SUMMARY. We have conducted a survey in the United Kingdom in order to evaluate the prevalence of anaemia and of iron deficiency in elderly male and female patients and in female patients of child-bearing age presenting to their general practitioner. Although haemoglobin values were approximately normally distributed, there was evidence of a bimodal distribution of serum ferritin values, particularly in elderly men. A correlation between serum ferritin and haemoglobin values was found to exist, both in the elderly and in younger women.The overall prevalence of anaemia in the study population was 14'3%, although it was higher for elderly men (20,1%) than either elderly women (13'7%) or women of child-bearing age (11'9%). Iron deficiency associated with anaemia was more common in elderly men (45,2% of those anaemic) than elderly women (28'6%) or women of child-bearing age (25'6%).This survey indicates that the prevalence of anaemia among certain groups of high risk patients presenting for consultation to their general practitioner is about 1 in 7, although the prevalence is as high as 1 in 5 in elderly men. Additional key phrases: ferritin; anaemia; general practiceAnaemia may be one of the more prevalent undiagnosed conditions in the community. Previous studies of haemoglobin concentrations have used random samples of the general population. H In these studies, the criteria for anaemia have ranged from a haemoglobin concentration below 12·5 to 13'Og/dL for men and 11·5 to 12·0g/dl, for women. Excluding pregnant women, the prevalence of anaemia was variously estimated as between 2·9% and 13·25% for men and 2·4% and 20·0% for women. Certain groups are thought to be at risk, specifically women of child-bearing age and the elderly.Serum ferritin concentrations correlate well with body iron stores, and can be used as indicators of iron deficiency or overload.':" Ferritin concentrations in women tend to be lower than in men, which is thought to reflect decreased iron stores through iron loss in menstruation.

No abstract

A maximum likelihood solution is obtained for the simple linear structural relation model where the underlying incidental distribution and one error variance are assumed known. Expressions for the asymptotic standard errors of the maximum likelihood estimates are obtained and these are verified using a simulation study.