Hypothesis-led ecological risk assessment of GM crops to support decision-making about product use

Ecological risk assessment assesses the probability that harmful effects to biodiversity will arise from using a particular genetically modified (GM) crop and the severity of those effects should they occur. Potential harmful effects to biodiversity include reductions in the abundance of species that are valued for their own sake or for the ecological services that they provide. Mechanisms by which harm could occur include exposure of species to toxic substances produced by the crop or displacement of species by the crop because it is able to colonise new habitats. Ecological risk assessments should contribute to decisions about whether to approve certain uses of the GM crop. They do this most effectively by being hypothesis-led, that is, testing hypotheses that the crop does not possess properties that indicate that ecological risk is unacceptable. These properties are defined after consideration of the aims of the decision-maker, which in most cases is a regulatory authority enacting government policy. Ecological risk assessment is ineffective when it tests null hypothesis that the GM crop is no different from a similar non-GM crop without first defining what differences should be regarded as important. Such data-led risk assessments offer no clear support for decision-making and tend to create controversy as decisions appear to be made arbitrarily. Current regulatory risk assessments for GM crops follow a mixture of hypothesis-led and data-led approaches.