non-intervention group (Bunn & Clarke, 1979).

96

I. V. E. Carlier et al.

Some caveats should be mentioned with respect to this study. The subjects were not assigned to groups randomly. Although the groups appeared to be readily comparable on the principal variables, other variables not tested here could have accounted for the group differences at 1 week post-trauma. In addition, our follow-up data at 6 months post-trauma were collected retrospectively, which may have produced some recall bias. As we focused on individual debrie ng, the results may not be directly comparable with data on group or couple debrie ng. In the future, studies of the ef cacy of group, as opposed to individual, debrie ng are needed. Also, random allocation research is indicated which can examine various types of debrie ng models and crisis intervention programmes. Acknowledgement
This study was supported by grants from the Dutch police forces.

References
American Psychiatric Association (1994). Diagnostic and statistic manual of mental disorders (4th ed.). Washington DC: American Psychiatric Association. Bernstein, E. M., & Putnam, F. W. (1986). Development, reliability, and validity of a dissociation scale. The Journal of Nervous and Mental Disease, 174, 727735. Bisson, J. I., Jenkins, P. L., Alexander, J., & Bannister, C. (1997). Randomised controlled trial of psychological debrie ng for victims of acute burn trauma. British Journal of Psychiatry, 171, 7881. Blacklock, E. (1997). Critical incident stress debrie ng. The Queensland Nurse, 16, 23. Blake, D. D., Weathers, F. W., Nagy, L. M., Kaloupek, D. G., Gusman, F. G., Charney, D. S., & Keane, T. M. (1995). The development of a clinical-administered PTSD scale. Journal of Traumatic Stress, 8, 7590. Bolwig, T. G. (1998). Debrie ng after psychological trauma. Acta Psychiatrica Scandinavica, 98, 169170. Brom, D., Kleber, R. J., & Hofman, M. C. (1993). Victims of traf c accidents: Incidence and prevention of post-traumatic stress disorder. Journal of Clinical Psychology, 49, 131140. Bunn, T. A., & Clarke, A. M. (1979). Crisis intervention: An experimental study of the effects of a brief period of counselling on the anxiety of relatives of seriously injured or ill hospital patients. British Journal of Medical Psychology, 52, 191195. Busuttil, A., & Busuttil, W. (1995). Psychological debrie ng. British Journal of Psychiatry, 166, 676681. Caplan, R. D., Cobb, S., French, J. P. R., van Harrison, R., & Pinneau, S. R. (1980). Job demands and worker health: Main effects and occupational differences. University of Michigan, Survey Research Center. Carlier, I. V. E., Fouwels, A. J., Lamberts, R. D., & Gersons, B. P. R. (1996). Post-traumatic stress disorder and dissociation in traumatized police of cers. The American Journal of Psychiatry, 8, 624629. Carlier, I. V. E., & Gersons, B. P. R. (1992). Development of a scale for traumatic stress incidents in police work. Psychiatrica Fennicac (Supplementum), 23, 5970. Carlier, I. V. E., & Gersons, B. P. R. (1994). Trauma at work: Posttraumatic stress disorder; an occupational health hazard. The Journal of Occupational Health and SafetyAustralia and New Zealand, 19, 254266. Carlier, I. V. E., & Gersons, B. P. R. (1995). Partial posttraumatic stress disorder (PTSD): The issue of psychological scars and the occurrence of PTSD symptoms. The Journal of Nervous and Mental Disease, 183, 107109. Carlier, I. V. E., Lamberts, R. D., & Gersons, B. P. R. (1997). Risk factors for posttraumatic stress symptomatology in police of cers: A prospective analyses. The Journal of Nervous and Mental Disease, 185, 498506. Carlier, I. V. E., van Uchelen, J. J., Lamberts, R. D., & Gersons, B. P. R. (1998a). Disaster-related posttraumatic stress in police of cers: A eld study of the impact of debrie ng. Stress Medicine, 14, 143 148.

Post-traumatic stress in police of cers

97

Carlier, I. V. E., van Uchelen, J. J., Lamberts, R. D., & Gersons, B. P. R. (1998b). Clinical utility of a brief diagnostic test for PTSD. Psychosomatic Medicine, 60, 4247. Cohen, S., & Hoberman, H. M. (1983). Positive events and social supports as buffers of life change stress. Journal of Applied Social Psychology, 13, 99125. Davidson, J. T. R., Book, S. W., Colket, J. T. et al. (1997). Assessment of a new self-rating scale for posttraumatic stress disorder. Psychological Medicine, 27, 153160. Davidson, J. T. R., Smith, R., & Kudler, H. S. (1989). Validity and reliability of the DSMIII criteria for post-traumatic stress disorder: Experience with a structured interview. The Journal of Nervous and Mental Disease, 177, 336341. Deahl, M. P., 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 debrie ng. British Journal of Psychiatry, 165, 6065. DiNardo, P. A., & Barlow, D. H. (1988). Anxiety Disorders Interview ScaleRevised. Albany, NY: Center for Phobia and Anxiety Disorders. Doctor, R. S., Curtis, D., & Isaacs, G. (1994). Psychiatric morbidity in policemen and the effect of brief psychotherapeutic intervention: A pilot study. Stress Medicine, 10, 151157. Draaijer, N., & Boon, S. (1993). The validation of the Dissociative Experiences Scale against the criterion of the SCID-D, using Receiver Operating Characteristics (ROC) Analysis. Dissociation, 6, 2837. Dyregrov, A. (1989). Caring for helpers in disaster situations: Psychological debrie ng. Disaster Management, 2, 123128. Eysenck, H. J. (1957). The dynamics of anxiety and hysteria. London: Routledge & Keegan Paul. Gersons, B. P. R., & Carlier, I. V. E. (1994). Treatment of work-related trauma in police of cers: Posttraumatic stress disorder and post-traumatic decline. In M. B. Williams & J. F. Sommer Jr. (Eds.), Handbook of post-traumatic therapy: A practical guide to intervention, treatment and research (pp. 325333). Westport, CT: Greenwood Press. Green, B. L. (1993). Identifying survivors at risk: Trauma and stressors across events. In J. P. Wilson & B. Raphael (Eds.), International handbook of traumatic stress syndromes (pp. 135144). New York: Plenum Press. Grif th, J., & Watts, R. (1994). The Kensey and Grafton bus crashes: The aftermath. East Lismore: Instructional Design Solutions. Hobbs, M., Mayou, R., Harrison, B., & Worlock, P. (1996). A randomised controlled trial of psychological debrie ng for victims of road traf c accidents. British Medical Journal, 313, 14381439. Horowitz, M. J., Wilner, N., & Alvarez, W. (1979). Impact of Event Scale: A measure of subjective stress. Psychosomatic Medicine, 41, 209218. Hyler, S. E., Rieder, R. O., Williams, J. B. W., Spitzer, R. L., Hendler, J., & Lyons, M. (1988). The personality diagnostic questionnaire:Development and preliminary results. Journal of Personality Disorders, 2, 229237. Hytten, K., & Hasle, A. (1989). Fire ghters: A study of stress and coping. Acta Psychiatrica Scandinavica, 80, 5055. Jenkins, S. R. (1996). Social support and debrie ng ef cacy among emergency medical workers after a mass shooting incident. Journal of Social Behavior and Personality, 11, 477492. Jensen, A. (1958). The Maudsley Personality Inventory (MPI). Acta Psychologica, 14, 314. Kenardy, J. A., Webster, R. A., Lewin, T. J., Carr, V. J., Hazell, P. L., & Carter, G. L. (1996). Stress debrie ng and patterns of recovery following a natural disaster. Journal of Traumatic Stress, 9, 3749. Lee, C., Slade, P., & Lygo, V. (1996). The in uence of psychological debrie ng on emotional adaptation in women following early miscarriage: A preliminary study. British Journal of Medical Psychology, 69, 4758. Lilford, R. J., Stratton, P., Godsil, S., & Prasad, A. (1994). A randomised trial of routine versus selective counseling in prenatal bereavement from congenital disease. British Journal of Obstetric Gynaecology, 101, 291296. Marmar, C. R., Weiss, D. S., Schlengen, W. E., Fairbank, J. A., Jordan, K., Kulka, R. A., & Hough, R. L. (1994). Peritraumatic dissociation and post-traumatic stress in male Vietnam theater veterans. American Journal of Psychiatry, 151, 902907. Matthews, L. R. (1998). Effect of staff debrie ng on posttraumatic stress symptoms after assaults by community housing residents. Psychiatric Services, 49, 207212.

98

I. V. E. Carlier et al.

McFarlane, A. C. (1988). The aetiology of post-traumatic stress disorders following a natural disaster. British Journal of Psychiatry, 151, 116121. Mitchell, J. T. (1983). When disaster strikes. Journal of Emergency Medical Services, 8, 3639. Mitchell, J. T., & Everly, G. S. (1996). Critical Incident Stress Debrie ng (CISD); An operations manual for the prevention of traumatic stress among emergency services and disaster workers (2nd ed., Rev. ed.). Ellicott City, MD: Chevron Publishing. Mollica, R. F., Caspi-Yavin, Y., Bollini, P., Truong, T., Tor, S., & Lavelle, J. (1992). The Harvard Trauma Questionnaire: Validating a cross-cultural instrument for measuring torture, trauma and posttraumatic stress disorder in Indochinese refugees. The Journal of Nervous and Mental Disease, 180, 111116. Robinson, R. C., & Mitchell, J. Y. (1993). Evaluation of psychological debrie ngs. Journal of Traumatic Stress, 6, 367382. Rose, S., & Bisson, J. (1998). Brief early psychological interventions following trauma: A systematic review of the literature. Journal of Traumatic Stress, 11, 697710. Spielberger,C. D., Gorsuch, R. L., & Lushene, R. E. (1970). STAI Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Turnbull, G., Busuttil, W., & Pittman, S. (1997). Psychological debrie ng for victims of acute burn trauma. British Journal of Psychiatry, 171, 582588. Turner, S. W., Thompson, J., & Rosser, R. M. (1993). The Kings Cross re: Early psychologicalreactions and implications for organizing a phase-two response. In J. P. Wilson & B. Raphael (Eds.), International handbook of traumatic stress syndromes (pp. 451459). New York: Plenum Press. Wagner, M. (1986). Trauma debrie ng in the Chicago Police Department. In J. Reese & H. Goldstein (Eds.), Psychological services for law enforcement (pp. 399403). Washington, DC: Federal Bureau of Investigation. Watts, R. (1994). The ef cacy of critical incident stress debrie ng for personnel. Bulletin of the American Psychological Society, 16, 67. Wessely, S., Rose, S., & Bisson, J. (1997). A systematic review of brief psychological interventions (debrie ng) for the treatment of immediate trauma-related symptoms and the prevention of posttraumatic stress disorder (protocol). The Cochrane Library, (CD-ROM). Oxford: Update Software Inc. Received 16 February 1999