Background. The impact of immigration on health services utilisation has been analysed by several studies performed in countries with lower levels of immigration than Spain. These studies indicate that health services utilisation is lower among the immigrant population than among the host population and that immigrants tend to use hospital emergency services at the expense of primary care. We aimed to quantify the relative over-utilisation of emergency services in the immigrant population. Methods. Emergency visits to Hospital del Mar in Barcelona in 2002 and 2003 were analysed. The country of origin, gender, age, discharge-related circumstances (hospital admission, discharge to home, or death), medical specialty, and variable cost related to medical care were registered. Immigrants were grouped into those from high-income countries (IHIC) and those from low-income countries (ILIC) and the average direct cost was compared by country of origin. A multivariate linear mixed model of direct costs was adjusted by country of origin (classified in five groups) and by the individual variables of age, gender, hospital admission, and death as a cause of discharge. Medical specialty was considered as a random effect. Results. With the exception of gynaecological emergency visits, costs resulting from emergency visits by both groups of immigrants were lower than those due to visits by the Spanish-born population. This effect was especially marked for emergency visits by adults. Conclusion. Immigrants tend to use the emergency department in preference to other health services. No differences were found between IHIC and ILIC, suggesting that this result was due to the ease of access to emergency services and to lack of knowledge about the country's health system rather than to poor health status resulting from immigrants' socioeconomic position. The use of costs as a variable of complexity represents an opportunistic use of a highly exhaustive registry, which is becoming ever more frequent in hospitals and which overcomes the lack of clinical information related to outpatient activity. © 2007 Cots et al; licensee BioMed Central Ltd.