Commentary: Measuring physical activity in Sub-Saharan Africa

Many have speculated that the rising incidence of non-communicable diseases in developing countries is, in part, attributable to changes to population physical activity patterns. However, the evidence supporting such assertions is relatively weak, largely because physical activity is a difficult behaviour to characterize in epidemiological studies. The increasing interest in attempting to assess it better has led to a proliferation of reports of the development and evaluation of physical activity questionnaires. 1 Almost all of these questionnaires have been developed for use in developed countries and their transfer to different cultural settings may not be appropriate. The report by Sobngwi et al. in this issue of the International Journal of Epidemiology, 2 in which a questionnaire was designed specifically for use in Sub-Saharan Africa, is therefore to be welcomed as an effort to develop an instrument for measuring this important exposure in a part of the world where changes in population energy expenditure could be having far reaching effects on disease patterns. Like many other groups, Sobngwi et al. find that their questionnaire has high retest reliability. Reliability is rarely a problem with physical activity questionnaires and the main issues usually concern validity. 3 The difficulty in these studies is the selection of the measurement method that is to be used as the comparison or gold standard. Physical activity is a multi-dimensional exposure. If the interest in a particular study were in assessing energy expenditure, for example, then ideally the comparison instrument should independently and accurately assess this particular sub-dimension. Selecting a comparison instrument which assesses a different but related dimension, such as aerobic fitness, may give rise to spurious validity and may eventually make it difficult to interpret the results of studies that use that questionnaire. 4 Previous validation studies have also run into problems when they have used subjective physical activity diaries to validate questionnaires. They may have claimed to have demonstrated validity when in reality they have shown correlated error. Unfortunately the best method for measuring energy expenditure in free-living individuals, the doubly labelled water technique, is extremely expensive and has only been used in few studies even in the developed world. 5 The next best thing then is a method which is objective and which has itself been compared to the optimal methods. Sobngwi et al. elected to employ heart rate monitoring and movement sensing as the comparison methods, a decision which makes the field work …