GM Crops Inquiry

Absolutely. I apologise for the late submission; I have just come back from the US.I thank the committee for the opportunity to attend today's meeting. I am a pathologist, specialising in toxicology. My group is particularly interested in matters that affect the developing child, the foetus and the infant.This morning, I want to address the issue of risk assessment, what it tells us and, more important, what it does not tell us. We all espouse the precautionary principle, but it is rarely invoked, mainly because evidence in the form of a risk assessment is usually produced to show that something is safe.The precautionary principle has been criticised in connection with stifling progress; that is covered on page 3 of the paper that I sent round. That comment appeared in Nature, and the next page of my paper features a letter that Professor Peter Saunders and I wrote to Nature. We pointed out that the precautionary principle was not about that. It is a tool for the use of, and for assistance to, decision makers in deciding whether the next step along a path ought to be taken. In the taking of that decision, consideration needs to be given to what the benefits are and what the costs are. That relates partly to the speed at which progress is made. Most of the things that I want to say this morning are essentially covered in our letter to Nature.Risk assessment is a process that considers various sorts of information and comes to some sort of probabilistic conclusion. We should realise that risk and hazard are not the same thing. If someone crosses the road, they might be hit by a bus; that is a hazard. The risk is the likelihood of them being hit each time that they cross the road. On page 6 of my paper, I present a summary of a risk assessment that was produced by Monsanto a few years ago. The bottom line says:"The overall risk of damage is assessed as low to effectively zero."That really means that the risk is non-zero. Everybody accepts that. Various things are missing from that statement, however. How long is it good for? Is it good for 10 minutes? Is it good for the duration of a growing season? Is it good for several generations? That is never explicitly stated.Page 7 deals with the four phases of risk assessment. First, people have to be able to identify hazards, because it is not possible to test for hazards that have not been thought about. That is usually the most difficult phase. Secondly, hazard assessment is required for those hazards that are identified. That is hard science, which requires experimentation. If that comes up with negative results, those have to be addressed. Thirdly, for food, we need exposure assessment. We need to ask who is eating what and how much of it. It is only when all those are in place that—fourthly—a risk assessment may be attempted. Without the first three steps, however, it is very difficult.Risk assessment was first devised by engineers to assess the structure of buildings, an area in which the problems are finite. However, the principle is now being applied to very complex systems, such as ecosystems, which contain many unknowns. When there are unknowns, people tend to replace data with models. At one extreme, it is possible to have risk assessments that are based totally on models, with no data. As well as the other extreme, we can have all the places in between.Page 9 covers hazard identification. Areas of potential hazard include genetic instability, horizontal gene transfer and pleiotropic, unpredictable effects, such as allergy and toxicity.For hazard assessment, we are presented with substantial equivalence as the test. However, that is a chemical test and we are really interested to know about the biological effects. That requires a biological hazard assessment, which is rather like the way in which a drug is tested. At the minute, however, the standard test is a simple one of chemical composition. If the product passes that test, it is then licensed—that is certainly the case in the United States.Page 12 of my paper deals with exposure assessment. No exposure assessment has been done, as far as we know. In the States, where the GM plants have been introduced, no one really knows who is eating what. Without that information, we have to ask what the risk assessment is based on.If subtle changes were being caused to people's health by GM plants, we would not know. What is worse, we would not have any way of detecting those changes, because we do not know where we are starting from and we do not know what the exposure level is. If GM products were acutely toxic, we would obviously know, but we accept that they are not acutely toxic. If, however, they were causing subtle changes at the level of allergy and so on—common things—we would not know. If thalidomide had caused cleft palate instead of a rather obvious malformation, the likelihood is that we still would not know about it, because cleft palate is a common condition. If one starts changing the rate of instance of common conditions, and one does not know the starting point and there is no exposure data, one cannot know whether something is causing a problem.I turn to the last slide, on page 18 of my paper. The toxicological aspect of novel foods is a tractable problem. We could develop the methods to test them adequately, but we are not doing so at the minute. Such testing is done in the pharmaceutical industry; it costs about $400 million to produce a new pharmaceutical. We take pharmaceutical products voluntarily—usually for a good indication, usually for a limited time and usually in thousandths or millionths of a gram, so the lifetime exposure is probably a few grams. With food, we do not have any choice. The dose is kilograms in a day or tonnes in a lifetime. There are very strong arguments for being rather cautious about this technology; we are not doing enough to test those things, and that needs to change.

I am particularly concerned that if we cause subtle changes to principal components of the food chain, the most likely place in which they will have an impact is development. Subtle changes to the composition would not be picked up by a test of substantial equivalence—biological activities would have to be looked for. We know that, at particular times in development, the body's hormones are controlling development in low parts per trillion, which is an incredibly low concentration. For instance, we know that some chemical pollutants, which each of us has in our bodies, are able to be bioactive and cause problems at low parts per trillion. Those things are being measured now.My first concern is that we need to test novel foods for biological activity in a developmental situation, which is the most critical time of life for being pushed off course. Currently, there are proposals to introduce vaccines and drugs into genetically modified plants to produce them for the pharmaceutical industry. If those genes get into the general gene population, there is a very definite prospect of them affecting development. My principal area of concern is at that level. Once a person is fully formed as an adult, it is rather more difficult to produce a toxicological effect.

It is always easier to do something than to do nothing. That is a basic statement. The precautionary principle states that the weight of evidence should be examined and that one should be prepared to act without knowing that something is absolutely, scientifically proven to be the case. It means being prepared to act on the balance of probabilities. When a new technology is started, there is very little or no evidence.In general, we have some working examples from the past. For instance, the green revolution, which involved the widespread use of chemicals in agriculture, had the net result that everyone sitting round this table has residues of several thousand chemicals that could not have been around when our grandparents were in their mothers' wombs.Now, people say, "Gosh, that is a problem. We should reduce that amount." That result could have been predicted, because we knew the chemical nature of those compounds. They were fat soluble and persistent, but if somebody had said at the beginning, "We should not develop the technology because this, this and that might happen," a storm of protest would have arisen, and the chance of precaution being used would probably have been low. We are at the beginning of a new technology. If we embrace it, we must accept that the genes will spread. That is a given.

We are not collecting the information to find that out. If changes are caused, the link will be difficult to demonstrate. The technology is powerful and potentially useful in medicine, but we must treat it with much respect. I understand that any changes that we make will be in perpetuity, because they will self-replicate.We have seen horizontal gene transfer. That is common sense. Any risk assessments that address the matter should assume that once genes are released into the environment, they will travel, although it is implied that that is not the case. Much more work must be done in the laboratory before general use is adopted and we use the planet as a test tube. Not enough homework has been done.

I am sorry. I sent it yesterday and I realise that it was late.

You quoted a statement of fact. If people do not know where they are starting from and do not monitor exposure to the substance in which they are interested, they cannot relate that to any changes that might happen in the pattern of disease. It is important to realise that the risk assessments that are being presented are, by and large, opinions. They are not based on adequate hard evidence for conclusions.We should undertake large-scale voluntary human feeding trials with such foods. That is what we do with pharmaceuticals—we test them in clinical trials. That is an option for considering human response. To look for allergies, we would be talking about big studies that involved several thousand people, but that could be done. However, at the moment, GM products are being released on to the market without any notion of who is eating what and what that is likely to cause. As I said, if a change is caused to something that is common anyway, the chance of finding a causal link is as near to zero as we would care to consider.

Not with the current database, no.

The British Medical Association is less sanguine than the chief medical officer. There is, therefore, diversity of opinion among the medical profession on that issue. The chief medical officer might well be right, and I hope that he is, but we do not know. I do not think that the way that we have approached the issue will tell us. The only way to find out about human allergy, for example, is to test the crops on humans. The surrogates that are used in sequence homologies of amino acids and peptides do not really test for that. Testing has to be done properly, and I do not think that we have done that yet.

Yes; that is a reasonable summary. The difficulty is not that we do not have the tools to approach the problem and to try to find out properly; the difficulty is that doing so will mean treating GM foodstuffs a bit more like pharmaceuticals, on a one-off basis, and that will be expensive. Of course, the developers do not like that idea. That is where the debate rests at the minute.

That is an example of something that we should be interested in finding out. The trials are a step along the way to full-scale introduction. If we have the trials without knowing the answer to the allergy question, we have taken more steps than we should have.Clearly it is more complex to test a food—which is a mixture of many thousands of different compounds—than it is to test a single drug. However, it should be possible to develop such techniques. That is what Dr Pusztai was charged with doing. He got £1.6 million of Scottish Office grant, in competition with other laboratories, to develop techniques of hazard assessment for novel foods. That work does not seem to be continuing, although it should be. We should be developing methods of assessing the toxicological nature of novel foods, but such work is not being funded further.

I have been prescribing insulin that has been produced by genetic modification techniques for decades, as have most other medics. That is an example of the use of recombinant DNA technology to produce medicines.Recombinant DNA technology as we know it now is a hit-and-miss technology. New technologies that are emerging will be much more precise and controllable. I briefly address that important issue on page 14 of my paper. The classical paradigm on which recombinant DNA technology is based is that DNA produces ribonucleic acid, which then codes for a protein—it is a simple mechanistic flow in that direction—but if you turn the page you will see that we have learned all sorts of things over the years that make that idea just too simplistic. There are overlapping genes; there are interrupted genes; there are genes that delocalise; there are multigenes that are on different chromosomes; and there are multigene families. Our understanding of how a genome is destabilised when a new piece of genetic information from another species is fired randomly into it is in its infancy.People are now trying to be much more selective about the way in which they put genes in. When they do that, many of the problems that have been associated with gene instability and unpredictable effects will diminish. My feeling is that there is still a lot more homework to do. We have to expect the unexpected. At the minute, that is standard, which is why issues have to be treated one at a time.The benefits that have been claimed for GM plants are that they will grow in saline conditions and so on, and that may well be the case, but as far as health safety goes, I do not think that we are doing the right testing. Substantial equivalence is a scam. People say that a potato has vaguely the same amount of protein and starch and stuff as all other potatoes, and therefore that it is substantially equivalent, but that is not a test of anything biological. We have to examine and test, which costs money, but I do not think that we can afford to play fast and loose with this technology. We need to take great care.

It depends what you are looking at. One could start to examine allergies in such areas, but the sites are small. I do not know what the population density is around them. It comes back to the point that if something is common, an enormous sample is needed to pick up a significant change. It may well be that such a sample would be unavailable. In developmental toxicology, we have animal-testing methods to examine such matters; for example we have developmental toxicity testing. That would be a first step.

It would be very helpful if such information were in the public domain. I do not know whether it is in the public domain, but any testing that is to be used for licensing a product that will become part of the food chain should be open to scrutiny. There is no reason why that should be commercially sensitive. I can understand that the method of introducing the transgene into the plant is commercially sensitive, but toxicological testing should be in the public domain so that it can be inspected and commented on.

Some information is in the public domain, including Pusztai's work, which was an example of testing for developmental toxicity. I know of one other paper that has been published, but much of the material is kept commercially confidential. If it is being presented for the purpose of licensing a new crop that could become part of the food chain, it should be in the public domain for scrutiny.

You would have to know where you start.

You have to look at the power of the experiment and to know that, you need to know the incidence of what you are interested in and the background population. If what you are interested in is normally rare, you might not need that many people to detect a change. However, if it is common, you will need a large sample to get the statistical power into the experiment. Statisticians could advise you on that.

Yes, through volunteer feeding trials.

There are large areas of uncertainty and ignorance. We do not know what has been tested and what the likely effects could be of X, Y and Z. There is a need for transparency.

Yes. The case for keeping commercial confidentiality on toxicological testing is weak. I can understand that there are sensitivities about the mechanism for producing the transgenic organism, but I do not see why the testing should be commercially confidential.

Good morning. As CMO, my concern is to ensure that the Executive is appropriately advised on all matters affecting health and health care in Scotland. I am thus concerned about any potential health effects of GM crops and their production. I welcome the committee's investigation into the issue and the opportunity to address the committee today.My role and concerns are shared by other United Kingdom CMOs with whom I work closely. My submission notes the findings of an investigation that was published in 1999 by the chief medical officer for England, Sir Liam Donaldson, and the UK Government's chief scientific advisor, Sir Robert May, that there was no evidence to suggest that GM food technologies were harmful to human health. I have seen no subsequent evidence to the contrary.The trials that are the subject of the committee's concern are not food trials. I emphasise and endorse what the minister has just said. No GM food enters the food chain of either humans or domestic animals as a result of the trials. The sole exception to that is the possibility of the introduction of pollen from the GM crop to honey. That has been specifically considered and covered in the submission that the committee received from the Advisory Committee on Releases to the Environment.I shall briefly summarise the responses that we have submitted to your four questions. The first question was whether the Executive should prevent GM crop trials from continuing on the ground that that is against the precautionary principle. Scottish ministers are advised on that by ACRE. That committee will not advise that trials should proceed where there is any reasonable concern over an effect on human health.The farm-scale evaluations are proceeding on the basis of the operation of the precautionary principles as follows: ACRE has concluded that there is sufficient evidence to conclude that growing those crops does not pose a threat to human health. However, the impact of herbicide-tolerant crops on farmland biological diversity is less clear. The carefully controlled studies are designed to address that uncertainty before any further decisions are made.We are now in the final growing phase of a three-year trial, which is an additional hurdle outwith the present regulations. The producers have voluntarily agreed that that phase should be undertaken in order to produce further information on that area of uncertainty.The risk assessment procedures for farm-scale evaluations build on knowledge found in previous phases of development in laboratory conditions and small-scale environmental releases. For this crop, you will know that the releases have been going on for about 14 years.There is a case-by-case assessment of both hazard and environment. The hazards to be assessed specifically include hazards to human health. I believe that the risk assessment by ACRE is robust and comprehensive and is consistent with EU guidance on the application of the precautionary principle. All of that is set out in detail in ACRE's submission to the committee. There is no evidence of any specific health effect that we should be monitoring. There is no evidence of any health effect in the workers or researchers who were involved in the previous phases of the trials. Neither is there evidence of any health effect in the populations around sites of commercial production of the crops in other countries. Those are the kind of conclusions that have led ACRE to believe that there is sufficient certainty that the crops do not pose a threat to human health.Local public health departments throughout Scotland consistently monitor the health status of their populations and are on the lookout for any new or emerging health hazards. I am sure that the committee is aware that that is important, particularly in the environment and context surrounding the consequences of 11 September last year.We have contacted all departments of public health in areas where farm-scale evaluations are being carried out and I can tell the committee that they report no unusual patterns of disease or of new or unexplained illness in their populations.

I acknowledge that, but there are certain key principles of health monitoring that have been well articulated by the World Health Organisation. The monitoring programme has to be specific, which means that we have to have an accurate definition of what we are monitoring and it has to be measurable, which means that we have to have systems that are capable of counting what we are trying to find. Further, the system has to be action-oriented, which means that the data that we collect must help to guide some action, as well as being realistic and timely.I heard the previous witness say that, had Thalidomide caused cleft palate, rather than phocomelia, no one would have noticed. I want the committee to know that we have excellent background monitoring systems in Scotland, particularly in relation to developing children and development in utero, and that we monitor congenital abnormalities. I have every confidence that any alteration of background patterns would be spotted. Perhaps Martin Donaghy would like to give the committee more information about that.

Yes. To pick up on what the CMO said, we have two levels of health monitoring. One level is general health monitoring, during which we try to pick up differences or different patterns of illness in the community. Every child who is born in a Scottish hospital has a return sent to the information and statistics division in Edinburgh and there is follow-up to find out about the child's development through our child health monitoring and surveillance programme. Therefore, we pick up problems with congenital abnormalities, clusters and increasing trends.The second type of monitoring happens when we look for something specific that has been emerging, such as a new illness. Recently, there has been concern about a new type of viral illness associated with birds, which was first seen in New York. We call such illnesses emerging illnesses. We target certain areas and collect evidence to monitor.In respect of genetic modification and the latter type of monitoring, if we wanted to institute a programme, we would need something specific to look for, as the CMO said, otherwise we could not do a count. Genetic modification is a technological process. Part of the regulatory framework includes ACRE's going through the specifics of which genes have been modified. If there was any concern that a specific gene could lead or relate to a potential health hazard, a trial should not go ahead. If we scanned the literature and found a potential health hazard, we could institute a programme as we do for other problems and new illnesses, but we have seen no evidence or indications at all of such hazards that would lead us to institute such a programme.Allergenicity gives rise to a range of health problems that we call atopic, such as asthma, eczema and rhinitis, or runny nose. In the past 10 to 15 years, there has been a large increase in such illnesses in the general population. Asthma in particular has been the subject of much intensive research. Some increases in asthma levels have been associated with exposure to plant allergens—that is, plants without any genetic modification. However, we have no specific evidence that links the genetic modification that is employed in the crop trials with specific increases in allergens.Members probably know that many of the crop trials involve oil-seed rape. Given the extent of the cultivation of conventional oil-seed rape, particularly in Scotland, it has been the focus of study over the past few years. There is some evidence that there has been a slight increase in rhinitis in particular as a result of oil-seed rape. From the evidence, it is thought that that might not be due specifically to the pollen or the allergen, but to the fact that those oil-bearing products can release volatile organic compounds into the air, which may cause rhinitis. That happens in only a very small minority of the population who are sensitive.We have no specific evidence on which we could develop a programme. There has been background monitoring. As the CMO said, we monitor every child in Scotland. We would pick up on increases in abnormalities to do with cleft palate, which a previous witness mentioned.

I will ask Martin Donaghy to comment. Margaret Jamieson flagged up a good example of the way in which public concern about a potential new hazard can develop over a long period. The best that we can say about the potential hazard from radio communications equipment is that a potential biological effect that we can now monitor has been postulated. The jury is still out on the case and activity continues on the matter.Margaret Jamieson mentioned a good example of how long-term concerns can develop, but I emphasise that what is going on in GM trials is of a different order: the trials are localised and farm-scale evaluations are designed and controlled to answer specific questions. Those questions are asked largely to enable us to get to the next stage of decision making. At that next stage of decision making—about the potential commercial development of the crops, which my colleagues will talk about—a different set of questions must be answered.

I will reiterate what Dr Armstrong said. The main difference between electricity pylons and GM crops is that the pylons are up. On GM crops, we are at the stage of going through a regulatory process, which involves assessment of scientific literature by ACRE. If there had been any potential hazards, the next stage—the crop trials—would not have been reached. We now have crop trials and the product of those trials is not going into the food chain, so people are not being exposed to it. The results of the trials will be evaluated and will be considered by ACRE. If there are negative results, the process will not move forward. The next stage will be to go through the next regulatory hurdle.We are at a different stage in relation to GM crops than we are with electrical pylons. A tight precautionary approach is being taken, which is designed to sift out problems before we make decisions on whether the products of the new GM technologies should be made available to the general population.

I am satisfied that ACRE has asked the right questions and that it has properly considered the appropriate evidence. I cannot say that there is no risk; of course there is a risk and it would be foolish of me to say that there is none. I remind members of the statement that was made by the previous witness—which I endorse—who said that we are not in a situation in which there is certainty, but in which we are ensuring that there is a proper balance of risks, benefits and costs.

People who suffer from allergies would not regard such changes as subtle—either one has an allergy or one does not have it.

I draw your attention to ACRE's evidence and that from the Royal Society, which make it crystal clear that it is not scientifically possible to prove a negative.

I certainly will. It is not the case that we are not looking for anything. As Martin Donaghy explained in detail, at least two levels of monitoring go on—background monitoring and specific monitoring. I have explained that if a specific effect were identified, we could monitor an area for that effect. For example, if it was identified that certain GM crops posed a threat of increased allergenicity that was likely to translate into higher levels of allergic reactions such as asthma, rhinitis and dermatitis through the simple mechanism of living near the crops—we are not talking about ingesting the crops—we would be able to monitor the area for those effects.

As I said, we must have a balance between risks, benefits and costs.

The precautionary principle refers to the need to take appropriate levels of action. I do not believe that such a level of action would be appropriate.

The background monitoring in all the areas in which the trials are being carried out has not detected any specific effects. We have asked the local public health departments to look for such effects. Each of those departments produces an annual report on the health status of their populations.

The trials, which started three years ago, were set up under a regulatory framework that asked Scottish ministers to authorise them on the basis of specific advice. The advice that ministers are required to take is set out in the legislative framework. That advice comes from a variety of sources, including the Advisory Committee on Releases to the Environment, the Health and Safety Executive, the Food Standards Agency and the Department of Health. All that advice indicated consistently that there was no discernible risk to human health that would require a specific health monitoring process. If there had been any evidence of such a risk, the trials would not have been authorised without the appropriate safeguards.

I seem to be going round in circles.

Before the crop trials take place, considerable research is carried out, particularly on animal models, to pick up whether there are any effects on mammalian systems. We examine the research on the nearest available equivalents to humans. That is the first level at which we look for effects. If there are effects, the trials go no further.Those research initiatives take place in laboratories; there is day-to-day contact with the workers who are involved in the initiatives and the Health and Safety Executive is involved. In some ways, those workers form what we call a sentinel population. If there were any demonstrable effects on the immediate population—that is, laboratory workers and researchers—such effects would be picked up and the trials would not go any further. By the time a crop trial is carried out and the evidence on the trial is submitted, work will have been done to identify potential effects. From our point of view, if there are potential effects on human health, there should not be a crop trial. ACRE is asked to screen out that possibility so that by the time of a crop trial there is no effect for testing to pick up.We are being asked, as a safety net, to find out whether there is a health effect. The problem with that is in working out what we should look for, because the effects might be subtle. Many hundreds, if not thousands, of tests could be carried out to find out whether the trials have an effect on people. The ethical point is that a trial should not proceed if we believe that there will be an effect on human health. However, if we raise the alarm unnecessarily, people will ask what we are looking for and we will have to reply, "We don't know, but there might be something." The evidence, however, shows that there is nothing specific to look for. To go back to our preliminary statement, we depend on our on-going health monitoring of the population.

It is important to say that the possibility of GM pollen's being in honey is not new; it was first considered by the Advisory Committee on Novel Foods and Processes—ACNFP—as long ago as 1991. That possibility has always existed. After specific concerns were expressed, the position was reassessed in 1999. That pre-dates the Food Standards Agency Scotland, so I am merely providing information from records that are in the public domain.Having been reassessed in 1999, the situation was brought to the agency's attention after we were set up in 2000. The ACNFP was last asked to examine the matter in 1991. The ACNFP is a very broad committee that contains toxicology experts and allergy experts as well as consumer experts and ethicists. That committee's view was that, although there could be GM pollen in honey, the levels would be so low that they would not present a meaningful risk to consumers.It might help if I put that in context. Research that was funded by the Ministry of Agriculture, Fisheries and Food a number of years ago tried to assess what amount of GM pollen might get into honey. We are trying to get people to understand that the amount—unless members want to talk in terms of nanograms, which I find quite difficult—is equivalent to one crystal of sugar in 28,000 1kg bags of sugar. The possibility is so tiny that, although it could exist through transfer from bees, it is not deemed to be a risk. As recently as last year, we asked the ACNFP to re-examine the matter.

On the question that the convener asked, let us leave pollen aside and consider that material—

The lady who wrote to you was specifically concerned that GM material could be entering the food chain through watercourses—

As I am sure members are aware, the expert advisory committee that was set up to advise ministers—the Advisory Committee on Releases to the Environment—is under the Office of Science and Technology's general guidance on the operation of such committees and is drawn from a very wide constituency. With specific reference to Nicola Sturgeon's question, I am sure that you are aware that the chairman of that committee, Professor Alan Gray, has considerable expertise in hydrology. He is director of the Natural Environment Research Council centre for ecology and hydrology. The Advisory Committee on Releases to the Environment is in the process of making a further appointment in Banchory from the NERC centre for ecology and hydrology. I want merely to underscore the point that, through those advisory committees, we make strenuous efforts to ensure that all possible threats to the environment, through all possible routes, are covered. Although they cannot give any guarantee that the environment is risk free, those committees exist to address those points.

I would like to return to the issue of pollen in general, regardless—

I hope that my answer will address your question.

The lady's question was about pollen going into the water and possibly into fish.

I will try to address all the points that have been made.Crop trials are not about food safety—Mac Armstrong has outlined the purpose and background of the trials. However, it is important to bear it in mind that ACRE's assessment process takes into consideration the possible implications of inhalation of pollen, accidental ingestion of material from crops and people's coming into contact with crops. At the forefront of our minds is the fact that the trials are being carried out for environmental reasons, but the implications of material getting into the food chain, being inhaled or coming into direct contact with people are addressed in the safety assessment that ACRE carries out. It is important to bear it in mind that we have not yet reached the stage of examining the crops from the point of view of food safety; however, that issue is not ignored in the safety assessment.

I can talk to you about substantial equivalence as it is used to asses GM foods, rather than—

The membership of ACRE includes Professor Janet Bainbridge, who is the chairperson of the Advisory Committee on Novel Foods and Processes and who gave evidence to the committee recently. The ACNFP is an expert scientific committee that advises the Food Standards Agency on novel foods, including GM foods. There is obviously cross-representation between ACRE and the ACNFP. Substantial equivalence is not in itself a safety assessment for a novel food or a GM food; it is a concept around which a safety assessment might be built. At the heart of substantial equivalence is the principle of taking a novel food—which would include a GM food—and comparing it with its conventional counterpart. That is the absolute starting point.Throughout the process of using substantial equivalence as a framework of safety assessment, we look at the similarities and differences—I stress the differences—between the new food and its conventional counterpart, which has a history of safe food use. Substantial equivalence looks at the differences; it does not just look at the similarities, and—

It covers various differences. Those might include the composition of the food, its nutritional value, its metabolism and so on. The safety assessment that then goes forward focuses on the health implications of the differences. It is important that members keep in their minds the fact that substantial equivalence does not just say, "Yes, they are broadly similar." It tries to identify the differences that might exist and to assess the potential—

Substantial equivalence is used widely in the world. It is—

It is not a test in itself. Substantial equivalence is a concept around which safety assessment—

The farm-scale evaluations are not food safety trials. They are environmental—

No, that is not true. Any environmental release, at whatever part of the regulatory process it takes place—I have said that the farm-scale evaluations are an additional hurdle that the Scottish Executive has introduced into the regulatory process—can be advised only on the basis of a risk assessment. The risk assessment has to go through the framework that is being described, which includes an assessment of health hazards.

Yes. The trials are not designed to test food safety; nonetheless, they are advised only on the basis that there is no hazard to human health.

I will not get drawn into that. It is not my role to say what commercial companies should or should not do. I will say, however, that your question perfectly illustrates the answer to a previous one, on the difference between the precautionary principle and the precautionary approach. The precautionary principle is not a principle about human health. It was stated in the Rio declaration on the protection of the environment. As a principle, it has been translated into a wide variety of contexts, and it is now used in relation to health—quite rightly so, in my view.However, the precautionary principle does not say that, until we have certainty, we must do nothing. This is illustrated at paragraph 8 of my written evidence, which states of the Royal Society of Edinburgh:"The RSE also referenced the European Commission communication on the use of the precautionary principle"—that is to say, the precautionary approach—"in the EU and internationally … which stated:‘Where action is deemed necessary, measures based on the precautionary principle should be, inter alia:proportional to the chosen level of protectionnon-discriminatory in their applicationconsistent with similar measures already taken'".That is the reason why the substantial equivalence idea—

Testing is mainly carried out by companies, university departments or research institutes that are commissioned to do that work.

The data go to ACRE. That does not apply only to GM foods; it is how all expert scientific advisory committees work. There are protocols about commercial companies submitting data to expert scientific committees, which often see data not just from commercial companies, but from universities, before publication. Therefore, those expert committees are privy to those details. Sometimes the commercial companies will publish data, sometimes they will not. Obviously, we try to promote openness in publication, but the only way—

Just to conclude, the only way in which the scientific committees can do their job—to properly assess the evidence—is by entering into agreements with commercial companies to obtain those data. Otherwise, they cannot properly assess the risk to the public.

There are toxicological tests on mammalian models. There are obvious ethical issues about using human beings as guinea pigs, so humans are not involved in those trials. There are toxicological tests on mammalian models, mainly mice, which are applied to the organisms that have been fed GM.

Some are and some are not.

With all due respect, I do not work for a commercial company. However, the usual reason is that companies will retain that information for competitive advantage. The Government's stance is to promote the publication of such data by companies, but those data are shared with the Government. If they were not shared, the expert committees could not do the job of assessing risk.

That type of approach is the same as that which is taken to the evaluation, and subsequent licensing, of new drugs. It is interesting that some of the companies involved develop both drugs and GM foods. The technology is the same; in fact, GMOs are used in pharmaceutical products already. This is a world in which commercial considerations apply, and the licensing procedures allow companies to disclose fully all the information that is necessary to satisfy the licensing authorities but not put it in the public domain at that instant.

The original question was about placing information in the public domain through means such as the media. The convener's point was about the access the Government has to that information. The Government has complete access. The differences between drug licensing and clinical trials were compared. To get permission to carry out clinical trials, drug companies submit research data on a commercially sensitive area, based on mammalian models, to the relevant committee. Once a drug trial takes place, everyone knows about it because it is in the public domain. The processes that are used in that area are the same as those that are used in many other areas, including the introduction of new food products to the market. The Executive is given access to data, but there are agreements about the release of commercially sensitive data to the public.

Yes, partly because the companies must be licensed to do the original experiments.

The answer to those questions is yes.

I make it absolutely clear that I no longer work for the BMA, although I was its secretary when the submission was written. The question is rather unfair. I do not need to explain to a parliamentary committee the democratic basis on which such organisations are run and the dispassionate role that the management of such an organisation is required to take when faced with a decision at an annual meeting that such a report should be written.I say for the record that I was profoundly disturbed by the line that the BMA report took and that I regard it as one of the poorest reports that the BMA has produced. The evidence is confused and the report consistently mixes up the precautionary principle's application to the proper process of evaluation of GMOs through the regulatory framework and the principle's application to the release of GMOs in commercially developed food. I have no hesitation in saying that I disagree profoundly with a number of points in the BMA's submission. If you wish, I will give the committee a point-by-point rebuttal of the BMA's evidence. I hope that you will not confuse my role as, in effect, the chief executive of the BMA when that report was written with a personal involvement with or endorsement of its conclusions.

From my experience, the BMA does not routinely consult all its members before it publishes such reports.

There is a difference. If I were to be brought to account for my stewardship of my role at the BMA, as I might be—I could, for example, be called to answer before an industrial tribunal or to answer on a health and safety or legal matter that related to my time there—I could not distance myself from my role in the process. That role is different from my professional opinion on the outcome of the process. I do my job as CMO as independently and professionally as I can and will defend the decisions that I make as CMO; I will remain accountable for them for the rest of my life.

I have explained that, in my position at the BMA, I was never required to endorse the line that was promoted in the report. Saying that sketchily states the relationship between the BMA's senior management and the process from which the report resulted. There is an overview of the process—that is the secretariat's role—but it is not possible to interfere with the content of the report. That is how the BMA is set up. My role was to ensure that the process was correct and that the report was produced; it was not to endorse the report, which I do not.

That is wrong: I deny that I must do so. That is not my role.

I co-ordinate the environment theme, which is one of the three main research themes at the SCRI.

I am a virologist by profession and have been a member of the Advisory Committee on Releases to the Environment and a biological safety officer.

In our written submission, we attempted to clarify our role. Our experience is not medical, but we have worked on issues of medical and clinical safety, particularly that of non-GM oil-seed rape as an allergen. We have been involved for many years in plant selection and breeding to reduce the chemicals that occur in the brassica, or cabbage, family that are potentially harmful to livestock and humans. We have interests in and contacts with health issues through our involvement in plant breeding and pollen movement and our study of chemicals emitted by oil-seed rape and other brassicas, but we have no direct clinical involvement or expertise.Our perception of the issue is coloured by our stance, which is outlined at the bottom of the first page of our submission. We give weight to reports and papers that are reviewed properly and which are open in their results and findings so that they can be reproduced. We apply a rigorous set of criteria, which we expect others to apply when they give opinions on matters of ecological or environmental health.I confirm that our role in the present farm-scale evaluations of GM, herbicide-tolerant crops is that of an impartial observer of their effects. We are not involved in the making or marketing of those crops, which in the Scottish context are oil-seed rape crops that are tolerant to a broad-spectrum herbicide, but we are involved heavily in the ecological field testing of them.

Our role has been to examine the rate and distance of cross-pollination between fields and other sources, such as the feral oil-seed rape plants that are seen commonly on road sides and field margins. Our view, which is based on considerable field research that has been conducted in several countries, including our own research, is that cross-pollination over distance is likely to occur. I cite some of the likely values that are based on the expected values from evidence that has been accrued so far. We might expect around one in 10,000 seeds in surrounding fields to be cross-pollinated hybrids.The Executive is funding us to obtain accurate measurements of cross-pollination in the fields surrounding the farm-scale evaluation GM sites. That work will be published next year. It will take time and we will have to go through the peer-review process. I cannot comment on that work in progress, but the likelihood is that genes will find their way, in low frequencies, from any field to other fields in the vicinity.

That is correct.

Yes. Our view is that we have to be case-specific on issues concerning GM crops. They are like other kinds of crops, in that they differ in species and in the type of modification—the bit of DNA that goes into them and its effect. We are against blanket statements and recommendations.On the basis of published evidence and the evidence of various learned bodies, including the committees that have examined the crops, the SCRI is satisfied that the main farm-scale evaluations of GM crops that are going on in the UK are safe for our staff who work directly with the crops—including me—and people who live in the neighbourhood. However, we cannot extend that statement to all eventualities that might occur in future.

As previous witnesses have pointed out, we can never prove that something is safe.

I do not believe so. We can never eliminate all possibilities. It is in the nature of things that evidence of safety is negative evidence—evidence of no risk. We can never eliminate the possibility of a risk that we have not thought of. That is the point that we are trying to make.

All hazards that can be identified should be considered and tested. I believe that that has happened with all the hazards that have been identified so far. It is difficult to say what more could be done. If someone identified another hazard—another problem that might arise—that could and should be considered and tested.

First, I think that antibiotic resistance markers have been thoroughly tested. Secondly, they are being phased out. Thirdly, they do not apply in this case—the materials do not have antibiotic resistance markers in them. As far as I am aware, all the tests on allergenicity that are possible have been done.

I believe that animal model tests have been done with the pollens.

You would probably have to ask ACRE about that.