Pages within the "home" section:

Changes in diet may present both potential risks and potential benefits to consumers. The balance of risk and benefit is of interest to food authorities developing food policy and consumer advice, to businesses developing new food products, and to consumers considering dietary changes.

In many cases, conventional risk assessment may show that adverse effects are unlikely. In other cases, a qualitative evaluation may be sufficient to conclude that either the risks or the benefits dominate. When this is not the case, it may be necessary to quantify not only the incidence of adverse and beneficial effects, but also the magnitude of their impact on health, their duration, and their impact on life expectancy. In addition, it may be helpful to combine these different dimensions of health impact into a single integrated measure such as disability-adjusted life years (DALYs). However, such assessments are difficult to conduct. Furthermore, all of the inputs required for the calculation are variable and uncertain, and it is important to assess the effect of this on the outcome. Finally, effective communication of risk-benefit information is challenging.

The QALIBRA project addressed these challenges by developing a flexible, general framework for quantitative risk-benefit assessment, including probabilistic approaches for quantifying variability and uncertainty; implementing the framework as user-friendly web-based software; conducting case studies with selected foods (oily fish and phytosterol-enriched margarines); and developing new strategies for risk-benefit communication.

The QALIBRA software integrates adverse and beneficial health effects using DALYs (as used in the WHO Global Burden of Disease surveys) or QALYs (Quality-Adjusted Life Years). These two measures are closely related but opposite in meaning: DALYs represent the number of healthy life years lost, whereas QALYs represent the number of healthy life years remaining. The basic calculation for DALYs is:


where YLD is years lived with a disease, W is a weight representing the severity of that disease on a scale where 0 = no effect and 1 = death, and YLL is the years of life lost due to early death from the disease.

The DALY calculation makes explicit the different types of information that are required to assess properly the net health impact of dietary change. This starts with estimates of intake of relevant adverse and beneficial foods or substances and of the corresponding dose-response relationships, as in a normal risk or benefit assessment. In addition, assessing overall health impact requires information on the severity of effects, which can be represented by DALY weights such as those published by WHO. Also required is the age of onset, duration, and probability of recovery or death associated with each disease. Information on these may be available from national health statistics, although their quality and relevance to the assessment is variable.

Ideally, DALYs (or QALYs) might be calculated by simulating disease histories over full lifetimes for the population under assessment, but this is hugely complex and requires explicit consideration of interactions between different health effects and with background diseases. QALIBRA uses a more practical approach, the “directly attributable health loss” method, which considers the health consequences of conditions starting in a single (average) year and ignores interactions. Consequently, the output is interpreted as a measure of potential health impact per year, and not as an estimate of actual health outcomes. This can be calculated for a single individual (e.g. with typical or worst case diet) for exploratory analyses, but is primarily intended to be applied to a representative sample of individuals as a measure of net health impact for the population.

The QALIBRA software allows the user to enter data on dietary intakes for one or many individuals, e.g. from existing models of dietary exposure, to represent variation of diet across the population. If other inputs (dose response, disease weights, duration etc.) vary with age, sex and other individual attributes, this can also be represented. Crucially, given the variability quality of these data, QALIBRA allows the user to quantify uncertainty in every input to the calculation, and uses Monte Carlo simulation to show the effect of those uncertainties as confidence intervals on the outputs, which may be DALYs or QALYs.

The flexible design of the software allows users to start with a simple, deterministic calculation for a single individual and gradually refine it to take account of variability and uncertainty, according to the needs of the risk-benefit question. It therefore offers a tool for the higher (quantitative) tiers of tiered approaches to risk-benefit assessment, such as those being considered by EFSA and the related EU project BRAFO.

The outputs generated by the QALIBRA software include tabular summaries of the potential annual health impacts, together with intermediate outputs to assist in interpretation of the results. Graphical outputs show how individual contributions to the overall impact vary across the population. The results include uncertainty intervals representing the effect of those uncertainties that have been quantified by the user. The additional impact of unquantified uncertainties must be considered qualitatively, including those due to the limitations of the model. Substantial expertise in the relevant disciplines, including toxicology, epidemiology, nutrition, modelling and statistics is essential to generate suitable inputs and to interpret the output.

To download a detailed report on the QALIBRA approach, including examples of QALIBRA assessment outputs, click here . You may also view the same material on our support pages here .