Background:Few studies have investigated the impact of caregiving on a national scale, or ethnic differences in caregiving processes. This is despite growing numbers of informal carers and demographic change in the UK. Individuals identifying as black African, black Caribbean, black British and black ‘other’ (BA_BAC) by UK Census ethnicity criteria (2001) are currently over-represented within psychiatric services in urban settings and are known to perceive statutory support to be inappropriate for their needs. Aim:Initially, to examine the mental and physical health consequences of caregiving in general. Secondly, the aim was to compare ethnic differences in the processes identified within cognitive models of caregiving in psychosis. The third aim was to assess the validity of a new measure of service users’ perceptions of the quality of caregiving relationships. Method:Primary outcome measures obtained from the Adult Psychiatric Morbidity Survey (APMS) 2007 were used to examine the physical and mental health impact of caregiving roles compared to no caregiving in a representative sample of households in England. In the second study, BA_BAC carers and carers who identified as white British or white ‘other’ (WB) were recruited through Early Intervention in psychosis services in London. The Camberwell Family Interview was used to assess Expressed Emotion and computerised self-report measurements were obtained of distress, appraisals of caregiving, explanatory models of illness and coping. Within and between groups comparisons were conducted cross-sectionally based on self-identified ethnicity. In the third study, a systematic review was conducted to identify items to measure perceived Expressed Emotion. An aggregated scale was completed by first-episode psychosis service-users who agreed to accompany their carers in participating in the studies, along with a clinical interview assessing symptoms. Results:25% (N=1883) of the APMS 2007 sample identified themselves as carers in study 1. They had poorer mental health, and higher psychiatric symptom scores than non-caregivers. A significant dose effect was found, with an observable decline in mental health above 10 hours per week. A twofold increase in psychiatric symptom scores in the clinical range was recorded in those providing care for more than 20 hours per week. In adjusted analyses, there was no excess of physical disorders in carers. In the first-episode psychosis sample (study 2), over one third of all carers (N=78) exceeded clinical thresholds of depressive symptoms. Sixty-five percent reported sleep problems and poor sleep was found to mediate the relationship between negative appraisals of caregiving and distress (p<.0001). BA_BAC carers (N=41) were less likely to report intrusive, overprotective behaviours in caregiving and they maintained higher levels of warmth in discussing their relative. Significant group differences were recorded in carers’ beliefs about the causes of illness and their coping strategies. In study 3, a measure of service-user perceived Expressed Emotion showed moderate to good reliability and validity, taking into account illness severity. Better performance was observed for the measure with BA_BAC service-user participants (N=12/44). Conclusions:Even at low levels of caregiving, (10 hours a week) there is observable impact on the mental health of carers. This is found from the initial stages of caring for someone with a mental health condition. Distress frequently reaches clinical thresholds, particularly for individuals providing the most care, and those trying to cope with complex needs. Strategies aimed at maintaining carers’ mental health must consider this adverse and often progressive impact. Ethnic group differences in responses to caregiving may affect reported well being in carers and their interactions with service users. This is relevant to the development of culturally-informed Family Interventions in psychosis and interventions aimed specifically at carers. Measurement instruments targeting service-user perceptions of caregiving relationships may have clinical applications, particularly for ethnically diverse groups.