When IVF treatment fails or ends in miscarriage, it is heartbreaking for infertile couples. It is also disappointing for their doctors, not least because the reason for many such failures is well understood.

In between half and three quarters of cases, the cause is an embryo with abnormal chromosomes. A test for identifying these problems, known as pre-implantation genetic screening (PGS), has existed for several years. But while some scientists claim it can lead to dramatic improvements in IVF success, professional bodies have been unconvinced. Both the British Fertility Society and the American Society for Reproductive Medicine (ASRM) currently advise against using PGS outside clinical trials. However, they may soon have to change their tune.

The main justification for this scepticism was a study from the Netherlands published 18 months ago, which found that PGS actually reduces success rates. The procedure involves removing a cell from an early embryo for genetic analysis, and this, it seemed, was causing excessive damage. The paper's impact was such that some doctors even declared PGS a dead end.

That was never a particularly sound conclusion - the Dutch team's methods have been criticised, and its lack of experience handling embryos could have accounted for the poor outcomes. And it is becoming less tenable all the time. At ASRM's conference in San Francisco this week, it was hard to escape the feeling that the mood is changing decisively. Advocates of PGS have always claimed that in their own experience, it works. They are now starting to produce data to match and exciting new technologies that promise still greater improvements to IVF.

Early PGS was a blunt instrument, because it could assess fewer than half of the 24 types of human chromosome. Abnormalities in the unseen chromosomes could not be detected. It also required invasive biopsy of a fragile early embryo, which is difficult to do without inflicting damage.

A newer technique called comparative genomic hybridisation (CGH) can screen all 24 chromosomes, at less risky points in embryo development. One approach is to biopsy five-day-old embryos called blastocysts, which have more than 100 cells and are more robust. Another is to use structures called polar bodies, which are discarded by eggs and can be examined without touching the embryo itself.

The CARE group of British clinics is already offering polar body CGH with success, and another team, from Oxford and Colorado, has data suggesting pregnancy rates as high as 80 per cent. There is still a need for randomised trials with control groups, but the indications look good.

As fast as CGH technology is being developed, however, it may be becoming obsolete. At a session on Monday, several teams presented details of newer and more powerful technologies that can screen chromosomes - and look for inherited diseases - with greater accuracy and simplicity. These use “gene chips” that can take an embryo's DNA fingerprint and provide much fuller information about its genetic quality. Which option proves best remains to be decided, and randomised controlled trials, again, are still required. The excitement that these papers generated, however, suggests that the scepticism towards PGS may not be justified for much longer.

It is important to note that this very scepticism has been a powerful driver of progress. The realisation that early applications of PGS were not working well spurred scientists to do better. The theory that selecting chromosomally normal embryos will improve success has always been sound, but data showed that the tools were not up to the job. So the response has been to build new ones.

It is just this sort of self-criticism and dispassionate assessment of evidence that best delivers real benefits for patients.

Mark Henderson is the Science Editor of The Times

Explore Health

• Expert Advice

• Health Features

• Mental Health

• Alternative Medicine

• Child Health

• Health Club