Evaluation des conséquences psychologiques des catastrophes environnementales : étude de faisabilité basée sur les innondations survenues dans le Vaucluse en 1992
Verger P, Rotily M, Baruffol E, Boulanger N, Vial M, Sydor G, Pirard P, Bard D
Sante 1999 Sep-Oct;9(5):313-8
Scientific studies of environmental disasters, whether human or natural in origin, have shown that the psychological impact of such events may be considerable and long-lasting. Several natural disasters have occurred in France, but their impact on public psychological health has not been assessed. In September 1992, there was a major flood in southeast France (Vaucluse), which caused 38 deaths. Four years later, we performed a pilot cross-sectional study to assess the feasibility of a larger epidemiological study to assess the psychological impact of this flood. Two affected towns were chosen for this study: Vaison-la-Romaine (VLR), where the flood was very sudden and 29 people were killed, and Bedarrides, where the water level rose more slowly. In Bedarrides, households were randomly selected from a list of victims (n = 100) and in Vaison-la-Romaine, households were randomly selected from the telephone directory (n = 140). Exposure to the flood was assessed by a series of questions, the answers to which were used to calculate an exposure score. The questionnaire also included psychometric scales for post-traumatic stress disorder (QE-PTSD), anxiety (Spielberger State-Trait Anxiety Inventory) and depression (Beck Depression Inventory). In Bedarrides, the participation rate was high: 69% of the selected households were successfully contacted and found to be eligible for inclusion and 74% agreed to a face-to-face interview. In Vaison-la-Romaine, 51% of the selected households were contacted and eligible and 50% agreed to the face-to-face interview. Our survey showed that exposure was multidimensional and that it was possible to calculate an exposure score suitable for the analysis of exposure-effect relationships. The PTSD scale was completed well by the interviewees, several of whom did not fill in the depression and anxiety scales correctly. However, analysis of the responses obtained showed that these tools had a high level of internal consistency. Cross-correlations between the various psychological scales used in this study were highly significant (p < 10-4). There was some degree of association between some psychometric scales and exposure. This pilot study shows that a cross-sectional study of the long-term psychological consequences of an environmental disaster could be carried out several years after the event but that the feasibility of such a study depends ultimately on its acceptance by the public and the relevant authorities. It underlines the need to collect exposure data immediately after the event and enabled us to identify and to adapt the tools required for this kind of evaluation. It should encourage public health decision-makers to support such evaluation and to improve the psychological and social support available to people exposed to floods.