Objective: To evaluate the use of a robust, cheap method for haemoglobin estimation by non-laboratory-trained personnel in a rural setting.
Design: Comparative study.
Setting: Tintswalo Hospital, Acornhoek.
Participants: 7 nursing sisters, 4 medical students, 2 lay persons.
Outcome measures: Haemoglobin estimates obtained with the colour scale were compared with the 'true Hb' values determined by the H x 3 Bayer-Technicon automated blood analyser.
Results: Although individuals varied in their abilities to use the colour scale, its performance was generally very good when measured against automated haemoglobinometry, as determined by bias and regression analysis and also in terms of its capacity to detect anaemia, as measured by sensitivity, specificity and positive and negative predictive values.
Conclusions: Haemoglobin estimates obtained with the World Health Organisation colour scale are generally reliable, although cognisance should be taken of individual variability. While the utility of the device in monitoring response to therapy remains to be seen, it promises to be a suitable method for mass screening for anaemia.