Applying a conceptual framework in countries where maternal mortality rates have dropped to under 100 per 100,000 successful births, the research focuses on two key questions concerning the organisation of service provision: where do women give birth and who delivers them? It does seem that a policy and a paradigm shift in favour of having skilled attendants at birth is taking place amongst some donors and multilateral agencies. Nonetheless, the evidence is not clear-cut.

The data for this study, drawn from internal government reports and Ministry of Health officials, yields four basic models of care, outlined in the table. The third column highlights country programmes where maternal mortality rates fit one of the models. Whilst a causal relationship between programme intervention and subsequent reduction of maternal mortality rates cannot be established, insights into success can.

Evidence unearthed suggestions that maternal mortality rates are:

• staggeringly high where non-professionals carry out home deliveries and never fall below 100 deaths per 10,000 women
• reduced 50/10,000 or below when deliveries are carried out by a professional and linked up with a strong referral system
• not necessarily reduced to under 100 by the fourth and arguably most advanced model: Mexico City rates are 100 plus per 10,000 even though over 90 percent of births take place in hospital.