[Is the analysis of blindness in developing countries at the occasion of a national census of interest to public health?].

Prevalence of blindness in developing countries is three to ten times higher than in developed countries. Needs for individual medical care are far from being met and this requires the implementation of comprehensive policies of prevention and treatment. Two distinct data sources are available to carry out these policies: direct ophthalmic surveys on representative population samples, and for some countries, a national census in which blindness cases are numbered out. In Togo, both sources are available; we have thus been able to analyse: data from the 1981 national census in which blindness is identified as a social handicap; and the results of 4 W.H.O. surveys conducted from 1982 to 1986, which measure blindness prevalence and analyse its causes. The analysis of both sources shows that: a national census cannot be a substitute for surveys which are the only means for determining blindness causes and implementing action programmes. nonetheless it provides exhaustive information, as well as data on geographical distribution and on age and sex pattern of blindness. It also allows the identification of high prevalence zones--not always spotted through a survey--and thus, the implementation of specific actions in affected zones. It appears, hence, that the question on blindness should be maintained in a national census; providing users bear in mind possible bias in blindness census, the fact that appreciation of blindness is a subjective exercise and also providing they don't yield to the temptation of drawing comparisons between blindness rates which don't cover the same realities.