Table 4 NOVA analysis-based identification of refinement measures

1

Confusion of formulation and processing

S1. Discrepancy between nature of classification parameters and the designation of classes

S2. Vague and over-inclusive definition of UPF

S3. No clear focus on nutrition value (e.g for specific “UPF-ingredients”)

M1. Clear definition, differentiation & separation of formulation & processing

M2. Reference to nutrition value as 1st target property for M1 measures

M3. In future: consider formulation & processing parameter(s) impact on palatability & satiation kinetics aspects as 2nd/3rd target properties (including aspects of e.g.: matrix volume, water & air content, H₂O-binding in modified NRF*)

2

Missing quantification of formulation and processing parameters

S4. Restriction to the qualitative characterization of food products (no thresholds for formulation, no levels of formulation and processing)

M4. quantifying levels of formulation (i.e. NRF*x.y.z) & processing (i.e. ΔNRF*x.y.z)

M5. enabling coupling of F- & P-parameters (e.g. FPFI^N) to simplify

M6. in the future: consideration of (i) matrix volume & water/air fractions incl. their change by processing (e.g drying/aeration/gelling); (ii) gastro-intestinal rheology

3

Weak evidence base (observational studies) for correlations between food classes & obesity/CNND disease risks

S5. No clear control of other risk factors for weight gain

S6. Risk of misclassification (i.e. single ingredient dependency for UPF)

M7. Base correlations with obesity & health risks on quantified F&P parameters taking nutrition value (& prospectively satiation kinetics) into account