Poor monitoring of drug safety is risking the health of millions in the Third World, according to an editorial in this week's British Medical Journal.

The article praises ongoing attempts by governments and industry to speed up the availability of new medicines to poor countries but warns that unacceptable safety short cuts are currently being taken.

With the developing world bearing almost 90% of the planet's disease burden, it is understandable, they say, that drug companies are coming under greater pressure to remove the legal and financial barriers to access. However, because most very poor countries do not have proper drug safety monitoring systems, researchers are having to observe how safe drugs appear in First World countries – then make potentially hazardous extrapolations.

The authors, from Liverpool University argue that data compiled from countries in the West cannot simply be used to measure the effects of drugs new to the developing world. This is because the incidence pattern and the severity of adverse reactions could differ greatly due to different genetic influences and the local environment.

The editorial claims that better collaboration and more pooling of data from different sources might be one solution. "In the short-term we need to make better use of ongoing or planned studies," said lead author Munir Pirmohamed, professor of clinical pharmacology at Liverpool University.

He said that given that the ability to detect an adverse drug reaction depends on its frequency and the total number of people taking the medicine, academic researchers, drugs companies and government scientists should collaborate more closely – and possibly develop a universal adverse reaction reporting system.

The authors note that pooling of data has already brought benefits. In one example, the risk of serious neurological problems in people taking the broad-spectrum anti-parasite medication ivermectin who were also infected with Loa loa encephalopathy was identified before mass treatment was started.

In the longer-term, they suggest that each country should have its own national pharmacovigilance system that should contribute to a global data-base such at the one held at the Uppsala Monitoring Centre in Sweden. They also note, however, that "in a climate where health resources are limited", funding such a system "would come second to other competing projects such as implementing a new vaccine programme". By Michael Day