Some parties have recently been championing home births. Where the mother is young, the pregnancy uneventful and if ambulance transfer to hospital can be made quickly in an emergency, a home birth might be an option.
But not many buy the idea. Out of 39,874 live births in Singapore last year, only 104 were home births. Still, that was the highest number of home births in the last 10 years, which saw between 60 and 80 annually.
The same is true in the United States, always a bellwether of social change: US home births rose by 41 per cent between 2004 and 2010. But in absolute terms, home births were only 0.79 per cent of all US births in 2010 compared to 0.56 per cent in 2004.
A study of 16,924 planned home births published last month in the Journal of Midwifery & Women's Health found that "planned, midwife-led home births for low-risk women" are very safe.
The study uses data from the 2004 to 2009 Midwives Alliance of North America (Mana) Statistics Project. Home deliveries of healthy babies were uneventful, there was no rise in complication rates and there was little need for transfer to hospital.
The study involved low-risk pregnancies, meaning mainly young healthy mothers with no complications such as hypertension or diabetes during preganncy. In contrast, women in Singapore tend to have their first baby at an older age, as the median age at first marriage for women in 2012 here was 28.
Conventionally, there are two measures of infant death rates. The first is the "neonatal mortality rate", which is defined as deaths from birth to four weeks of life for every 1,000 live births. This is used the world over.
This study, however, covered infant deaths from birth up to six weeks of life, instead of four. No one else uses such a statistic. And the study showed that the death rate of babies from birth to six weeks was 1.61 for every 1,000 low-risk, live births at home.
By contrast, the death rate of babies from birth up to four weeks of life, for all live births in the US - both in hospital and at home - was around four for every 1,000 live births.
Why did the study resort to such an unusual measure of infant mortality? Perhaps statistical manipulation turned up the best rates for midwives if the period covered was stretched from birth to six rather than four weeks.