Guide to Risk Assessment and Biosafety in Biotechnology, GRABB

An Initiative of the United Nations Environment Programme (UNEP)
SECTION:

RISK ASSESSMENT

TITLE:

Current Procedures for Applying Risk Assessment

BY:

Mark Segal, Julian Kinderlerer, Gilbert Howe, Morris Levin

LABEL: PRA UPDATED: 31 Dec 1997

Contact: IRRO Secretariat

IMPORTANT NOTICE:

The following is a Draft and is considered incomplete. It does not represent the views of any member of any of the governments cited, except where official documents are included through linkage. It represents the views of an author and as such, there may be some errors in interpretation of methods with which he is not directly experienced. Future versions will include contributions from those with such direct experience, and such errors will be corrected and coverage made more complete


CONTENTS

Introduction

Determining the Purpose for an Assessm Purpose for an Assessment

Determining the Scope of an Assessment

Full Assessments-Iterative Use of Data

Reasons for limiting the scope

Implicit inclusion of elements in simplified risk assessments

Accounting for exposure

Familiarity

Available Review Resources

Legal Influences

Format for Integration of Risk

Parallel Processing

Rationale for parallel review

Variations on the parallel approach

Sequential Analysis

Framework for Risk Management (U.K.)

Rationale for sequential review

Variation on sequential analysis

Advisory Committees/Peer Review

Data availability

Disclosure and Proprietary Information

Summary

 


Introduction

The risk from releases of organisms to the environment is usually represented by some version of the general paradigm:

 risk = /(hazard ,exposure)

 where hazard means the potential for adverse effects arising from intrinsic features of the organism and exposure means the potential that a sensitive receptor, whether individual, population or system, will encounter the organism. In risk assessment, sometimes both of these elements are explicitly expressed, and at other times one or the other may not appear to be considered. However, all risk assessments for releases of organisms will somehow have taken both hazard and exposure into account at re into account at some time during the analysis process.

This chapter will not provide a primer on the elements that comprise a risk assessment. The primary source for that information is in the unit on Concepts and Theory of Risk Assessment. The purpose for this chapter is to illustrate that risk assessment may take many forms, while adhering to the general model shown above, and that how and why one uses risk assessment will determine what version of this model is most useful.

This essay describes various influences that affect the way a risk assessment is designed and presents some approaches that may be used to combine the elements of a risk assessment into a unified analysis. Assessments are driven by the needs of the assessor, and these needs can vary, depending on who requires the assessment, what is the subject of the assessment, how the release will take place, whether there are specific legal requirements and many other issues.

D="b2">Determining the Purpose for an Assessment

There may be many reasons to perform a risk assessment for a release of an organism:

Mandatory legal requirements

Sponsoring institution needs

 In many countries, there are mandatory legal requirements that assessments be performed prior to release. These assessments may be done directly by a governmental entity, or the government may require one to be prepared by the individual(s) proposing the release, or by a committee of experts on behalf of the government. In some many cases the laws that concern releases may be administered by completely different departments and agencies within the same national government. Often these legal requirements may be in conflict. In the U.K., for example, just the consideration of genetically modified organisms involved so many different components of the federal government, an effort to coordinate these elements was demanded, and a Coordinated Framework was issued in tas issued in the Federal Register, in 1986. In the European Union, Directive (90/220) has an effect on coordinating oversight by member countries. Elements of government that typically have purview over releases of organisms are those associated with biotechnology, public health, agriculture, forestry, scientific experimentation and environmental control. For a resource on the various regulatory approaches to biotechnology assessment, one can use BINAS (http://uba.cenargen.embrapa.br/binas/regulations.html).

 Even if there is not a legal requirement to do so, a sponsoring institution, such as a university or company, may ask an investigator to do the assessment. The institution may want such an assessment to protect itself from problems, such as lawsuits, should something go wrong with a release, or it may simply need risk information to prepare its own risk-benefit analysis to verify that a project is worth doing. Even non-governmental organizations (NAOS) ma(NAOS) may desire their own assessments if they are considering challenging a proposed release.

Determining the Scope of an Assessment

The risk assessment processes described in previous chapters (links) show the full range of issues that need to be considered when a comprehensive assessment is required. While there is often a presumption that a risk assessment must be a complicated document which provides involved scenarios, the extent of an assessment, the level of detail, the issues considered, all can vary remarkably among assessments performed for similar purposes. Although there may be many cases where complex assessments are necessary, often there may be only a few issues that require resolution, and simplified assessments may be used in these cases.

Full Assessments-Iterative Use of Data

At one end of the spectrum, assessments of releases that are widespread and include organisms with a high likelihood of survival in many different ny different habitats require a comprehensive approach.

Tolerance for error will usually be lower than that for small scale field release experiments, or, especially, for enclosed system use with very limited opportunity for intentional release. When non-indigenous organisms are to be introduced into a region for the first time, such comprehensive assessments are usually required by regulation. The more unusual organisms, or those with more widespread the application, usually demand that a broad review of potential for effects on humans and the environment be done.

The chapters on risk assessment methodology cover many of the issues that need to be addressed. Often governments provide guidance on what to consider for review. In Canada, under CEPA there are specific data requirements related to a limited set of categories of uses of organisms in biotechnology. The data requirements are spelled out in regulations so that a noilier knows immediately what types of data will have to be provided to rovided to the government in order for assessment to proceed. Canadian Food Inspection Agency (CFA) provides guidance for review of released organisms that may be classified as biofertilizers under the Fertilizer Act (http://www.cfia-acia.agr.ca/english/actsregs/fert/fertrege.html) Organisms used as pesticides in the U.K. are subject to review by the U.K. Environmental Protection Agency using the Federal Insecticide Fungicide and Rodenticide Act (FIFRA) (http://www.law.cornell.edu/uscode/7/ch6.html) and Food Quality Protection Act (FQPA) (http://www.epa.gov/opppsps1/fqpa/) . Lists of testing requirements are provided (http://www.epa.gov/docs/OPPTS_Harmonized/885_Microbial_Pesticide_Test_Guidelines/Series/) which provides formal Guidelines under FIFRA for review of all microbial pesticide products. The US EPA also provides informal guidance on the components that might be neet be needed for a full risk assessment for biotechnology products, through Points to Consider (cite)http://www.epa.gov/opptintr/biotech/ptcbio.htm .

 

Reasons for limiting the scope

While complex cases may require the full use of risk assessment tools, such effort may not be needed in many other cases. There are various reasons for limiting the scope of an assessment, but such limits are possible only because elements of a full assessment are taken account in an implicit way. Reasons for reducing the scope of a risk assessment include:

 

Implicit inclusion of elements in simplified risk assessments

Even though an assessment might have a limited scope, it does not mean that all elements of risk are not considered. Rather, it generally means that the risks associated with some elements of release are assumed to be at a certain acceptable level based on evidence gathered long before the formal risk assessment has begun. As mentioned at the beginning, risk encompasses impacts on public heath and on the environment, and arises from exposure and hazard. This is often represented by:

risk = exposure x hazard

 but is more appropriately expressed as:

risk = /(exposure, hazard)

 If either element of risk is constrained, one needs onlyd, one needs only to account for the other element to be able to evaluate the total risk. Thus, for organisms that are familiar, or are directly derived from familiar organisms that are known to have very limited hazard potential , the hazard component of risk is constrained and one may concentrate on exposure. Likewise, processes and procedures that are commonly used are assumed to provide limits to exposure to any organism, whenever those processes or procedures are employed. This constrains exposure, and permits concentration of resources on questions of hazard. By using familiar organisms under limited exposure conditions, very simplified assessments can be considered.

 

Accounting for exposure

One reason for limiting the scope of a risk assessment might be that the amount of exposure to a released organism has been accounted for by other means prior to initiation of the formal risk analysis. Assessments for releases of organisms that could have could have global impact are generally comprehensive, while those where release is only incidental to production, and thus likely limited to the site of production, are often very simplified. Small scale field research is usually treated differently than releases of commercial products. Some governments implicitly address these limits to exposure by providing regulatory exemptions to certain types of products containing organisms that might otherwise be subject to full scale risk analysis. (US EPA,1997) (http://www.epa.gov/docs/fedrgstr/EPA-TOX/1997/April/Day-11/t8669.htm) . Organisms for use in contained systems of for experimental field application may have expedited reviews provided that conditions of use are shown to restrict exposure to the organism.

As an example, in the U.K. under the Toxic Substances Control Act (TSCA), in an indirect fashion, the scale of the intended release has an influence on the complexity of the review. In general, as the s as the scale increases the scope of the assessment tends to be more complex. Specifically, genetically modified micro-organisms in the fermentation industry may be exempt from review even when used commercially. These exemptions are established for uses of such organisms under contained conditions, where there has been a history of safe use of the recipient organism and where genetic modifications are well characterized and limited. (http://www.epa.gov/docs/fedrgstr/EPA-TOX/1997/April/Day-11/factsheet/fs001.htm)

While simplified reviews may be used for cases of limited exposure, they may also occur where release is predicted to be extensive, but where such release is an intended component of use. In the case of biological pesticides, a desired element may be broad scale distribution of the product organism. Under such circumstances, reviews may assume the broadest exposure possible and concentrate on the intrinsic hazard potential of the released organism.

 

 

Familiarity

Another reason for altering the scope of an assessment could include use of organisms that have frequently been released without apparent adverse outcomes. (US EPA,1997) (http://www.epa.gov/docs/fedrgstr/EPA-TOX/1997/April/Day-11/factsheet/fs001.htm This issue is sometimes addressed in terms of familiarity (US National Research Council, 1989) with the principal organism, or recipient organism in the case of genetic engineering. This option requires that extensive knowledge of an organism=s intrinsic characteristics has been compiled and analyzed some time previous to the initiation of a specific case review, or that sufficient long-term experience, supported by adequate documentation of low hazard potential, shows that the use of organism under well defined conditions is unlikely to pose a significant hazard.

In the above U.K. TSCA example, microorgani microorganisms used for commercial fermentation are exempted from review, in part because of an established history of safe use. Although this system would appear to eliminate the use of risk assessment for the exempt micro-organisms, in actuality a risk assessment based on a category of micro-organisms was performed prior to publishing the exemption requirements. These reviews were based on characterization of familiar micro-organisms, coupled with an understanding of the effects of common production practices on likely exposure during manufacture, the point where greatest release potential was understood to exist. The exemptions for listed microorganisms were based on the conclusion that any such organism would be used under conditions assumed during the course of the categorical assessments that formed scientific support for the exemptions. Thus, any new organism that is actually produced in compliance with these conditions would therefore match the risk considerations of the class of organisms that were at were considered as part of the categorical risk assessment, and further review would be redundant.(cite) (http://www.epa.gov/opptintr/biotech/rulesup.htm) . In a related way, for some other well studied microorganisms, there are exemptions under TSCA for those research uses that are not contained, but are released to the environment. In these cases, experience with field experimentation was taken into account to show that hazard potential would be low.

The U.K. Department of Agriculture (USDA) has addressed this issue similarly, for plants. After reviewing, and seeing the results of, modified crops, it determined that certain categories of plants need not be regulated further. New regulations were developed that enabled the Animal and Plant Health Inspection Service (APHIS) to extend determinations of non-regulated status to organisms that resemble those for which and initial determinatetermination, based on data review, was made (http://www.aphis.usda.gov/BBEP/BP/) .

 

Available Review Resources

In some programs, limits to resources or time available for performing reviews may contribute to decisions on whether, and how, the scope of a review for a specific release is defined. For most industrialized countries, the size of the review resource, in people or costs assigned, is not an issue; necessary resources are found. However, while the amount of resources assigned to a case review should not be the determining factor in how a review is done, nevertheless, there are practical reasons for desiring to be efficient in using analytical resources for reviews of releases of organisms. Excessive expenditure of human resources during reviews may cause delays in decisions that adversely affect the submitter and the industry, and reduce public confidence in government. If time limits for review are legally enforceable, exrceable, expending too much effort on trivial cases might lead to rushing through more complex cases. Rather than apply equal effort to all cases, a ranked approach is sometimes necessary to ensure that demands on the review system are not likely to overwhelm the available staff. Those cases that are truly novel, or that involve very high exposure potential, or use potentially hazardous organisms can thus receive priority use of assessment resources.

 

Legal Influences

An influence on whether to limit review may be the legal consequences of misinterpreting or misapplying an assessment when developing a risk management decision. Laws and regulations may thus dictate the extent of an assessment. For example, the US EPA approval of a commercial product under TSCA may permit unrestricted use, potentially including use for purposes other than those initially proposed, unless the US EPA negotiates restrictions with the maker of the product. The final stage risk assessmentk assessment for such a product therefore must consider that the organism will have the widest plausible dispersal as a result of manufacture and use, even if the manufacturer only plans a limited production initially. This is done, because once an approval is given without restriction, there is no recourse for the government, should production and use exceed the bounds considered when the risk assessment was made. In practice, if there is any chance of wider use than anticipated, and described, by the manufacturer, a restriction limiting the manufacture and use to that anticipated by the submitter will usually be negotiated into a binding agreement between the government and manufacturer of the organism. Thus, to ensure that the highest quality decision is made, it may be necessary to perform a comprehensive review for all cases where the consequences of making an error, scientific, legally or politically, are too great for the organization to risk. Even apparently trivial cases with high visibility mbility may require greater than average scrutiny. Doing so may be good public policy, even if not scientifically justified. The use of simplified review is thus often limited to cases that are straightforward, or for which a consensus exists, legally and politically, as well as scientifically, that fully comprehensive assessments are not justifiable.

 

Format for Integration of Risk

There are two distinct approaches to integrating hazard and exposure components into a risk assessment, both of which should reach the same conclusions using the same facts. The selection of one or the other methods depends on many factors, including available human and information resources, who has responsibility for an assessment, regulatory issues and governmental philosophy. This discussion makes no value judgements about either method, but rather, offers the reader alternatives that may be used based on a particular set of circumstances. The two general assessment concepts can be cls can be classified as parallel processing versus sequential analysis.

 

Parallel Processing

Parallel processing describes the simultaneous and somewhat independent evaluation of different components of an assessment, prior to merging the results for completion of the risk analysis. A parallel processing approach is used at the US Environmental Protection Agency for analysis of biotechnology product microorganisms subject to TSCA, released to the environment. In this approach, different discipline specialists are usually assigned to specific component analyses that form part of a risk assessment. These component analyses are done independently of each other and are later integrated into a risk assessment, by a technical integrator. The component analyses include:

 Product Identification(I. e. confirmation of the identity of the organisms used to develop the product organism, and identity of the tity of the product organism itself)

Construct Analysis (Review of the methods to genetically engineer the organism)

Hazard Analyses(Separate ecological and human health reviews)

Exposure Analyses(Reviews of both worker and general population exposure)

Risk Integration(Combining the components to draw summary conclusions)

 Usually a different analyst, who specializes in one or another of the review types, is assigned to each component. Despite the assignment of responsibility to single reviewers, frequent interaction among all members of the review team takes place throughout the process. The technical integrator has overall control of the review process, ensuring that critical information is exchanged among the staff and verifying that key statements from each analysis are incorporated into the final summary, or integrated risk assessment. The independent reports of each reviewer form background documents, but the integrated risk assessmed risk assessment is what is ultimately used for risk management decisions, which are made by senior staff and managers.

 

Rationale for parallel review

A simultaneous risk characterization component review provides a few advantages for certain situations. It separates the risk characterization phase from the risk management phase and also permits separation of hazard and exposure components of a review. It allows for independent assessment of the relevant components, until there is a need to integrate the results of each components assessment. It thus ensures that there is complete documentation of all aspects of a risk characterization, uninfluenced by the outcomes of other aspects. It also ensures that the risk management decision process does not unduly influence interpretations of science during the risk characterization phase. This may be important if there are likely to be legal challenges to the analyses that form the basis of a risk management decigement decision. Such challenges must be taken very seriously under U.K. environmental law. This also allows for development of expertise in a particular discipline; provides for consistency over time and across assessments, and permits development of institutional memory that can be applied to subsequent cases. Additionally, an advantage for this approach is that it permits the government to complete complex reviews within the 90 day time frame allowed by law. As is parallel processing for computers, there is no waiting for one task to be completed, before any new task can be taken on.

Each team member, having a designated area of expertise for the case at hand, does as much as possible for one case and moves on to another case or to some other assignment. The review components are scheduled to be completed at specified times during the risk analysis. Occasionally, the technical integrators may have to wait a bit for some pieces to come together, but they are not idle, usually being able to work conc work concurrently on several cases. Thus each assigned case can be completed within 90 days provided all needed data are supplied in a submission. This parallel approach is mostly used for full scale assessments that are needed for products nearing commercialization. However, new regulations now provide shorter review periods for certain limited assessments.

Variations on the parallel approach

A comparable, but not identical, situation can be found in Canada. In that country there is a transitional approach taking place as new biotechnology regulations that went into effect in September, 1997, are implemented. Both Health Canada and Environment Canada have responsibilities for biotechnology under the Canadian Environmental Protection Act (CEPA). Thus, unlike the U.K. EPA, which does both health and environmental risk characterizations, two independent departments must coordinate review in Canada. Notification goes to Environment Canada, but copies are sent to review groups ew groups in both departments for simultaneous independent review. At Environment Canada, a team approach will be used initially, as implementation of the new regulations begins in 1997, but it is hoped that a single reviewer will ultimately be able to perform the environmental reviews. If issues arise in either the health or environment review, lead reviewers meet to discuss control options.

CEPA, while similar in scope to the U.K. TSCA, differs from TSCA in that preestablished data requirements are listed in schedules (http://www.ec.gc.ca/cceb1/eng/nsneregejan1597b.html#SCHEDULE XV ) that vary based on the scope and location of intended manufacture or import. In addition, CEPA regulations cover microorganisms and organisms other that microorganisms and are not limited to genetically modified organisms, like the U.K. TSCA. Also, rather than a single 90 day period, there are different time limits for rmits for reviews that take place at certain stages of product development. For example , for those microorganisms introduced into a site for experimental field study (100 ha maximum), notification of the government is required 90 days prior to release; for those rereleased into the original site of isolation there is a 30 day review waiting period; most other releases require 120 days before release may take place. This differentiation of review intensity reflects an understanding that most biotechnology products will undergo stages of development beginning at the laboratory and moving through field research to full scale marketing. The Canadian regulations thus appear to be designed so that reviewers will have an opportunity to see the data on a released organism at each of the major stages of development. Thus, an iterative review process is implied by the nature of data requirements and review times specified in regulations.

Even situations that don=t requONT>t require full scale analysis can utilize parallel, team based approaches.

Using a different set of laws (the Federal Insecticide Fungicide and Rodenticide Act [FIFRA] and the Food Quality Protection Act [FQPA]) (http://www.epa.gov/pesticides/regleg.htm ) the US EPA also regulates pesticide products in an analogous manner. Pesticide reviews are not limited to genetically engineered microorganisms, as are TSCA reviews, and may include both microorganisms and plants. The relevant statutes permit the establishment of specific initial data requirements, whereas TSCA does not allow for these, and there are not the statutory time frames imposed by TSCA. The government has some self-imposed time constraints to ensure reasonable responses to its regulatory customers. Initial phases of the pesticide registration process (Notifications) are usually dealt with by a single reviewer, since the data requirements are limited and the focus of an assessment is on proper design of initi of initial field research. Later stages (Experimental Use Permits and Registration actions) generally involve team based reviews comparable in scope to those employed for TSCA. The components of the risk assessments are basically the same for each, even though the assessments are carried out by personnel in different offices working under different laws. However, since there are to published data requirements for pesticidal products, some components of the assessment may not require specific tests to be submitted and analyzed if they are not appropriate. http://www.epa.gov/oppfod01/oppinfo/

In both the U.K. and Canada, for organisms that may be released to the environment when used in biotechnology, consideration of risk management is withheld until a completed risk assessment (risk characterization) is available.

 

Sequential Analysis

Contrasting somewhat with the approaches used in North Ameri North America, the UK uses a sequential approach. The components of risk assessment are comparable to those in risk assessment that use a parallel approach. The differences lie in the order in which the components are considered and whether all elements are utilized in a formal sense to reach a decision. The outcome of these approaches should be the same, whether done sequentially or in parallel.

In this approach, some elements of risk characterization are given priority, and if the results of such analysis indicate that risk may be too high, measures for reduction of risk are immediately considered, rather than waiting for all elements of the risk characterization to be completed. Thus, risk management is considered at the same time as risk characterization takes place.

This system, like all others currently in use, considers risk first and determine actions that may be necessary based on risk. This results in the following iterative approach:

 

Fra"b16">Framework for Risk Management (U.K.)

 

Rationale for sequential review

Advantages to this assessment process are that non-essential components can be identified and eliminated from review, or limited in scope more easily than a parallel assessment. Concomitantly, this approach gets quickly to the critical considerations ofconsiderations of a review. It may simplify the assessment and shorten review time. More likely, it will quickly identify key areas that require additional data and facilitate the iterative aspects of the review process. Fewer people need be assigned under these circumstances; only the experts needed to work on key concerns will be utilized. However, it is not as easy to document, fully, the basis for a decision, since some elements may never get considered. Should an additional concern, not considered early on, crop up late in an assessment, considerable backtracking may be needed, and additional iterations may have to be introduced.

 

Variation on sequential analysis

The Canadian Food Inspection Agency (CFA) regulates released organisms that may be classified as biofertilizers under the Fertilizer Act (http://www.cfia-acia.agr.ca/english/actsregs/fert/fertrege.html) Product safety assessments include a screening stage aEM> stage and a follow-up stage. CFA uses a form of sequential analyses by first assigning the review to a Biotechnology Officer at the screening stage. Three questions are considered to determine the level of assessment required for a specific product:

(1)Is this product similar to other products (familiar) that are already in the market?

(2)Is the product substantially equivalent to other products that have already been approved?

(3)For familiar and substantially equivalent products, does the product (or commodity) require regulation under existing legislation?

Depending upon answers to these questions and initial review of data, the Biotechnology Officer may determine that certain aspects of human or environmental safety require more investigation or substantiation. In this case, a preliminary data requirements package will be identified, and communicated to the applicant by the Biotechnology Officer. Applicants are expected to submit responses to the requirements, requirements, which may include surrogate data or requests for waivers. The follow-up assessment is coordinated by the Biotechnology Officer. Health Canada and Environment Canada are consulted as needed (e.g.where residues in food or feed crops are anticipated). The Biotechnology Officer at CFA is responsible for assimilating the individual aspects of the assessments (e.g. health safety, environmental safety, exposure), making a determination on the safety of the product, and communicating the decision to the applicant. Revisions of safety assessments are conducted when an applicant intends to amend a pattern of use or formulation, or when significant new information becomes available. Thus, there is an iterative sense to the assessment process for Canadian biofertilizers.

 

Advisory Committees/Peer Review

Regardless of the wealth a country, few governments can afford to keep a permanent staff of trained experts in all disciplines who can address al address all the issues that might arise in every case of environmental release of non-indigenous organisms. Even the most sophisticated government agencies rely on experts outside government to assist on the more complex cases that they face. The manner in which experts are employed can vary significantly form one location to another, but the concept of using the best available expertise is a key to good risk assessment practice.

Where scientists are employed directly by a government, such as in the U.K., experts are often used to provide formal, or informal, peer review functions. For example, as the U.K. Environmental Protection Agency began to develop its regulatory framework for biotechnology product review, it established a Biotechnology Scientific Advisory Committee (BSAC) to assist it. This committee, comprised primarily of academic research scientists, with some assistance from government scientists not directly associated with regulatory affairs, provided guidance on overarching issues that ues that could affect whole categories of assessments, and it also reviewed analyses of specific product applications provided by government staff scientists. In the latter case, the BSAC served initially as an informal peer review committee. While a law, the Federal Advisory Committee Act, requires formal chartering of most committees whose purpose is to advise the US government, there are sunset provisions that require renewal of such charters. The BSAC was seen as a committee that could give advice on scientific matters, but had no power to make decisions on behalf of the government. After expiration of the charter of the BSAC, the peer review function was assigned to a different standing advisory committee. Since, these committees are purely advisory, the government is not compelled to act on their advice. Also, most members of such committees come from academic research backgrounds, and they may be prone to raising issues based on intellectual curiosity, which sometimes is not directly pertinent to decis to decisions that the government must make based on law and regulation. Many regulatory decisions must be made based on legitimate non-scientific considerations, and these committees need to be insulated from such influences to ensure the best quality scientific advice. In most cases, the advisory committees have confirmed that government scientists were able to perform risk assessments in an excellent manner, but occasionally, members have raised issues that required some additional consideration by the government.

A somewhat different approach provides a stronger advisory committee which can provide advice that is much closer to a regulatory function than is true in the U.K. In the U.K. the Advisory Committee on Releases to the Environment (ACRE), has been formed to give advice to various ministries, on the environmental and human health safety issues relating to the release of genetically modified organisms into the environment (http://www.shef.ac.uk/~doe/). ACRE is aACRE is a statutory, independent expert committee which is appointed by the Secretary of State under the Environment Protection Act 1990. Many countries in Latin America have established committees using panels of expert from government and academic institutions. In Mexico, the National Committee for Agricultural Biosafety analyzes all requests for importation, interstate transport, controlled release and deregulation of transgenic plant varieties or products. The advantages of the advisory committee approach are that it utilizes existing authority, it may be the easiest form of regulatory mechanism to implement, and because the majority of the group members may be research scientists there is a certain measure of credibility attached. There is, of course, an expense for operating the committees. In most cases, the members are volunteers using their own time to serve. This may be a disadvantage because the members may choose to take more time to review a case than would otherwise occur if government vernment scientists did the review under enforceable regulatory time constraints. Commercial applicants often criticize the use of advisory committees for providing unnecessary A delays when the lack of time constraint allows protracted review without firm decisions. Finally, in some sense, governments abdicate some responsibility when placing decisions in the hands of appointed committees. While committee members generally have expertise in fields relevant to the cases examined, they may not always have the ability to project their interpretations beyond the narrow focus of their individual disciplines. They frequently do not integrate their observations into the context of risk assessment paradigm described above (link to p. 3). This privatization of the assessment process may be difficult to reconcile with existing legal authority in a few cases, but may be welcomed by the public in others.

s.

Data availability

As indicated above, data used in reviews varies with the type of analysis and the laws and regulations governing a release. The wide nature of the environment is such as to make it extremely difficult to produce a >flow-scheme= for an assessment of risk. What almost all countries have done is to produce a series of questions regarding the organism, vectors and predecessors and the site into which it is to be released. These questions allow a case-by-case estimate of risk. The way countries gather data needed to answer these questions can be very different. Some, like the Canadian CEPA review or the U.K. pesticides review, have specified data requirements that must be met before a release is considered. Whether in the form of schedules of data (CEPA, http://www.ec.gc.ca/cceb1/eng/nsneregejan1597b.html#SCHEDULE XV ) or tier testing schemes (FIFRA), submitters must show that they have considered specific questions and gathered appropriate data. Others, like the U.K. TSCA biotechnology review, provide non-binding guidance for submitters. The US EPA uses an informal question set in the form of points to consider (http://www.epa.gov/opptintr/biotech/ptcbio.htm). Submitters are encouraged to review the points to consider to determine which ones are relevant to their particular case, since no single case will have features that demand evaluation of all of the points listed. During TSCA review, if some issues have not been addressed, the US EPA can contact the submitter and ask for additional data, suspending the review period if the submitter agrees, which generally occurs. For those countries whose review approaches are iterative, determination of data needs may be left entirely to the noilier. Only aftenly after initial review takes place, does a case specific data set evolve. Later iterations refine the data needs until all questions have been satisfactorily answered.

A firm set of data requirements is generally preferred by most commercial concerns, because they can use these as a predictive guide to both cost and time for producing a product. However, no company wants to see an overly broad set of requirements that ask for information not relevant to the products they make.

Most governments prefer flexible data requirements for biotechnology and biosafety, because there are too many variables associated with environmental release to permit predicting which data will be critical for every case. However, if data requirements are too loosely drawn, or do not exist at all, this causes problems for the government evaluators because most initial submissions will almost invariably lack some key data element. Eventually, experienced submitters learn to provide the data most relevant for their oror their organisms, but the learning curve is usually steep.

There is no perfect answer to how much data should be required to enable risk characterization of a released organism. Most governments have made it a policy to encourage first time applications to contact them well in advance of an intended release so that specific guidance can be provided.

Disclosure and Proprietary Information

Access to the process of decision making is not an automatic consideration. While many advisory committees are covered by so-called Sunshine Laws, that are intended to allow public participation, there are often limits to this access. Commercial concerns that are not covered by patent protection have a need to limit access to their research by their competitors, yet many research activities are covered by regulations that require that risk assessments be completed. Those who do the assessments have a responsibility to ensure that confidential information is legitimately protected. protected. One approach to this problem is for review authorities to limit the kinds of information that can be withheld from public scrutiny. Usually these limits cover such things as information on product identity, intended production volume, methods and sites of production, methods of application and other items that affect commercialization. Data on potential health and environmental effects are usually thought to be subject to public access. While some data is often withheld during the period of review, once an analysis is complete and a decision is made, public access to the decision process is not assured. Private concerns guard their intentions very closely until a product has successfully been introduced. Thus, requests for most field testing of commercial products may reasonably be expected to be accompanied by requests for limited access to supplied data.

This approach, however, can lead to suspicion among certain sectors of the public, and may be counterproductive in cative in cases where there is little concern. The U.K. Environmental protection Agency encourages is biotechnology applicants to consider holding public meetings at field test sites in order to reassure local authorities and private citizens. Withholding data may make the public fear far worse consequences than are noted by risk reviewers.

Summary

Risk assessment methods need not be identical in every case for the assessments to be effective. As long as the general components comprising a risk assessment are considered at some stage in a review process, most reviewers will reach the same conclusions for identical cases. Complex assessments can be expected to be required for cases that involve first-time applications or organisms never previously reviewed. The use of generic assessments, to consider certain features of a group of organisms that do not vary with the specifics of a case, is one approach that can result in simplified assessments for cases that involve tt involve these organisms. Performing analyses in parallel can save considerable time, but may require a resource level beyond the capabilities of many less-sophisticated review establishments. Sequential reviews may work better where experience staff are in short supply and can sometimes shorten and simplify a review process, provided all the key issues are identified early, and no revisiting of issues is required. Very specific, firmly formulated data requirements could provide good guidance for submitters, but the complex nature of environmental interactions of released organisms makes these very difficult to construct. Most reviewers prefer case specific data specifications developed after an initial consultation with the prospective submitter. Because each of these factors can influence the way the general risk assessment paradigm is applied in a specific case, no single, best method, can be said to exist. Nevertheless, aless, as long as reviewers keeps in mind that all assessments are variations on the theme, risk=f(hazard,exposure), risk characterizations for very similar cases done by different individuals and organizations should provide comparable results.

 US EPA .1997. Microbial Products of Biotechnology; Final Regulation Under the Toxic

Substances Control Act; Final Rule. Federal Register 62: 17909-17958. April 11, 1997

 US National Research Council. 1989. Field Testing Genetically Modified Organisms. National Academy Press p. 5