IVF clinics routinely mislead patients by "cherry-picking" their best results, new research suggests.

Jack Wilkinson, a medical statistician at the University of Manchester, said success rates should be banned from websites unless an agreed system for reporting is brought in.

He said: "Our study shows that success rates are advertised on 67% of IVF clinic websites, and many of these may be highly misleading, because clinics can cherry-pick their results from a dizzying array of options.

"The concern is that clinics can always construct figures that show their own performance in the best possible light, while making competitors look bad.

"This is not helpful to patients, who may struggle to understand differences in what each clinic reports and may be misled into making comparisons on the basis of incomparable results.

"Even if clinics are acting in good faith, the current system does not produce clear or reliable information for the people that need it."

He continued: "There is a strong incentive to selectively report success rates in a way that exaggerates performance.

"Just as advertising of prescription drugs to patients is not permitted, a ban on advertising IVF should be considered if there is no binding standard of reporting."

The research, published in the journal BMJ Open, analysed data for 53 websites reporting success rates.

It found pregnancy rates were reported more frequently (by 83% of clinics) than birth rates of a live baby (51%). Live birth rate is more important because the pregnancy rate covers women who will go on to miscarry.

The team found 33 different ways of reporting pregnancy rates and nine different ways of reporting birth rates.

Both these sets of results could then be split into subgroups of patients, such as only featuring those who were younger.

Some 72% of clinics published results without saying how many patients they applied to, while 26% did not give the age of patients - despite a mother's age being highly likely to influence the results.

A spokesman for the Human Fertilisation and Embryology Authority said: "Patients have told us that they find researching fertility clinics overwhelming, particularly when it comes to understanding statistics.

"That's why many use our Choose a Fertility Clinic service (http://guide.hfea.gov.uk/guide/), which shows the same statistics across all UK clinics so patients can make like-for-like comparisons using information which has been independently checked.

"We expect clinics to follow our requirements to present data fairly, accurately, and in a manner that helps patients make the best choices for their future treatment."

Professor Adam Balen, chairman of the British Fertility Society, said: "We encourage all clinics to present accurate and reliable data that is not misleading and reflects the true results as monitored and presented by the HFEA on their website.

"Data should also include live birth rates for different age groups and multiple pregnancy rates."