1

Nutrition Security and Optimal Dietary Intake in Taiwan

Yi-Ting Liu1, Chia-Sheng Hsu2, Ching-Cheng Chang1,2, Shih-Hsun Hsu1

1Department of Agricultural Economics, National Taiwan University, Taiwan. 2Institute of Economics, Academic Sinica, Taiwan.

Abstract

The rapid urbanization and income growth have led to changes in the global food

systems and nutrition transition in many developing countries. The prevalence of

obesity and dietary-related diseases has become an urgent issue in developing national

health policies. Taiwan is a rapidly aging society and its highest prevalence of obesity

is observed in adults above 65 years old. This paper aims to find the optimal age-

specific dietary intake pattern for both sexes based on the data from Nutrition and

Health Surveys in Taiwan (NAHSIT). Goal programming approach is adopted to

minimize the gap between observed diet and the dietary recommendation intakes (DRIs)

from the public health authorities. Food consumption constraints is added to prevent

diet plans from including unreasonable pattern and to minimize impacts on household

food expenditures. Our result suggests that lack of calcium intake is a common problem

for all age groups and both sexes in Taiwan and a shift from meat and oil products to

more fish, nuts and vegetables is required. The age- and gender-specific results also

show how goal programming modeling can be used to translate the DRIs into

economically acceptable food plans.

Keywords: Goal programming, optimization, food intake pattern, Nutrition and Health

Survey in Taiwan (NAHSIT)

2

Nutrition Security and Optimal Dietary Intake in Taiwan

Introduction

In recent years, rapid urbanization and income growth in many developing

countries have led to changes in food systems resulting in a global nutrition transition.

Nutrition transition refers to recent global shifts in dietary patterns towards higher

intakes of saturated fats, sugars and refined foods, and lower intakes of high-fiber foods,

driven by technological advances that have made energy dense, nutrient-poor foods

cheaply available on global food markets (Ghattas, 2014). Popkin and Gordon-Larsen

(2004) indicated that modern societies seem to be converging on a diet high in saturated

fats, sugar, and refined foods but low in fiber, which is called food westernization.

Westernization of Asian diet has changed the food production and consumption in

an irreversible pattern (Pingali, 2006). Taiwan was among the fastest food

westernization countries in Asia. Using household survey data, Peng et al. (2015a),

Misra and Khurana (2008), and Pan et al. (2011) found that Taiwanese consumed more

meat, sweetened beverage, and oil but less vegetable and fruit leading to persistently

rising obesity prevalence rate. In 2016, the Health Promotion Administration (HPA)

announced that Taiwan had the highest obesity and overweight ratio (44% of adults) in

all Asian countries. In addition, the diseases related with obesity such as cardiovascular

diseases and diabetes had become the 2nd and 5th of the ten leading causes of death in

Taiwan.

Furthermore, Taiwan will enter an “aged society” as its proportion of elderly

reached 14 percent in 2018. The statistics of Ministry of the Interior indicate that its

dependency ratio has exceeded 100% in 2017, meaning elder generation is more than

younger one. Chang et al. (2011) found that the physical and mental health significantly

led to inadequate nutrient intakes of the elderly in Taiwan, and they suggested that not

only younger generation but elderly have problems for malnutrition. Therefore, the

paper aims to study the optimal dietary intake pattern for Taiwan taking into account

the compounding effect of age and gender. A mathematical goal programming model

and the Nutrition and Health Survey in Taiwan (NASHIT) data representing the dietary

habits of adults over 19 years old in Taiwan are adopted in our empirical results.

Linear programming model has been widely used to find optimal diets (Soden &

Fletcher, 1992, Briend et al., 2003, Darmon et al. 2006). Masset et al. (2009) and Okubo

et al. (2015) used goal programming to find the optimal dietary pattern for Americans

and Japanese. Masset et al. (2009) focused on preventing serious cancer problem and

Okubo et al. (2015) focused on the adjustment of dietary culture to prevent chronic

disease for Japanese. We follow the similar approach and divide the NASHIT data into

3

eight sub-groups by age and sex to provide suitable dietary recommendations for each

age and sex group. The optimal results will be used to show how programming models

can be used to develop practical dietary guidelines for men and women across different

age groups by translating the DRIs into economically acceptable food plans.

Material and Methods

Input data

Input data of this study mainly sources from two databases, including the data of

NAHSIT survey in Taiwan 2005-2008 and food and nutrient database (FDA, 2015).

The NAHSIT data is 24 hours dietary recall data obtained through face-to-face

interviews in each survey. In 2005-2008, Taiwan’s government conducted the third

NAHSIT survey targeting children aged 0-6 years (including 1,443 children) and adults

aged over 19 years (included 3,116 adults aged 19-64 years and 1,545 elderly aged 65

years and over), which had 6,104 total participants. The raw data of NASHIT survey

shows all food items ate by one participant during 24 hours. Before data processing, the

total number of food item for all participants are 172,337, and the total number of food

item for above 19 age of participants are 138,025. Considering that children dietary

intake of children is simple and the obesity rate of adult is getting higher year by year,

so this study aims to analyze the dietary intake of people who are above 19 years old.

Finally, the total effective subjects above 19 years old are 2,908 which is constructed

by 1,446 females and 1,462 males. Subsequently, we use food categories (see Appendix

Table 1) to divide data into 12 food groups and 48 subgroups. In order to compare

dietary habit with dietary recommendation intakes (DRIs), we separate data into 8 sub-

groups by 4 age region (9-30 years old, 31-50 years old, 51-70 years old and over 71

age) and sex. Although NAHSIT survey also includes nutrient information, however, it

doesn’t show uniform criteria of each nutrient. Therefore, we employ the uniform

criteria defined by food and nutrient database to cooperate with NAHSIT data, hence

we can transfer food form into nutrient form.

4

Food and Nutrient database provides uniform criteria which shows how much

nutrients are contained for each kind of food per 100 grams (see Appendix Table 2 and

Appendix Table 3). Food and Nutrient database contains 1,333 kinds of food and 23

nutrients but we only choose 11 nutrients in order to in accordance with the nutrients of

FBS (food balance sheet). Therefore, we compare the food name of food and nutrient

database with the food name of NASHIT data, and sort the food name of NASHIT data

by the food name of food and nutrient database, hence we can transfer the food

consumption of NAHSIT data into nutrient form by uniform criteria of Food and

Nutrient database. Finally, we represent the nutrient profile of NAHSIT data (hereafter

nutrient profile) with 6 food groups and 16 subgroups which is the same with the

categories of food and nutrient database. Therefore, we get 8 nutrient profiles for each

sub-group which respectively contains the consumption pattern of energy and 11

nutrients. We simply describe the processing of data with Figure 1.

Figure 1. The diagram of data processing

The process of establishing nutrient profile is mainly constructing of the following

steps. For modeling purposes, we collapsed food into 16 sub-groups based on similarity

of nutrient composition firstly (Gao et al.2006). However, some food items of NASHIT

data usually are not consumed by people, and to avoid those affecting the final result of

data processing, we eliminate the food item which the frequency of occurrences is less

than 1 % of total food item numbers from NASHIT data.

Second, through mapping the representative food item from NAHSIT survey and

Food and Nutrient database in Taiwan, we transfer food item into the form of nutrients.

Therefore, we assign respectively a weight to every nutrient from each representative

food item that corresponded to the percentage consumption of its sub-group, and then

we use a bottom-up method to category into 16 sub-groups (Marcoe et al., 2006).

Finally, nutrient profiles are represented by nutrient content of one serving of food

from each food subgroup. They are calculated separately for each sex and age group

NASHIT

raw data NASHIT

data

8 sub-groups &

2,908 samples

Eliminate 1%

outlier of food

item

Food and Nutrient

database

Nutrient by

age & sex

8 sub-groups

of nutrition

intake

5

and are used as input data for our diet model. Table 1 reports the average dietary intake

per day of each sub-group. According to Taiwan food guides (FDA, 2012), suggesting

that every Taiwanese should eat Fruit 2-4serving, vegetable 3-5 serving, cereal/roots 6-

15 serving, soybean/fish/meat/egg 3-8 serving, dairy 1.5-2 serving, oil/nuts 4-8 serving

per day. However, we can see that Table 1 shows people who are in the range of 19-30

years old in both sex and greater than or equal to 71 years old has a low vegetable intake.

Furthermore, males of both 19-30 and 31-50 age groups are over the

soybean/fish/meat/egg intake suggested by DRIs. On the other hand, Huang et al. (2003)

indicated that the prevalence rate of lactose intolerance is 95% in Asia due to Asian do

not have persistent dairy intakes. Therefore, Table 1 also shows that all sub-groups are

insufficient dietary intake of dairy which is consistent with the characteristic of dietary

intake in Asia.

We simply summarize three points of averaged dietary intake habit for Taiwanese

from Table 1. First, people under 19-30 years old have higher intake of snacks than

traditional principal food of rice and wheat. Snacks include bread, sandwich, hamburger,

and instant noodles, and those food usually are not homemade cuisine. Therefore, we

can infer younger generation of Taiwanese change their sources of staple food due to

the lifestyle of eating out. Second, Taiwanese generally have more intakes of beef and

pork than other kinds of meat and use them to be the main protein source in

soybean/fish/meat/egg category. However, people eat more fish and less beef and pork

as the age increases, meaning both the main protein source and dietary preferences will

change with aging. Nut intake also gets higher with aging, and reaches the peak at the

age of 51-70 on both female and male sexes. Third, Taiwanese have more vegetable oil

intakes than animal oil intakes no matter in which sub-groups.

On the view of nutrition intake, Table 2 shows that people are normal intakes of

vitamin C, vitamin A, Iron, and Phosphorus nutrient on the whole. However, some

nutrients are under or over DRIs. First, energy intake of people under 70 years old are

over DRIs, but above 70 years old cannot achieve DRIs. Furthermore, for macronutrient,

people in each age are both high intakes of protein and fat, and people above 70 years

old doesn’t reach the carbohydrate intake of DRIs. For micronutrients, Table 2 indicates

all people are insufficient intake of calcium compared with DRIs, and also insufficient

intake both of niacin and vitamin B1 for the 71+ age group.

Therefore, we provide some points for the result of talbe2. First, both of younger

and middle age group need more energy intakes to meet energy demand for growing up

and physical labor. On the other hand, with age increasing, elderly reduce their energy

intake due to the declining of basal metabolic rate and the decreasing of activity. Second,

we can notice that the 71+ age group has the problem of insufficient nutrition intake,

implying malnutrition problems of elderly has to be improved in Taiwan. Third, there

6

is a serious problem for lacking calcium intake in Asia countries, especially for Taiwan

and Japan. Chen et al. (2016) told that the main food such as milk, bread, cheese,

sardines are the main calcium source in western countries, however, the main food such

as rice are few nutrient of calcium in the eastern countries. Therefore, Taiwanese cannot

intake enough calcium from the main food and has to replenish it with other sources,

such as dairy and dark vegetable. However, Taiwanese has a low intake of calcium due

to the low dietary of dairy products and vegetable even though they are the other sources

of calcium nutrient. Additionally, the nutrition intake of Table 2 also infers that dietary

intake of fruit and dairy is insufficient, on the contrast, dietary intake of bread, instant

noodle, sandwich and hamburger are gradually rising (Wu et al. 2011).

Further, we respectively compare dietary habit of Taiwan to the similar studies of

Japan and USA and report it in Table 3. Moving from Taiwan to Japan, we can see that

the energy intakes of Japanese are higher than Taiwanese. For the comparison with

macronutrients, Taiwan’ males have higher of protein and fat than Japan’s males,

however, Taiwan’s females are lower intakes of protein and fat than Japan’s females in

the 51-70 age group. Although grain is main food both in Taiwan and Japan, however,

we can see Japanese are more intakes of Carbohydrate and lower intakes of Iron,

vitamin A, and vitamin C intake than Taiwanese. Japanese have a high intake of refined

grain, fish, seaweed, soybean product, salt and low intake of fat. Taiwanese have a high

intake of meat and low intake of dairy and fruit (Wu et al., 2011; Okubo et al., 2015),

which is meaning different dietary intake result in different nutrition intake in the two

countries even they have similar main food.

Additionally, we also contrast dietary intake of Taiwan with the USA to understand

the difference between Taiwan and Western countries. Comparison with the energy

intake of two countries, people in the USA have the higher intake of energy than Taiwan.

Moreover, USA's macronutrient intake such as protein, fat, and carbohydrate are also

higher than Taiwanese. For the intake of vitamin, there is lower intake both of vitamin

A and vitamin B in the USA than in Taiwan.

In sum, we can understand the dietary habit in Taiwan from above narrated,

Taiwanese are similar diet habit with Japanese, however, the less intake of both energy

and carbohydrate are in Taiwan. Moreover, Taiwanese have more sufficient intake of

minerals and vitamins than in Japan and the USA, excepting calcium. People still face

the problem of malnutrition even the income level of Taiwan has reached the standard

of developed countries. Therefore, we will design an optimal dietary intake model of

Taiwan to find out the useful policy implications for optimal dietary intake, and make

contributions for improving malnutrition of Taiwanese.

7

Table 1 Average dietary intake per day per capita by age and sex

Male Female

M19-30 M31-50 M51-70 M71+ F19-30 F31-50 F51-70 F71+

Cereal 14.17 14.25 12.43 10.85 11.78 9.73 9.64 8.77

Rice, Wheat 6.77 8.28 8.73 8.39 5.55 4.84 6.87 6.97

Sugar 0.25 0.30 0.27 0.25 0.36 0.19 0.20 0.17

Snack, Other 7.15 5.67 3.43 2.21 5.87 4.7 2.57 1.63

Soy/fish/meat/egg 9.21 8.91 7.81 5.00 6.88 6.21 4.86 3.87

Poultries 1.59 1.00 0.74 0.40 1.05 0.7 0.28 0.29

Livestock 3.35 3.31 2.54 1.65 2.62 1.95 1.39 1.11

Fish 1.79 2.00 2.37 1.65 1.14 1.79 1.57 1.36

Egg 0.79 0.75 0.55 0.35 0.51 0.49 0.36 0.24

Soybean 0.99 1.52 1.25 0.70 0.93 1.08 1.06 0.73

Other 0.70 0.33 0.36 0.25 0.63 0.2 0.2 0.14

Oil 6.69 6.64 5.59 4.49 5.25 5.24 4.33 3.33

Plant oil 3.53 3.86 3.19 2.67 2.52 2.81 2.55 1.95

Animal oil 0.11 0.09 0.13 0.15 0.05 0.13 0.29 0.16

Nut 0.17 0.26 0.54 0.36 0.14 0.27 0.34 0.21

Other 2.88 2.43 1.73 1.31 2.54 2.03 1.15 1.01

Fruit 0.93 1.62 1.72 1.74 1.11 1.78 1.92 1.13

Dairy 0.22 0.31 0.51 0.74 0.31 0.32 0.60 0.59

Vegetable 2.31 3.17 3.81 3.52 2.12 3.19 3.68 2.89

8

Table 2 Nutrition intake by age and sex

Energy Protein Fat Carbohydrate Calcium Phosphorus Iron VitB1 VitB2 Niacin VitC VitA

Kcal/day g/day g/day g/day mg/day mg/day mg/day mg/day mg/day mgNE/day mg/day μgRAE/day

M19-30 2348.00 94.73 86.74 277.25 538.38 1269.54 18.21 1.41 1.65 23.75 164.00 943.93

DRI 2150 60 48-71 269-389 1000 800 10 1.2 1.3 16 100 600

M31-50 2451.28 99.06 88.39 287.43 615.95 1334.74 18.28 1.41 1.52 24.59 183.22 1099.48

DRI 2100 60 47-70 262-315 1000 800 10 1.2 1.3 16 100 600

M51-70 2107.06 90.92 71.38 258.86 679.15 1258.46 17.20 1.36 1.49 22.58 204.40 1210.18

DRI 1950 55 43-65 243-292 1000 800 10 1.2 1.3 16 100 600

M71 1679.28 69.75 54.39 227.88 703.91 1048.59 13.38 1.12 1.47 16.92 170.71 1259.58

DRI 1900 60 42-63 237-285 1000 800 10 1.2 1.3 16 100 600

F19-30 1867.67 76.06 70.83 234.65 489.63 1035.48 14.84 1.10 1.23 18.72 135.97 669.19

DRI 1650 50 37-55 206-247 1000 800 15 0.9 1 14 100 500

F31-50 1707.99 72.72 61.49 216.74 558.52 1048.53 15.57 1.10 1.24 19.97 178.63 1050.49

DRI 1650 50 37-55 206-247 1000 800 15 0.9 1 14 100 500

F51-70 1568.73 66.62 50.64 216.31 651.73 1033.67 14.18 1.10 1.31 16.67 193.01 1178.00

DRI 1600 50 36-53 200-240 1000 800 10 0.9 1 14 100 500

F71 1296.60 54.65 40.13 182.62 569.62 828.05 10.67 0.87 1.14 12.78 127.10 1064.59

DRI 1500 50 33-50 187-225 1000 800 10 0.9 1 14 100 500

Source: This study.

9

Table 3 Comparison nutrient intake with Japan and USA

Country Taiwan Japan USA Taiwan Japan USA Taiwan USA Taiwan Japan USA Taiwan Japan USA Taiwan USA

Sex Male Female

Age 31-50 30-49 30-49 51-70 50-69 50-69 71+ 70+ 31-50 30-49 30-49 51-70 50-69 50-69 71+ 70+

Energy (kcal) 2451.28 2391.00 2571.00 2107.06 2457.00 2373.00 1679.28 2022.00 1707.99 1856.00 1929.50 1568.73 1898.00 1745.50 1296.60 1604.00

Protein (g) 99.06 81.30 101.95 90.92 89.50 91.80 69.75 80.80 72.72 66.80 74.40 66.62 74.90 68.65 54.65 60.00

Fat (g) 88.39 62.43 96.00 71.38 65.52 92.85 54.39 80.90 61.49 59.80 75.10 50.64 55.00 69.85 40.13 62.90

Carbohydrate

(g) 287.43 316.80 302.00 258.86 334.15 272.00 227.88 237.00 216.74 249.63 233.50 216.31 269.04 205.50 182.62 201.00

Calcium (mg) 615.95 544.00 1093.00 679.15 641.00 1005.00 703.91 940.00 558.52 526.00 888.50 651.73 656.00 821.00 569.62 809.00

Phosphorus

(mg) 1334.74 1197.00 1678.00 1258.46 1350.00 1519.50 1048.59 1576.00 1048.53 1015.00 1251.50 1033.67 1173.00 1165.50 828.05 1069.00

Iron (mg) 18.28 8.30 16.35 17.20 10.30 16.30 13.38 16.50 15.57 7.40 12.75 14.18 9.50 12.55 10.67 12.40

VitB1 (mg) 1.41 1.10 - 1.36 1.10 - 1.12 - 1.10 0.87 - 1.10 0.96 - 0.87 -

VitB2 (mg) 1.52 1.40 - 1.49 1.60 - 1.47 - 1.24 1.20 - 1.31 1.50 - 1.14 -

Niacin (mgNE) 24.59 20.90 34.00 22.58 23.10 29.00 16.92 25.40 19.97 15.90 22.40 16.67 18.00 21.00 12.78 18.40

VitC (mg) 183.22 92.00 78.75 204.40 138.00 80.85 170.71 97.50 178.63 90.00 74.55 193.01 150.00 72.65 127.10 87.40

VitA (μgRAE) 1099.48 628.00 666.50 1210.18 786.00 666.00 1259.58 759.00 1050.49 584.00 570.00 1178.00 682.00 602.00 1064.59 654.00

Source: NAHSIT (2005-5008); Okubo et al. (2015); USDA (2014).

10

Goal programming model

In 1957, Charnes and Cooper were first to mention goal programming and this tool

first applied in management (Feng et al. 1989). Anderson and Earle (1983) indicated

that goal programming is presented as a method of achieving nutritional balance in

selected diets. In recent years, many experts used goal programming to design optimal

dietary intake (Ferguson et al., 2006; Masset et al., 2009; Okubo et al., 2015). According

to previous studies (Ferguson et al., 2006; Okubo et al., 2015; Masset et al., 2009), the

objective function minimizes the deviation in food intake between the observed and

optimized food intake patterns met the nutritional recommendations with a little change

as possible from the reported food intake.

In this study, we follow the approach suggested by Ferguson et al. (2006), Masset

et al. (2009) and Okubo et al. (2015) and define the objective function as follows:

Minimize Y = ∑|(𝑋𝑗𝑜𝑝𝑡

− 𝑋𝑗𝑜𝑏𝑠)/𝑋𝑗

𝑜𝑏𝑠|

16

𝑗=1

, (1)

where Y denotes the objective function to minimize, 𝑋𝑗𝑜𝑝𝑡

denotes the serving of food

subgroup j in optimized food subgroup intake pattern, and 𝑋𝑗𝑜𝑏𝑠 denotes the mean

serving of food subgroup in the observed food intake pattern across the whole

population. Therefore, we try to minimize the summation of all deviation from dietary

intake to optimal intake, Y, hence we define new decision variable to be non-negative

and representing positive and negative deviation from the observed food per serving,

and the definitions are as the following:

If 𝑋𝑗𝑜𝑝𝑡

< 𝑋𝑗𝑜𝑏𝑠, then 𝑑𝑗

− = (𝑋𝑗𝑜𝑏𝑠 − 𝑋𝑗

𝑜𝑝𝑡)/𝑋𝑗

𝑜𝑏𝑠 and 𝑑𝑗+=0, (2)

If 𝑋𝑗𝑜𝑝𝑡

> 𝑋𝑗𝑜𝑏𝑠, then 𝑑𝑗

− = 0 and 𝑑𝑗+=(𝑋𝑗

𝑜𝑝𝑡− 𝑋𝑗

𝑜𝑏𝑠𝑡)/𝑋𝑗𝑜𝑏𝑠, (3)

If 𝑋𝑗𝑜𝑝𝑡

= 𝑋𝑗𝑜𝑏𝑠, then 𝑑𝑗

− = 0 and 𝑑𝑗+=0, (4)

Subject to: 𝑑𝑗+ − 𝑑𝑗

− = (𝑋𝑗𝑜𝑝𝑡

− 𝑋𝑗𝑜𝑏𝑠)/𝑋𝑗

𝑜𝑏𝑠 (5)

Equation (2), (3), and (4) are summarized in equation (5). Because the decision variable

is redefined, therefore the new linear function Y’ is expressed as the summation of the

deviational variables and the objective function is rewritten as the following.

Minimize 𝑌′ = ∑ 𝑑𝑗+ + 𝑑𝑗

−

16

1

(6)

11

Constraints

We can make the nexus between food intake and nutrition intake through different

nutrient constraints, hence our optimal dietary model not only can minimize the gap

between observed and optimized food intake pattern but also provide the desirable

intake level of energy and nutrients.

Table 4 shows constraints of each food intake. We set both upper and lower limit

of food intake to ensure that the solution of programming model will not excess the

normal quantity of food consumption. These bounds of limit are derived from the actual

distribution for dietary intake of NAHSIT data by each sub-groups. We have to calculate

how many serving of food are consumed by everyone under each sub-group. Then we

can know the real distribution for food consumption in different sub-groups, thus

respectively take the 95th quantile and 5th quantile of six food types to be the upper

and lower limit for six food type in different sub-groups.

We apply nutritional constraints suggested by HPA (2016) to ensure the optimal

solution of our model for each food would be equal to or greater than the desired value

which was based on DRIs for Taiwan adults. Energy constraints are defined by different

level (lower, low, moderate, and high) of physical activity in accordance with FDA

(2012). We believe that most of Taiwanese are at low level of physical activity, hence

we define the energy constraints of our model with a low level of physical activity.

According to the DRIs of Taiwan, if a person engages in low level of physical activity,

whose energy constraint will be set at 2150 kcal/day for men of 19-30 year olds; 1650

kcal/day for women of 19-30 year olds; 2100 kcal/day for men of 31-50 year olds; 1650

kcal/day for women of 31-50 year olds; 1950 kcal/day for men of 51-70 year olds; 1600

kcal/day for women of 51-70 year olds; 1900 kcal/day for men 71+ year olds and 1500

kcal/day for women of 71+ year olds.

For the constraints of nutrients (protein, calcium, phosphorus, iron, vitamin A,

vitamin B1, vitamin B2, niacin and vitamin C), we follow the Recommended Dietary

Allowance (RDA) or Adequate Intake (AI) to define the lower limit of constraint for

each nutrient. However, we also define the upper limit of constraint for some nutrients

(protein, calcium, phosphorus, iron, vitamin A, niacin and vitamin C) followed the

Upper Intake Level (UL) to avoid some health problems caused by intake over of

nutrient. Therefore, we report all constraints of nutrient intake in Table 5.

12

Table 4 Food constraints by both sex and each age group Unit: Serving

Age 19-30 31-50 51-70 71+

Food group Subgroup Male (Lower/Upper)

Female (Lower/Upper)

Male (Lower/Upper)

Female (Lower/Upper)

Male (Lower/Upper)

Female (Lower/Upper)

Male (Lower/Upper)

Female (Lower/Upper)

Cereal 3.54/22.68 2.56/20.65 3.48/22.07 2.21/15.60 3.43/20.91 2.17/15.41 3.69/19.99 2.56/14.95 Rice 0.00/15.24 0.00/10.97 0.00/15.76 0.00/10.90 0.00/16.46 0.00/13.38 0.00/16.74 0.00/14.06 Sugar 0.00/0.86 0.00/1.17 0.00/1.17 0.00/0.60 0.00/1.11 0.00/0.78 0.00/0.95 0.00/0.42 Other 0.00/15.09 0.00/14.06 0.00/11.91 0.00/9.30 0.00/8.95 0.00/5.97 0.00/5.92 0.00/4.53

Oil 0.00/14.29 0.00/11.18 0.00/13.94 0.00/10.20 0.01/12.36 0.00/11.03 0.00/10.58 0.00/8.20 Plant oil 0.00/11.72 0.00/8.83 0.00/13.14 0.00/8.80 0.00/10.87 0.00/8.33 0.00/9.03 0.00/6.30 Animal oil 0.00/0.39 0.00/0.27 0.00/0.55 0.00/0.70 0.00/0.65 0.00/0.99 0.00/0.89 0.00/0.83 Nut 0.00/0.01 0.00/0.48 0.00/1.33 0.00/1.60 0.00/2.17 0.00/1.98 0.00/1.72 0.00/1.09 Other 0.00/8.67 0.00/7.72 0.00/6.57 0.00/4.90 0.00/5.12 0.00/2.26 0.00/3.98 0.00/3.45

Soy/Fish/Meat/Egg 1.34/18.28 0.49/13.40 1.43/16.57 0.46/13.20 0.47/14.82 0.00/11.06 0.29/11.40 0.00/9.41 Poultries 0.00/6.16 0.00/4.51 0.00/3.95 0.00/2.80 0.00/2.86 0.00/1.97 0.00/2.56 0.00/1.59 Livestock 0.00/10.34 0.00/6.49 0.00/9.56 0.00/5.90 0.00/7.22 0.00/4.67 0.00/5.40 0.00/5.17 Fish 0.00/4.87 0.00/3.70 0.00/6.31 0.00/5.90 0.00/7.30 0.00/5.65 0.00/6.35 0.00/4.03 Egg 0.00/2.70 0.00/1.69 0.00/2.31 0.00/2.00 0.00/1.99 0.00/1.23 0.00/1.33 0.00/1.15 Soybean 0.00/4.53 0.00/3.29 0.00/5.36 0.00/4.40 0.00/4.89 0.00/4.47 0.00/3.33 0.00/2.53 Other 0.00/1.79 0.00/1.45 0.00/1.36 0.00/0.90 0.00/1.20 0.00/0.67 0.00/0.91 0.00/0.57

Dairy Dairy 0.00/0.84 0.00/1.05 0.00/1.60 0.00/1.40 0.00/2.70 0.00/2.40 0.00/2.88 0.00/2.58

Vegetable Vegetable 0.02/6.56 0.02/5.66 0.21/8.33 0.18/8.30 0.29/10.64 0.35/9.35 0.09/8.44 0.07/7.99

Fruit Fruit 0.00/3.48 0.00/4.54 0.00/5.49 0.00/5.30 0.00/6.02 0.00/5.57 0.00/6.30 0.00/3.96

Source: FDA (2012).

13

Table 5 Nutrient constraints by both sex and each age group

Nutrient Unit/day Male Female

19-30

(Lower/Upper)

31-50

(Lower/Upper)

51-70

(Lower/Upper)

71+

(Lower/Upper)

19-30

(Lower/Upper)

31-50

(Lower/Upper)

51-70

(Lower/Upper)

71+

(Lower/Upper)

Energy kcal 2150 2100 1950 1900 1650 1650 1600 1500

Protein g 60 60 55 60 50 50 50 50

Fat % of energy 20-30% 20-30% 20-30% 20-30% 20-30% 20-30% 20-30% 20-30%

Carbohydrate % of energy 50-60% 50-60% 50-60% 50-60% 50-60% 50-60% 50-60% 50-60%

Calcium mg 1000/2500 1000/2500 1000/2500 1000/2500 1000/2500 1000/2500 1000/2500 1000/2500

Ph mg 800/4000 800/4000 800/4000 800/4000 800/4000 800/4000 800/4000 800/4000

Iron mg 10/40 10/40 10/40 10/40 15/40 15/40 10/40 10/40

Vitamin A μgRAE 600/3000 600/3000 600/3000 600/3000 500/3000 500/3000 500/3000 500/3000

VitB1 mg 1.2/- 1.2/- 1.2/- 1.2/- 0.9/- 0.9/- 0.9/- 0.9/-

VitB2 mg 1.3/- 1.3/- 1.3/- 1.3/- 1/- 1/- 1/- 1/-

Niacin mgNE 16/35 16/35 16/35 16/35 14/35 14/35 14/35 14/35

Vitamin C mg 100/2000 100/2000 100/2000 100/2000 100/2000 100/2000 100/2000 100/2000

Notes: “-” denotes no upper bound of nutrient intake is defined.

Source: HPA (2016).

14

Empirical Results

Optimal dietary intake

Table 6 shows the result of optimized nutrient intake for each sub-group satisfied

with all nutritional constraints. As mentioned before, we know that some sub-groups

don’t achieve nutrition goal in observed food intake pattern, including 3 nutrients for

male sub-group of 19-30 year olds; 5 nutrients for female sub-group of 19-30 year olds;

4 nutrients for male sub-group of 31-50 year olds; 5 nutrients for female sub-group of

31-50 year olds; 4 nutrients for male sub-group of 51-70 year olds; 3 nutrients for

female sub-group of 51-70 year olds; 4 nutrients for male sub-group of aged 71 and

over; 3 nutrients for female sub-group aged 71 and over. Therefore, we can see that

calcium is the most difficult to achieve the nutrient goal for each sex and age group,

and high fat intake and low carbohydrate intake from the past intake pattern. Through

the linear programming, we can derive an optimized food intake pattern satisfied all

nutrient constraints of DRIs from the optimal dietary intake model. Therefore, we can

respectively suggest an adjustment of food intake to meet the DRIs’ requirement of

macronutrient and micronutrient of DRIs for each sex and age group.

Figure 2 graphs the observed food intake pattern and optimized dietary intake. We

can see that the observed intake of cereal are higher than the optimized intake of cereal

for the most of sub-groups, meaning most of Taiwanese should have less intake of rice

or other cereal to reach the optimized intake of cereal, excepting the male group of 51-

70 year olds, and the female group both of 19-30 year olds and 71+ year olds.

Additionally, we suggest that Taiwanese not only should decrease intake of oil but

increase intake of vegetable to meet the optimized intake for a healthy dietary intake.

For the intake of Soy/fish/meat/egg, the different difference between observed and

optimized intake are found out from the different sub-groups. The result indicates that

male group of 19-30 year olds and 31-50 year olds should decrease the intake of

Soy/fish/meat/egg. For female, both 31-50 year olds and 71+ year olds should increase

the intake of Soy/fish/meat/egg. We should use the increasing of soybean and fish

consumption to make up for the decreasing of poultry and livestock consumption and

to reach a better dietary of health. Moreover, we also can see that Taiwanese should

increase dairy intake especially for the group of 51 and over year olds. The details of

optimized dietary intake are reported in Appendix Table 4.

According to Nutrition Information (2017), cereal and oil/nut are the main source

of energy for Taiwanese. Nuts are rich in nutrients such as vitamin B group, vitamin E,

mineral, fiber and unsaturated fat. These nutrients are helpful to prevent cardiovascular

diseases and metabolic diseases. However, oil is rich in fat which could cause

cardiovascular diseases, especially saturated fat. Therefore, we should eat more nuts

and less oil intake for a healthy fat intake. Dairy is not only rich in calcium but also

helpful to digest and absorb protein, carbohydrate, fat, a variety of vitamins (especially

15

vitamin B2), minerals, but it lacks vitamin C and iron. However, the prevalence rate of

lactose intolerance is 95% in Asia because we don’t have persistent dairy intake (Huang

et al. 2003). Therefore, vegetable is another choice to be the source of calcium, and is

rich in vitamins, minerals, fiber, and phytochemicals. Especially, the dark vegetable has

high amounts of calcium nutrient and can be supplement with calcium nutrients for us.

Fruit is also rich in vitamin A, vitamin C, mineral and fiber, therefore we can have more

fruits to meet the demand of those nutrients in daily life. Finally, soy/fish/meat/egg are

the source of protein, soybean is the source of vegetable protein, and fish/meat/egg are

the source of animal protein. Furthermore, the average fat content of fish is lower than

meat and the proportion of fatty acids is healthier. Food of meat contains a considerable

number of saturated fatty acids or processed seasoning made of frozen food, both would

be adverse health of cardiovascular disease, that is, choosing lean and fresh meat should

be more appropriate. In addition, egg has the best quality of protein among all foods.

In summary, our optimized food intake pattern is reasonable, and we believe it will

make contributions for suggestions of healthier dietary intake. We find out that

Taiwanese should increase intake of carbohydrate nutrient mainly sourced from rice.

Also, the optimized intake of Soy/fish/meat/egg indicates that we should take the

healthier protein from fish and soybean to replace meat. Furthermore, dairy is the best

source of calcium, and elderly are easy to get osteoporosis caused by lacking calcium

(Lin et al. 2013). Therefore, the optimized dairy intake pattern also suggests that elderly

in Taiwan should increase dairy intake, and it would be the benefit of improving

osteoporosis. On the other hand, increasing of the optimized vegetable intake will

replenish the lacking calcium of Asian caused by lactose intolerance.

16

Table 6 Optimized nutrient intake patterns

Age 19-30 31-50 51-70 71+

unit/day Male Female Male Female Male Female Male Female

Energy kcal 2150.00 1650.00 2100.00 1650.00 1950.00 1600.00 1900.00 1500.00

Protein g 94.13 73.49 85.25 90.85 91.41 72.61 78.15 70.69

Fat g 71.00 55.00 70.00 55.00 65.00 53.00 64.00 50.00

Carbohydrate g 289.52 221.61 288.78 207.00 255.54 212.41 259.18 199.37

Calcium mg 1000.00 1000.00 1000.00 1000.00 1000.00 1000.00 1000.00 1000.00

Phosphorus mg 1434.88 1431.38 1549.06 1461.81 1549.33 1249.04 1371.77 1280.79

Iron mg 23.77 20.27 18.98 25.51 19.13 15.23 20.94 22.97

Vitamin A μgRAE 1648.54 1008.95 1285.98 1441.80 1211.76 1120.84 1657.05 3000.00

Vitamin B1 mg 1.20 0.90 1.21 1.20 1.20 1.25 1.67 1.64

Vitamin B2 mg 2.14 1.85 2.55 2.06 2.09 2.31 3.32 3.42

Niacin mgNE 20.37 14.00 19.39 21.44 19.55 14.00 16.00 14.00

Vitamin C mg 169.77 116.02 158.49 136.85 151.44 137.76 183.87 187.74

Source: This study.

17

Table 7 Optimized food intake pattern

Age 19-30 31-50 51-70 71+

Food group Sub-group Male Female Male Female Male Female Male Female

Cereal 12.69 10.37 12.87 6.48 10.31 6.97 10.29 7.90 Rice 6.93 5.11 8.40 4.33 7.74 5.95 8.60 6.63 Sugar 0.25 0.36 0.28 0.19 0.27 0.20 0.25 0.15 Other 5.51 4.90 4.19 1.96 2.30 0.81 1.44 1.12

Oil 2.24 0.15 1.65 1.53 3.37 2.54 3.40 2.65 Plant oil 0.00 0.00 0.00 0.00 1.80 1.46 2.28 1.67 Animal oil 0.11 0.04 0.09 0.13 0.12 0.29 0.14 0.16 Nut 0.19 0.11 0.21 0.21 0.41 0.30 0.26 0.15 other 1.94 0.00 1.35 1.19 1.04 0.49 0.72 0.67

Soy/Fish/Meat/Egg 7.94 7.48 5.66 4.84 7.41 7.11 4.29 4.77 Poultries 0.51 0.34 0.88 0.57 0.66 0.25 0.39 0.28 Livestock 1.40 0.00 0.00 1.00 0.00 1.25 1.59 1.04 Fish 1.51 1.36 1.58 1.52 4.46 1.33 1.56 1.50 Egg 0.77 0.50 0.74 0.47 0.53 0.34 0.34 0.23 Soybean 2.94 3.29 1.40 3.69 1.18 0.97 0.68 0.69 Other 0.35 0.17 0.24 0.16 0.28 0.15 0.21 0.12

Dairy Dairy 0.10 0.20 0.23 0.25 0.41 1.71 0.93 0.63

Vegetable Vegetable 6.57 5.66 5.10 8.26 4.22 3.50 5.95 7.06

Fruit Fruit 0.73 0.98 1.45 1.64 1.61 1.77 1.69 1.03

Source: This study.

18

Figure 2 Food Intake Pattern between Observed and Optimized

Source: This study.

19

FBS and optimized dietary intake

Subsequently, we compare 2015 FBS with average optimal food intake pattern.

From Table 7, we can see that our food system provides too much cereal, oil/nut,

soy/fish/meat/egg, dairy and fruit but provide less vegetable due to the average of

individual economic behavior. The minus deviation also means food loss and waste

from the viewpoint of health dietary intake, if Taiwanese changes dietary intake in

accordance with the optimal dietary intake and doesn’t adjust food supply system.

Table 7 Comparison of optimized food demand and real food supply

Food group Optimized 2015 FBS Deviation

Cereal 9.74 13.69 -3.96

Oil/nut 2.19 14.06 -11.87

Soy/Fish/Meat/Egg 5.68 7.85 -2.17

Dairy 0.56 0.64 -0.08

Vegetable 5.79 2.91 2.88

Fruit 1.36 2.01 -0.65

Source: This study; COA (2015)

From the macro sight, we represent the data of 2015 FBS with six food groups and

four nutrients, and also the same for optimized food intake. Therefore, Table 8 shows

the structure of each macronutrient for six food groups. We can see that cereal is the

main source of energy from 2015 FBS, but soy/fish/meat/egg from optimized food

intake. Additionally, the main source of protein is soy/fish/meat/egg no matter from the

sight of FBS or optimal dietary intake. On the other hand, the main source of fat from

the sight of FBS is oil/nut, but soy/fish/meat/egg from the sight of optimized food intake.

That is due to the optimized food intake pattern suggests Taiwanese should eat less

meat and oil/nut but fish. However, the main source of carbohydrate is still cereal no

matter from the sight of FBS or optimal dietary intake.

Table 8 Contribution of each macronutrient for six food category

2015 FBS optimized food intake

energy protein fat carbohydrate energy Protein Fat carbohydrate

Cereal 37% 23% 2% 75% 28% 12% 14% 48%

oil/nut 26% 5% 58% 2% 9% 5% 14% 7%

soy/fish/meat/egg 26% 60% 35% 4% 53% 70% 64% 36%

vegetable 3% 4% 1% 5% 0% 0% 0% 0%

Dairy 3% 6% 3% 3% 8% 13% 8% 6%

Fruit 5% 2% 1% 11% 1% 0% 0% 2%

Source: This study.

In summary, we provide some policy implications for food system both on demand

and supply side. Firstly, from the side of food supply, we find out that Taiwan’s food

20

system provide too much food of less nutrient-density, such as oil and red meat.

Although these are necessary for daily life, we still can eat less of these and use healthier

food such as fish and vegetable to replace them. Therefore, Taiwan government should

introduce some rules of health food supply to provide more health food for Taiwanese.

Secondly, from the side of food demand, empirical results found that people eat

more red meat and oil caused them absorb too much fat. Also, calcium insufficient

because Taiwan population drink less dairy. Therefore, government should promote eat

less fat density food in order to prevent rise obesity prevalence rate. It is a linchpin that

Taiwan's government needs to encourage people to eat more vegetables which are good

calcium sources for their sufficient calcium intakes.

Conclusion

In recent years, rapid social and economic change result in dietary habit change. This

situation may cause malnutrition to Taiwan population and make obesity prevalence

rate higher than before. As time change, people eat less dairy and vegetable but eat

more oil/nut and meat. According to these investigation, to make people eat healthier,

we design an optimal dietary intake model.

From many years ago, linear programming is a tool which people always estimate

dietary intake but goal programming can narrow the scope of feasible solution to make

optimal dietary intake pattern more precisely. Therefore, we put NASHIT survey 2005-

2008 data into our diet model. The results show that our optimized food intake pattern

is reasonable. People become eat healthier food such as intake more fish, soybean,

vegetable and dairy and eat less oil and red meat. Also, all nutrients achieve their goal

by adjustment of dietary intake pattern.

According to these results, we provide two policy recommendations. First,

government should product or import healthier food for Taiwan people and promote

Taiwan people eat less fat density food to prevent obesity rate rise. Secondly, because

of lactose intolerance, government should encourage people to eat more vegetable

because vegetable is also a good calcium source for people.

Research Limitation

Our data only obtain form 2005-2008 because the newest data do not issue and other

NAHSIT survey data are about children; therefore, we use data in 2005-2008. If the

newest data issue, we can use the same model and calculate again. The stability of food

intake pattern and change in dietary habits during then time gap between data collect

and analysis might have slightly influenced the conclusion.

21

Reference

Anderson, A. M., Earle, M. D., 1983. Diet planning in the third world by linear

and goal programming. Journal of the Operational Research Society, 9-16.

Briend, A., Darmon, N., Ferguson, E., Erhardt, J. G., 2003. Linear programming:

a mathematical tool for analyzing and optimizing children's diets during the

complementary feeding period. Journal of pediatric gastroenterology and

nutrition 36(1), 12-22.

Chang, H. H., Nayga, R. M., Chan, K. C., 2011. Gendered analysis of nutrient

deficiencies among the elderly. Journal of Family and Economic Issues 32(2),

268-279.

COA, 2015. Food balance sheet 2015. Council of Agriculture, Taiwan. Retrieved from

http://agrstat.coa.gov.tw/sdweb/public/book/Book.aspx

Darmon, N., Ferguson, E. L., Briend, A., 2006. Impact of a cost constraint on

nutritionally adequate food choices for French women: an analysis by linear

programming. Journal of nutrition education and behavior 38(2), 82-90.

FDA, 2012. Taiwan Food Guides. Food and Drug Administration, Ministry of

Health and Welfare, Taiwan. Available at

https://health99.hpa.gov.tw/media/public/pdf/21733.pdf (in Chinese).

FDA, 2015. Food and Nutrient database. Food and Drug Administration, Ministry

of Health and Welfare, Taiwan. Available at

http://www.fda.gov.tw/TC/siteList.aspx?sid=284 (in Chinese).

Feng, C.M., Ngai, C., Hung, C.H., 1989. A multiobjective programming model for

transportation and resources use. Journal of Management, 27-40.

Ferguson, E. L., Darmon, N., Fahmida, U., Fitriyanti, S., Harper, T. B.,

Premachandra, I. M., 2006. Design of optimal food-based complementary

feeding recommendations and identification of key “problem nutrients” using

goal programming. The Journal of nutrition 136(9), 2399-2404.

Gao, X., Wilde, P. E., Lichtenstein, A. H., Tucker, K. L., 2006. The 2005 USDA

Food Guide Pyramid is associated with more adequate nutrient intakes within

energy constraints than the 1992 Pyramid. The Journal of nutrition 136(5),

1341-1346.

Ghattas, H., 2014. Food Security and Nutrition in the context of the Global

Nutrition Transition. Technical paper. Rome: FAO.

HPA, 2016. Dietary Reference Intakes (DRIs). Health Promotion Administration,

Ministry of Health and Welfare, Taiwan. Available at

http://www.hpa.gov.tw/Pages/Detail.aspx?nodeid=544&pid=725 (in

Chinese).

22

HPA, 2017. How serious of obesity rate in Taiwan? Health Promotion

Administration, Ministry of Health and Welfare, Taiwan. Available at

http://obesity.hpa.gov.tw/TC/faqContent.aspx?id=67&chk=84b32626-2bca-

430c-982b-784a9b53fe69&param=pn%3D3 (in Chinese).

Huang, J.T., Ye, C.H., 2003. Lactose intolerance, Primary Medical Care & Family

Medicine 18, 146-150 (in Chinese).

Lin, S. C., Koo, M., Chen, M. W., Tsai, K. W., 2013. Prevention and care of

osteoporosis in the elderly. Tzu Chi Nursing Journal 12(4), 58-62.

Masset, G., Monsivais, P., Maillot, M., Darmon, N., Drewnowski, A., 2009. Diet

optimization methods can help translate dietary guidelines into a cancer

prevention food plan. The Journal of nutrition 139(8), 1541-1548.

Marcoe, K., Juan, W., Yamini, S., Carlson, A., Britten, P., 2006. Development of

food group composites and nutrient profiles for the MyPyramid Food

Guidance System. Journal of nutrition education and behavior 38(6), S93-

S107.

MOI, 2017. Aging index broken hundred Taiwan elderly for the first time more

than children. Ministry of Interior, Taiwan. Available at

https://udn.com/news/story/8042/2333220 (in Chinese).

MOHW, 2017. Statistical result of ten leading cause of death. Ministry of Health

and Welfare, Taiwan. Available at http://www.mohw.gov.tw/cp-16-33598-

1.html (in Chinese).

Misra, A., Khurana, L., 2008. Obesity and the metabolic syndrome in developing

countries. The Journal of Clinical Endocrinology & Metabolism 93, s9-s30.

Nutrition Information (2017). Academia Sinica, Taiwan. Available at

https://gao.sinica.edu.tw/health/howeat_one.html (in Chinese).

Okubo, H., Sasaki, S., Murakami, K., Yokoyama, T., Hirota, N., Notsu, A., Date,

C., 2015. Designing optimal food intake patterns to achieve nutritional goals

for Japanese adults through the use of linear programming optimization

models. Nutrition journal 14(1), 57.

Pan, W. H., Wu, H. J., Yeh, C. J., Chuang, S. Y., Chang, H. Y., Yeh, N. H., Hsieh,

Y. T., 2011. Diet and health trends in Taiwan: comparison of two nutrition

and health surveys from 1993-1996 and 2005-2008. Asia Pacific journal of

clinical nutrition 20(2), 238-250.

Peng, C. J., Lin, C. Y., Guo, H. R., 2015a. A comparison of food supply from 1984

to 2009 and degree of dietary westernization in Taiwan with Asian countries

and world continents. BioMed research international, 12pages.

Peng, C. J., Lee, M. S., Wahlqvist, M. L., Pan, W. H., Lee, W. C., Lin, C., Guo, H.

R., 2015b. Needs-based food and nutrient security indices to monitor and

23

modify the food supply and intakes: Taiwan, 1991–2010. Food Policy 57,

142-152.

Popkin, B. M. (2006). Global nutrition dynamics: the world is shifting rapidly

toward a diet linked with noncommunicable diseases. The American journal

of clinical nutrition 84(2), 289-298.

Popkin, B. M., Gordon-Larsen, P., 2004. The nutrition transition: worldwide

obesity dynamics and their determinants. International journal of obesity 28,

S2-S9.

Soden, P. M., Fletcher, L. R., 1992. Modifying diets to satisfy nutritional

requirements using linear programming. British Journal of Nutrition 68(03),

565-572.

USDA, 2014. National Health and Nutrition Examination Survey (NHANES)

2013-2014, U.S. Department of Agriculture (USDA). Retrieved from

https://www.ars.usda.gov/ARSUserFiles/80400530/pdf/1314/Table_1_NIN

_GEN_13.pdf

24

Appendix Table 1. 12 Food Group and 48 Subgroup

Food group code Food name Subgroup code Subgroup name

A Cereal/Roots A1 Rice

A2 Wheat

A3 Roots

A4 Legume

B Oil B1 Plant oil

B2 Animal oil

B3 Nut

C Poultries C1 Chicken

C2 Duck

C3 Other poultry

D Livestock D1 Pork

D2 Beef

D3 Other livestock

E Fishes E1 Fresh-water fishes

E2 Salt-water fish

E3 Fish haslet and fish product

E4 Other fish

F Other

proteins

F1 Egg

F2 Dairy

F3 Soybean

G Vegetables G1 Dark green vegetable

G2 Light green vegetable

G3 Shoots

G4 Melons

G5 Beans

G6 Mushrooms

G7 Other vegetables

G8 Pickled vegetables

G9 Kelp

H Fruits H1 Fresh Fruit

H2 Fruit product

H3 Fresh fruit juice

I Snacks I1 Bread

I2 Pastry

Other snacks I6

25

Food group code Food name Subgroup code Subgroup name

I3 Candy

I4 Beverage

I5 Processed juice

J Alcohol J Alcohol

K Seasoning K1 Sugar

K2 Salt

K3 Soy sauce

K4 Other seasonings

L Others L1 Instant food

L2 Sandwich and hamburger

L3 Dumpling

L4 Soup

L5 Others

Source: National Food Consumption Database

26

Appendix Table 2. Six Food Group and Corresponded Subgroup

Six food group Group

code

Subgroup Formula

Cereal/Roots A Rice, Wheat, Roots, Legume One serving of Cereal/Roots

=carbohydrate/15

K1 Sugar

I, L

Bread, Pastry, Snack, Candy,

Beverage, Processed juice, Instant noodles, Sandwich,

Hamburger, Dumpling, Soup,

Other

Soybean/

Fish/

Meat/

Egg

C Poultry One serving of

Soybean/

Fish/Meat/Egg

=protein/7

D Livestock

E Fish

F1 Egg

F3 Soybean

I1, I2, I6,

L

Bread, Pastry, Snack, Instant

noodles, Sandwich, Hamburger,

Dumpling, Soup, Other

One serving of

soy/fish/meat/egg=

[protein-(servings

of Cereal×2)]/7

Oil/Nuts B1 Plant oil One serving of

Oil/Nuts=fat/5 B2 Animal oil

B3 Nut

I1, I2, I6,

L

Bread, Pastry, Snack, Instant

noodles, Sandwich, Hamburger,

Dumpling, Soup, Other

One serving of Oil

=[fat-(serving soy/fish/meat/egg

×5)]/5

Fruits H One serving of Fruit

=energy/60

Dairy F2 One serving of

Dairy

=protein/7

Vegetables G One serving of

Vegetable

=energy/25

Source: NAHSIT (2005-2008); this study

27

Appendix Table 3. Definition of One Serving of Food

Food definition

One serving of cereal/root Cereal=1/4 bowl of rice, 1/2 bowl of

Porridge, 1/2 bowl of rice products (rice

noodles, winter powder),2 slices of toast

(thin)

Root=1/2 bowl of sweet potato (or potato,

taro, lotus root, yam, pumpkin)

One serving of vegetable =Not cooked edible part of about 100 g =

cooked vegetables a dish (diameter 15 cm

about the size of the disc) or about half bowl

One serving of fruit =Fruit is equivalent to a fist or about 1 rice

bowl cut all kinds of fruit

One serving of dairy = 240c.c low-fat or skim milk

= 240c.c low-fat low-sugar or sugar-free

yogurt

= 3 tablespoons of low fat or skim milk

powder

= 2 Low-fat cheese (45g)

One serving of soy/fish/meat/egg Soy=Soybeans 20 g or 1 dried bean curd or

traditional tofu (80g) (thick; 2 cells) (thin; 3

grid) or tender tofu about 1/2 box

Fish= fish 37.5g (about half palm, about 1 cm

thick) or shrimp 37.5g

Meat=Poultry, livestock 37.5g (about half

palm , about 1 cm thick)

Egg=1 egg

One serving of oil/nut Oil = 1 tablespoon vegetable oil (about 5 g)

Nut= 1 tablespoon of peeled nuts (about 10 g)