The child survival strategy in developing countries has been driven largely by a selective disease control approach, which focuses on a few specific and inexpensive technical interventions directed toward conditions such as acute dehydrating diarrhoea and the immunizable diseases. Conceptually, this approach of designing health programmes in poorer developing countries by considering diseases one at a time is inadequate, because it fails to take into account the fact that children are typically afflicted with multiple disease conditions concurrently as well as sequentially. Furthermore, this technology-centred approach automatically eliminates from consideration a range of interventions for highly prevalent conditions such as low birth weight, parasitic infestations, or vitamin A deficiency, which may not be considered as direct ‘causes’ of death but, in fact, have a major indirect contribution to mortality by making children more frail. In this paper we develop an analytical model which demonstrates how multiple disease conditions interact through the mechanisms of competing risks and production of frailty to produce the high mortality levels witnessed among children in developing countries. This model permits an assessment of the demographic impact of different combinations of disease control interventions on reducing infant mortality. In terms of health policy, the analytical model demonstrates that there are multiple routes to improving child survival in developing country populations. Important in this context is the evidence that a modest reduction in several risk factors simultaneously - for example by improving household sanitation and personal hygiene - has the potential for producing substantial improvements in infant and child survival. The model does confirm a role for selective primary health care, but the criteria for selecting diseases for intervention should relate not only to their anticipated direct impacts on mortality, but also to their indirect effects on mortality through leaving large numbers of afflicted survivors more frail, and therefore at greater risk of dying of other disease conditions.