SUMMARYThe GHQ-30 was used to assess psychological disturbance among staff working in four long-stay wards, four homes for the elderly mentally infirm, four elderly persons' homes and four private nursing homes. Rates of disturbance were consistent across all 16 units. Rates of CGHQ caseness were low, compared with published findings from a representative UK community sample. A strong relationship was demonstrated between staff psychological disturbance and aggression from residents over the last week. Disturbed staff were more likely to perceive a lack of support at work, and to report 'shouting back' at aggressive residents. No relationship was found between rates of disturbance and a range of demographic and work-related factors. Implications for service provision and future research are discussed.KEY woms-Care staff, residential homes, stress, psychological distress, aggression.There is increasing concern about the quality of residential care afforded to elderly people in Britain (International Journal of Geriatric Psychiatry, 1986). Staff working in these settings are often poorly trained or untrained, and there are particular concerns at this time, as the number of long-stay psychogeriatric beds is diminishing. Thus psychiatric support for the most demanding cases is under threat.In recent years stress among informal carers for dementia sufferers has been extensively studied, Morris e t al. (1988) providing a comprehensive review. Studies show high stress levels in carers, and a complex relationship between stress and other factors such as severity of dementia, level of problem behaviour, coping mechanisms used, impact of support services and satisfaction derived from the caring role.It follows that professionals caring for the elderly in institutions where a high proportion have dementia (Mann e t al., 1984) may also suffer from stress, and that staff in long-stay wards, who cope with the most demanding and behaviourally disturbed residents, may have the highest stress levels. There is evidence (Hilton et al., 1989; Lam el al., 1989) of greater dependence, disability and disturbance among long-stay patients than residents in local authority homes. However, there is a paucity of research in this area, particularly regarding large-scale, generalizable surveys (Sutherland and Cooper, 1990). Studies by Norfolk and Stirton (1985) and Benjamin and Spector (1990) demonstrated high levels of stress symptoms in psychogeriatric nursing staff, but were flawed by preselection bias and lacked a clear definition of stress 'caseness'. Livingstone and Livingstone (1984) used the GHQ-60 in their study of nurses in three selected hospital settings, finding consistent results across the three institutions. No attempt was made in these studies to relate findings to general population samples.Various authors have attempted to analyse the CCC 0885-6230/94/05038