Health Canada Decision-Making Framework for Identifying, Assessing, and Managing Health Risks - August 1, 2000

Previous       Table of Contents            Next

2.2 Assess Risks and Benefits

Assess risks using biological, chemical, and physical data from scientific studies; integrate information related to risk factors (e.g. social, cultural, ethical considerations, economic status), and risk perceptions, where this information is demonstrated to have an impact on the level of risk. Assess benefits in a similar manner.

This step involves assessing the health risks (both known and potential) that may result from exposure to a specific agent.. Where appropriate, such as in the evaluation of a therapeutic agent, the step also involves assessing the health benefits (known and potential) related to the agent, and examination of risks relative to benefits. Where possible, both risk and benefit assessment should be undertaken in a multi-disciplinary fashion, taking into account all available, scientifically credible information.

2.2.1 Assess Risks

Risk assessment must be conducted distinctly from other activities. Appropriate mechanisms must be in place to ensure that there is no interference with the scientific assessment of risk.

Taking a Broad Approach

Risk assessment involves determining the likelihood that a specific adverse health effect will occur in an individual or population, following exposure to a hazardous agent. This is typically accomplished by examining physical, chemical, and biological data obtained from scientific investigations, such as those conducted in laboratories (e.g. toxicology or microbiology studies), and those involving human populations when available (e.g. epidemiological investigations, clinical trials). Risk assessment involves recognizing that a hazard exists (hazard identification - is it harmful?), defining its characteristics (hazard characterization - how harmful is it?), considering the extent of exposure to the hazard (exposure assessment - what levels are humans exposed to?), and comparing current or predicted levels of exposure to a measure of the potential of the agent to induce adverse health effects (risk characterization, a summary and integration of the scientific analyses from the preceding tasks).

It is important to include all relevant scientific data in the assessment of health risks. Failure to evaluate all relevant data may limit the ability of the management team to identify and analyze an appropriate range of
potential risk management options, and to select the strategy that will be most effective, have the least unintended negative effects, and be undertaken at a reasonable cost.

The value of using a broad approach to risk assessment stems from the recognition that a variety of different factors or determinants may influence our health, in addition to the "physical" environment, both natural (air, water, food, soil) and human-built, and that health effects (known and potential) should be examined both directly and indirectly. It also involves considering the outcomes for specific populations in addition to risks to whole populations, including maximally exposed individuals. It further involves considering the perspectives and knowledge of a range of interested and affected parties to the extent possible and appropriate for a given risk situation.

Thus risk assessment involves examining and integrating information on risk factors (such as gender, age, ethnic origin, social situation, economic conditions, education, culture or personal convictions), when following critical examination, there is a demonstrated influence on the level and/or likelihood of risk for specific populations. Such an approach may be used for example, when determining different levels of exposure to food contaminants, which may result from different consumption patterns that occur due to social/cultural practices or economic status. It is important for Health Canada to acknowledge the influence of various risk factors on health, even if is ultimately decided that they are best addressed by other departments. In order to bring together all the relevant information, the risk assessment team may need to include experts from a variety of disciplines, the nature of which may vary from risk to risk. The extent to which a broad approach can be taken during risk assessment, may be limited by existing legislation.

The Link With Risk Management
Risk assessment is a key part of the decision-making process, not only because it provides an estimate of the level of risk, but because it can help to identify possible options for risk management. For example, examining information on a range of exposures and how changing the exposures would affect the level of risk, helps to identify and analyze potential risk management options and thereby contributes to policy development. While risk assessment must be conducted separately from risk management, in order to maintain scientific integrity, the two processes must be linked: risk management goals are used to focus risk assessments, while the results of risk assessment provide critical information for risk management.

Assess Risks - General Tasks

• Identify Hazards.
  
  
• Characterize Hazards.
  
  
• Assess Exposures.
  
  
• Characterize Risks.

Identify Hazards
Although hazards are identified in a preliminary way during issue identification, this is undertaken in more detail during risk assessment. Typical activities in the identification of hazards includes:

• identifying the agent(s) causing the adverse health effect(s);
  
  
• collecting relevant scientific data;
  
  
• determining the relative weight of studies having different results;
  
  
• determining the relative weight of different types of studies (e.g. epidemiology, toxicology);
  
  
• examination of the scientific data for evidence of a relationship between the agent(s) and the adverse health effect(s);
  
  
• identifying the mode and mechanism of action of the agent(s);
  
  
• identifying those dose levels that are, and are not, associated with adverse health effects (e.g. for toxicology studies, No Observed Adverse Effect Levels [NOAELs] or Lowest Observed Adverse
  Effect Levels [LOAELs]);
  
  
• determining the critical effects associated with exposure to the agent;
  
  
• determining the significance of a positive finding in studies having different routes of exposure compared to the population(s) at risk;
  
  
• deciding if the studies have any data limitations that might affect their interpretation or invalidate their results;
  
  
• for nonhuman studies, ensuring that adequate protocols, a sufficient number of animals, and appropriate dose levels have been used, and determine how different metabolic pathways or rates should be
  considered;
  
  
• considering sources of uncertainty and other limitations, and how may these impact upon the hazard identification;
  
  
• deciding the overall weight of evidence taking into account the quality of the data; and
  
  
• identifying the hazard(s) of concern.

Characterize Hazards
Hazard characterization is a process that involves qualitatively and/or quantitatively evaluating the adverse health effect(s) that humans may experience under expected levels of exposure to the agent(s) under study. Traditionally, hazard characterizations have focused on physical health effects, and have relied on data from toxicology and epidemiology studies and in some cases, from surveillance; more recently, emotional and mental health effects are starting to be explored. As scientific data are often incomplete or not available, estimations must often be supplemented with more qualitative approximations. Since most exposures tend to be at low, chronic, levels, hazard characterization often requires extrapolation of data from studies involving high level of exposure (i.e. exposure in occupational settings or in laboratory studies).

In order to characterize hazards it may be necessary to determine a number of factors, including:

• which critical health effects are associated with exposure to the agent;
  
  
• for which of these effects data are adequate to characterize exposure-response;
  
  
• what dose-response models should be used to extrapolate from observed to relevant doses (i.e. when the potency of the agent to induce effects does not fall within or near an observable range);
  
  
• how the dose-response relationship should be extrapolated (e.g. using best estimates or upper confidence limits);
  
  
• whether traditional data analysis should be used or whether an alternative approach should be used;
  
  
• whether there is a need to take into account interactions between agents, and if so how to do this;
  
  
• whether certain human populations are likely to be more sensitive to exposure then others (susceptible populations);
  
  
• how to deal with differences in exposures between study populations and the population for which risk estimates are required;
  
  
• how to deal with differences in physiological characteristics between study populations and the population for which risk estimates are required;
  
  
• for nonhuman studies, what mathematical models and assumptions to use to extrapolate results to humans;
  
  
• sources of uncertainty and other limitations, and how these may impact upon the hazard characterization;
  
  
• a threshold of exposure for the induction of the critical effect by the agent, taking into account the quality of the data; and
  
  
• the nature, severity, and reversibility of the known or potential adverse effects in humans at expected levels of exposure.

Assess Exposures
Exposure assessment is a process used to develop a qualitative and/or quantitative estimate of the magnitude, frequency, duration, route and extent of human exposure to an agent. In other words, the purpose of an exposure assessment is to calculate the dose of a hazardous agent to which one or more populations or subpopulations are exposed. This activity is key to the risk assessment process because without exposure there is no risk. Exposure assessment may include a number of the following steps.

• characterize the exposure pathway to the extent possible [see the Characterizing the Exposure Pathway section that follows);
  
  
• determine whether exposures are source-specific (e.g. for radiation), or medium-specific (e.g. for consumer products), from point or disperse sources, or whether a combination of sources and media
  are relevant;
  
  
• consider the physical and chemical properties of the agent;
  
  
• identify the location(s), point(s) of contact, and pattern(s) (e.g. seasonal) of exposure;
  
  
• determine how to estimate the size and nature of the populations likely to be exposed;
  
  
• determine whether certain segments of the population are exposed to the agent at higher levels than others;
  
  
• determine what method should be used to assess exposures (e.g. deterministic, probabilistic, scenarios;refer the Box that follows);
  
  
• examine exposure data when available (e.g. through monitoring);
  
  
• in cases where exposure data are not available, predict exposure based on data for related agents as well as on exposure simulations;
  
  
• determine how to extrapolate exposure measurements from the study population to the population(s) of interest;
  
  
• determine how to take into account various factors that may affect exposure, including the time and duration of exposure;
  
  
• if there is a need to consider interactions between agents, examine exposure for each of these agents;
  
  
• document sources of uncertainty and other limitations, and how may these impact upon the exposure assessment;
  
  
• determine the overall weight of evidence taking into account the quality of the data;
  
  
• estimate the likelihood of exposure; and
  
  
• estimate exposure levels.

Characterizing the Exposure Pathway
Before exposure can be assessed, it is necessary to characterize the exposure pathway, which describes how a hazardous agent reaches an individual or population. This involves obtaining information on: the source from which the agent originates; environmental media which carry the agent to individuals or populations of humans (e.g. food, air, water, soil, consumer products); the location, which is the point where contact between the agent and humans occurs (e.g. the home, workplace, recreational sites); the target population(s) or subpopulation(s), the people who are exposed to the agent (e.g. a swimmer who bathes in a contaminated river); and one or more route(s) of exposure, which are the means of entry into the human body. Examples of routes of exposure include: ingestion, which includes swallowing food, water, soil, and other substances; inhalation, which includes breathing in a gas, vapour or airborne particles; skin contact, which may involve corrosion caused by skin irritants or skin penetration by agents such as ionizing and nonionizing radiation; through the intravenous, intramuscular, intraperitoneal, subcutaneous, or intradermal routes, as in the case of drugs.

The Use of Modeling
For some agents, particularly those involving voluntary exposure, such as prescription drugs, exposure assessment is relatively straightforward. But for other agents, such as environmental or food contaminants, an exposure assessment is usually based on considerable uncertainties. It is often not possible to measure exposures directly; rather they must frequently be predicted, for example by using monitoring data and mathematical modelling and reconstructing historical exposure patterns.

There are two broad types of mathematical models used in exposure assessment: those that predict exposure to the agent, and those that predict the concentration of the agent. Exposure models can be used to estimate the exposures of populations based on small numbers of representative measurements. Models that predict concentration can be combined with information on human time-activity patterns to estimate exposures. Modeling may be done on long-term and short-term exposures, both of which have limitations. For example, in long-term exposure modeling, changes may occur in natural levels of exposure over time and in activity patterns of exposed persons; in short-term modeling, there are difficulties in modeling concentrations that vary widely over time.

As with modeling, extrapolation of results can lead to uncertainties in exposure assessments. Sometimes exposures of particular groups of individuals, such as occupational workers, are used to estimate exposures in other populations. Uncertainties may result from the extrapolation of data from high to low doses, because adverse effects observed at high doses may not be seen at lower ones. An important aspect of exposure assessment is to determine which groups in a population may be exposed, as well as which groups may be especially sensitive. Another concern is how to deal with the effects of exposure to multiple agents, which may have similar adverse health effects.

Examining Information on Risk Factors
A variety of risk factors can influence the level of exposure experienced by specific subpopulations. Where appropriate, information on social, cultural, ethical, economic, and other risk factors, as well as risk perceptions, must be collected and analyzed to determine how exposure may be affected. Information that meets an acceptable level of scientific rigor is then integrated with other exposure-related information to develop more comprehensive exposure estimates.

Characterize Risks

A sound risk management decision is based on a careful analysis of the weight of scientific evidence that supports conclusions about the risks of an agent to human health.

When characterizing risks, investigators determine whether exposure to a hazardous agent poses a significant risk to human health, by comparing information obtained through hazard characterization and exposure assessment. Risks are usually characterized in terms of their potency (for chemicals or radiation, in terms of a dose-response relationship), pathways of exposure, and reasons for variation in response among exposed populations. Risk characterization often involves no additional scientific information, but requires judgement, for example, when interpreting data related to population groups with varied sensitivity and different exposures.

To be truly useful, risk characterization must be accurate, balanced, and informative. This requires "getting the science right and getting the right science"[National Research Council, 1996]. It involves the use of reliable technical and scientific input from a range of disciplines, including biological, chemical, physical, economic, social, and behaviourial sciences; sound scientific analyses; and providing opportunities for discussion and deliberation, recognizing that this needs to be much more extensive in some situations than in others [National Research Council, 1996].

The success of risk characterization depends on conducting a systematic analysis that is appropriate to the issue, that carefully considers scientific uncertainties, related assumptions, and potential impacts on decisionmaking, and that responds to the health-related needs of interested and affected parties. Success also depends on discussions or deliberations that formulate the risk issue, guide analyses, seek the meaning of analytical findings and uncertainties, and improve the ability of interested and affected parties to understand and participate effectively in the decision-making process.

Requirements of Risk Characterization [U.S. NRC]

• Get the science right: Ensure that the underlying analysis meets high scientific standards in terms of measurement, analytic methods, databases used, plausibility of assumptions, and consideration of both the magnitude and the nature of uncertainty, taking into account limitations that may result from the level of effort expended on the analysis.
  
  
• Get the right science: Ensure that the analysis addresses the significant risk-related concerns of public officials and the spectrum of interested and affected parties. Set priorities for assessment so as to emphasize the issues most relevant to the decision.
  
  
• Get the right participation: Ensure that there is sufficiently broad participation so that important, decision-relevant information enters the process, that all important perspectives are considered, and that legitimate concerns about inclusiveness and openness are addressed.
  
  
• Get the participation right: Ensure that the process used for risk characterization satisfies both decision makers and interested and affected parties, and is responsive to their needs, to the extent possible. Ensure that the information, viewpoints, and concerns of all parties are adequately represented and taken into account, that parties are adequately consulted, and that their participation can potentially affect the way risk issues are defined and understood.
  
  
• Integrate information in accurate, balanced, way: Ensure that the risk characterization presents the state of knowledge, uncertainty, and disagreement about the risk situation, and reflects the range of relevant knowledge and perspectives. The risk characterization should strive to satisfy interested and affected parties that they have been adequately informed within the limits of available knowledge. It should also consider and reflect the limitations of scientific knowledge (e.g. various kinds of uncertainty).

[Source: Adapted from - National Research Council (1996). Understanding Risk: Informing Decisions in a Democratic Society. National Academy Press, Washington, D.C.]

Involving Other Technical Specialists, Policy Makers, and Interested and Affected Parties
Although scientists play the lead role in risk characterization, policy makers, other technical specialists, and interested and affected parties should also have opportunities for involvement. Risk characterizations provide a key source of information for risk management decision-making, and consequently play an important role in ensuring that risk management goals are met. Policy makers and interested and affected parties can help to ensure that the characterizations have focused on the correct risk issue and have answered the health-related questions of primary concern. Other technical specialists, particularly economists, can help to ensure that the characterizations provide the type of information that they need to perform further analyses (e.g. comparison of risks and benefits). The manner and extent of involvement will depend on many factors as noted in the Identify the Issue and Its Context section above [National Research Council, 1996].

A summary of some of the tasks involved in risk characterization follows.

Review the Hazard and Exposure Information
This involves examining, summarizing and integrating information obtained through hazard identification, hazard characterization, and exposure assessment. Among the factors to consider are the quality, completeness, and relevance of the information, and the nature and impact of uncertainties and other limitations related to the information and any analyses that are conducted.

Generate a Quantitative Estimate of the Risk
In order to produce a risk estimate, quantitative information on exposure (and if available, dose), from the exposure assessment, is combined with information on the dose-response relationship obtained through hazard characterization. The process of developing a quantitative risk estimate will differ, depending on the type of risks being considered - carcinogens and "noncarcinogens" (agents that do not cause cancer or for which there are insufficient data on carcinogenic potency), microbial pathogens, etc.

Consider Statistical and Biological Uncertainties and Their Impacts
Risk estimates often contain a some level of uncertainty. Uncertainties may result from: the limited availability of scientific data, on for example, exposure or intake rates; long time delays between exposure and effect; the need to extrapolate data to predict the health consequences of human exposures; difficulties in determining appropriate mathematical models for extrapolation; simultaneous exposures to a variety of different agents (making it difficult to determine the effects of a single agent); and judgements made at each step of the process.

It is important to consider the nature, sources, and levels of uncertainties related to the risk estimates, and how these may impact upon the risk assessment, and to document this information. It is also important to determine whether the uncertainties are "acceptable", or whether analyses need to be repeated using better data or better techniques in an attempt to reduce the uncertainties. Both uncertainty analyses and individuals' interpretations of what uncertainties mean, can be strongly affected by the social, cultural and institutional context of a decision.

Uncertainties that result from the incompleteness and unavailability of scientific data frequently require scientists to make inferences, assumptions, and judgements in order to characterize a risk. Making judgements about risk based on scientific information is called evaluating the weight of the evidence. Risk characterizations based on scientific data, should include not only plausible conclusions about the characteristics of the risk (based on available information), but also evaluations of the weight of evidence that support the conclusions, descriptions of major sources of uncertainty, and alternative views.

Uncertainties related to potential health effects, dose-response relationships, and exposure, have increasingly led to the use of a range or distribution of risk estimates rather than a single value. Single numerical estimates of risk can give the misimpression of precision, be easily misinterpreted and be misused in the absence of information which puts them into context. Using a distribution indicates the likely maximum and minimum risks for different individuals and the relative likelihood of intermediate risks between these extremes.

Generate a Qualitative Description of Uncertainty
This involves preparing a summary of the uncertainties that have been noted throughout the risk assessment process, and explaining the potential impacts of the uncertainties on the risk estimates in a nontechnical manner, which is understandable to the risk management team and to interested and affected parties. Among the general uncertainty issues to be addressed are the following:

• For what purpose was the assessment conducted and what are the potential implications of the results of the assessment?
  
  
• How much is known about the capacity of the agent to cause adverse health effects in laboratory animals (if relevant) and humans?
  
  
• How much is known about the biological mechanisms and dose-response relationships underlying any effects that are observed in the laboratory and/or in epidemiological studies?
  
  
• How much is known about the pathways, sources, patterns, and magnitudes of human exposure and number of persons likely to be exposed?
  
  
• How much is known about susceptible subgroups and their likelihood of exposure?
  
  
• What do other risk assessors, decision-makers, and interested and affected parties need to know about the primary conclusions and assumptions and about the balance between confidence and uncertainty in the assessment? What are the strengths and limitations of the assessment?

Dealing with Uncertainty - Some Health Canada Examples

The method for dealing with uncertainty depends on number of factors, including the nature of the agent being examined:

• For diseases, public health decisions are often based on the best available information, in consultation with appropriate stakeholders. Where possible, statistical inferences are used to assess uncertainty/confidence levels. In some cases, statistical re-sampling methods through simulation are used. In extremely difficult cases, scenario analysis combined with qualitative information may be used.
  
  
• For radiation, if the risk is significant, then the uncertainty provides a range for the estimated number of deaths/injuries due to the radiation exposure. In some cases, standard dose-response relationships are based on the mean value and ignore the uncertainty in the data. For practical purposes, advice is often based on the mean value of the risk, as long as the risk is significant.
  
  
• For Priority Substances (under the Canadian Environmental Protection Act), confidence and/or uncertainty in a data set are reflected in the manner in which or the extent to which the data are used. Qualitative statements concerning uncertainty are always included; where data permit, uncertainty and variability are characterized quantitatively.
  
  
• For food additives, uncertainty/confidence level in data are considered through the use of appropriate safety factors or mathematical models. Equally important is the nutritional value of the
  food.

Identify Which Population Group(s) Should Be the Primary Target of Risk Management Efforts
This involves determining which population or populations are at greatest risk (known or potential) and thus which should be the focus of risk management efforts.

Perform a Risk Comparison
Risk characterizations often include some form of risk comparison, which is a way to combine frequency estimations with some estimates of the significance (or severity) of the health effects. Two increasingly common methods used to compare risks are risk ranking and risk prioritization. Risk ranking is useful for comparing hazards that cause a similar effect in a single medium, such as carcinogens found in drinking water. Risk prioritization involves using specific criteria, such as the exposure levels compared to the potency to induce cancer, to determine the priority for action.

Examine the Weight of Evidence
This involves determining and examining the weight of the scientific evidence, in a qualitative way, order to determine whether there is support for the conclusions about risk. It may also involve: determining whether other agents might cause the same type of effects; examining the contribution that a particular agent makes, relative to those having similar types of effects in the affected population(s), or subpopulation(s); determining how the risk is distributed in relation to other risks to which the affected population(s), or subpopulation(s) are exposed; and examining the effects of risk interactions (combined exposure to two or more agents or conditions, such as immune status, genetic risk factors).

Determine Whether Additional Data Must Be Collected
If the data and methods used for analysis are not adequate based, for example, on scientific standards, or if no evidence exists (e.g. there is no statistical significance), it may be necessary to conduct additional studies or repeat the analysis using different methods or data. Results of the analysis may reveal that additional information must be collected to properly address the issue (this does not however, preclude use of a precautionary approach, involving implementation of an interim risk management strategy while further data are collected). If peer review is required, it is necessary to identify the reviewers, and then to obtain and consider their comments. In cases where there are legislated timeframes for completion of assessments, as for Priority Substances under CEPA, it is often not possible to collect additional data or repeat analyses; in such cases it is usually indicated that better data might help to reduce uncertainty.

Present the Risk Assessment to the Risk Management Team
Risk assessments may be presented using a variety of methods; the choice of method may be a function of the legislative mandate. It may be useful to provide a table indicating the estimated level of risk for the exposed population by route of exposure, as well as a full characterization of the risk, including a discussion of uncertainties, a discussion of the comparability and consistency of similar but different risks (e.g. for the average individual versus the most exposed individual), and the extent to which professional judgements have been used to deal with sources of uncertainty and their potential impacts. Risk assessment should also be made available to interested and affected parties, taking into account the need to keep some information confidential (e.g. drug formularies).

2.2.2 Assess Benefits

The inclusion of benefit assessment (and consequently the comparison of risks and benefits) as part of the decision-making framework, is not intended to imply that benefits (known or potential) must be assessed
in every situation, but rather that it should be undertaken in a consistent and systematic manner in situations where it is appropriate to do so.

In general, benefit assessment should be attempted when it is difficult or impossible for consumers to judge the benefits associated with exposure to an agent and to compare them with the associated risks. For example, it is often necessary to evaluate the benefits of a specific product (e.g. a drug or medical device), when a claim is made that a product improves health, in order to put the risk associated with that product into the proper context of overall health. There are however, instances where benefit assessment is not necessary or possible, such as where the level of risk is deemed to be minimal or "de minimus", where it is not ethical to consider benefits because it might imply that a product is being endorsed, or where legislative mandate does not allow benefits to be assessed.

In cases where it is appropriate to compare risks and benefits, the comparison should be done using a societal perspective, unless dealing with a situation in which only an individual is affected (e.g. special release of an unapproved drug). A population or sub-population should not be placed at risk for the benefit of others.

Technical specialists (in this case, economists) play the lead role in benefit assessment and in making risk/benefit comparisons. However, there is a role for other participants to play, including scientists responsible for the risk assessment, policy makers, and interested and affected parties. Like risk assessments, benefit assessments and risk/benefit comparisons provide a key source of information for risk management decision-making, and consequently play an important role in ensuring that risk management goals are met. Policy makers and interested and affected parties can help to ensure that assessments are focused on the benefits of most relevance, and that appropriate consideration is given to specific populations and equity issues. Other technical specialists, particularly scientists, can provide guidance in the use of risk assessment results in risk/benefit comparisons, and can flag additional risk information needs. The manner and extent of involvement will depend on many factors as noted in the Identify the Issue and Its Context section above.

Assess Benefits - General Tasks

• Collect and Assess Information on Benefits.
  
  
• Prepare a Risk/Benefit Comparison.

Brief descriptions of these tasks as well as some related considerations are provided below. Further information related to benefit assessment may be found in the draft Guidance Document on Socioeconomic Analysis.

Collect and Assess Information on Benefits
In order to assess benefits (known or potential), specialists:

• identify the type(s) of benefits to be examined;
  
  
• identify the measures to be used;
  
  
• collect and analyze the benefit information;
  
  
• determine how to deal with uncertainty; and
  
  
• summarize the benefit information.

Identify the Type(s) of Benefits to be Examined
The first task involved in benefit assessment is to determine what types of benefits are to be examined.

These may include direct health benefits (e.g. relief of disease symptoms), or indirect health benefits (e.g. economic, social, or cultural impacts). An important part of this determination is identifying the perspective to be used for the analysis (e.g. specific interested and affected parties), and the nature and size of the population(s) that would benefit.

Identify the Measures to Be Used
Once the type(s) of benefits have been determined, it is necessary to identify the measures for the benefit assessment, and for reporting of the results (e.g. effectiveness, efficiency, quality of life, dollar values). Like risks, benefits may be assessed qualitatively or quantitatively, depending on the nature of the information available. The extent to which benefits are assessed, as well as the specific considerations taken into account, vary depend upon factors such as the issue being addressed, the context in which it is being considered, and the nature and amount of information that is available. Once measures have been identified, it is necessary to select the methodology to be used (e.g cost-benefit analysis), as well as any modeling techniques to be used.

Collect and Analyze the Benefit Information
This task involves collecting and analysing the benefit-related information. Information may be collected through various means, for example through socioeconomic analysis, or for therapeutic products, through the results of clinical trials. One of the first things to examine once the information has been collected and analyzed, is the adequacy of the data and methods used for the analyses, as well as whether the analyses have addressed the appropriate concerns. If the data or methods are not of high quality or are not relevant, it may be necessary to conduct other studies or reanalyze the data. Another item for consideration is whether any analyses should be reviewed by third-party experts, and if so, who the third parties should be.

Determine How to Deal with Uncertainty
As with risk assessments, benefit assessments are frequently subject to uncertainty. Given this, it is important to identify the nature, sources and level of uncertainty, both in terms of the benefit data themselves and in terms of the analyses that are conducted. As well, it is important to determine the potential impacts that the uncertainty will have on the benefit assessment. If the level of uncertainty is not acceptable, it may be repeat the analyses using better data or better techniques.

Summarize the Benefit Information
The final task in benefit assessment involves summarizing and integrating the information in a fair and balanced manner, similar to what is undertaken during risk characterization. The resulting benefit assessment summary should include any assumptions, uncertainties, and judgements, and should be written in a nontechnical format, suitable not only for risk managers, but for interested and affected parties.

Prepare a Risk/Benefit Comparison
In order to complete the risk/benefit assessment, specialists:

• examine risk and benefit data; and
  
  
• present the risk/benefit comparison to risk managers.

Examine Risk and Benefit Data
This involves integrating, analysing, and comparing the results of the risk and benefit assessments. Risks, benefits, and any associated costs must be evaluated in terms of the needs, issues, and concerns of interested and affected parties. In the case of therapeutic products, such as drugs, the risk-benefit profile of the agent may be compared with that of alternative therapeutic agents.

When comparing risks, benefits, and costs, consideration can be given to individual versus collective risk and benefits, who benefits relative to who bears the risk (as different parties may be involved), and freedom of choice versus risks and benefits to society as a whole. As with risk assessments, benefit assessments can benefit from peer review, especially when they are complex.

Present the Risk/Benefit Comparison to Risk Managers
Risk/benefit comparisons may be presented using a variety of methods, depending on the type of analytical techniques used. It is useful to summarize technical results in an easily understandable manner, to explain the methodology and criteria used, to discuss uncertainties, assumptions, and their potential impact on analyses and on decision-making. Risk/benefit comparisons should also be made available to interested and affected parties, taking into account the need to keep some information confidential (e.g. drug formularies).