NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Institute of Medicine (US) Food Forum. Nutritional Risk Assessment: Perspectives, Methods, and Data Challenges, Workshop Summary. Washington (DC): National Academies Press (US); 2007.
Nutritional Risk Assessment: Perspectives, Methods, and Data Challenges, Workshop Summary.
Show detailsDuring the final session, moderated by Robert L. Buchanan, representatives of five sectors summarized key points from their perspectives, as follows: Catherine E. Woteki, industry; Barbara O. Schneeman, government; Gregory Paoli, risk assessment; Darlene Adkins, consumers; and Laina Busch, public health. The meeting ended with a lively open discussion involving many individuals, as did three of the earlier sessions. To capture the richness of the final presentations and of the open discussions throughout the meeting, this chapter organizes the points made under two major categories: the risk assessment process and consumer perspectives.
THE RISK ASSESSMENT PROCESS
Several participants described risk assessment as a systematic process for organizing and evaluating existing knowledge and its limitations. The workshop presenters pointed out the complexity of the process and issues of uncertainty that need to be addressed. Gregory Paoli emphasized that risk assessment is a tool that can be used to manage both uncertainty and complexity. Risk managers find the product of a well-conducted risk assessment useful in their decision making: the process lays out the evidence, the assumptions that have been made, and the limitations of the evidence.
The Dietary Reference Intake (DRI) process introduced the use of the risk assessment model for assessing nutritional risk in the United States and Canada. According to Barbara Schneeman, the work on DRIs demonstrated benefits to the use of the risk assessment model for nutrition and the potential value of improving the process. Catherine Woteki indicated that much can be learned from the DRI process, from recent applications of risk assessment to microbiological risks, and from work on the evidence-based review process—each of which evolved over the past decade.
Ways of Thinking About Risk
The workshop presenters examined aspects of different types of risk. Fundamental questions related to the use of nutritional risk assessment in setting DRIs include, How do you state the problem? and What are the endpoints? Among the possible endpoints of a nutritional risk assessment are the prevention of nutrient deficiency diseases, the maintenance of the body pool or of body stores of nutrients, the maintenance of body function, the reduction of risk factors for chronic diseases, chronic disease prevention, and the prevention of toxicity.
One challenge may be defining the context for risk: To whom does the risk or benefit apply? How do genotypes fit in? Another challenge is addressing the risk (and the benefit) for the population as a whole. Should a nutrient risk assessment weight a public health issue more heavily than a nutrient shortfall that is identified on the basis of a biochemical parameter? How does one risk trade off from another? Well-conducted nutritional risk assessments can provide key information to risk managers who need to develop policies related to food fortification and dietary supplements.
What are the challenges and opportunities for risk assessments that involve dietary patterns rather than a single nutrient? Alice Lichtenstein suggested that a model that captures dietary behavior combined with one or two validated biomarkers might be a useful approach to addressing the risks associated with dietary patterns.
Participants raised many other questions, such as the following: Does a probability of benefit (as is implied by health claims on food products) create other types of risk by implying that one food should be eaten in place of another? Instead of addressing each topic in isolation, how can the full range of options be understood and presented to enable the full impact of proposed policies to be considered by risk managers? Joseph Rodricks indicated that there are tools for examining the risks of excess together with the risks of inadequacy. Several participants emphasized that the process of finding answers for many of these questions is only in the early stage.
Considerable discussion focused on distinctions between risk–risk assessments and risk–benefit assessments. Barbara Petersen considered the differences in the two approaches to be semantic. A number of participants stated that a risk–risk approach is essential to nutrient risk assessment; others indicated that communication that uses a risk–risk terminology can be very problematic to consumers because consumers think in terms of risk and benefit.
Expanding Applications of Risk Assessment in Setting Dietary Reference Intakes
Suzanne Murphy challenged the group to consider the extent to which risk assessment approaches would be applicable to establishing nutrient intake recommendations. Is it appropriate to use one method to set Estimated Average Requirements (EARs) and an entirely different method to set Tolerable Upper Intake Levels (ULs)? In particular, could some aspects of models of risk assessment be used to set EARs, and could some aspects of setting requirements be used to set ULs? Barbara Schneeman and Gregory Paoli both noted that the approach used to set EARs incorporates elements of risk assessment. Schneeman pointed out that the provisional definition of nutritional risk developed by the Codex Committee for Nutrition and Foods for Special Dietary Use incorporates the probability of an adverse health effect associated with either the excessive or the inadequate intake of a nutrient or some other food component.
Considering the obstacles to studies involving high doses of substances, several participants pointed out the need for further work in developing models for estimating the distributions of risk. Richard Forshee suggested that priority be given to nutrients for which the difference between the amount specified for the Recommended Dietary Allowance and the amount specified for the UL is small. Sanford Miller suggested that Barbara Petersen’s Beneficial Utility Index model merits a closer look for use for nutritional risk assessment that addresses both low and high intakes of a nutrient.
Using Tools in Nutritional Risk Assessment
Among the nutritional risk assessment tools discussed were biomarkers, methods for collecting dietary data, and evidence-based reviews.
Biomarkers
The term biomarkers actually covers three concepts for different types of indicators: (1) biomarkers for nutrient intake, (2) biomarkers for susceptibility to disease, and (3) biomarkers of effect (biomarkers that indicate a causal relationship between intake of the nutrient and the occurrence of the disease).
Dietary Data Collection Methods
New tools are in development for the collection of accurate dietary data and their appropriate analysis. The U.S. Department of Agriculture has an ongoing program for updating and expanding food composition data. Considerable concern was raised that bias is likely to be present in estimates of dietary intakes (including supplement intakes) and that the bias could affect estimates of the percentage of the population with inadequate or excessive intakes. Major sources of bias include (1) the underreporting of dietary intakes and (2) overages in the labeled contents of supplements and fortified foods, especially for labile nutrients such as vitamin C. For both folate and vitamin A, estimates from U.S. dietary data indicate that large proportions of the population have inadequate intakes, but laboratory data do not support that finding.
From a different perspective, dietary data may be the most accurate data available to risk assessors. Information on the accuracy of the reporting of supplement use is not yet available. To understand dietary data better, Barbara Petersen suggested the examination of data from consumers with reported high nutrient intakes and the use of scaling methods once estimates of the extent of the bias are available.
Evidence-Based Reviews
Alice Lichtenstein expressed the view that finding ways to apply the evidence-based review approach to the DRI process would require considerable work. Challenges include data limitations and the extensive resources needed for such reviews. In the establishment of DRIs, a major role for evidence-based review may be to frame the nutritional risk assessment questions and identify whether the evidence for answering the questions exists. Evidence-based reviews set criteria in advance and clearly define the process for examining the scientific literature.
Using Data from Observational Studies Versus Data from Randomized Controlled Trials
The workshop speakers and other participants expressed different points of view regarding the value of data obtained from observational studies in comparison with data from randomized controlled trials. For example, Alice Lichtenstein noted the discordance between data from the two types of studies, whereas Julie Mares expressed the view that the interpretation of the observational data may have been in error. It is inappropriate to use findings from observational studies of food–disease relationships to draw conclusions regarding supplements, for example.
Addressing Uncertainty
The handling of uncertainty was a topic of considerable discussion among the workshop participants. Scientific uncertainty usually represents variability around a point estimate. Several participants stated that the uncertainty factors used in the DRI process to set ULs do not reflect scientific uncertainty. Instead, the uncertainty factors are more like safety factors, in that they consider the severity of the effect or they attempt to provide a way to extrapolate from one type of data to another. A number of participants expressed concern that bias and risk management decisions entered into the setting of the uncertainty factors. These participants called for a reexamination of the use of the uncertainty factor in setting ULs.
Several participants noted the importance of addressing uncertainty more explicitly with regard to setting both the EAR and the UL. Barbara Petersen’s method of expressing uncertainty by the addition of upper and lower limits to her efficacy and toxicity curves (Figure 3-4 in Chapter 3) serves as one example of how to do this. Joseph Rodricks mentioned that there are some formal tools for weighing uncertainties and the knowledge of adverse effects versus other considerations.
Maintaining Appropriate Separation Between Risk Assessment and Risk Management
Several participants underscored the importance of clear communication between risk assessors and risk managers at the problem formulation stage, before the risk assessment begins. Laina Bush argued that the economists and other social scientists involved in regulatory impact analysis should also be involved in the problem formulation step. Risk assessors need to understand what kinds of information the risk managers require for their decision making.
The development of probabilistic models of the distribution of risk instead of determination of a cutoff point might help address concerns about the inappropriate incorporation of risk management decisions into the risk assessment. Robert Buchanan suggested that the risk assessment entail a clear description of the relationship between exposure and risk and that a second group (the risk managers) decide the level of intake that is safe. That is, given a distribution of risk and other relevant information, the risk manager would decide the point on the distribution curve that should be targeted. The report Dietary Reference Intakes: Applications in Dietary Planning (IOM, 2003) gives an example of this approach with regard to deciding what point to target on the distribution around the EAR.
Catherine Woteki pointed out that federal agencies make risk management decisions using information provided by the DRI process, an approach that is somewhat analogous to the separation of risk management and risk assessment. Numerous participants emphasized that value judgments that occur in the risk assessment process need to be clearly identified.
Possible Next Steps
Workshop to Explore the Risk Assessment Model
Catherine Woteki suggested that holding a workshop of several days’ duration, as briefly described below, could serve as a useful next step:
- Conduct a trial risk assessment of a single nutrient.
- Select a nutrient and a risk assessment model or models in advance.
- Explore the application of the model(s) to the nutrient: define questions and endpoints, identify available and needed data, and identify the assumptions to be used.
- Evaluate the results.
- Determine data and resource needs.
- Address the appropriateness of the risk assessment model or models to nutrition in terms of biology and public health.
Several participants supported the potential value of such an approach.
Methodological Research and Data Collection
Methods Gregory Paoli emphasized that there is a great need for methodological research—that is, research on the mathematical properties of the problem conducted with hypothetical data. Such research would allow one to learn, for example, how important it is to have data on the standard deviation and the tails of the distribution. Are such data really needed to obtain a robust estimate of risk? Risk assessors can use this knowledge to guide the selection of the data to be collected. Economists have developed methods that can be used to compare different outcomes, and these could be informative for the nutritional risk assessment process.
Other potentially useful methods include (1) the exploratory analysis of large data sets on supplements to search for toxic effects and (2) the use of pharmacokinetic or toxicokinetic modeling with animal studies and in vitro data. Joseph Rodricks stated that, in the chemical area, toxicokinetics can be especially helpful in providing a basis for methods to extrapolate across species and across doses and an understanding of variability in populations. Thus, he suggested that toxicokinetic methods be considered for use with nutritional risk assessment. Several participants emphasized the importance of developing and using animal models more extensively, with the ferret–carotene model given as an example of how one can obtain valuable information on dose–response. The use of such models would require research on methods for the extrapolation of the findings from those studies to humans.
The Beneficial Utility Index described by Barbara Petersen was viewed as a promising new model by some, especially if it can adequately incorporate susceptibility, duration of exposure, and delayed response. Joseph Rodricks suggested the exploration of methods that consider quality-of-life factors, such as the disability-adjusted life years method, to find useful quantitative approaches to the comparison of different health endpoints in nutritional risk assessment.
Gregory Paoli pointed out that, in other fields of study, considerable time and effort are expended on characterizing and testing risk assessment methods before they are used as a basis for decision making. Several participants called for the development of more robust methods before a crisis arises.
Data collection Several presenters and other participants noted the urgent need for data, including data that could be useful for methodological research. Much emphasis was placed on the need for dose–response data. Dietary Reference Intakes Research Synthesis, Workshop Summary (IOM, 2007), is a useful source of information on data needs.
Some participants called for the purposeful collection of data on the adverse effects of nutrients and other food components and for investigators to build plans for the collections of such data into their applications for federal funding. Robert Buchanan suggested a two-part plan: (1) the development of a framework that would allow data from different studies to be combined for analytical purposes and (2) action to convince a group of researchers to collect their data using that framework. A challenge in grant applications would be the incorporation of a consistent framework for data collection while leaving room for individual creativity in experimental design. Consideration would need to be given to the length of the observation period needed to obtain an accurate assessment of adverse events.
CONSUMER PERSPECTIVES
Darlene Adkins expanded on comments made during the discussions, underscoring the point that uncertainties and complexities that challenge risk assessors and managers lead to challenges to consumers as well. Consumers find it difficult to make sense of the glut of nutrition information that they receive—information that comes from a very wide range of sources, with much of it unsolicited. Kathy Wiemer, for example, noted that even nutrition professionals tend to perceive, incorrectly, that any intake that is above the UL is dangerous. John Milner and Darlene Adkins pointed out that individuals vary in their responsiveness to the benefits or the risks from dietary changes and that this needs to be considered in consumer education efforts.
Communicators need to provide consumers with messages that are credible, reliable, objective, and unbiased. Ways need to be found to present the complexities in an understandable manner—one that allows the consumer to make sound decisions that consider the facts in the light of the consumer’s own situation. In particular, Darlene Adkins called for a reassessment of the nutrition label, an exploration of how consumers can be equipped to handle more complete nutrition information, and ways to provide useful nutrition information about food consumed at food establishments. To accomplish these tasks more effectively, both Darlene Adkins and Laina Bush called for research on consumer behavior and factors that influence their decision making.
References
- IOM (Institute of Medicine). Dietary Reference Intakes: Applications in dietary planning. Washington, DC: The National Academies Press; 2003. [PubMed: 25057648]
- IOM. Dietary Reference Intakes Research Synthesis: Workshop summary. Washington, DC: The National Academies Press; 2007.
Footnotes
- 1
Special thanks were given to Kathleen Koehler of the Office of the Assistant Secretary for Planning and Evaluation of the U.S. Department of Health and Human Services, Marianne Miliotis of the Food and Drug Administration, Michael D. McElvaine of the Office of Risk Assessment and Cost–Benefit Analysis of the U.S. Department of Agriculture, Suzanne Harris and Stephen Olin of the International Life Sciences Institute, and Maria Oria of the Institute of Medicine’s Food and Nutrition Board for their work in helping to organize the workshop
- PubMedLinks to PubMed
- Perspectives on Challenges and Solutions: Summary Remarks and Suggested Next Ste...Perspectives on Challenges and Solutions: Summary Remarks and Suggested Next Steps - Nutritional Risk Assessment
Your browsing activity is empty.
Activity recording is turned off.
See more...