Allan Pacey and Chris Barrett, of Sheffield University's department of obstetrics and gynaecology, have developed a computer training program that they believe will eventually show that the declining figures are not a sign of reduced male fertility but of inadequate training of laboratory technicians.
About 300 laboratories in the UK carry out semen analysis, but evidence is growing that their standards and results vary wildly.
Doubts about the accuracy of sperm counts have existed since 1995, when the British Andrology Society (clinicians and scientists involved in the study of sperm and the male reproductive system) published the results of a pilot study in which portions of the same semen sample were sent to 20 laboratories throughout the country for analysis.
"The results were pretty alarming," said Dr Pacey, a member of the society. "The centre that got the lowest sperm count found three million sperm per millilitre, while the the highest counted 240 million per mil, so the difference was huge."
Further studies have suggested these disparities are frequent. Dr Pacey says that statistics indicating that human sperm levels are falling must therefore be viewed cautiously, and that, with a low sperm count being considered as anything below 20 million sperm per millilitre, it is possible that some men have been told they are infertile when they are not, and vice versa. He believes the problem lies in the poor standard of training that currently exists.
"We're a specialist fertility centre," he said, "but if you went to some other hospitals in Britain, or around the world, the semen analysis is sometimes carried out by someone who hasn't been trained. Sometimes it's a urologist, sometimes a haematologist. I once heard of a dermatologist who did it."
Semen analysis, which involves studying sperm motility (swimming movement) and morphology (its overall shape) as well as count, is a complicated and painstaking process. There are more than a dozen separate steps necessary to obtain a sperm count alone, from "fixing" or killing the sperm sample so that it can be viewed under a microscope to measuring its correct volume. One mistake in any of these stages can mean that the final count is grossly inaccurate.
The British Andrology Society holds training courses but places are limited and, with many clinics short-staffed, few can afford to send technicians anyway.
"At the moment, all there is is the World Health Organisation manual which people should try to follow," says Dr Diane Critchlow, organiser of the UK National Quality Control Assessment Scheme which is currently assessing laboratories responsible for semen analysis. "We've found that a lot of people don't even have a copy of the manual in the lab, which is a bit frightening."
The program developed by Dr Pacey and Dr Barrett contains digitised images of sperm as they would be seen through a microscope, as well as video footage that leads the technician step by step through each stage of the analysis. Any errors are pointed out either en route or at the end, when the computer supplies the correct answer and says if and where the analyst went wrong.
The hope is that this will provide an accessible and relatively cheap training system that will result in consistent standards of analysis. The program, which has taken Dr Pacey and Dr Barrett two years to develop, and for which they received no funding, will be presented to the World Health Organisation in Geneva next month. If it receives the WHO's approval it could soon be available on CD-Rom all over the world.