The Consultation recognized that increased scientific, legal and political demands are being made on the standards, guidelines and other recommendations elaborated by Codex. This is, in part, due to:
- increased consumer interest in food safety
- the WTO's SPS and TBT Agreements
- harmonization initiatives
- calls for increased scientific rigour
- the need for transparency
- shrinking national regulatory resources.
To respond to these increasing demands, the greater application of risk assessment in the Codex decision-making process is essential. However, the generic framework for risk assessment described here represents a structural goal and may not be able to be fully utilized when adopting a risk assessment approach for some classes of hazards in food, such as biological hazards. In this respect, the Consultation recognized that Codex must be "technology forcing" if necessary data are to be developed.
An important principle is the functional separation of risk assessment from risk management. However, certain interactive elements are essential for a systematic risk assessment process. These elements may include ranking of hazards in the hazard identification step and risk assessment policy issues. Where risk management issues may intrude in risk assessment, the decision-making process should be transparent.
The broad mandate of a risk assessment approach to food safety encompasses a range of activities in addition to elaboration of standards, guidelines and other recommendations. Examples are the design of import and export inspection systems, acceptance/rejection criteria for foods, monitoring and surveillance programmes, development of information needed to formulate efficacious management strategies, and the overall allocation of food safety regulatory resources proportional to all classes of hazards in food. In the future strategic plan for utilization of risk assessment, this broad mandate should be addressed whenever appropriate in the Codex system.
Finally, the Consultation recognized that additional consultations would need to be convened regarding specific issues in risk assessment, as well as generally addressing the topics of risk management and risk communication. Nevertheless, the Consultation concluded that implementation of its recommendations would contribute significantly to the ability of Codex to meet its responsibilities of protecting consumers and facilitating international trade in food in a more consistent and open manner.
8.1.1 Scientific risk assessment should be the basis for Codex risk management decisions involving health and safety aspects of food standards. An important principle in this regard is the functional separation of risk assessment from risk management, while recognizing the interactive elements that are essential for a pragmatic risk analysis approach.
8.1.2 In regard to chemical hazards, the Codex should assure harmonized approaches to the risk assessment of food additives, contaminants, and residues of pesticides and veterinary drugs, particularly in the assessment of exposure.
8.1.3 Codex should encourage the development of risk assessment for biological hazards with the recognition that scientific understanding and knowledge are not currently adequate to quantitatively assess risk in most instances.
8.1.4 To meet its obligations under the SPS Agreement, Codex must be "technology forcing" to develop the necessary scientific information for the risk assessment of chemical and biological hazards.
Risk assessment of chemical hazards in foods usually results in the selection of risk management options to ensure that foodborne risks to consumers are not appreciable (notionally "zero"). This approach to food safety needs to be carefully examined with respect to the intent of the SPS Agreement and the concepts of "acceptable risk" and equivalence. For chemical hazards not evaluated by means of an authentic quantitative risk assessment model, issues of equivalence may take the form of a comparison of equivalent margins of safety above this notionally "zero" risk baseline.
8.2.1 Exposure assessments for food additives, contaminants, and pesticide and veterinary drug residues should be considered an integral part of the Codex risk assessment procedure for these substances. Because this is primarily a scientific task, exposure assessments should continue to be carried out by the JECFA/JMPR. Where necessary, exposure assessment should be expanded to take into account differences in dietary patterns, both within and between countries, and include estimates of intake by especially vulnerable groups.
8.2.2 CAC should request all Member States to make available dietary exposure data, including information on levels of chemicals in various foods and intakes of those foods by their population. Where such information is not available, CAC should encourage countries to develop appropriate food contamination monitoring programmes consistent with national priorities and to obtain dietary intake information for the general population and, if feasible, sub-groups of interest.
8.2.3 The methodology and guidelines currently used for predicting the dietary intake of pesticide residues should be reviewed with a view to obtaining more accurate estimates of human exposure.
8.2.4 Consideration of exposure scenarios for acute and chronic adverse health effects should be applied wherever appropriate in risk assessment of residues of veterinary drugs.
8.2.5 For contaminants for which adequate data exist and no threshold for adverse health effects can be established, JECFA should be requested to provide a quantitative estimate of the health risks associated with specified levels of intake, including attendant uncertainty.
8.2.6 The process by which the JMPR derives MRLs should be made more transparent.
8.2.7 To promote the transparency and credibility of the risk assessment process and to facilitate review of the proceedings when necessary, it is recommended that decisions be thoroughly documented and that all significant supporting data and other information be archived. This information should be available to Member States and to appropriate oversight international organizations.
8.2.8 To promote the quality and consistency of toxicological and other data, FAO and WHO should encourage the use of standardized test protocols and minimum data requirements that have been or will be recommended by recognized international expert groups.
8.2.9 WHO should review criteria for establishing safety factors, benchmark doses and generic cross-species scaling factors, taking into account the efforts that are being made by other international groups.
8.2.10 Scientific data are often necessary to depart from default assumptions, particularly in the evaluation of carcinogenic risk. Codex should encourage efforts to develop scientific criteria to help resolve differences in the requirements for and interpretations of such data.
Biological hazards in foods continue to pose significant risk and have a high profile internationally. There is a need to reduce the risk to the minimum which is technically feasible and practical. International standards and guidelines developed to address risk reduction must be transparent and outcome oriented. To facilitate this, risk assessment techniques must be applied to determine the significance of hazards and be used as a tool to evaluate risk management strategies such as HACCP. The utilization of risk assessment techniques to provide an estimate of potential adverse health effects will be an essential component of the process for establishing international trade policies. Additional information must be developed in order to facilitate quantitative approaches to risk assessment for biological agents. At present however, a qualitative approach is the only one available for accomplishing such a risk assessment. Therefore, the CAC should produce an overall strategy and implementation plan to address the following recommendations.
8.3.1 The standards, processes and procedures relating to biological hazards and contained within Codex standards and codes of practice should be based on sound science and quantitative risk assessment to the maximum extent possible. This would imply an analysis of individual biological hazards in a broad range of foods, rather than the study of multiple risks associated with single foods.
8.3.2 Where Codex produces standards or codes of practice that contain processes and procedures, the intended outcome of the processes or procedures in terms of food safety should be clearly stated.
8.3.3 Guidance should be provided to enable assessment of equivalence of alternate processes or procedures that meet the intended outcome.
8.3.4 Consideration should be given to a means of comparing relative risk of different options being considered to control a hazard. Overall minimization of risk of adverse effects should be the goal and CAC should consider not only relative risks of biological origin but all potential risks.
8.3.5 A review should be undertaken of the risk analyses implied by the use of 2- and 3- class sampling plans for microbiological end-product specifications, especially in the light of the wider use of HACCP based systems and the improved process controls which result.
8.3.6 Specific research directed towards identifying and characterizing biological hazards of concern should be encouraged by the CAC to enable more quantitative risk assessment.
8.3.7 Quantitative methods of risk assessment should be developed for biological hazards to facilitate and improve the application of HACCP.
Many sources of both uncertainty and variability exist in the process of risk assessment of foodborne hazards to human health. Explicit consideration should be given to uncertainty and variability in the risk assessment process so that these may be taken into account in the formulation of risk management policies.
8.4.1 The limitations of the risk characterization methods presented here make clear that risk managers should be aware of the uncertainty in risk estimates and include this awareness in their decisions and their communications of risk to the public.
8.4.2 This situation suggests the need to consider carefully the uncertainties of model assumptions and inputs so that effort is directed at those components having the largest contribution to overall variance in model predictions.