Skip to main navigationSkip to main content
The University of Southampton
Mathematical Sciences

Statistical Methodology for Growth Hormone Misuse in Competing Athletes

Statistical methodologies developed at the University of Southampton and led by Professor Dankmar Böhning have refined a diagnostic biomarker test to detect growth hormone misuse in sport. These refinements have sustained and galvanised the use of the test by the World Anti-Doping Agency (WADA).

Context

An Olympic  skier

“Fairness” in competitive sports is a widely accepted paradigm. The International Fair Play Committee states: “Fair play means not cheating by taking drugs or doping. Anyone who does this ruins the game for everyone else.”

Athletes competing with their naturally given resources is widely accepted, whereas the application of stimulating substances to gain advantages in competitions is considered as unfair.

Research challenge

The current test for detecting growth hormone was developed and validated at the University’s Faculty of Medicine in the Growth Hormone 2004 Project (led by Professors Peter Sönksen and Richard Holt) and was first used at the 2012 London Olympic and Paralympic Games. The test enabled growth hormone to be detected several weeks after doping – a vast improvement on the previous window of 24-48 hours.

Two Russian powerlifters returned a positive test and were disqualified, having never been caught by the previous GH test.

The detection of doping using the test relies on decision limits, which are compared with the results of the test to determine if it is positive. These decision limits are set by the World Anti-Doping Agency (WADA) to maintain a 1 in 10,000 false positive rate. This is seen as a robust and appropriate level of reliability in situations where legal conflicts with athletes under suspicion of growth hormone misuse arise.

In order to ensure this strict false positive rate was achieved, Professor Dankmar Böhning of Southampton’s School of Mathematical Sciences was invited to lead the statistical modelling of a test results dataset five times larger than was used for previous decision limits.

Böhning, with Professor Wei Liu, used distribution theory to refine the decision limits, and provided an adjustment to counteract a positive age effect in males that was revealed in the larger dataset. 

Impact

The new decision limits based on Böhning’s analysis of the increased dataset were first incorporated into WADA guidelines in July 2015. Since then, the refined test has been used on more than 30,000 occasions.

Since the new decision limits were introduced in 2015, there have been no positive test results. This is a clear indication of the influential effect of the test in preventing doping with growth hormone.

Additionally, the statistical improvements to counteract the positive age effect in males was introduced in 2016, a year of strong growth in use of the test, demonstrating further the effectiveness of Böhning’s work.

These refinements have helped to sustain the high level of reliability of this test and its continued use in all world sports subject to the WADA code.

Key Publications

List of all staff members in
Staff MemberPrimary Position
Dankmar BöhningProfessor of Medical Statistics
Wei LiuProfessor of Statistics
    Share this case study Share this on Facebook Share this on Twitter Share this on Weibo
    Privacy Settings