High education is associated with low fat and high fibre, beta-carotene and vitamin C

Norsk Epidemiologi 2000; 10 (1): 57-62
57
High education is associated with low fat and
high fibre, beta-carotene and vitamin C –
Computation of nutrient intake based on a short food frequency
questionnaire in 17,265 men and women in the Tromsø Study
Bjarne Koster Jacobsen1 and Hugo Nilsen2
1Institute of Community Medicine, University of Tromsø, Tromsø
2Department of Clinical Nutrition, University Hospital of Tromsø, Tromsø
Correspondence: Bjarne K. Jacobsen, Institute of Community Medicine, University of Tromsø, N-9037 Tromsø, Norway
ABSTRACT
Educational level has been correlated to the intake of several nutrients. In a population-based study
including 17,265 men and women aged 25-69 years, the intake of nutrients were calculated based on 37
questions about food habits. In this paper, we present results from the dietary survey with emphasis on
the relationships between dietary habits and educational level. Compared to subjects with low formal
education, subjects with high educational level have less fat in their diet and more dietary fibre, beta-
carotene, vitamin C and alcohol (p-value for linear trend < 0.001). Our results confirm that high educa-
tion is associated with healthy food habits and relatively higher alcohol consumption. There is a need for
efforts in order to change the food habits of the less educated.
NORSK SAMMENDRAG
Personer med lang utdanning har ofte et bedre kosthold enn personer med kortere utdanning. I denne
undersøkelsen har vi estimert inntaket av en rekke næringsstoffer basert på 37 spørsmål om kostvaner
som ble stilt til personer som tok del i Tromsø-IV-undersøkelsen (1994/95). Vår studie inkluderer 17 265
menn og kvinner i Tromsø i alderen 25-69 år. Vi presenterer resultater fra denne kostholdsundersøkelsen
med vekt på relasjoner mellom kostvaner og utdanningslengde. Sammenlignet med personer med kort
formell utdanning, har personer med lang utdanning mindre fett i kosten og høyere inntak av fiber, beta-
karoten, vitamin C og alkohol (p < 0.001). Resultatene bekrefter at personer med lang utdanning har et
helsemessig gunstigere kosthold, men et høyere alkoholinntak, enn personer med kort utdanning.
Funnene understreker behovet for målrettede tiltak for å utjevne sosiale forskjeller i kostvaner i Norge.
INTRODUCTION
food items in self-administered questionnaires, not to
calculate the intake of nutrients. This focus has given
It is a rather consistent finding in nutritional epidemio-
interesting information about relationships between
logy that higher socio-economic class, often measured
level of education and food and alcohol habits (2,9).
as high level of education (1), is associated with a
Detailed information about the nutrient intake of an
healthy diet. This has been demonstrated in Norway
individual can be obtained by applying dietary survey
(2-5) and in other countries like Scotland (6), the US
methods like the dietary history, repeated 24-hour re-
(7) and Finland (8). There have, however, been
calls, direct recording (or even weighing) of the foods,
exceptions. In Finland, Roos and co-workers could not
or a comprehensive food frequency questionnaire (10).
confirm that subjects with higher education had lower
Unfortunately, financial and logistic constrains usually
energy from fats or from saturated fats (8).
make it impossible to use these dietary survey methods
In order to show relationships between socio-
in large population samples.
economic class and food habits as well as nutrient in-
It is, however, known that information about the
take in the general population, large population studies
frequency of intake of a limited number of food items
are needed. In previous population studies in Tromsø,
explains a large part of the between-person variation in
the ambitions have primarily been to include informa-
nutrient intake (11-15). This means that there is a rela-
tion about the frequency of intake of some important
tively high correlation coefficient between the nutrient

---

B.K. JACOBSEN AND H. NILSEN
58
intake estimated from a few food items and that esti-
holic beverages were included, but not coffee, tea and
mated from a comprehensive dietary survey method
“light” (sugar free) beverages, as they do not contri-
covering the intake of many food items.
bute to the energy intake.
Thus, several research groups have estimated ener-
Missing values, albeit few for each subject, were
gy and nutrient intake from relatively short question-
substituted with a value. This value was either the mo-
naires. In a previous Norwegian study, the intakes of
dal value based on the distribution from the respon-
nutrients were calculated based on 31 food frequency
dents, or the most likely value based on the rest of the
questions (16). We have in a similar way calculated
answers. For example: If the respondent stated not to
the nutrient intake of more than 17,000 men and wo-
eat bread, but failed to answer the question about the
men in Tromsø, Norway, based on information from
number of slices of bread with cheese, we assumed
two self-administered questionnaires included in the
that the respondent does not eat bread with cheese. Si-
fourth Tromsø Study (1994/95).
milarly, if the respondent had indicated the consump-
There were therefore two aims of this study. The
tion of whole fat milk, but did not answer the question
first aim was to present relationships between level of
about skimmed milk, we assumed that the subject did
education and the intake of fats, dietary fibre, beta-
not drink skimmed milk. It should be kept in mind that
carotene and vitamin C, relationships of interest for
substitutions for missing values only were performed
preventive medicine. The second aim was to describe
if at least 31 out of 34 questions had been answered.
how we estimated the intake of nutrients in the fourth
The nutrients were estimated based on the 33 food
Tromsø Study. This is of interest as the dietary survey
frequency questions, the information about type and
is the basis for our study.
amount of table fat, the types of fats used in the
cooking and type of bread used. Portion sizes for
different types of foods (e.g., slice of bread, dinner
SUBJECTS AND METHODS
courses, vegetables) were estimated for each gender
based on data from previous dietary history surveys in
The fourth Tromsø Study took place in 1994 and 1995.
northern Norway (17,18). The intakes of food groups
It included all subjects aged 25 and above (born in
like meat, lean fish, fatty fish, fish products (e.g., fish
1969 or earlier) in the municipality of Tromsø, Nor-
patties) and processed meat products (e.g., sausages)
way, with approximately 60,000 inhabitants. The great
were calculated using recipes. These recipes reflected
majority of the population lives in the city of Tromsø.
the usual intake in northern Norway of the food items
Only subjects aged 25-69 are included in the present
included in the food group. The nutrient intakes were
study. The invitation to participate in the survey came
calculated using the Norwegian food composition
with a questionnaire about alcohol habits (teetotaller,
table (19). In the case of missing values, data from the
and how much beer, wine and spirits the respondent
corresponding Swedish food composition table were
consumed in a normal two week period). There were
applied (20).
also questions about type of table fat, educational level
We computed the total energy intake, fats (total
and two questions about physical activity in leisure
fats, saturated-, monounsaturated-, and polyunsatura-
time (low and high intensity, respectively).
ted fats as well as n-3 fatty acids), protein, carbo-
A total of 24,358 men and women aged 25-69 at-
hydrates, sugar, cholesterol, fibre, iron, calcium, beta-
tended the screening (76 % of the eligible population).
carotene, thiamine, riboflavin, folate and vitamins B6,
At the screening, attenders were given another 4-page
B12, A, C, D and E. The intake of nutrients from cod
questionnaire. This questionnaire included a question
liver oil and supplements were not included.
about physical activity at work. There were also 35
About 72 % of the subjects who got the question-
questions about food habits, including the amount of
naire returned it with at least 31 of the 34 questions an-
spread fat, type of fats used in the cooking and type of
swered. Thus, nutrients were calculated for 8486 men
bread used. There were two types of frequency ques-
and 9132 women. Subjects with very low or very high
tions. The first type included daily consumption of 14
energy intake (< 1 and > 99 percentile) were excluded
different food items used daily (e.g., glasses of milk or
from the present analyses. The analyses presented
slices of bread per day) with six consumption cate-
therefore include 8316 men and 8949 women, or about
gories from “0” to “More than 6”. The second type
53 % of the population aged 25-69 in Tromsø. Due to
concerned weekly consumption of 18 different food
missing values for other variables, the number of
items/food groups (e.g., lean fish, yoghurt or carrots),
subjects included in some of the analysis is somewhat
also with six consumption categories from “Never” to
lower.
“About daily”. The questionnaires are available upon
The statistical analyses, mainly analysis of vari-
request from the authors.
ance, were performed using the SAS package (21).
In order to be included in the computation of ener-
gy and nutrients, the respondent had to have answered
RESULTS
at least 31 of 34 questions: The question about type of
table fat used and the frequency of use of 33 food
The energy intake and the proportion of energy from
items that contribute to the energy intake. Thus, alco-
fat, protein, carbohydrates and alcohol according to

---

EDUCATION AND NUTRIENT INTAKE
59
age and gender are given in Table 1. Fats contributed
were 7.4 mg and 156 mg. When represented as abso-
31 % of the energy, protein 17 %, carbohydrates 51 %
lute intake of nutrients (not per 10 MJ), we found for
and alcohol 1 %. Energy from sugar represented 8 %
both men and women highly significant relationships
of the total energy intake. The energy intake was
between the two nutrients and educational level.
inversely related to age and lower in women than in
men. Table 2 shows positive statistically significant
age-adjusted relationships between physical activity at
work and leisure and the total energy intake.
Table 2. Total age-adjusted energy intake per day according
The age-adjusted relationships between level of
to physical activity in leisure (low intensity and high inten-
education and the intake of fats and alcohol (as percent
sity) and work (type of work). Tromsø 1994/95.
of the energy intake), dietary cholesterol and fibre (as

gram per 10 MJ) are displayed in table 3. The intake of
Men
Women

fats and cholesterol (in women) was inversely related
Total
Total
Number
energy
Number energy
to educational level, whereas the intake of alcohol and
of
intake
of
intake
dietary fibre was positively related to the length of
subjects
(MJ/day)
subjects (MJ/day)

education (p < 0.001). Mean intake of dietary choles-
Low intensity1
No activity
856
9.2
649
6.6
terol and fibre per 10 MJ were in men 0.30 gram and
< 1 hour/week
1423
8.9
1324
6.5
27.0 gram, respectively. The corresponding figures
1-2 hours/week
2847
9.1
3307
6.7
from women were 0.31 gram and 27.7 gram. The total
? 3 hours/week
3150
9.3
3649
6.8
p for linear trend
< 0.001
< 0.001
dietary cholesterol intake (not as gram per 10 MJ) was
statistically significantly (p < 0.001) related to educa-
High intensity2
No activity
2817
9.0
4265
6.7
tional level in men, but not in women. The relationship
< 1 hour/week
2099
9.1
2073
6.7
was, however, not strong as the mean intake differed
1-2 hours/week
2042
9.2
1912
6.8
< 15 mg/day between the highest and lowest educatio-
? 3 hours/week
1333
9.5
641
6.8
nal levels (results not shown).
p-value for linear trend
< 0.001
0.002
The associations between length of education and
Type of work
two nutrients related to consumption of fruits and
Sedentary work
3655
8.9
3353
6.6
Requiring walking
1706
9.2
2416
6.8
vegetables (beta-carotene and vitamin C) are shown in
Requiring walking and lifting
1470
9.5
1818
6.9
table 4. Both for men and women, we found that the
Heavy manual labour
629
10.1
84
7.2
intakes of the two nutrients were positively correlated
p-value for linear trend
< 0.001
0.001

to educational level. Mean intake of beta-carotene and
1Low intensity is physical activity that do not make the respondent breathless
and sweating.
vitamin C per 10 MJ were in men 4.3 mg and 136 mg,
2High intensity is physical activity that makes the respondent breathless and
respectively. The corresponding figures in women
sweating.
Table 1. Mean (SD) total energy intake, percent energy from all fats, polyunsaturated fats (PUFA), and saturated fats (SFA),
protein, carbohydrates and alcohol, according to age and gender. Tromsø, 1994/95.

Age Number of
% energy from

group subjects Energy (MJ) All fats PUFA SFA protein carbohydrates alcohol

Men
25-34
2435
9.4 (2.1)
32.5 (6.1)
5.7 (2.1)
13.3 (2.9)
16.4 (2.3)
49.5 (6.2)
1.5 (1.7)
35-44
2494
9.3 (2.1)
30.8 (5.9)
5.3 (1.9)
12.6 (2.8)
16.5 (2.2)
51.1 (6.1)
1.6 (1.8)
45-54
2015
9.0 (2.0)
29.8 (5.7)
5.0 (1.7)
12.2 (2.7)
16.8 (2.2)
51.8 (6.1)
1.6 (2.0)
55-64
1037
9.1 (1.9)
28.7 (5.7)
4.8 (1.6)
11.8 (2.8)
16.9 (2.3)
53.1 (6.1)
1.1 (1.6)
65-69
335
8.7 (1.9)
28.2 (5.6)
4.6 (1.6)
11.7 (2.9)
17.1 (2.3)
53.6 (6.2)
0.9 (1.5)
All
8316
9.2 (2.0)
30.7 (6.0)
5.3 (1.9)
12.6 (2.9)
16.6 (2.3)
51.1 (6.2)
1.5 (1.8)
Women
25-34
2977
6.9 (1.6)
33.3 (5.9)
5.4 (2.1)
14.2 (2.9)
16.4 (2.3)
49.4 (5.6)
1.0 (1.3)
35-44
2710
6.8 (1.6)
32.6 (5.8)
5.3 (2.0)
13.8 (2.8)
16.6 (2.3)
49.8 (5.7)
1.1 (1.4)
45-54
1994
6.6 (1.6)
31.1 (5.9)
4.9 (1.7)
13.3 (3.0)
16.9 (2.3)
51.0 (6.0)
1.0 (1.5)
55-64
931
6.7 (1.6)
30.4 (5.6)
4.7 (1.6)
13.0 (2.9)
17.2 (2.4)
51.7 (5.8)
0.8 (1.5)
65-69
337
6.5 (1.6)
31.3 (5.7)
4.7 (1.4)
13.5 (3.0)
17.4 (2.4)
51.0 (6.0)
0.4 (0.9)
All
8949
6.7 (1.6)
32.2 (5.9)
5.1 (1.9)
13.7 (2.9)
16.7 (2.4)
50.2 (5.8)
1.0 (1.4)

---

B.K. JACOBSEN AND H. NILSEN
60
Table 3. Age-adjusted relationships between level of education and the percentage of energy from all fats, polyunsaturated
fats (PUFA), saturated fats (SFA) and alcohol as well as the intake of dietary cholesterol and fibre per 10 MJ. Tromsø,
1994/95.

Level of
Number of
% energy from
gram cholesterol gram fibre

education
subjects
All fats
PUFA SFA
alcohol
per 10 MJ
per 10 MJ

Men
7-10 years
2035
30.9
5.4
12.6
1.1
0.30
27.0
11 years
2502
30.5
5.3
12.4
1.3
0.30
27.4
12 years
661
29.8
5.1
12.1
1.6
0.30
27.8
13-15 years
1573
29.7
5.0
12.2
1.5
0.30
28.1
? 16 years
1533
29.1
4.5
12.3
1.7
0.30
28.5
p-value for linear trend
< 0.001
< 0.001
< 0.001
< 0.001
0.7
< 0.001
Women
7-10 years
2350
32.3
5.3
13.6
0.6
0.32
27.5
11 years
2495
32.2
5.2
13.6
0.9
0.32
27.8
12 years
962
31.8
5.0
13.6
1.0
0.32
28.1
13-15 years
1543
31.3
4.8
13.4
1.1
0.31
28.9
? 16 years
1581
30.7
4.4
13.4
1.3
0.31
29.7
p-value for linear trend
< 0.001
< 0.001
0.002
< 0.001
0.001
< 0.001

Table 4. Age-adjusted relationships between level of education and the daily intake of beta-carotene and vitamin C (mg per
10 MJ). Tromsø, 1994/95.

Men
Women


Level of
Number of
mg beta-carotene mg vitamin
Number of
mg beta-carotene mg vitamin
education
subjects
per 10 MJ
C per 10 MJ
subjects
per 10 MJ
C per 10 MJ

7-10 years
2035
3.8
133
2350
7.1
154
11 years
2502
4.3
136
2495
7.6
158
12 years
661
4.6
140
962
7.7
156
13-15 years
1573
4.8
144
1543
7.9
164
? 16 years
1533
5.1
147
1581
8.2
165
p-value for linear trend
< 0.001
< 0.001
< 0.001
< 0.001

DISCUSSION
applying local, gender specific serving sizes in the
computations, and not national averages for men and
The nutrients included in the analysis (energy, fats,
women combined, improves the method. The propor-
dietary cholesterol and fibre, beta-carotene and vita-
tion of energy from fats, protein, carbohydrates and
min C) were selected for two reasons. Firstly, these are
alcohol and mean intakes per 10 MJ of dietary choles-
important nutrients with regard to risk of major chro-
terol, fibre and vitamin C are quite similar to data from
nic diseases like cardiovascular diseases and cancer.
the Norkost surveys, which are intended to be repre-
Information about social gradients in the intake of
sentative for the Norwegian population aged 16-79
these nutrients is therefore important. Secondly, it is
(5,22). We found, however, in women 6 % higher
known from previous studies that the intakes of at least
percentage of the energy from fats and in men 18 %
some of the nutrients are related to educational level
higher intake of vitamin C per 10 MJ than in the
(2-9). Thus, demonstrating these relationships also in
Norkost surveys (30.4 % and 115 mg, respectively, in
this study population may tend to validate nutrient
the Norkost survey in 1997) (5).
intake data (predictive validity).
There are, however, also several limitations in our
data. Some food groups (e.g., fast food meals like hot
Nutrient intake
dog and hamburger) were not included in our ques-
The dietary survey method used (self-administered
tionnaire and the information about added fat (e.g., to
questionnaire) has the advantage that information from
fish dishes) is limited. This may be a problem in parts
a large population can be collected. We believe that
of the population with dietary habits that differ from

---

EDUCATION AND NUTRIENT INTAKE
61
the majority, for example subjects who get a relatively
fore expected to find positive relationships between
larger proportion of the energy from fast food. Fur-
level of education and the intake of beta-carotene and
thermore, the validity of the nutrient information will
vitamin C. This was confirmed. These associations, as
by necessity differ from one nutrient to the next, de-
well as the relationship found for dietary fibre, are also
pending on how well the food items we have included
in accordance with national (4,5) and international (6-
in the questionnaire contribute to the variation in the
8) data.
nutrient intake.
Thus, subjects with relatively low level of educa-
Our main purpose is, however, to rank the persons
tion have dietary habits that may put them at increased
according to their nutrient intake. The significant rela-
risk of chronic diseases like cardiovascular diseases
tionships between energy and physical activity and
and cancer. Although this is an expected finding, it is
between educational level and the intake of nutrients
still a cause for concern. The national health promo-
indicate that we are able to do so. Thus, although the
tion policy aims at decreasing health differences with-
calculation of nutrient intake is based on limited die-
in the population. The overall quality of the Norwe-
tary information, several arguments support the vali-
gian diet has in some respect (e.g., the percentage of
dity of this method of assessing the relative nutrient
energy from fats) improved the last 30 years, but our
intake of individuals in a population.
findings along with other studies show that educational
gradients persist. Johansson and coworkers (4) have
Relationships with level of education
previously shown that although educational level is a
predictor for healthy food habits, the degree of
We found significant relationships between the level
attention paid to having a healthy diet was a stronger
of education and the intake of some nutrients. Our
predictor. Adjusted for this variable, the impact of
results are, generally speaking, in accordance with
educational level on the intake of fruits and vegetables,
other findings from Tromsø (2,9), national Norwegian
fibre and percent energy from fats was reduced. The
data (3-5), and some international data (e.g., 6,7). Of
variable of interest may therefore not be years of
particular interest and importance is the inverse rela-
education, but the interest in eating to keep healthy.
tionship between educational level and the proportion
This calls for amendments of the information strate-
of energy from fats. The significant inverse relation-
gies applied in the Norwegian nutrition policy in
ship between educational level and the percentage of
general, and in the health education directed toward
energy from polyunsaturated fat may be somewhat
persons with low educational level particularly.
surprising. However, some national data may indicate
In conclusion, we find in this study of more than
a similar trend (5).
17,000 Norwegian men and women aged 25-69 that
We found in men no relationship between dietary
persons with higher education have more healthy die-
cholesterol intake per 10 MJ and level of education.
tary habits than persons with less education. This con-
This was also found in the Norkost survey (5). The
firms previous studies. We believe that our findings
statistically significant inverse relationship (p < 0.001)
underline the need for special efforts in order to
between level of education and the actual dietary
change the food habits of the less educated.
cholesterol intake (not per 10 MJ) in men, but not in
women, is in accordance with national data (5).
Norwegian men and women with long education
ACKNOWLEDGEMENT
have higher consumption of fruits and vegetables than
The study was carried out in cooperation with the National
men and women with less education (2-5). We there-
Health Screening Service, Oslo, Norway.
REFERENCES
1. Liberatos P, Link BG, Kelsey JL. The measurement of social-class in epidemiology. Epidemiol Rev 1988; 10:
87-121.
2. Jacobsen BK, Thelle DS. The Tromsø Heart Study: Risk factors for coronary heart disease and level of
education. Am J Epidemiol 1988; 127: 923-32.
3. Hjartåker A, Lund E. Relationship between dietary habits, age, lifestyle, and socio-economic status among
adult Norwegian women. The Norwegian Women and Cancer Study. Eur J Clin Nutr 1998; 52: 565-72.
4. Johansson L, Thelle DS, Solvoll K, Bjørneboe G-EAa, Drevon CA. Healthy dietary habits in relation to social
determinants and lifestyle factors. Br J Nutr 1999; 81: 211-20.
5. Statens råd for ernæring og fysisk aktivitet. Norkost 1997. Landsomfattende kostholdsundersøkelse blant
menn og kvinner i alderen 16-79 år. Rapport nr. 2/1999. Oslo: Statens råd for ernæring og fysisk aktivitet,
1999 (in Norwegian).

---

62
B.K. JACOBSEN AND H. NILSEN
6. Bolton-Smith C, Smith WCS, Woodward M, Tunstall-Pedoe H. Nutrient intakes of different social-class
groups: results from the Scottish Heart Health Study (SHHS). Br J Nutr 1991; 65: 321-35.
7. Shimakawa T, Sorlie P, Carpenter MA, Dennis B, Tell GS, Watson R, Williams OD. Dietary intake patterns
and sociodemographic factors in the Atherosclerosis Risk in Communities Study. Prev Med 1994; 23: 769-80.
8. Roos E, Prättälä R, Lahelma E, Kleemola P, Pietinen P. Modern and healthy?: Socioeconomic differences in
the quality of diet. Eur J Clin Nutr 1996; 50: 753-60.
9. Jacobsen BK. Frequency of alcohol use and the level of education. J Intern Med 1989; 225: 417-22.
10. Willett W. Nutritional epidemiology. Oxford: Oxford University Press, 1990.
11. 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.
12. Overvad K, Tjønneland A, Haraldsdòttir J, Ewertz M, Jensen OM. Development of a semiqualitative food fre-
quency questionnaire to assess food, energy and nutrient intake in Denmark. Int J Epidemiol 1991; 20; 900-5.
13. Haile RWC, Hunt IF, Buckley J, et al. Identifying a limited number of foods important in supplying selected
dietary nutrients. J Am Diet Assoc 1986; 86: 611-6.
14. Stryker WS, Salvini S, Stampfer MJ, Samson L, Colditz GA, Willett WC. Contributions of foods to absolute
intake and between-person variation of nutrient consumption. J Am Diet Assoc 1991; 91: 172-8.
15. Mosdøl A, Holmboe-Ottesen G, Løken EB, Solvoll K, Johansson L, Thelle DS. Contributions of food
categories to absolute nutrient intake and between-person variation within a representative sample of 2677
Norwegian men and women. Nor J Epidemiol 2000; 10 (This issue).
16. Gaard M, Sandvin O, Løken EB. Calculation of nutrient intake based on dietary data from 2nd cardiovascular
disease survey in Finnmark, Sogn og Fjordane, and Oppland. 1977-83. Technical report 1. Oslo: The Cancer
Registry, 1996.
17. Nilsen H, Utsi E, Bønaa KH. Dietary and nutrient intake of a Sami population living in traditional reindeer
herding areas in north Norway: comparisons with a group of Norwegians. Int J Circumpolar Health 1999; 58:
120-33.
18. Grimsgaard S, Bønaa KH, Hansen JB, Nordøy A. Highly purified eicosapentaenoic acid and docosahexaenoic
acid in humans have similar triacylglycerol-lowering effects but divergent effects on serum fatty acids. Am J
Clin Nutr 1997; 66: 649-59.
19. Rimestad AH, Blaker B, Flåten A-M, Nordbotten A. Den store matvaretabellen, 1995. Oslo: Universitets-
forlaget, 1995 (in Norwegian).
20. Statens Livsmedelverk. Energi och näringsämnen. Uppsala: Statens Livsmedelverk, 1993 (in Swedish).
21. SAS Institute. SAS/STAT Guide. 6th Version. Cary: SAS Institute, 1987.
22. Johansson L, Solvoll K, Bjørneboe G-EAa, Drevon CA. Dietary habits among Norwegian men and women.
Scand J Nutr 1997; 41: 63-70.

---