FoodEx2
For MCRA, FoodEx 2 introduces the following points of attention:
Reading and dealing with FoodEx 2 coded data sets
Reading and dealing with food facets
Reading and exploiting food hierarchy data
Reading and dealing with FoodEx 2 codes
All data entities that contain foods data are potentially affected by the introduction of FoodEx 2. In MCRA, the following data tables are adapted to allow for input of full FoodEx 2 food codes:
Foods
Consumptions
Concentrations
For these tables, the food code is allowed to be the complete FoodEx 2 food code and automatically recognized as such. As an example, Table 13 shows how the FoodEx 2 coded consumptions should be provided to the system. On important note: the maximum field length of the food code is 50. This means that there is a maximum of five facets that can be specified for a food.
Individual |
DayOfSurvey |
Food |
Amount |
FoodSurvey |
|---|---|---|---|---|
14233701 |
1 |
A011R# F28.A07GX |
153.43 |
FS01 |
18843004 |
1 |
A011R# F28.A07GX |
125.23 |
FS01 |
34025701 |
1 |
A011R# F28.A07GX |
153.60 |
FS01 |
14720005 |
2 |
A011R# F28.A07GX |
105.00 |
FS01 |
49174010 |
1 |
A011R# F28.A07GX |
140.00 |
FS01 |
62794010 |
1 |
A011R# F28.A07GX |
67.00 |
FS01 |
61392002 |
1 |
A011P# F28.A07GL$F28.A07KQ |
104.72 |
FS01 |
61281231 |
1 |
A011P# F28.A07GL$F28.A07KQ |
109.72 |
FS01 |
Reading and dealing with facets data
Within MCRA, the following facets related aspects are accounted for:
Reading facets data
Dealing with facets
Facets in concentration data
Facets in food conversion
Using facets as processing factors
Using hierarchy data in the output
Reading facets data
To incorporate input of facets data in MCRA, two tables Facets and FacetDescriptors are introduced as optional tables of the Foods data group. The table for Facets and table for Facet descriptors.
Within MCRA, the facets of FoodEx 2 coded foods, consumptions, and concentrations are automatically linked to the provided facets and facet descriptors. Also, the facet descriptor names are added automatically to the foods containing these facets.
Dealing with facets
The introduction of food facets allows for much more detailed specifications of consumption and concentration data. However, it introduces the problem of deciding on which level of detail the exposure assessment should be performed. That is, should concentration models be generated on the level of foods-without-facets or on the level of foods-with-facets? E.g., should the concentrations of clementine peeled (A01CE#F28.A07LC) and clementine unprocessed (A01CE#F28.A0C0S) be modelled separately or should one model be constructed for clementine (A01CE)? Treating all clementine’s as equal may yield over-simplified conversions, whereas treating all separately may lead to many concentration models based on only few measurements. In MCRA, no implicit grouping of concentrations of equal foods with different facets is applied. If concentrations are provided for both clementine peeled (A01CE#F28.A07LC) and clementine unprocessed (A01CE#F28.A0C0S), then these are modelled separately. Another question is whether the order of the facets is relevant or not. E.g., is A0BYV#F02.A06GF$F03.A06HY the same as A0BYV#F03.A06HY$F02.A06GF? Regarding this matter, MCRA considers the facet order to be important. I.e., A0BYV#F02.A06GF$F03.A06HY is not the same as A0BYV#F03.A06HY$F02.A06GF.
Facets in food conversion
For conversion of foods-as-eaten to modelled foods, MCRA considers foods with different facet strings as different foods. I.e., there is no implicit conversion of foods-with-facets to foods-without-facets and also the order of the facets is important. However, as it is realistic to convert food-with-facets to the base food without facets, an additional (explicit) conversion step remove-all-facets is added that converts foods with facets to the base foods. I.e., the action is “remove all”. There is no conversion step for “stripping off one facet at a time”. The reason for this is that there is no good way of deciding which facet to strip off first. This new conversion step is somewhat equivalent to the already existing default processing conversion step (step 6), and is therefore implemented as step 6b of the conversion algorithm. Particular rules followed by this step:
Conversion of food-with-facets to food-without-facets.
Using facets that reveal processing data
Facets containing processing information, such as part-consumed-analysed (F20) and processing technology (F28) could be integrated with processing data. As an example, consider clementine peeled (A01CE#F28.A07LC). This could be linked to clementine (A01CE), with processing type removal of external layer (A07LC). Linking to processing data could be achieved by entering processing data using the facet codes. As an alternative to the current processing factor tables, a facet-based processing factors table is defined for processing facets. That is, the codes for food processed and unprocessed are implicitly defined for FoodEx 2.
FacetCode |
Substance |
FoodCode |
ProcNom |
ProcUpp |
ProcNomUncUpp |
ProcUppUncUpp |
|---|---|---|---|---|---|---|
A07LC |
SubstanceX |
A01CE |
0.5 |
0.6 |
0.05 |
0.06 |
F28.A07GV |
SubstanceX |
A0BY |
0.2 |
0.1 |
0.03 |
0.04 |
Note that in the example, the facet code could be specified as the full facet code, or just the code of the facet descriptor. As a more elaborate example consider
French fries from cut potato (A0BYV#F02.A06GF$F03.A06HY$F04.A00ZT$F28.A07GR)
For this food code, the substring of the processing facet is extracted from the list of facets.
A0BYV#F02.A06GF$F03.A06HY$F28.A07GR$F04.A00ZT with processing facet link A07GR
A0BYV#F02.A06GF$F03.A06HY$F04.A00ZT
In MCRA, a table FacetProcessingFactors is introduced that allows for specification of processing factors by means of facets. This table has the following structure:
Column name |
Key |
Required |
Type |
Size |
Description |
|---|---|---|---|---|---|
idProcessingType |
Yes |
Yes |
String |
5 |
The facet code of this processing factor definition. May be specified as full facet code, i.e., facet code plus facet descriptor code, or as the facet descriptor code. |
idFood |
Yes |
Yes |
String |
200 |
The food code |
idCompound |
Yes |
No |
String |
50 |
The substance for which this processing factor is defined. |
Nominal |
No |
Yes |
Double |
Nominal value (best estimate of 50th percentile) of processing factor (defines median processing factor) |
|
Upper |
No |
Yes |
Double |
Upper value (estimate of 95th percentile or “worst case” estimate) of processing factor due to variability |
|
NominalUncertaintyUpper |
No |
Yes |
Double |
Upper 95th percentile of nominal value (Nominal) due to uncertainty. A standard deviation for uncertainty of the nominal value (Nominal) is derived using the nominal value (Nominal) and upper 95th percentile (NominalUncertaintyUpper) |
|
UpperUncertaintyUpper |
No |
Yes |
Double |
Upper 95th percentile of upper value (Upper) due to uncertainty. From the nominal value (Nominal), upper value (Upper) and the specified uncertainties of these values (NominalUncertaintyUpper and UpperUncertaintyUpper, respectively) the degrees of freedom of a chi-square distribution describing the uncertainty of the standard |
The integration with the food conversion algorithm is as follows: Conversion step 2 (processing) is extended with a step 2c (processing facet) that attempts to match facets of a food code to processing data provided in the processing facets table. The following important rules are followed:
Processing factors can be defined for base-food-code/facet-code combinations and translate as food-with-processing-facet to food-without-processing-facet.
If multiple processing facets are present in the food-as-eaten code, then the last processing facet is used first for conversion.
Facet processing factors can be specified using the full facet code (i.e., facet-code plus facet-descriptor-code) or just the facet descriptor code. If both are specified for the same food, the full facet code is used.
Facet processing factors can be defined substance-specific, and non-substance-specific. Processing factors that are defined substance-specific always precede non-substance specific processing factors.
Processing factors defined by a food-processed/food-unprocessed combination precede processing factors defined through facets.
Weight reduction factors for processing factors defined for facets should be included in the food translation table and should match exactly.