Download Development of Data-based Semi-quantitative

Jpn J Clin OncoI1998;28(11)679-687
Development of Data-based Semi-quantitative Food Frequency
Questionnaire for Dietary Studies in Middle-aged Japanese
Shinkan Iokudome', Masato Ikeda2, Yuko Tokudome3, Nahomi Imaeda4, Ikumi Kitagawa 5 and
Nakako Fujiwara 1
1Department of Public Health, Nagoya City University Medical School, Nagoya, 2Department of Occupational Health
Economics, University of Occupational and Environmental Health, Kitakyushu, 3Nagoya Bunri College, Nagoya,
4Nagoya City Personnel Health Management Center, Nagoya and 5Nagoya Seirei Junior College, Seto, Japan
Background: We designed a data-based semi-quantitative food frequency questionnaire to
clarify the relationship between food intake and lifestyle-related diseases among middle-aged
Japanese.
Methods: A total of 351 middle-aged individuals were recruited to a one-day weighed diet record
survey in 1994. In all, 586 foods were consumed. Intakeof 31 nutrients including energy, protein,
fat, carbohydrate, vitamins, minerals and dietary fiber by food was computed by multiplying the
weight of food consumed by its nutrient content. First, 252 foods with up to 90 cumulative 0/0
contribution to nutrientintakewere selected. Of these, foods having apparently the same/similar
nutrient content were combined into 206 foods by research dietitians. Next, 183 foods with up to
0.90 cumulative multiple regression coefficient and 90 cumulative % contribution were chosen.
At this stage an additional food grouping was made.
Results: Finally, 102 foods/recipes were included in the questionnaire: rice (2 items), bread and
noodles (11), eggs, milkand dairyproducts (10), soybean, soybean products andotherbeans (7),
meat including beef, pork and chicken (12), fish (5), other fish, shellfish and fish products (10),
green-yellow vegetables (8), other vegetables and mushrooms (7), edible roots (2), seaweeds
(3), seeds (2), fruits (8), beverages (7) and confectioneries (8). The frequencies were classified
intoeight categories. Portion size was calculated for the respective foodslargelyfrom the one-day
weighed diet record.
Conclusions: The developed semi-quantitative food frequency questionnaire substantially
coveredthe intakeof 31 nutrients andmaybecompetentto rankmiddle-aged Japanese efficiently.
Key words: weigheddiet record - contribution analysis - multipleregression analysis - semi-quantitative
food frequency questionnaire
INTRODUCTION
As is well known, the Japanese enjoy the longest life expectancy
in the world, which may be accounted for by genetic predisposition, modest lifestyle, improvements in sanitary environment and
a well-organized health care/medical/welfare system. Three-
Received February 2,1998; accepted August 10, 1998
For reprints and all correspondence: Shinkan Tokudome, Department of
Public Health, Nagoya City University Medical School, Mizuho-ku,
Nagoya 467-8601, Japan. E-mail: [email protected]
Abbreviations: SQFFQ, semi-quantitative food frequency questionnaire;
WDR, weighed diet record; MRA, multiple regression analysis; CA,
contribution analysis; TDF, total dietary fiber; SDF, soluble dietary fiber;
lDF, insoluble dietary fiber; SFA, saturated fatty acid; MUFA,
mono-unsaturated fatty acids; PUFA, polyunsaturated fatty acid; EPA,
eicosapentaenoic acid; DHA, docosahexaenoic acid; RDA, Recommended
Dietary Allowances
fifths of deaths are due to chronic diseases, such as cancer, heart
disease and cerebrovascular disease. Among lifestyle factors,
smoking, diet and exercise may be the top three. As with smoking
and physical exercise, food intake is controllable. However,
unlike either smoking, which is always harmful, or physical
exercise, which is beneficial if done moderately, diet has two
profiles. Food is primarily requisite and favorable, but excessive
or imbalanced intake may be deleterious. Thus, further research
is needed to elucidate the association between food intake and
health/disease, including cancer, on the basis of a dietary
questionnaire specifically and systematically developed in Japan.
Recently, we designed an evidence-based semi-quantitative
food frequency questionnaire (SQFFQ) (1-3) on the basis of a
one-day weighed diet record (WDR) according to multiple
regression analysis (MRA) and also contribution analysis (CA).
Although calibration/validation and reproducibility studies in
terms of food list, intake frequency and portion size are currently
in progress, it seemed informative to report the procedures
680
Development of data-basedSQFFQ
undertaken for preparing this SQFFQ and the foods largely
contributing to nutrient intake of interest and serving to classify
individuals efficiently into a tertile/quartile/quintile categorization.
SUBJECTS AND METHODS
SUBJECTS
A total of351 (171 males and 180 females) middle-aged Japanese
living in Aichi, Mie and Gifu Prefectures, in Central Japan,
parents of student dietitians attending a Junior College and
majoring in nutrition, participated in a one-day WDR survey
during October-November 1994. The mean ages ± standard
deviations for males and females were 50.2 ± 5.2 and 46.9 ± 4.4,
respectively.
WEIGHED DIET RECORD
The WDR was carried out on a weekday. Weight measurement
was done before cooking if food was prepared at home, otherwise
after cooking. Foods were weighed individually; while foods
such as soup were divided by the number of household members.
Not only completeness but also accuracy ofWDR were reviewed
by the respective students and research nutritionists.
NUTRIENTS OF INTEREST
The following 31 nutrients were selected: energy, protein, fat,
carbohydrate, total dietary fiber (TDF) [including soluble dietary
fiber (SDF) and insoluble dietary fiber (IDF)], minerals (including potassium, calcium, magnesium, phosphorus, iron, zinc and
copper) and vitamins (including carotene and vitamins A, C, D
and E). Fat was divided into saturated fatty acids (SFA),
mono-unsaturated fatty acids (MUFA) (including oleic acid),
polyunsaturated fatty acids (PUFA), n-6 PUFA, n-3 PUFA and
cholesterol. n-6 PUFA was subdivided into linoleic acid
(l8:2n-6) and arachidonic acid (18:3n-6) and n-3 PUPA was
divided into a-linolenic acid (18:3n-3), eicosapentaenoic acid
(EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3).
SELECTION OF FOODS/RECIPES
In all, 586 foods were consumed by the subjects. The nutrient
intake by food was computed by multiplying the food intake (in
grams) or serving size by the nutrient content per gram of food as
listed in the Standard Tables of Food Composition, Version 4 and
the Follow-up of the Standard Tables of Food Composition,
Version 4 (4,5). First, modified CA was applied to nutrients of
interest as follows (6,7). We defined the percentage contribution
of nutrient k by food i as the arithmetic mean of the individual %
contribution of nutrient k by food (lPCjik), which was estimated
by
586
ire;
= QjiD ik/
I
Q;iDik x 100,
i=1
586
if
I
i=l
Q;iD ik = 0 then Il'C, = 0 was assumed.
351
Percentage contribution of nutrient k by food i =
I
IPC jik/ 351,
j=l
wherej=l, ... ,351 subjects,i= 1, ... ,586foods,k= 1, ... ,31
nutrient factors, Q = grams of foods consumed and D = nutrient
content per gram of food. We selected 252 foods with up to 90
cumulative % contribution, except for 80% for potassium and
calcium. Those foods having apparently the same/similar nutrient
contents were combined into 206 foods by research nutritionists,
irrespective of the different cooking processes or appearances:
cooked bread (e.g. pre-cooked bread, sandwich, pizza); beef (e.g.
steak, grilled beef, Sukiyaki, Shabu-shabu); blue-skinned fish
(e.g. mackerel, sardines, horse mackerel, yellowtail, Spanish
mackerel); bone-edible small fish (e.g. Wakasagi, Shishamo,
Mezashi, Shirasuboshiy; other green-yellow vegetables (e.g.
Komatsuna, green pepper, garland chrysanthemum, leek, okra);
other vegetables (e.g. cucumber, lettuce, bean sprouts, onion,
eggplant, Chinese cabbage); edible roots (e.g. potato, sweet
potato, taro, yam) and so on. Next, not only CA but also forward
MRA (8) were performed by adopting total intake of a specific
nutrient as the dependent variable and overall amounts of nutrient
from 206 foods as the independent variables for the 351
individuals (9-11). We chose 183 foods with up to 90 cumulative
% contribution and 0.90 cumulative multiple regression coefficient/cumulative R2. However, seasonings, such as salt and Shoyu
(soy sauce), were excluded from the questionnaire and some
foods, such as brown rice, kiwi fruit, strawberries, potato chips
and milk powder, were intentionally included. At this phase we
did additional food grouping. Finally, 102 foods/recipes were
included in the SQFFQ.
Rice is a Japanese staple food and substantially provided most
nutrients. Rice intake was inquired into in a special manner, that
is, the intake frequency and portion size according to the size of
rice bowl for breakfast, lunch, supper and meal/snack between
meals, respectively.
Supplementary questions regarding fat and oil intake were also
prepared. Intake frequency of vegetable oil, mayonnaise, dressing,
butter, margarine, lard, etc. and foods/recipes cooked with such fat
and oil were asked about separately by dish. The brand names of
fat and oil consumed at home were further surveyed to assess fatty
acids intake. Sea foods such as fish, shellfish and seaweeds are also
part of the typical Japanese diet and they are major sources of n-3
PUFA such as EPA and DHA; therefore, food grouping was left
rather loose in order to ascertain the intake of such nutrients.
Other information was collected on the type of breakfast (e.g.
Japanese or Western type), the intake frequencies of fish and
meat, lean and fatty meat, and chicken skin, the frequency of
eating out and the type of food (e.g. Japanese, Chinese or
Western) along with sweetness and saltiness.
In addition, a separate lifestyle questionnaire included smoking, physical exercise and sleeping hours, inquiries with regard to
intake of alcohol, vitamin and mineral supplements, and functional (or designer) foods. The proportion of individuals taking
vitamin and mineral supplements, however, was very small in this
study population, and intake of vitamin and mineral supplements
was not taken into consideration.
Jpn J Clin OncoI1998;28( 11)
Table 1. Intake of macro-nutrients by the 351 subjects
Table 2. Number of foods contributing to 31 nutrients with up to 90
cumulative % contribution and 0.90 cumulative R2 by the 351 subjects
Male (171)
Female (180)
Mean (min-max)
Mean (min-max)
Energy
2163 kcal (1045-3619)
1785 kcal (956-3169)
Energy
Protein
83.5 g (33.1-149.1)
70.4 g (35.5-123.4)
Fat
62.0 g (22.4-146.2)
Cumulative % contribution
Cumulative R2
78
45
Protein
82
59
57.7 g (17.2-118.8)
Fat
61
41
35
18
50
25
Animal origin
24.7 g (0-84.1)
22.2 g (0-83.5)
Carbohydrate
Vegetable origin
30.5 g (4.2-84.1)
30.6 g (4.5-76.5)
SFA
6.8 g (0-31.7)
4.8 g (0-33.5)
Marine origin
Carbohydrate
281.6 g (112.3-606.2)
681
239.3 g (109.0-470.6)
INTAKE FREQUENCY
Intake frequencies were classified into eight categories: never or
seldom, 1-3 times per month, 1-2 times per week, 3-4 times per
week, 5-6 times per week, once a day, twice a day and three or
more times a day. For beverages, including coffee, green tea,
black tea and oolong tea, questions on intake frequency were left
open-ended when the answer was more than 5-6 times per week.
PORTION SIZE
The meanportion size (or average serving size) was calculatedfor
the respectivefoods from the one-day WDR and typical/standard
values and/or natural units from the literature were also taken into
account.Portion size was questioned whether the size was half as
much, the same, 1.5times larger or twice the averageserving size.
An open-ended response was also prepared.
RESULTS
STUDY SUBJECTS
Table 1 shows the intake of energy, protein, fat and carbohydrate
by study subjects. More macro-nutrients were consumed by
males than females. These nutrients were slightly less than the
national average in the respective age groups (12). Energy intake
was somewhat less than the Recommended Dietary Allowances
(RDA) in either sex (13). Protein intake was, however, slightly
higher than the RDA. Approximately 60 g of fat on average or
25% of the fat-energy ratio was consumed, which was slightly
above the RDA.
MUFA
47
30
PUFA
43
20
Cholesterol
35
8
Oleic acid
40
23
16
n-6 PUFA
30
Linoleic acid
33
16
Arachidonic acid
29
19
n-3 PUFA
42
16
a-Linolenic acid
29
9
EPA
50*
10
DHA
28
11
Vitamin C
27
10
Carotene
18
2
Vitamin A
29
5
Vitamin E
62
24
Vitamin D
25
6
Potassium (K)
90
42
Calcium (Ca)
76
24
Magnesium (Mg)
49
27
Phosphorus (P)
94
54
Iron (Fe)
90
37
Zinc (Zn)
38
Copper (Cu)
40
TDF
36
SDF
29
16
IDF
32
27
29
*Stopped at 85% partly because 51 subjects did not consume foods/recipes providing EPA.
SELECTION OF FOODS/RECIPES
As mentioned,586 foods were initially consumed by the subjects.
Firstly, CA was used and 252 foods with up to 90 cumulative %
contribution were chosen for 31 nutrients. Those foods having
apparentlythe same/similarnutrient contents were combined into
206 foods by the research nutritionists. Next, 183foods satisfying
not only the 90 cumulative % contribution but also 0.90
cumulative R2 were selected. The mean number of foods by CA
was 46.7 with a range of 18 for carotene to 94 for phosphorus,
which was significantly greater than the mean by MRA of 21.6
with a range of 1 for zinc and copper to 59 for protein (Table 2).
LIST OF FOODS/RECIPES
According to the categorizationscheme of the Standard Tablesof
Food Composition (4,5), 102 foods/recipes were included in the
SQFFQ as follows: rice (2 items), bread and noodles (11), eggs,
milk and dairy products (10), soybean, soybean products and
other beans (7), meat including beef, pork and chicken (12), fish
(5), other fish, shellfish and fish products (10), green-yellow
vegetables (8), other vegetables and mushrooms (7), edible roots
(2), seaweeds (3), seeds (2), fruits (8), beverages (7) and
confectioneries (8) (see Appendix).
Developmentofdata-based SQFFQ
682
Table3. Percentage contribution, cumulative % contribution and cumulative R2 of the top 20 foods for energy
Rank
Food
Cumulative
Rank
Food
% contribution
Cumulative R2
Well-milled rice
31.0
0.35
% contribution
Well-milled rice
31.0
2
White bread
35.6
2
Beer
2.0
0.43
3
Chicken egg
39.9
3
White bread
4.6
0.49
4
Milk (whole)
42.4
4
Beef (chuck loin)
0.8
0.52
5
Salad oil (mixed)
44.7
5
Chinese noodles
1.2
0.55
6
Udon/Soba (Japanese noodles)
46.8
6
Sake
l.l
0.58
7
Beer
48.8
7
Arare (rice cake cubes)
0.4
0.61
8
White sugar
50.4
8
Wheat flour
1.4
0.64
9
Tofu (soybean curd)
52,1
9
Milk (whole)
2.4
0,66
10
Wheat flour
53.4
10
Mayonnaise (egg yolk type)
1.2
0.68
II
Vegetable oil (mixed)
54,7
II
Pork (ground meat)
0.6
0,70
12
Pork (Boston butt)
56,0
12
Chicken egg
4.4
0.71
13
Mayonnaise (egg yolk type)
57,2
13
Mayonnaise(whole egg type)
0.6
0.73
0,3
0,74
14
Koji-miso (soybean paste)
58.5
14
Dorayaki (bean-jam pancake)
15
Chinese noodles
59.7
15
Mirin (sweet Sake used for seasoning)
0.4
0.75
16
Sake
60,8
16
Bread crumbs
0.4
0,76
17
Salmon
61.8
17
Udon/Soba (Japanese noodles)
2.1
0.77
18
Potatoes
62,8
18
Salad oil (mixed)
2.3
0.78
19
Beef (chuck loin)
63.6
19
Beef (sirloin)
0,2
0,79
20
Abura-age (fried soybean curd)
64,S
20
Under-milled rice
0,2
0.79
Table4. Percentage contribution, cumulative % contribution and cumulative R2 of the top 20 foods for protein
Rank
Food
Cumulative
Rank
Food
% contribution
Cumulative R2
Well-milled rice
14.7
0.16
% contribution
Well-milled rice
14,7
2
Chicken egg
23,2
2
Round herring (salted and dried, round)
0,3
0.21
3
White bread
27,2
3
Chicken egg
8.6
0.26
4
Tofu (soybean curd)
30.8
4
Yellowfin tuna
1.0
0,31
5
Salmon
34,0
5
Shoyu (soy sauce)
2,3
0,35
6
Milk (whole)
37,1
6
Dried squid
0,2
0.39
7
Koji-miso (soybean paste)
39,6
7
Pork (ground meat)
1.0
0.43
8
Coffee drink (canned)
0,2
0.47
Koji-miso (soybean paste)
2,5
0.50
Chicken (breast)
0.5
0,52
8
Shoyu (soy sauce)
9
Mushi-Kamaboko (steamed fish paste)
43,8
9
Pork (Boston butt)
45.7
10
10
41.9
11
Pacific saury
47.4
II
Milk (whole)
3.1
0,54
12
Beef (chuck loin)
48,8
12
Chinese noodles
l.l
0,56
13
Udon/Soba (Japanese noodles)
50.2
13
White bread
3.9
0,58
14
Shiba-shrimp (prawn)
51.4
14
Under-milled rice
0.1
0,60
IS
Beef (flank)
52,5
15
Chicken (breast), sasami (fillet)
0,7
0.61
16
Chicken (wing)
53.5
16
Chicken (wing)
l.l
0.63
17
Chinese noodles
54,6
17
Salmon
3,2
0.64
18
Beef (inside round)
55.6
18
Lettuce
0.1
0.65
19
Chicken (thigh)
56.7
19
Squid
0.8
0,66
20
Abura-age (fried soybean curd)
57.7
20
Beef (sirloin)
0.4
0,67
Jpn J Clin OncoI1998;28( 11)
Table5. Percentage contribution, cumulative % contribution and cumulative R2 of the top 20 foods for fat
Rank
Food
Cumulative
Rank
Food
% contribution
Cumulative R2
% contribution
Chicken egg
10.3
Pork (ground meat)
1.2
0.10
2
Salad oil (mixed)
18.0
2
Mayonnaise (egg yolk type)
4.0
0.18
3
Vegetable oil (mixed)
22.5
3
Pork (Boston butt)
3.0
0.25
4
Milk (whole)
26.8
4
Chicken egg
10.3
0.31
5
Mayonnaise (egg yolk type)
30.8
5
Salad oil (mixed)
7.7
0.36
6
Well-milled rice
34.7
6
Beef (sirloin)
0.5
0.41
7
Tofu (soybean curd)
38.3
7
Safflower oil
1.8
0.45
8
Pork (Boston butt)
41.3
8
Vegetable oil (mixed)
4.5
0.48
9
Abura-age (fried soybean curd)
43.8
9
Beef (chuck loin)
2.2
0.53
10
Margarine
46.2
10
Milk (whole)
4.3
0.56
11
White bread
48.5
11
Mayonnaise (whole egg type)
1.9
0.59
12
Beef (chuck loin)
50.7
12
Margarine
2.3
0.61
13
Koji-miso (soybean paste)
52.7
13
Sausage (Vienna)
1.5
0.64
14
Mayonnaise (whole egg type)
54.6
14
Butter
1.2
0.66
15
Safflower oil
56.4
15
Butter peanuts
0.4
0.67
16
Pacific saury
58.1
16
Beef (ground meat)
1.1
0.69
17
Salmon
59.6
17
Pork (belly)
1.1
0.70
18
Sausage (Vienna)
61.1
18
Peanuts (dried)
0.2
0.72
19
Beef (flank)
62.6
19
Bacon
0.7
0.73
20
Corn oil
63.9
20
Beef (flank)
1.4
0.75
Table 6. Percentage contribution, cumulative % contribution and cumulative R2 of the top 20 foods for carbohydrate
Rank
Food
Cumulative
Rank
Food
% contribution
Well-milled rice
50.5
0.45
Udon/Soba (Japanese noodles)
3.1
0.51
Cumulative R2
% contribution
Well-milled rice
50.5
2
White bread
57.1
3
White sugar
60.2
3
Chinese noodles
1.8
0.57
4
UdonrSoba (Japanese noodles)
63.3
4
White bread
6.6
0.62
2
5
Wheat flour
65.5
5
Coffee drink (canned)
0.9
0.67
6
Chinese noodles
67.3
6
Wheat flour
2.2
0.70
7
Potatoes
68.9
7
Arare (rice cake cubes)
0.6
0.74
8
Milk (whole)
70.4
8
Mandarin orange
1.4
0.77
9
Kaki (Japanese persimmon)
71.9
9
Dorayaki (bean-jam pancake)
0.5
0.78
10
Mandarin orange
73.3
10
An-pan (bean-jam bun)
0.7
0.80
11
Apple
74.6
11
Macaroni
1.0
0.82
12
Beer
75.9
12
Short cake
0.6
0.83
13
Macaroni
77.0
13
Kaki (Japanese persimmon)
1.5
0.84
14
Banana
77.9
14
Under-milled rice
0.3
0.86
15
Coffee drink (canned)
78.8
15
Sweet potatoes
0.6
0.87
16
Onion
79.5
16
Manju (steamed bun with bean-jam filling)
0.4
0.88
17
An-pan (bean-jam bun)
80.2
17
Buckwheat flour
0.5
0.90
18
Pre-cooked Chinese noodles
80.9
18
Pre-cooked Chinese noodles
0.7
0.91
19
Shoyu (soy sauce)
81.6
19
Potatoes
1.6
0.91
20
Bread crumbs
82.2
20
Pumpkin
0.5
0.92
683
684
Development of data-based SQFFQ
FOODS/RECIPES FOR MACRONUTRIENTS
The percentage contribution, cumulative % contribution and
cumulative R2 of the top 20 foods/recipesfor energy, protein, fat
and carbohydrate are shown in Table 3-6, respectively.
One-third of the energy was contributed by well-milled rice,
followed by white bread, chicken egg, milk (whole) and salad oil
(mixed) according to CA (Table 3). Well-milledrice, beer, white
bread, beef (chuck loin) and Chinesenoodles were selectedin that
order by MRA.
CA revealed that approximately 15% of total protein was
derived from well-milled rice, followed by chicken egg, white
bread, tofu (soybean curd) and salmon(Table4). Well-milledrice,
round herring (salted and dried, round), chicken egg, yellowfin
tuna and Shoyu (soy sauce) were chosen by MRA.
For fat, chicken egg was ranked at the top based on CA,
followed by salad oil (mixed), vegetable oil (mixed), milk
(whole) and mayonnaise (egg yolk type) (Table 5). Pork (ground
meat), mayonnaise (egg yolk type), pork (Boston butt), chicken
egg and salad oil (mixed) were selected by MRA.
More than half the carbohydrate was supplied by well-milled
rice, followed by white bread, white sugar, Udon/Soba (Japanese
noodles) and wheat flour on the basis of CA (Table 6).
Well-milled rice, Udon/Soba, Chinese noodles, white bread and
coffee drink (canned) were chosen by MRA.
efficientin categorizingindividuals,but unsuitablefor computing
the absolute nutrient level.
Substantial foods selected by MRA were covered by those
chosen by CA; rather, specific foods were only chosen by MRA.
The ranking was very different by each type of analysis.
Well-milled rice, a Japanese staple food, is an exception; it was
not only the greatest source of variance contribution but also the
largest source of absolute contribution to energy, protein and
carbohydrate. Chicken egg, for example, was ranked third in
energy by CA but twelfth by MRA. On the other hand, Chinese
noodles were a minor source of absolute contribution to energy
but a major source of variance contribution.
It appeared odd that Shoyu (soy sauce) was chosen as a major
provider of protein according to CA and MRA. Shoyu, however,
is often used for seasoningof foods, including meat and fish, and
seemed to have contributed substantially to absolute intake and
variance in intake of protein. We know of an analogous
observation where popcorn, cholesterol-freein itself, was picked
as a major contributor to cholesterol, which may be plausibly
explained by the fact that popcorn is ordinarily seasoned with
butter in the USA (1).
Table 7. Percentage coverage of 31 nutrients by the
Energy
PERCENTAGE COVERAGE OF NUTRIENTS
BY
SQFFQ
The percentage coverage of 31 nutrients by the SQFFQ was
computed assuming all 586 foods as unity (Table 7). As
mentioned, fat and oil intake in supplementary items in the
SQFFQ and alcohol intake in a lifestyle questionnaire were taken
into account. The average was 94% with a range from 85% for
iron to 99% for a-linolenic acid.
DISCUSSION
Admittedly, the sample size was not largeenough and the one-day
WDR was done in a specific seasonand in a selected area in Japan.
We were naturally unableto estimate within-individual variation on
the basis of a one-day WDR. The present data therefore were not
necessarily accurate or complete in terms of quality and quantity.
The averageintakeof macro-nutrients, however, was somewhat less
than the national data published (12) and slightly greater than the
RnA (13); hence the profiles of study subjects were unlikely to be
very different from those of the general population.
In choosing foods/recipes, there are two contrasting methods
(14-15): one is based on CA (6,7) and the other on MRA (9-11).
Each method has its particular advantagesand disadvantages.The
former approach is based on absolute intake and is useful for
making energy adjustments. Hence this procedure is especially
suitable for studies to clarify the association with absolute
nutrient intake but often inappropriate for categorizing individuals. The latter is based on the variance of nutrient intake. The
cumulative R 2 was explained by a significantly smaller number
of foods than cumulative % contribution, that is, MRA may be
91
Protein
89
Fat
93
Carbohydrate
91
SFA
96
MUFA
97
PUFA
98
Cholesterol
97
Oleic acid
96
n-6 PUFA
98
Linoleic acid
98
Arachidonic acid
97
n-3 PUFA
98
a-Linolenic acid
99
EPA
96
DHA
96
Vitamin C
93
Carotene
95
Vitamin A
95
VitaminE
96
Vitamin D
96
Potassium (K)
87
Calcium (Ca)
91
Magnesium (Mg)
86
Phosphorus (P)
89
Iron (Fe)
8S
Zinc (Zn)
93
Copper (Cu)
94
TDF
93
SDF
92
IDF
93
SQFFQ
Jpn J Clin OncoI1998;28(11)
Seasonings, such as salt and Shoyu, were selected by CA and
MRA for certain minerals including sodium and magnesium.
However, they were excluded from the SQFFQ because it is
difficult to weigh seasonings accurately and measure actual
intake. On the other hand, brown rice, kiwi fruit, strawberries,
potato chips and powdered milk were not consumed by the
subjects because some of them have seasonal variations, but they
were included because they are important sources of vitamins,
minerals and fat.
The percentage coverage of selected nutrients by the SQFFQ
was 94% on average with a range from 85% for iron to 99% for
a-linolenic acid. The coverage of certain nutrients may be
overestimated, owing in part to the incompleteness of the
composition table. The values were more than 81 %, mathematically computed by multiplying 0.90 by 0.90 (that is, we selected
foods according to 90 cumulative % contribution twice),
substantially because foods chosen for other nutrients have
covered a specific nutrient and supplementary questions in terms
of fat and oil intake eventually attained a high coverage of fat,
fatty acids and related nutrients. Consequently, the SQFFQ
developed seemed efficient not only in estimating the absolute
intake of nutrients, except for certain minerals, but also in
categorizing individuals by variance in intake.
Although the reproducibility of dietary questionnaires has
often been evaluated in Japan (16-18), few validation studies
have been carried out (19-22). To our knowledge, few data-based
Japanese SQFFQs have been developed. Questionnaires were
frequently prepared by selecting foods on the basis of the nutrient
content from the food composition tables according to researchers' interest, by citing historical questionnaires or by
referring to National Nutritional Surveys. Such questionnaires
appeared unsatisfactory for elucidating specific hypotheses. They
were mostly designed to ask only about frequency and seemed
inefficient in calculating nutrient intake and values adjusted for
energy intake. Japanese dietitians have been concerned about
calculating food and nutrient intake and giving advice on whether
a client's nutrient intake is deficient or excessive. Epidemiologists, nutritionists and biostatisticians should collaborate to
prepare a valid and reliable SQFFQ to be applied to a particular
population under a specific hypothesis.
APPENDIX
LIST OF FOODS/RECIPES INCLUDED IN A SEMI-QUANTITATIVE
FOOD FREQUENCY QUESTIONNAIRE
RICE
1. Well-milled rice
2. Brown rice, under-milled rice with embryo, rice boiled with
barley
685
6. Pre-cooked bread, sandwiches, pizza
7. Japanese noodle [Udon (wheat noodle)]
8. Japanese noodle [Soba (buckwheat noodle)]
9. Spaghetti, macaroni au gratin
10. Pre-cooked noodles (dried snack noodle)
11. Chinese noodles [Ramen (Chinese noodle), fried noodle,
Champon (Chinese-style hotchpotch with noodle), etc.]
12. Okonomiyaki (Japanese pancake, a thin, flat, unsweetened
oil fried with bits of vegetables, meat or shellfish)
13. Mochi (Japanese rice cake)
EGGS. MILK AND DAIRY PRODUCTS
14. Eggs
15. Milk (whole)
16. Milk (condensed)
17. Milk (low fat)
18. Milk powder
19. Lactic acid beverage
20. Yogurt
21. Cheese
22. Ice cream (>8% milk fat)
23. Other ice cream
SOYBEAN, SOYBEAN PRODUCTS AND OTHER BEANS
24. Tofu (soybean curd) for Miso soup
25. Tofu, Yakko tofu (tofu cut in cubes), tofu steak, Mabo-dofu
26. Nama-age (fried tofu), Agedashi-dofu (lightly fried tofu),
Ganmodoki (fried tofu paste)
27. Abura-age (fried tofu)
28. Natto (fermented soybean), soybean
29. Koji-miso (soybean paste) (Miso soup, vinegar Miso)
30. Koya-dofu (frozen and dried tofu)
MEAT
31. Ground chicken [e.g. meatball, Soboro (powdered
chicken)]
32. Beef and pork ground together (e.g. hamburger, rolled
cabbage)
33. Ground beef (e.g. hamburger, meat sauce)
34. Ground pork (e.g. dumpling)
35. Chicken
36. Beef (e.g. steak, grilled beef, Sukiyaki, Shabu-Shabuy
37. Other beef (e.g. fried with vegetables, cooked with
vegetables)
38. Pork
39. Ham
40. Sausage (except fish sausage)
41. Bacon
42. Liver
FISH
BREAD AND NOODLES
3. White bread, soft rolls
4. Croissant
5. Bun, bean-jam bun
43. Salmon, trout
44. Eel
45. Blue-skinned fish (e.g. mackerel, sardines, horse mackerel,
yellowtail, Spanish mackerel)
686
Development ofdata-based SQFFQ
46. Red-meat fish (e.g. tuna, bonito)
47. White-meat fish (e.g. red bream, sea bream, cod, flatfish)
OTHER FISH AND SHELLFISH
48. Bone-edible small fish [e.g. Wakasagi (fresh-water smelt),
Shishamo (smelt), Mezashi (salted and semi-dried sardines),
Shirasuboshi (boiled and semi-dried whitebait)]
49. Cod roe, salmon roe
50. Canned tuna
51. Oyster
52. Shellfish (e.g. short-necked clam, corbicula)
53. Dried cuttlefish/squid
54. Cuttlefish, squid, octopus
55. Shrimp, Crab
56. Fried fish paste products (e.g. Satsuma-age, Agemono)
57. Fish paste products (e.g. Kamaboko, Chikuwa)
GREEN-YELLOW VEGETABLES
58.
59.
60.
61.
62.
63.
64.
Spinach
Pumpkin/squash
Carrot
Broccoli
Tomato
Vegetable juice, tomato juice
Other green-yellow vegetables (e.g. Komatsuna, green
pepper, garland chrysanthemum, leek, okra)
65. Pickles (e.g. pickled green-yellow vegetables, Takuan)
FRUITS
80.
81.
82.
83.
84.
85.
86.
87.
Citrus fruits (e.g. orange, tangerine, mandarin orange)
Fruit juice
Kaki (Japanese persimmon)
Banana
Apples
Strawberries
Kiwi fruit
Other fruits
BEVERAGES
88.
89.
90.
91.
92.
93.
94.
Coffee drink (canned)
Coffee (regular, instant)
Black tea
Sugar for coffee and black tea
Green tea
Oolong tea, Chinese tea
Other refreshments
CONFECTIONERIES
95.
96.
97.
98.
99.
100.
101.
102.
Kasutera (sponge cake)
Sembei (rice crackers), Arare (rice cake cubes)
Japanese style confectioneries (Manju, etc.)
Potato chips
Doughnut
Chocolate, chocolate cake
Cookie
Cake (e.g. short cake, cream puff)
OTHER VEGETABLES
66.
67.
68.
69.
Cabbage
Daikon (Japanese radish), turnip
Burdock, bamboo shoot
Other vegetables (e.g. cucumber, lettuce, bean sprouts,
onion, eggplant, Chinese cabbage)
70. Kiriboshi-daikon (dry strips Japanese radish)
MUSHROOMS
71. Shiitake (mushroom) (dried)
72. Shiitake (raw), Shimeji (champignon), etc.
EDIBLE ROOTS
73. Potatoes (e.g. potato, sweet-potato, taro, yam)
74. Kon-nyaku (devil's tongue)
SEAWEEDS
75. Toasted laver, toasted and seasoned laver
76. Hijiki (brown algae), Kombu (kelop)
77. Wakame (seaweed)
SEEDS
78. Sesame
79. Peanut, almond
Acknowledgments
The study was partly sponsored by a Grant-in-Aid from the
Ministry of Education, Science, Sports and Culture (06454242).
The authors thank Ms Y. Kubo and Ms Y: Ito for their technical
assistance in preparing the manuscript.
References
1. Willett W. Nutritional Epidemiology. New York: Oxford University Press,
1990.
2. Margetts BM, Nelson M. Design Concepts in Nutritional Epidemiology. New
York: Oxford University Press, 1991.
3. Thompson FE, Byers T. Dietary assessment resource manual. J Nutr
1994; 124:S2245-S317.
4. Resources Council, Science and Technology Agency, Japan. Standard Tables
of Food Composition in Japan, 4th ed. Tokyo: Resources Council, Science
and Technology Agency, Japan, 1982 (in Japanese).
5. Resources Council, Science and Technology Agency, Japan. Follow-up of
Standard Tables of Food Composition in Japan. Tokyo: Ishiyaku Shuppan,
1992 (in Japanese).
6. Block G, Dresser CM, Hartman AM, Carrol MD. Nutrient sources in the
American diet: quantitative data from the NHANES II survey. I. Vitamins and
minerals. Am J EpidemioI1985;122:13-26.
7. Freudenheim JL, Krogh V, d' Amicis A, Scaccini C, Sette S, Ferro-Luzzi A, et
al. Food sources of nutrients in the diet of elderly Italians: I. Macronutrients
and lipids. Int J EpidemioI1993;22:855-68.
8. SAS Institute. SAS/STAT User's Guide, Version 6, 4th ed., Vol. 2. Cary, NC:
SAS Institute, 1990.
9. Hankin ill, Stallones RA, Messinger HB. A short dietary method for
epidemiologic studies. m. Development of questionnaire. Am J Epidemiol
1968;87:285-98.
Jpn J Clin OncoI1998;28(1l)
10. Byers T, Marshall J, Fiedler R, Zielezny M, Graham S. Assessing nutrient
intake with an abbreviated dietary interview. Am J Epidemiol
1985;122:41-50.
11. Overvad K, Tjenneland A, Haraldsd6ttir J, Ewertz M, Jensen OM.
Development of a semiquantitative food frequency questionnaire to assess
food, energy and nutrient intake in Denmark. Int J Epidemiol1991 ;20:900-5.
12. Health Promotion and Nutrition Division, Health Service Bureau, Ministry of
Health and Welfare. Recommended Dietary Allowances for the Japanese, 5th
revision. Tokyo: Daiichi Shuppan, 1996.
13. Health Promotion and Nutrition Division, Health Service Bureau, Ministry of
Health and Welfare. Status of National Nutrition. Results of National
Nutrition Survey in 1994. Tokyo: Daiichi Shuppan, 1996 (in Japanese).
14. Stryker WS, Salvini S, Stampfer MJ, Sampson L, Colditz GA, Willett WC.
Contributions of specific foods to absolute intake and between-person
variation of nutrient consumption. J Am Diet Assoc 1991;91:172-8.
15. Mark SD, Thomas DG, Decarli A. Measurement of exposure to nutrients: an
approach to the selection of informative foods. Am J Epidemiol
1996;43:514-21.
16. Fukao A, Shimizu H, Maezawa M, Hisamichi S. The reproducibility of
answer to the questionnaire for food intake frequency in a survey. Nihon
KoshuEiseiZasshi 1990;37:347-52 (in Japanese).
687
17. Ozasa K, Watanabe Y, Higashi A, Liang H, Hayashi K, Shimouchi A, et al.
Reproducibility of a self-administered questionnaire for dietary habits,
smoking and drinking. Nihon Eiseigaku Zasshi 1994;48:1048-57 (in
Japanese).
18. Fujiwara N, Tokudome S. Reproducibility of self-administered questionnaire
in epidemiological surveys. J EpidemioI1997;7:61-9.
19. Nakamura M, Aoki N, Nasu K, Kondo 1. A comparison between a food
frequency and amount questionnaire and 7-day diet record with weighing.
Nihon KoshuEiseiZasshi 1994;41:682-92 (in Japanese).
20. Takatsuka N, Kawakami N, Kawai K, Okamoto Y, lshiwata K, Shimizu H.
Validation of recalled food intake in the past in a Japanese population. J
EpidemioII996;6:9-13.
21. Tsubono Y,Takamori S, Kobayashi M, Takahashi T, Iwase y, Iitoi Y,et al. A
data-based approach for designing a semiquantitative food frequency
questionnaire for a population-based prospective study in Japan. J Epidemiol
1996;6:45-53.
22. Date C, Yamaguchi M, Tanaka H. Development of a food frequency
questionnaire in Japan. J EpidemioI1996;6:S131-S6.