The 12th Asian Network Symposium on Nutrition

Opening Remarks 

Mr. YOSHITAKE Toru

Keynote Speech

Nutrient Profiling to Support Healthy Diet Policy Implementation in South East Asia

Dr. Angela de Silva

Regional Adviser Nutrition and Health for Development, WHO Regional Office for South East Asia

Abstract（73KB）

Slides（1027KB）

YouTube

Lecture 1

The Development and Implementation of the Japanese Nutrient Profile Model

Dr. TAKIMOTO Hidemi

Abstract（71KB）

Slides（786KB）

YouTube

Lecture 2

Development of a Nutrient Profile Model for Processed Foods in Japan: A Localization Strategy Considering Nutritional Standards and Food Culture

Dr. TAKEBAYASHI Jun

Abstract（90KB）

Slides（3345KB）

YouTube

Lecture 3

Development of a Japanese Version Nutrient Profile Model for Dishes and Its Applicability Across Asia

Dr. TOUSEN Yuko

Abstract（78KB）

Slides（2259KB）

YouTube

Dr. TAKIMOTO Hidemi

To Dr. Angela de Silva

What are your views on investors evaluating or valuing food companies based on their nutritional performance?

To Dr. Angela de Silva

In Indonesia, many foods consumed daily are sold in traditional markets without nutrition labels. Is nutrient profiling primarily intended for packaged and ultra-processed foods? How might countries such as Indonesia adapt the model for informal food systems?

Additionally, could nutrient profiling be used as a supportive tool to encourage healthier reformulation of traditional and local foods, rather than solely as a regulatory measure?

To Dr. Angela de Silva

To Dr. Angela de Silva

To any of the panelists

Could you please tell us about FOPNL in your country?

To any of the panelists

Could you please tell us about FOPNL in your country?

In Japan, Front-of-Pack Nutrition Labeling (FOPNL) is voluntary, but the Consumer Affairs Agency (CAA) published national guidelines to standardize the format. This revised format focuses on displaying the amounts of energy, protein, fat, carbohydrates, and sodium (salt equivalent) per serving, along with their percentages relative to Nutrient Reference Values (NRVs) for adults.

Our Japanese NPM (NPM-PFJ/DJ) takes a different approach, as it is specifically designed to encourage industry reformulation by providing a holistic health evaluation through ratings. Although there is no direct link between the Japanese FOPNL guidelines and our Japanese NPM at present, I believe our model will complement the Japanese FOPNL in the future to further enhance consumer health.

To Dr. TAKEBAYASHI Jun

Are there any published papers or reports available on the previous research related to nutrient profiling, especially the development of a Nutrient Profile Model (NPM) for processed foods in Japan?

Thank you for your interest. The development and methodology of our nutrient profiling models were published in 2024.

1. Takebayashi, J., Takimoto, H., Okada, C., et al. (2024). Development of a nutrient profiling model for processed foods in Japan. Nutrients, 16(17), 3026.
https://doi.org/10.3390/nu16173026

2. Tousen, Y., Takebayashi, J., Okada, C., et al. (2024). Development of a nutrient profile model for dishes in Japan version 1.0: A new step towards addressing public health nutrition challenges. Nutrients, 16(17), 3012.
https://doi.org/10.3390/nu16173012

To Dr. TAKEBAYASHI Jun

Regarding the scoring system, could the score be tailored to specific health conditions? For example, non-communicable diseases (such as diabetes, hypertension, and kidney disease), infectious diseases (such as diarrhea, dengue hemorrhagic fever, and HIV), nutritional status (such as stunting, underweight, and obesity), or specific age groups?

Thank you for your very important question. The NPM-PFJ is primarily designed as a public health tool to prevent non-communicable diseases (NCDs) and obesity in the general population.
The scoring algorithm specifically addresses NCD risks by penalizing high sodium (for hypertension) and high energy/sugars (for diabetes and obesity), while rewarding dietary fiber, protein, and FVNL. To ensure the model is tailored to the Japanese population, the reference values were adapted based on multiple national standards such as Dietary Reference Intakes (DRIs) for Japanese (2020).

To Dr. TOUSEN Yuko

I have a question regarding the dish-based NPM. Have you tested, or are you planning to test, its validity by comparing the judgments of the Japanese NPM (NPM-JP) with those of the Health Star Rating (HSR) system, which uses a similar algorithm, and the Choices International NPM, which is product-group specific and threshold-based?

Thank you very much for your question.
For the processed-food version of the Japanese NPM (NPM-PFJ 1.0), we have already confirmed its validity by examining the correlation with the Health Star Rating (HSR). Using 668 processed foods from the Standard Tables of Food Composition in Japan, we observed a very strong positive correlation between the NPM-PFJ final scores and HSR (Pearson’s r = 0.939, p < 0.01), which supports the robustness of the model.
For the dish-based version as well, we have conducted a similar validation using HSR, and we obtained a comparable level of correlation to that seen in the processed-food version.
Regarding comparisons with other international NPMs such as the Choices International system, we have not yet carried out a full analysis, but we would like to consider such evaluations as needed in the future.

To Dr. TOUSEN Yuko

I have a question regarding how consumers distinguish between nutrients of concern—particularly “positive” (beneficial) nutrients and “negative” (nutrients to limit)—when evaluating mixed dishes.

For example, a creamy curry may be high in salt, while also being rich in vegetables and legumes. In such cases, it can be challenging to assess both the beneficial and less desirable nutritional aspects simultaneously, rather than focusing on only one side (e.g., nutrients to limit or nutrients to encourage).

What are your thoughts on this issue?

Thank you very much for your question.
Mixed dishes often contain both nutrients to limit, such as sodium, and nutrients to encourage, such as vegetables and legumes. For this reason, the Japanese dish-based NPM is designed to evaluate both aspects together.
In particular, seasonings and cooking oils are difficult to assess as individual processed foods, but by evaluating them within a dish, we can assess their nutritional impact based on the actual amount people consume.
The model first gives points for nutrients to limit, and then subtracts positive points for vegetables, protein, and dietary fiber. This allows us to reflect the overall nutritional balance of the dish.
For example, even if a curry contains high sodium, the vegetables or legumes in it still contribute positive points. In this way, the model captures both strengths and weaknesses at the same time, and helps support balanced food choices.

To Dr. Takebayashi and Dr. Tousen

Both NPM-PFJ and NPM-DJ are based on the Australian Health Star Rating (HSR) system. Could you please share the rationale for choosing HSR as the base model, rather than Nutri-Score or the WHO SEARO model?

We selected the Health Star Rating (HSR) as our base model because its transparent and traceable algorithm allowed for direct recalibration using Japanese standards, such as the DRIs (2020), NRVs (2015), and Health Japan 21 goals. While Nutri-Score and the HSR shares a similar scoring/rating algorithm, HSR provides a greater number of food categories with specific nutrient criteria, making it more adaptable to the complexities of Japanese food culture. Regarding the WHO SEARO model, we did not use it as a direct reference for our scoring system because it is a threshold-based model. However, we consulted the WHO SEARO model indirectly when establishing our specific food category settings. 

To Dr. TAKEBAYASHI Jun

How can a Nutrient Profile Model be implemented in a country where a complete food composition table is not available?

Is it still possible to apply this type of model under such circumstances? If so, how should we begin developing this approach given these limitations?

The absence of a comprehensive food composition table presents a significant obstacle to developing a Nutrient Profile Model (NPM). Even for our NPM-PFJ version 1.0, our reliance on the Standard Tables of Food Composition in Japan represents a major limitation, as missing data values directly compromise the model's accuracy.
Because a scoring model calculates a final score by balancing all evaluated components—including energy, saturated fat, sugars, sodium, protein, fiber, and FVNL—missing data has a substantial impact. The evaluation relies heavily on the specific interaction between risk-increasing "base points" and risk-reducing "modifying points".
Consequently, in contexts where reliable food composition data is scarce, a threshold model focused on individual nutrient criteria may be more practical than a scoring model. While scoring models are designed to rank foods, threshold models focus on classifying them, making them a more feasible starting point when complete nutritional data is unavailable for the majority of products.

To Dr. TAKEBAYASHI Jun

Could you please clarify whether NPM-PFJ is used both for calculating the Health Star Rating and for guiding food reformulation efforts in Japan? It would be helpful to understand how the model supports these two complementary objectives.

The NPM-PFJ is designed to bridge the gap between nutritional evaluation and product improvement. As a scoring tool aligned with Japanese standards, it provides a quantifiable metric for a product's 'healthiness.' This evaluation serves as a direct catalyst for reformulation by illustrating how specific adjustments—such as sodium reduction—directly improve a product's rating. For instance, lowering the salt content in bread can objectively raise its score from 3.0 to 3.5, offering manufacturers a clear incentive to optimize their recipes. Ultimately, evaluation and reformulation are not separate goals but complementary drivers intended to enhance the nutritional quality of processed foods in Japan.

To Dr. Tousen and Dr. Takimoto

It is very interesting and encouraging to learn that Japan has developed a dish-based NPM. Given the wide variation in recipes across restaurants and households, could you kindly explain how the model is designed to accommodate such variability?

Thank you very much for your question.
As you mentioned, recipes for dishes vary widely across restaurants and households. For the development of the Japanese dish-based NPM, we first used standard dishes commonly eaten in Japanese households to establish the thresholds. These provide a neutral and culturally appropriate reference.
Of course, many types of dishes exist, and their intended use can differ. Food culture also varies greatly. Therefore, in our research, we are collecting a wide range of actual dish data and evaluating them using the Japanese NPM to examine how well the model can accommodate real-world variability.
In the future, as we accumulate more data, we hope to clarify the range of variation within each dish category and consider refining the thresholds or the scoring algorithm, depending on the purpose of use. This will help the model better reflect the diversity of dishes consumed in daily life.