Anaemia in Dar-Es-Salaam and methods for its investigation

Abstract

The aetiology of the gross anaemia occurring in 204 patients admitted to the Muhimbili Hospital, Dar-es-Salaam, was investigated by means of a series of experiments. In addition 204 pregnant women attending an antenatal clinic were investigated on a more limited scale. The investigation depended on the estimation of the Hb concentration only, a rise in Hb level following the administration of a given factor being taken as an indication of deficiency in that factor. The basic design employed was that of a factorial experiment, although such procedures as fractional replication, confounding and the formation of blocks were also employed. Only for iron was a main effect demonstrated; no effect for protein, folic acid, cyanocobalamin, ascorbic acid or pyridoxine was elicited, and all interactions between iron and these other factors were found to be non-significant. In addition, treatment with anthelmintic showed no main effect or interaction with iron or protein. No difference in the speed of response could be detected for oral and intravenous iron, and at the Hb levels encountered here there appeared to be no benefit in giving more than 400 mg. ferrous sulphate daily by mouth. Investigations carried out on 360 prisoners indicated that hookworm infection was a more important predictor of Hb concentration than was diet as estimated by the stay in prison. What effect could be attributed to diet appeared to be extremely slow and this perhaps also accounted for the lack of benefit derived from the hospital diet in those admitted; hookworm infection was almost universal in the hospital patients and the load was heavy. Factorial designs proved appropriate for this type of investigation. Within-patient variation seemed to preclude the use of designs depending on the formation of blocks based on a previously determined Hb-producing capacity. Although the response to treatment was dependent on age in some experiments, this was not always the case. A time-response curve which was linear over a large part of the range perhaps accounted for the failure to increase precision by correcting for the initial Hb level.