Two-year-old Isabel Peake is getting used to her new baby sister, Zoe, at home at Alresford in Hampshire. "Grumpy baby," she says, when Zoe cries. But Isabel will never know what her sister looks like because she is blind, as a result of contracting bacterial meningitis at birth.

The infection was caused by a bacterium called Group B Streptococcus (GBS), which is carried by between 20 and 30% of the population in their genital or intestinal tracts. It is harmless to adults but potentially catastrophic for babies. Although most pregnant women have never heard of it, GBS is now the commonest cause of severe infection in newborn infants in the UK.

As well as being blind, Isabel has cerebral palsy. She can't walk, crawl or feed herself and will need care for the rest of her life. All this could probably have been prevented had her mother, Leesa Yeo, lived in the US, Canada or many other EU countries because she would have been routinely screened for GBS. If a pregnant woman is known to be GBS-positive, the onset of the disease in the baby can generally be prevented by giving the mother antibiotics during labour.

But in this country, women are rarely told about the condition, let alone tested for it. Yeo's introduction to GBS came via a leaflet in hospital, after Isabel's diagnosis. Only at this point did she discover how it could have been prevented.

It was too late for her little girl, but fully in keeping with official guidelines from the Royal College of Obstetricians and Gynaecologists (RCOG) which don't support routine screening and recommend that antibiotics are only offered to women in certain "high-risk" categories - for example, in cases of premature labour or when the mother has already had GBS-infected babies.

GBS is sometimes detected incidentally during pregnancy via the standard NHS test used to detect infections, the high vaginal swab (HVS). Unfortunately, this is notorious for producing false negatives for GBS in women who are carriers. According to research published in the journal Obstetrics and Gynaecology in 1996, the error margin in women who are carriers may be as high as 50%. Yet since 2003, there has been a simple, accurate alternative known as an enriched culture medium (ECM) test available from a private laboratory, which costs £32.

It is still not understood why some children born to mothers who are carriers become infected and others don't, but it is estimated that out of the 700 children who get GBS infections each year, between 70 and 100 will die.

The numbers are small in the context of 700,000 live births a year, but for two couples in Oxfordshire, the statistics mask the horror of what must be every expectant parent's worst nightmare. On September 6 2002, Sue and Nick McKeown and Alison and Craig Richards were at the John Radcliffe Hospital, Oxford, awaiting the births of their first babies.

The McKeown's daughter, Teresa, was born at 6.30am after a long and difficult labour. Then, in an instant, the McKeowns' world disintegrated. "I just remember hearing the midwife hitting the wall and saying, 'Code blue'," says Sue. "You know they aren't good words. There was no crying - just absolute silence."

Teresa's body was no match for the colonies of bacteria, which had caused septicaemia and pneumonia. She died in her parents' arms, at 12 hours old. Afterwards the couple were told she had been "riddled" with GBS.

As Teresa was losing her battle for life, Alison Richards was in another delivery suite giving birth to her son, Owen, following an 18-hour labour. "He was delivered and put straight on to my chest," she says. "That's when I said, 'Something's wrong.' His eyes were closed and he wasn't making any noise." Owen, who had had a healthy heartbeat throughout the labour, never took a breath on his own. The post-mortem revealed an "overwhelming infection" of GBS.

Returning home, the McKeowns and the Richards checked their pregnancy books for references to GBS. "It was only mentioned in one - an American book that was more than 10 years old. That made me really angry," says Sue.

Catherine Greenwood, a consultant obstetrician at the John Radcliffe, was instrumental in getting the hospital's policy on GBS changed six months after the events of that day. "Everyone in the unit found it incredibly shocking that these two babies, who had been apparently fine in labour, had died, and that gave me the impetus to push for a change in policy at the hospital, she says."

Certain women, such as those wishing to have home births or on immuno suppressants drugs, are now screened using a test which has a higher level of accuracy than that used previously. More women who test positive are offered antibiotics and since these changes, no babies have died from GBS infection in the hospital.

But what about the rest of the country? Jane Plumb, who founded the charity Group B Strep Support (GBSS) in 1996, following the death of her baby, Theo, says the RCOG guidelines don't go far enough. "We want the ECM test to be made available on the NHS to all women regardless of their risk status, with all those who test positive offered antibiotics during labour. Only 60% of babies who contract GBS are born to high-risk mothers - research suggests a screening would prevent 80-90% of babies becoming infected."

Why won't the Department of Health change the guidelines? "We can only introduce routine screening and treatment when we have clear evidence that it will do more good than harm," says a spokesperson. "There is insufficient evidence that treating those carrying the organism with intravenous antibiotics during labour would be beneficial, and this would represent a major shift in the way maternity care is provided. There are risks of serious reactions to penicillin and other antibiotics. There are also concerns about the development of antibiotic resistance."

So it seems paradoxical that the National Electronic Library for Health, the NHS's own website, used by staff to treat patients, states that when women are "incidentally" found to have GBS during pregnancy, "It is important to inform the woman of her carrier status and to record this information in her antenatal notes so appropriate action may be taken at the time of delivery. During labour, intravenous antibiotic prophylaxis should be discussed."

If it is deemed necessary to discuss GBS with women discovered to be carriers "incidentally", why not test everyone? Expense is one reason. As Greenwood notes, "But, of course, you can't convince any woman who has lost a baby that cost effectiveness is what matters."

However, questions of screening and prevention are academic if women don't even know about GBS. But despite the Department of Health's policy that all pregnant women should be "well informed about all possible risks to their baby, including GBS," there is no consensus among ante-natal health professionals on how much information to give out. Earlier this year, one major hospital agreed to let the charity GBSS leave leaflets in its ante-natal clinic, only for a member of staff to request they remove them for fear of frightening women unnecessarily.

Leesa Yeo says she is too busy to think about blame, but she can't understand why all women aren't informed about such a potentially devastating infection. "I just feel frustrated that we are routinely tested and given information about comparatively rare conditions such as syphilis, Aids and hepatitis B - why not GBS?"

Group B Strep Support Group: www.gbss.org.uk