Results
Description of the Sample
Of the 224 women who returned the survey packet, 223 (99.6%) completed the French IES-R. Table 3 presents the demographics of the study sample and the mean scores on the IES-R. Spotty missing values (0.3%) for the completed IES-R were replaced by the group mean for the item.
Internal Consistency
Pearson correlations between the subscales and the total score were high
and were significant at the 0.01 level (Table 4). Cronbach’s a coefficients were also high, suggesting that the French IES-R has good internal consistency:
intrusion subscale a = 0.86, avoidance subscale a = 0.86, hyperarousal
subscale a = 0.81, and total score a = 0.93.
Convergent Validity
To evaluate the convergent validity of the French IES-R, Pearson correlations were calculated between the IES-R and the threat subscale of the Ice Storm Questionnaire (intrusion r = 0.29, avoidance r = 0.22, hyperarousal r = 0.23, total r = 0.29) and between the IES-R and the total score on the GHQ (intrusion r = 0.28, avoidance r = 0.27, hyperarousal r = 0.44, total r = 0.37). Although modest, all correlations were significant at the 0.01 level.
Factorial Structure of the IES-R
To assess the construct validity of the French IES-R, a principal components analysis (PCA) was performed on the 22 items of the scale. While 5 components had eigenvalues > 1.0, Catell’s scree test was used to determine the number of components to be extracted. The scree plot comprised a large first factor (eigenvalue = 8.6), followed by 2 other factors (eigenvalues = 2.1 and 1.5) and 19 smaller factors, which levelled off. The scree test was ambiguous: it allowed for 2 or 3 components to be extracted. Thus, in accordance with the theoretical structure of the IES-R proposed by Weiss and Marmar, it was decided to extract 3 factors (5). A PCA using an orthogonal (varimax) rotation was then performed with the 3-factor forced solution. Factor loadings ³ 0.50 were considered significant. The solution, which explained 56% of the variance, generated a hyperarousal factor (items 2, 4, 10, 14, 15, 18, 21), an avoidance factor (items 8, 11, 12, 13, 17, 22), and an intrusion factor (items 1, 3, 5 ,6, 9, 16) (Table 1). Items 7, 19, and 20 did not load on any of the 3 factors. A PCA was subsequently performed without these items, increasing the variance explained to 62%. Removing these items from the analysis did not alter the results. The 2-factor solution included an avoidance factor and an intrusion-arousal factor.
Discussion
This study assessed the internal consistency, convergent validity, and construct validity of a French translation of the IES-R in a sample of future mothers exposed to the 1998 ice storm.
First, we showed that the French translation of the IES-R possesses good internal consistency. Our alpha coefficients, ranging from 0.81 to 0.93, compare with the findings of Weiss and Marmar (5) for the original English version of the IES-R and with Maercker and Schützwohl (6) for the German version of the scale.
Although test–retest data were not available for this sample, we did obtain data from another francophone sample. The sample consisted of 39 individuals (60.5% women; mean age 35.4 years, 55.6% married, 30.6% single) living in the area surrounding Paris. All participants experienced a traumatic event in 2001 and were at their first or second visit to a psychologist or psychiatrist when they were asked to participate in the study. After giving their informed consent, participants completed the French version of the IES-R at the time of the initial assessment and again at 3 months. Satisfactory test–retest reliability was obtained in this sample: intrusion r = 0.73, avoidance r = 0.77, hyperarousal r = 0.71, total scale r = 0.76, with P < 0.01 for all correlations.
Unexpectedly, the correlations between the French IES-R, threat perception, and the GHQ were fairly low. Perhaps the low correlations between the threat perception subscale and the IES-R owe to the low variance in the women’s responses on the threat subscale. Further, the responses on the GHQ-28 may have also affected the correlations: according to the women’s responses, about 30% met criteria for depression, but a clinical psychologist found that only 2 were actually suffering from depression. Many of the somatic indicators of psychological symptoms addressed in the GHQ could be associated with simply being pregnant or with caring for an infant (for example, sleep problems).
Exploring the Factorial Structure of the IES-R
Our results suggested that a 2- or 3-factor solution was acceptable for the IES-R. We chose to report a 3-factor solution, explaining 56% of the variance, because it is consistent with the proposed theoretical structure of the scale. Other studies on the psychometric properties of the scale have reported different factor structures. For instance, Weiss and Marmar (5) found a 1-factor solution that explained 49% of the variance. Conversely, Maercker and Schützwohl (6) reported a 4-factor solution (intrusion, avoidance, hyperarousal, and avoidance and numbing), which explained 63.8% of the variance. To sum up, it remains unclear at this stage whether the IES-R is composed of 1, 2, 3, or 4 factors. This ambiguity is also reflected in the factor structure of the DSM-IV criteria for PTSD, which have yielded 2-, 3-, and 4-factor solutions (13). More studies will be required before a satisfactory answer to this important theoretical issue is found .
|
Table 3 Description of the sample (n = 223)
|
|
|
% (n)
|
Mean
|
SD
|
Range
|
|
Marital status
|
|
|
|
|
|
Never married
|
3.6 (8)
|
|
|
|
|
Common-law
|
55.6 (124)
|
|
|
|
|
Married
|
40.4 (90)
|
|
|
|
|
Divorced
|
0.4 (1)
|
|
|
|
|
Hollingshead classes
|
|
|
|
|
|
Lower class
|
2.7 (6)
|
|
|
|
|
Lower middle class
|
3.6 (8)
|
|
|
|
|
Middle class
|
30.9 (68)
|
|
|
|
|
Upper middle class
|
48.6 (107)
|
|
|
|
|
Upper class
|
14.1 (31)
|
|
|
|
|
Age
|
|
29.1
|
4.7
|
18–41
|
|
Days without power
|
|
14.9
|
8.9
|
0–45
|
|
Days without telephone
|
|
4.4
|
8.4
|
0–34
|
|
Threat perception (ice storm questionnaire)
|
|
1.5
|
1.4
|
0–7
|
|
General Health Questionnaire
|
|
6.6
|
6.0
|
0–26
|
|
The Impact of Event Scale-Revised
|
|
|
|
|
|
Intrusion subscale
|
|
5.7
|
5.4
|
0–24
|
|
Avoidance subscale
|
|
3.9
|
5.2
|
0–26
|
|
Hyperarousal subscale
|
|
2.3
|
3.6
|
0–18
|
|
Total score
|
|
11.8
|
12.4
|
0–58
|
Item 7 (“I felt as if it hadn’t happened or wasn’t real”), item 19 (“Reminders of it cause me to have physical reactions, such as sweating, trouble breathing, nausea, or a pounding heart”), and item 20 (“I had dreams about it”) did not load on any of the 3 factors. Compared with the theoretical 8 intrusion, 8 avoidance, and 6 hyperarousal items, the PCA performed in this study resulted in 6 intrusion, 6 avoidance, and 7 hyperarousal symptoms. This could be owing to the fact that item 7 measures a dissociative-like symptom, and items 19 and 20 were not highly endorsed by the women in this study.
Finally, items 2, 5, and 14 loaded on factors that differed from those in the theoretical model. Item 2 (“I had trouble staying asleep”) should have loaded on the intrusion factor but instead loaded on the hyperarousal factor. This item has also been problematic for Weiss and Marmar (5), who found that it correlated equally on the intrusion and hyperarousal subscales. This led the German translation to include item 2 in the hyperarousal subscale. Item 5 should have loaded on the avoidance factor but instead loaded on the intrusion factor. This may be because the first part of the statement, “I avoided letting myself get upset” taps into avoidance, whereas the second part, “when I thought about it or was reminded of it” taps into intrusion symptoms. Item 14 (“I found myself acting or feeling like I was back at that time”) should have loaded on the intrusion factor but actually loaded on the hyperarousal factor. It is possible that the word “feeling” was interpreted as a manifestation of hyperarousal symptoms.
|
Table 4 Correlations between the Impact of Event Scale-Revised subscales
and total score
|
|
|
Intrusion
|
Avoidance
|
Hyperarousal
|
|
Intrusion
|
—
|
—
|
—
|
|
Avoidance
|
0.62
|
—
|
—
|
|
Hyperarousal
|
0.69
|
0.56
|
—
|
|
Total score
|
0.90
|
0.86
|
0.83
|
|
All correlations are significant at the 0.01 level (2-tailed)
|
In conclusion, the French IES-R has proven to be a reliable measure for posttraumatic stress symptoms in our sample of future mothers exposed to the 1998 ice storm. Given the popularity of the IES, the translation and validation of the IES-R into French fills an important gap in our ability to assess PTSD symptoms among French-speaking populations.
Funding and Support
This work was supported by a grant from the McGill University Stairs Memorial Fund to Dr King and research fellowships from the Fonds de la recherche en santé du Québec (FRSQ) to Dr King and to Dr Brunet, as well as by a graduate student fellowship from the FRSQ—Fonds pour la formation de chercheurs et l’aide à la recherche (FRSQ-FCAR) to Annie St- Hilaire.
References
1. Green BL, Solomon SD. The mental health impact of natural and technological disasters. In: Freesy JR, Hobfoll SE, editors. Traumatic stress: from theory to practice. New York: Plenum Press; 1995. p 163–80.
2. Horowitz M, Wilner N, Alvarez W. Impact of Event Scale: a measure of subjective stress. Psychosom Med 1979;41:209–18.
3. Joseph S. Psychometric evaluation of Horowitz’s Impact of Event Scale: a review. J Trauma Stress 2000;13:101–13.
4. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington (DC): American Psychiatric Association; 1994.
5. Weiss DS, Marmar CR. The Impact of Event Scale - Revised. In: Wilson JP, Keane TM, editors. Assessing psychological trauma and PTSD. New York: Guilford Press; 1997. p 399–411.
6. Maercker A, Schützwohl M. Erfassung von psychischen Belastungsfolgen: Die Revision der Impact of Event-Skala (assessing effects of traumatic events: the Impact of Event Scale-Revised). Diagnostica 1998;44:130–41.
7. King S, Barr RG, Brunet A, Saucier J-F, Meaney M, Woo S, and others. La tempête de verglas: une occasion d’étudier les effets du stress prénatal chez l’enfant et la mère (The ice storm: an opportunity to study the effects of prenatal stress on the baby and the mother). Santé mentale au Québec 2000; XXV:163–84.
8. Green BL, Goodman LA, Krupnick JL, Corcoran CB, Petty RM, Stockton P, and others. Outcomes of single versus multiple trauma exposure in a screening sample. J Trauma Stress 2000;13:271–86.
9. Goldberg DP, Hillier VF. A scaled version of the General Health Questionnaire. Psychol Med 1979;9:139–45.
10. Banks MH. Validation of the General Health Questionnaire in a young community sample. Psychol Med 1983;13:349–53.
11. Shevlin M, Hunt N, Robbins I. A confirmatory factor analysis of the Impact of Event Scale using a sample of World War II and Korean veterans. Psychol Assess 2000;12:414–7.
12. Spurrell MT, McFarlane AC. Life-events and psychiatric symptoms in a general psychiatry clinic: the role of intrusion and avoidance. Br J Med Psychol 1995;68:333–40.
13. Asmundson GJ, Frombach I, McQuaid J, Pedrelli P, Lenox R, Stein MB. Dimensionality of posttraumatic stress symptoms: a confirmatory factor analysis of DSM-IV symptom clusters and other symptom models. Behav Res Ther 2000;38:203–14.
Author(s)
Manuscript received March 2002 and accepted April 2002.
1. Assistant Professor, Psychiatry, McGill University and Douglas Hospital Research Centre, Montreal, Quebec.
2. Student, McGill University and Douglas Hospital Research Centre, Montreal, Quebec.
3. Unité de psychiatrie et psychotraumatisme, Hôpital Tenon, Paris.
4. Professor, Psychiatry, McGill University and Douglas Hospital Research Centre, Montreal, Quebec.
Address for correspondence: Dr A Brunet, Douglas Hospital Research Centre, 6875 LaSalle Blvd, Verdun, QC H4H 1R3
e-mail: alain.brunet@douglas.mcgill.ca
1 | 2 | 3
|
