Abstract
Public health policy differs from programme insofar as the former is the expression of goals at a higher decision-making level (international, regional, national or provincial) and the latter involves the execution of intervention measures at the community or individual level. It has recently become fashionable to speak of "evidence-based" policy. There is now ample evidence to suggest that early nutritional influences on chronic disease risk in later life are contributing to the acceleration of the overall worldwide epidemic of obesity and non-transmissible diseases. In developing countries, in which 80% of the world's population resides, the opportunities for preventive policy must be balanced against needs, cost and effectiveness considerations and the intrinsic limitations of policy execution. Not everyone in the population is at risk of suffering from any given negative condition of interest, nor will everyone at risk benefit from any given intervention. Hence, decisions must be made between universal or targeted policies, seeking maximal cost-efficiency, but without sowing the seeds of either discrimination or stigmatization with a non-universal application of benefits. Moreover, although large segments of the covered population may benefit from a public health measure, it may produce adverse and harmful effects on another segment. It is ethically incumbent on policy makers to minimize unintended consequences of public health measures. With respect to the particular case of mothers, fetuses and infants and long-term health, only a limited number of processes are amenable to intervention measures that could be codified in policy and executed as programmes.