The interesting study reported by Reeves et al (Reference Reeves, Sauer and Stewart2001) draws attention to mental health service provision for ethnic minority elders. However, their findings could be misleading as they repeat common errors of cross-cultural research.

By definition, African and Caribbean elders are not a homogeneous group. Neither are they synonymous with ‘African—Caribbean’. As a population, they are of different history, ethnicity and culture. Furthermore, as migrants from different geographical regions of the world it is important that their different identities are appreciated, especially in their ‘third age’. Unlike the melting pot of second and third generations, these elders maintain distinct values that influence their social and helpseeking interactions. Migration pathways between the groups are diverse as well, ranging from long-term to recent, academic pursuit to meeting labour needs and the ‘culture-shocked’ to the assimilated.

There are also fundamental problems with defining cases by place of birth. ‘African-born’ includes Algerians, Egyptians, East African Asians and White South Africans. Similarly, ‘Caribbean-born’ persons of African, Asian and mixed-race provide a richly heterogeneous population of elders. How does one draw meaningful scientific conclusions?

The authors did not clarify the proportion of subjects from each of the African and Caribbean groups in the study. This may influence analysis and outcomes. One must also make a clear distinction between onset of illness and contact with services. First contact above age 65 years does not necessarily imply late onset of illness as alternative care pathways and help-seeking patterns may prevail (e.g. years of Pentecostal church attendance). Psychotic symptoms may go undisclosed for many years, particularly among groups suspicious or mistrusting of mental health services. On the other hand, reports of witchcraft or communication with ancestors, previously culturally sanctioned, may be mistaken for psychotic (or schizophrenia-like) experiences.

The authors argue that referral bias by primary care and community physicians is unlikely, as evidenced by low contact rates for anxiety disorders and depression. This is of concern, however, as other researchers (Reference AbasAbas, 1996; Reference ShahShah, 1998) have reported underdiagnosis of these disorders in ethnic minority populations, with a focus on psychotic and behavioural over affective symptoms.

Although Reeves et al opine that rates may be influenced by social isolation, physical ill health and social exclusion, this was not supported by evidence from their study.

The conclusions of this study are by no means generalisable and highlight ethnic and cultural confusion, as well as the neglect of depression and anxiety disorders, in research involving African and Caribbean elders. In the words of an elderly African, ‘if the heart is too heavy with sorrow, it may disturb the mind’. As clinicians we must not ignore this cry.