The massive government research project, organised by the fledgling Food Standards Agency, will try to identify areas where a lot of GM foods are being bought.
Public health figures for these areas will then be analysed to see if there are any increases in health problems in those areas.
However, some scientists say it will be "virtually impossible" to produce meaningful research in this way.
The research is expected to start in July, and last only 18 months.
Consumption patterns will be matched to data about the number of birth defects, or increased cases of cancer, diabetes and other diseases.
A spokesman for the Food Standards Agency said: "We are contracting market research firms to get in touch with the supermarkets and establish this information.
"This will be compared with information held by health authorities on conditions such as Down's syndrome."
Vegetarian ingredient
The ingredients it will search back for are foods containing GM maize and soya, and chymosin, an ingredient of some vegetarian cheeses and pizzas.
But members of the scientific community have pointed out that the study cannot hope to produce valid results when there are so many other factors which might affect health and which are not taken into consideration.
Professor Will Waites, a professor of food microbiology at the University of Nottingham who speaks on behalf of Cropgen, a scientific body linked to the GM food industry, was critical.
He said: "It took decades to prove the link between smoking and ill health, and that was fairly obvious.
"I wouldn't expect them to find anything meaningful in 18 months."
Tim Marshall, a senior lecturer in epidemiology at Birmingham University, with much experience of setting up scientific studies, said that the research was likely to be "worthless".
He said: "It might be interesting to identify different dietary patterns in different parts of the country, but I thought that had been done already.
"I think what they are trying to do is impossible. It's a question that needs answering, but I don't think it can be answered.
"So much else is going on in these patients that there is no way you would be able to calculate a cause and effect relationship, or a non-cause and effect relationship."
GM testing
The Food Standards Agency is also embarking on a series of tests to determine whether supermarket food claiming to be GM-free lives up to its description.
More than 20 laboratories are to be involved in the testing programme.
However, leading scientists have already warned that current testing to distinguish between GM-free and GM foods is unreliable and misleading.