Critics argue the study does not compare like with like
There may be serious risks for the baby when mothers who chose a home birth are transferred to hospital, a study says.
Researchers writing in the journal BJOG examined all home births over a decade.
While the death rate for births at home was low, if the mother transferred to hospital the risk of losing the baby soared to eight times above average.
Home birth advocates say the study does not consider women who develop problems in hospital. And the deaths of babies numbered just 65, from 10,752 cases.
Less than 2% of women currently opt for a home birth, but the government has promised all prospective mothers the choice by 2009.
A team at the National Collaborating Centre for Women's and Children's Health looked at data on all home births in England and Wales between 1994 and 2003.
There were broken down into three sub-sections: planned home births, unplanned home births - when a mother intended to go to hospital but was caught unawares, and a "transferred group" - when women who had planned a homebirth ended up giving birth in hospital.
The death rate among babies whose mothers planned - and did - give birth at home was lower than the average for all births taken together - including those in a hospital, midwife-led unit, as well as at home.
This, the study's authors suggest, was not surprising, given that on the whole only women with few risk factors are likely to be able to book a homebirth and they were thus a self-selecting group.
But when complications did arise for women in this group and she needed to be transferred to hospital, her risk of losing her baby was nearly eight times higher than the national average - and 12 times higher than had she given birth at home.
On average, around 15% of women who have chosen a home birth will be transferred - although the figure is significantly higher for first-time mothers and significantly lower for those who have already had children.
"Essentially women who opt for a home birth face either a very successful, satisfying outcome, or a potentially disastrous one - there isn't the greyer area that you see with hospital births," says Professor Philip Steer, the editor of the BJOG.
"Women's decisions about where and how they give birth must be respected, but it is very important that they are not misled about the risks and are in a position to make an informed choice."
However, the authors conceded that the actual data about transfers was taken from a multitude of studies and was inevitably inconsistent.
"Transfers" did not refer exclusively to women who were rushed to hospital in labour, but included any woman who, having initially chosen a home birth at 12 weeks, ended up giving birth in hospital.
The reasons were also unknown, potentially varying from a change of heart or the desire for more effective pain relief to a major complication.
And the actual numbers are small. In the 11-year period studied, sixty-five babies died after their mothers were transferred to hospital out of a total of 10,752. Childbirth in the UK, experts stress, remains very safe.
The National Childbirth Trust, which wants all women to have the choice of a home birth, says the study should have looked at the number of babies who died in hospital after their mothers developed complications.
"You're not comparing like with like," says Mary Newburn, head of policy at the charity. "The study is seriously flawed."
The Royal College of Midwives said such the research was "welcome but inconclusive".
"The nature of transfer means that complications are anticipated or evident. This principle applies irrespective of place of birth, whether this is at home, in a birth centre or in hospital," said Dame Karlene Davis.
A spokesperson for the Department of Health said: "The department welcomes this article, whose findings will add to the much larger and more detailed study we have already commissioned on safety of place of birth."