Response to the Food Standards Agency consultation on Nutrient Profiling

Submitted on behalf of IFR by
Richard Faulks

February 2005

Background

On 1 February 2005, a consultation forum was held at the Institute of Food Research to discuss the FSA proposals on nutrient profiling. This session was hosted at the instigation of IFR Enterprises Ltd, as a result of discussion at a cluster meeting of the Food & Health Network, IFR’s industry forum (information on the F&HN can be found at www.foodandhealthnetwork.com).

IFR provided an impartial venue for this discussion forum to allow the industrial representatives (some of whom are F&HN members) the ability to discuss their views and air their concerns over the FSA’s proposed Nutrient Profiling. They had particularly expressed a wish to hear the IFR’s scientific opinion. All of those present agreed that IFR should provide feedback to the FSA covering the range of issues discussed. All food industry representatives who attended the meeting will also receive a copy of the conclusions of the meeting.

The conclusions below represent a consensus of views from the group in response to the FSA’s consultation document. Also attached is an Excel Spreadsheet applied to a range of foods from the food tables, using FSA Model SSCg3d.

Conclusions

Method

There was strong criticism of the subjective approach of pre-allocating foods into 3 groups based on a subjective assessment of their ‘healthiness’ and then designing an empirical model that ‘correctly’ allocates them into the preconceived framework.

Stigmatisation

The fixed banding will stigmatise specific foods and food groups which have an important nutritional role in sectors of the population. e.g. cheese and cheese products. Some recognised composite foods, although they can be moved some way towards being healthier, cannot be moved sufficiently in their recognised form to move from one category to a lower category.

Innovation

If foods cannot be moved to a lower category without too much difficulty then little or no attempt will be made to make them healthier. It was felt that reductions in fat, salt and sugar could be made on a limited scale and this would be of benefit, but may not be sufficient to change category and would therefore offer the industry no incentive for movement.

Consumer reaction

There is little or no evidence that nutrient profiling, however indicated to the consumer, will achieve the objective of moving consumers to choose healthier options. Without evidence to indicate that nutrient profiling will provide long term changes in consumer attitudes and food choices the scheme is unlikely to succeed, while at the same time increasing manufacturing costs in compliance. This will be reflected in higher retail prices and/or a less competitive manufacturing base.

Choice of Nutrients

Although there was general agreement on the relationships between some nutrients and non-communicable disease it was felt that some important factors were overlooked, specifically, dietary fibre (NSP + Resistant starch), cholesterol, mono-unsaturated fatty acids, and ‘anti-oxidant’ phytochemicals. NME sugar was felt to be an over-complicated definition, and ’added sugar’ (as already defined, e.g., sucrose, corn (starch) syrups, honey, fruit product concentrates etc) is entirely satisfactory and understood by the consumer.

It was unclear if added components (either as replacement or fortification) could be included, as opposed to endogenous content.

Physical properties

It was recognised that the physical structure of the food plays an important role in the diet in modulating consumption, digestion and delivery of nutrients, e.g. Glycaemic Index. Thus two foods my be similarly categorised but one of these foods may offer health benefits due to modulated release of nutrients, promotion of satiety, and/or prolongation of suppression of appetite. This would not be promoted via the proposed system and would therefore inhibit innovation to provide a healthier product through modification of physical characteristics, rather than simply composition.

Basis of Expression

This was probably the most contentious issue. It was felt that many food items would be stigmatised, particularly those eaten in small quantities and which are so important in an acceptable diet because they add appeal, e.g. sauces, pickles, condiments, olives, dressings, sugar, syrups etc.

The conflict seems to have arisen because there are 2 perspectives (1) a scientific description of the food and (2) the role of the food in its contribution to health. The ultimate aim is improved health through healthier choices (made by individuals) and to achieve this, the role of the food in the individual diet is key. However, because every individual is different it is likely that a different pattern of ‘food risks’ will occur in each case. At this stage it is difficult to reconcile scientific description with individual (or population) risk.

The resolution of choice appears to be to integrate portion size and to re-express the amount of nutrient supplied on portion size rather than on 100g or an energy basis. This approach however may not be appropriate across communities where ‘food risk’ comes from different dietary patterns. e.g. it may apply in the UK but not in France. Such an approach also runs the risk of being misinterpreted, e.g. a ‘low risk’ or ‘green’ food so allocated because of the very small portion size normally consumed may lead consumers to believe they can eat more.

Industrial Preferred Option

The preferred option appears to be one based on the proportion of the Guideline Daily Amount (GDA) of the nutrient supplied by a portion of the food. From a nutritional perspective this is very complicated for the consumer mainly because it requires a reasonable level of numeracy to understand and would require some ability to literally or conceptually add up numbers over the day. Consumers could also misinterpret the information and assume that the amount indicated as the GDA is what the food should contain and therefore a ‘healthy option’ would be the one where the nutrient content was nearest the GDA.

The GDA system for nutrient profiling has, however, the merit that it can be normalised to energy or portion size and simply allocated a percentage (or colour code) for each nutrient. e.g. re-expressed using an easy to understand scale as in signposting.

Why is this food in this category?

Where a food scores highly it is not apparent why this is the case. The high score could arise because of high energy, saturated fat, added sugar and/or salt and the consumer would not know. This is very important to many consumers who wish to avoid specific food components. Conversely, a ‘less healthy food’ could be rehabilitated simply by adding iron, calcium or n-3 PUFA’s while retaining the health risk associated with, for example, high salt.

The broad banding suggested may well result in 2 comparable products as indicated by category, but the consumer does not know how close or distant they are from the next category so will be unable to choose the healthier option. The manufacturer therefore has no incentive to reduce the concentration of nutrients which pose the health risk.

Risk associated with various food components

The risks associated with different food components are not the same. For example there is probably less risk from eating too much sugar than from eating too much salt or fat. Secondly the combined risk of two components may be much more that the simple sum of the risks. This again brings into focus the different concepts of using food composition as a surrogate of risk and real food risk. The relationship is far from clear.

Rate of change of risk

The model SSCg3d assigns a linear risk to each component and offsets that risk linearly with another component. This begs the question: e.g. can the risk of high blood pressure induced by too much salt be offset by increasing calcium content? Furthermore, is this relationship linear?

It is quite clear from the literature that risk is not normally linear with intake; rather it tends to increase exponentially with intake. Thus a small reduction in intake at the upper end of the scale may reduce risk by 50% whereas the same reduction at the lower end of the scale may not reduce risk significantly.

The model proposed needs to integrate the risks and composition in a more realistic way.

Transparency

The FSA’s proposed method was considered to be totally opaque to the consumer who would need a lot of education and instruction if they were to fully grasp the meaning of the proposed categories. In the event that transparency will be lost by using this method then a more meaningful (and complex) method would be a preferred option.

Analysis of some foods from the Food Tables

At the request of some of the industrial membership IFR set up a simple Excel spread-sheet to assign foods on the basis of the Model SSCg3d. It has not been possible to cover all the entries; however, the following points have emerged from the table.