Acute health eVects of the Sea Empress oil spill

Study objective—To investigate whether residents in the vicinity of the Sea Empress tanker spill suVered an increase in self reported physical and psychological symptoms, which might be attributable to exposure to crude oil. Design—Retrospective cohort study; postal questionnaire including demographic details, a symptom checklist, beliefs about health eVects of oil and the Hospital Anxiety and Depression and SF-36 mental health scales. Setting—Populations living in four coastal towns on the exposed south Pembrokeshire coast and two control towns on the unexposed north coast. Patients—539 exposed and 550 unexposed people sampled at random from the family health services authority age-sex register who completed questionnaires. Main results—Adjusted odds ratios for self reported physical symptoms; scores on the Hospital Anxiety and Depression and SF-36 mental health scales, in 1089 people who responded out of a possible 1585 (69%). Conclusions—Living in areas exposed to the crude oil spillage was significantly associated with higher anxiety and depression scores, worse mental health; and self reported headache (odds ratio = 2.35, 95% CI 1.56, 3.55), sore eyes (odds ratio = 1.96, 95% CI 1.06, 3.62), and sore throat (odds ratio = 1.70, 95% CI 1.12, 2.60) after adjusting for age, sex, smoking status, anxiety, and the belief that oil had affected health. People living in exposed areas reported higher rates of physical and psychological symptoms than control areas. Symptoms significantly associated with exposure after adjustment for anxiety and health beliefs were those expected from the known toxicological eVect of oil, suggesting a direct health eVect on the exposed population. (J Epidemiol Community Health 1999;53:306–310) On the evening of the 15 February 1996, the Sea Empress laden with more than 130 000 tonnes of light crude oil, ran aground on rocks at the entrance of Milford Haven harbour in south west Wales. Over the next week an estimated 72 000 tonnes of crude oil and 360 tonnes of heavy fuel oil were released into the sea, contaminating around 200 km of coastline (fig 1) and causing strong smells and complaints of symptoms from residents of the coastal towns. Similar oil spills from the tankers Exxon Valdez and Braer were followed by increases in physical and psychological symptoms in the general population. We undertook a population based retrospective cohort study to assess the acute physical and psychological health impact on the exposed population. Methods This study was commissioned four weeks after the incident. The exposed group was defined as people living on the most exposed south coast of Pembrokeshire (Milford Haven, Pembroke Dock, Tenby, and Saundersfoot) and the control group as similar but unexposed populations living on the north coast (Aberaeron and Fishguard) (fig 1). Sample size calculations were based on the prevalence of symptoms in the control group reported in the Braer study. A sample of 814 in both exposed and control areas was required for 80 per cent power to detect a statistically significant (p<0.05) doubling of prevalence of symptoms in the exposed group against a background prevalence of 3 per cent. To allow for non-response, random samples of 1000 adults (of both sexes aged 18 to 65 years old) were selected from each of the exposed and control areas using the Dyfed-Powys family health services authority age-sex register. A questionnaire, with a retrospective health diary covering the four weeks immediately after the incident, was devised that also incorporated a symptom check list used in the Braer investigation, the Hospital Anxiety and Depression Scale (HAD), the mental health profile of the SF36 as well as a series of questions relating to beliefs about the eVects of the oil spill on the environment, employment, health, and finances. The HAD Scale is widely used in clinical settings and was included in the study because of the marked excess of cases of clinically relevant anxiety and depression after the Exxon Valdez oil spill. The SF-36 mental health scale was included as population normative baseline data of SF-36 health scores was available from the Pembrokeshire health status study undertaken in 1994 in randomly sampled adults from North (n=659) and South Pembrokeshire (n=597). Questionnaires were mailed to people on the 4 April (seven weeks after the grounding of the tanker). Non-responders were sent a second questionnaire one month later. Nonresponders to the second questionnaire were contacted by telephone or visited at home by two of us (MT and DE) between 21 May and 18 June. Statistical analysis was carried out using SPSS for Windows and Epi Info Logistic. Crude odds ratios for each self reported J Epidemiol Community Health 1999;53:306–310 306 Iechyd Morgannwg Health R A Lyons Gwent Health Authority J M F Temple D L Fone University of Wales Swansea D Evans Welsh Combined Centres for Public Health S R Palmer