Relative validity of a short qualitative FFQ 737

.........................................................................................................
European Journal of Public Health, Vol. 23, No. 5, 737–742
ß The Author 2012. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.
doi:10.1093/eurpub/cks096 Advance Access published on 26 July 2012
.........................................................................................................
Relative validity of a short qualitative food frequency
questionnaire for use in food consumption surveys
Willem De Keyzer1,2, Arnold Dekkers3, Veerle Van Vlaslaer4, Charlene Ottevaere1,
Herman Van Oyen5, Stefaan De Henauw1,2, Inge Huybrechts2,6

1 Department of Nutrition and Dietetics, Faculty of Health Care Vesalius, University College Ghent, Ghent, Belgium
2 Department of Public Health, Ghent University, University Hospital 2 Blok A, Ghent, Belgium
3 National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands

Downloaded from https://academic.oup.com/eurpub/article/23/5/737/444280 by CNR user on 30 November 2023

4 Department of Applied Sciences and Education, Plantijn Hogeschool—University College, Antwerp, Belgium
5 Unit of Epidemiology, Scientific Institute of Public Health, Brussels, Belgium
6 International Agency for Research on Cancer (IARC), Dietary Exposure Assessment Group, 150 Cours Albert Thomas,
69372 Lyon Cedex 08, France

Correspondence: Willem De Keyzer, Department of Nutrition and Dietetics, Faculty of Health Care Vesalius,
University College Ghent, Keramiekstraat 80, B-9000 Ghent, Belgium, tel: +32 9 321 21 38, fax: +32 9 220 17 26,
e-mail: willem.dekeyzer@hogent.be

Background: The aim of the present study was to assess the relative validity of a self-administered qualitative food
frequency questionnaire (FFQ) applied in the Belgian food consumption survey. Methods: Comparison of food
consumption data from an FFQ with 7-day estimated diet records (EDR) was made in a sample of 100 participants
(aged 15–90 years). The FFQ comprised a total of 50 foods. Both FFQ and EDR foods were categorized into 15
conventional food groups. Results: De-attenuated Spearman rank correlation coefficients between the FFQ and the
EDR ranged from 0.16 for potatoes and grains to 0.83 for alcoholic beverages, with a median of 0.40 for all 15
food groups. The proportion of participants classified in the same tertile of intake by the FFQ and EDR ranged from
32% for potatoes and grains to 76% for alcoholic beverages. Extreme classification into opposite tertiles was <10%
for milk and soy products, alcoholic beverages, fried restgroup foods and fats. Conclusion: Notwithstanding the
short nature and the absence of portion size questions, the FFQ appears to be reasonably valid in both genders and
across different age categories for most food groups. However, for the food groups bread and cereals, potatoes
and grains, and sauces, estimates should be interpreted with caution because of poor ranking agreement.
.........................................................................................................

Introduction (EDR). For 15 food groups, the performance of the FFQ to rank
individuals according to intake and agreement with a 7-day EDR will
n 2004, the first food consumption survey was performed in be evaluated. In addition, more extensive analyses will be performed
IBelgium. 1
During this survey, a representative sample of the to investigate associated measurement error structures.
Belgian population aged 15 years and over was recruited from the
national register. The individual level of food and nutrient intake
was assessed by two non-consecutive 24-h recalls using EPIC-Soft. A Methods
quantitative food frequency questionnaire (FFQ) was used to study
Study design
the adequacy of food intake in different subgroups of the
population. Furthermore, subgroups at risk for a deficient or Using a cross-sectional study design, food intake assessed with a
excessive intake of specific foods or nutrients were identified. An 7-day EDR was compared with food intake assessed with the short
extensive overview of the methods used in this first Belgian food FFQ. During a first visit, participants were provided with a general
consumption survey is given elsewhere.2 questionnaire comprising socio-demographic and anthropometric
Short FFQs satisfy many conditions to be used as dietary questions and a paper-based FFQ. After 2–6 weeks, a second visit
assessment instrument in the context of epidemiological studies was planned, during which both questionnaires were returned to the
because of their inexpensiveness and low burden for participants.3 researchers. Furthermore, a 7-day EDR was provided, and instruc-
Also, in the context of nutritional surveillance, they have potential to tions were given for completion. During a final visit, the EDR was
serve as a quick measure for long-term usual food intake and iden- collected and checked for completeness by a dietitian. Any
tification of non-consumers both in adults and children.4–7 For both remaining quality issues were discussed with the participant and
purposes (epidemiological and surveillance), however, it is clarified or corrected.
Data collection was performed in Flanders from October 2005 to
paramount that validity of the instrument is assessed and taken
April 2006. This study was conducted according to the guidelines
into account during interpretation of results in future use.
laid down in the Declaration of Helsinki, and all procedures
Data from the Belgian food consumption survey indicated that
involving human subjects were approved by the regional Ethics
the response rate of a short FFQ was higher compared with the 24-h
Committee of Ghent University Hospital. A written informed
recall interviews, which is very likely to be due to the lower
consent was obtained from all participants.
respondent burden. Because of these advantages, the FFQ is being
used as a quick screening tool to assess different aspects in the diet of
our Belgian population. Therefore, it is important to evaluate the Participants
validity of this short FFQ. Hence, the aim of the present study is to To resemble best the target population of the food consumption
assess its validity compared with 7-day estimated diet records survey (i.e. nationally representative), different age categories were
738 European Journal of Public Health

included and equality in gender was pursued. In total, three age values that are indicated by the FFQ tertiles. Kruskal–Wallis one-way
categories were recruited: adolescents and young adults (15–29 analysis of variance was used to determine whether differences of
years), adults (30–59 years) and elderly (60+ years). For those means between tertiles were statistically significant. Third, intakes
categories, a different approach for recruiting participants was assessed with the EDR were classified into tertiles, and agreement
performed. However, a convenient sample was drawn from the between both methods was assessed using the weighted  statistic,
population. (i) In adolescents, a multi-stage sampling was calculated with a linear set of weights and CI of 95%.13 This analysis
performed. Firstly, five secondary schools providing both general was not performed for food groups for which more than 33.3% of the
education and vocational training were contacted in the region of participants had a zero consumption either for the FFQ or EDR.
Ghent. Four schools agreed to participate in the study. Subsequently, Fourth, the percentage classified into the correct or adjacent tertile
parent’s permission was asked by written request. Because selection and the percentage grossly misclassified (lowest tertile for one
of classes and communication with parents were performed by the method and highest tertile for the other) was calculated. Finally,
school’s administration, the number of invited participants is agreement between the FFQ and the EDR at the individual level
unknown. (ii) Young adults and adults invited for participation was assessed using mean difference and standard deviation of the
were acquaintances and family of students and researchers. (iii) mean difference, visually represented by a Bland and Altman
Elderly were recruited via social service centres. Elderly living in a plot.14 To correct for non-normal distributions, a Box-Cox trans-

Downloaded from https://academic.oup.com/eurpub/article/23/5/737/444280 by CNR user on 30 November 2023

residential care setting were excluded given the more limited formation was performed on food group intake data from both in-
freedom in food choices. A total of 233 adults (>18 years) were struments (EDR and FFQ) before plotting. The grand mean between
invited accordingly. The participants did not receive any incentive both methods is plotted (dotted lines) including the 95% CI around
for their participation. the mean (error bars at the right). The fitted regression line is also
plotted with the 95% CI (dash-dot lines) and 95% CI for new obser-
The food frequency questionnaire vations (dashed line). To quantify the error measurement structure,
two additional tests were performed on the Bland and Altman plots.
The FFQ under study was a self-administered qualitative question-
The first one tests whether the zero difference line (solid line) is
naire comprising 50 food items. These food items were either
outside of the grand mean 95% CI. The second one tests whether
individual foods (e.g. soft drink) or an aggregation of similar
the fitted regression line’s slope is significantly different from zero.
foods (soy products containing soy milk, soy drinks and soy
Statistical tests were performed using IBM PASW Statistics program
desserts). The FFQ is a data- and experience-based version,
version 18.0.0 (SPSS Inc., an IBM company, IL, USA), two-tailed and
adopted from the Health interview survey in Belgium,8 to be used
alongside a repeated 24-h recall using EPIC-Soft in the Belgian P < 0.05 was considered as statistically significant. The construction
national food consumption survey.9 The frequency categories used of the Bland and Altman plots including the tests for difference and
in the FFQ were as follows: never, less than 1 day per month, 1–3 slope were created using the S-PLUS statistics program.
days per month, 1 day per week, 2–4 days per week, 5–6 days per
week, 1 time a day, 2–3 times a day and more than 3 times a day. The Results
usual food intakes derived from the FFQ were calculated by multi-
plying the frequency of consumption with a standard portion size In total, 156 participants agreed to participate in the study repre-
for each food item.10 The same standard units were used to calculate senting a response rate of 55% among adults and elderly. The
portion sizes of estimated foods during data entry of the EDR. response rate for the adolescents could not be calculated because
cluster sampling was used and the total number of adolescents
Estimated dietary record eligible for inclusion from all schools was unknown. Almost all par-
ticipants completed the FFQ (n = 155); however, only 100 (64%)
Structured open-ended diaries containing pre-defined food groups
were able to complete all 7 days of the food record. Characteristics
(including the option ‘other food items’) at six food occasions
of participants are shown in table 1.
(breakfast, lunch, dinner and three snacks) were provided to all
Twenty-six percent of the participants were between 15 and 29
participants. All participants were informed on how to complete
years old and 22% were 30–59 years old. These percentages agree
the food record. The diary also contained a written example for
with those from the food consumption survey in 2004, where 33%
future reference. During a 7-day period, all consumed foods and
from the sample was between 15 and 29 years, 19% between 30–59
drinks had to be reported with notification of date and place of
years and 48% >60 years old.9 Forty-four percent of men and 36% of
consumption, estimated consumed quantity expressed as a
women were overweight or obese (BMI  25.0 kg/m2). These body
household measure, unit or weight, specification and if present a
characteristics also agree with those from the Belgian food consump-
brand name. Separate forms were included to report home-made
tion survey where mean BMI was 25.0 kg/m2 for men and 24.2 kg/m2
recipes, so name of dish, total quantities of all ingredients and
for women and percentage of men and women being overweight or
fraction of dish consumed could be stated. Only participants with
obese was 47% and 35%, respectively.9
zero missing dietary records were included in the analysis.
In a previous version, crude correlations were also presented.
De-attenuated correlations of consumed foods between both
Data and statistical analysis methods are presented in table 2. For the food groups fish,
For almost all food groups, consumption data were not normally alcoholic beverages and fried restgroup foods, correlation coeffi-
distributed. Therefore, only non-parametric tests were used during cients are calculated on consumers only given the high number of
analysis. First, Spearman rank order correlations of food group zero consumptions for these foods during the 7-day recording of
intakes between the 7-day EDR and FFQ were calculated for all par- dietary intake. For the total sample, a strong de-attenuated correl-
ticipants and stratified by gender and age category. In addition, ation coefficient (0.70) was found for alcoholic beverages. The
because day-to-day variation in intake of most specific foods is de-attenuated correlation coefficients were moderate (0.40–0.69)
generally high, de-attenuated correlation coefficients were for beverages, fruit, milk and soy products, cheese, restgroup
calculated to correct for within-person variation in the EDR.11 drinks, fried restgroup foods and fats. Fair (0.20–0.39) de-attenuated
Secondly, to assess measurement error of the FFQ, ‘actual values correlation coefficients were found for vegetables, fish, meat and
for surrogate categories’ were calculated12 as follows: participants eggs, restgroup foods and sauces. Finally, weak de-attenuated cor-
were assigned to tertiles according to food consumption estimated relations were found for bread and cereals, and potatoes and grains.
by the FFQ, then the mean food intake in each tertile was calculated Large differences in de-attenuated correlation coefficients were
using data from the EDR. This gives an indication of the ‘true’ (EDR) found between men and women. For both, strongest correlations
 Relative validity of a short qualitative FFQ 739

Table 1 Distribution of age and body mass index categories of participants

All 15–59 years 60 years

Malea Female Malea Female Male Female

(n = 44) (n = 56) (n = 25) (n = 23) (n = 19) (n = 33)

Mean age (SD) 50 (23.0) 56 (23.4) 32 (12.9) 30 (12.7) 73 (7.6) 74 (6.1)

Mean BMI (SD) 25.6 (4.0) 24.7 (4.1) 24.7 (4.1) 22.3 (3.0) 26.8 (3.6) 26.4 (4.0)
BMI category (n, %)
Under-/ normal weight (BMI < 25.0 kg/m2) 24 (56) 36 (64) 16 (67) 21 (92) 8 (42) 15 (46)
Overweight (BMI: 25.0–29.9 kg/m2) 14 (33) 13 (23) 6 (25) 1 (4) 8 (42) 12 (36)
Obesity (BMI  30.0 kg/m2) 5 (11) 7 (13) 2 (8) 1 (4) 3 (16) 6 (18)

a: For one participant body length was missing so BMI could not be calculated.

Downloaded from https://academic.oup.com/eurpub/article/23/5/737/444280 by CNR user on 30 November 2023

Table 2 De-attenuated Spearman rank correlation coefficients of The degree of misclassification associated with categorized intakes
food group intakes between the FFQ and 7-day EDR assessed by the FFQ compared with the 7-day EDR was examined as
the proportion of participants classified in the same, adjacent or
All Gender Age category (years) opposite tertile (table 3). For restgroup drinks ranking into tertiles
was not possible because more than 33.3% of participants did not
Food group (n = 100) Male Female 15–59 60+(n = 52)
(n = 44) (n = 56) (n = 48)
consume any food from this food group during the 7-day EDR. The
proportion of participants classified in the same tertile was 50% or
Beveragesa 0.40 0.57 0.24 0.30 0.49 higher for beverages, milk and soy products, cheese, alcoholic
Bread and cereals 0.16 0.07 0.29 0.36 0.08 beverages, fried restgroup foods and fats. Extreme misclassification
b
Potatoes and grains 0.16 0.41 0.25 0.06 0.08 into the opposite tertile did not exceed 10% for milk and soy
Vegetablesc 0.36 0.59 0.22 0.23 0.42
Fruit 0.52 0.60 0.48 0.56 0.41
products, alcoholic beverages, fried restgroup foods and fats.
Milk and soy productsd 0.65 0.61 0.71 0.58 0.77 Results from the weighted  statistic showed good agreement
Cheese 0.51 0.56 0.49 0.66 0.33 (0.61–0.80) for alcoholic beverages; moderate agreement (0.41–
Fishe 0.22 0.34 0.38 0.09 0.46 0.60) for milk and soy products, and fats; fair agreement (0.21–
Meat and eggsf 0.35 0.29 0.36 0.33 0.30 0.40) for beverages, vegetables, fruit, cheese, fish, meat and eggs,
Alcoholic beveragesg 0.83 0.81 0.71 0.94 0.82
h and fried restgroup foods. For bread and cereals, potatoes and
Restgroup foods 0.37 0.32 0.48 0.30 0.60
Restgroup drinksi 0.59 0.65 0.43 0.67 0.18
grains, restgroup foods and sauces,  coefficients were poor (0.20).
Fried restgroup foodsj 0.65 0.80 0.58 0.60 0.71 Observation of the Bland and Altman plots and associated tests
Saucesk 0.29 0.30 0.26 0.26 0.33 suggested intake-related bias for beverages, vegetables, milk and soy
l
Fats 0.68 0.60 0.65 0.65 0.56 products, meat and eggs, and restgroup foods (table 4 and
Supplementary Annex 1). Constant additive error was present for
For fish, alcoholic beverages and fried restgroup foods, correlations
are based on consumers only.
bread and cereals, potatoes and grains, cheese, alcoholic beverages,
a: All drinks (including fruit and vegetable juices and non-sugared and restgroup drinks. For bread and cereals, overestimation by the
soft drinks, excluding milk, soy drinks and drinks from restgroup). FFQ was observed, whereas for all other foods listed underestimation
b: Potatoes (excluding fried potatoes and fries), rice and pasta. errors were present.
c: Raw and cooked vegetables including legumes.
d: Milk, buttermilk, chocolate milk, milk added to coffee or tea,
yoghurt, soy drinks and desserts. Discussion
e: Fish, shellfish and fish products. The purpose of the current FFQ is to be used in the context of
f: Meat, meat products, poultry, game, offal, eggs and vegetarian
nutritional surveillance (i.e. food consumption surveys), comple-
products (tofu, Quorn, tempeh).
mentary to a repeated 24-h recall interview, providing policy
g: Wine, beer and spirits.
h: Sweets and candy bars, chocolate, biscuits and pastry.
makers with a quick instrument to screen consumption behaviour
i: Sugared soft drinks, sports drinks and energy drinks. of the population and to detect zero consumers necessary for proper
j: Fries, baked potatoes and crisps. estimation of food intake distributions from more detailed dietary
k: Cold sauces like mayonnaise and ketchup. interviews. The present qualitative FFQ showed a fair-to-moderate
l: Butter, margarine, low-fat margarine and lard. agreement in ranking participants towards their food intake
compared with a 7-day EDR. Considerable differences were found
across gender and age categories with respect to the FFQ’s ability to
correctly rank participants according to their usual intake.
were present for alcoholic beverages and milk and soy products in Acceptable ranking of participants by the FFQ was demonstrated
women. The weakest correlation was present for potatoes and grains for beverages, fruit, milk and soy products, cheese, alcoholic
in men and vegetables in women. For all age categories, alcoholic beverages, fried restgroup foods and fats. According to results
beverages yielded strongest de-attenuated correlation coefficients. In reported in other studies,15, 16 correlation coefficients between
the youngest age category, the weakest correlation coefficient was FFQ and EDR-based estimates tended to be lower for vegetables
present for beverages, whereas for the middle and oldest age compared with fruit. According to Wakai,16 a possible explanation
category, there was a negative correlation for potatoes and grains, could be that frequency of fruit consumption is easier to report then
and bread and cereals in the oldest age category only. vegetables because fruits are more often consumed as raw foods
Table 3 includes mean food group intakes based on the 7-day whereas vegetables are more frequently part of cooked dishes and
EDR for the FFQ tertiles. For all food groups, significant differences, therefore not integrally recalled. In addition, fruit is frequently
indicating good ranking of participants, were found between ranks consumed as a single food item and comes in natural or typical
of intake, except for bread and cereals, potatoes and grains, units, whereas vegetables are often sliced or cut which makes them
restgroup foods and sauces. more difficult to quantify. On the other hand, in a Mediterranean
740 European Journal of Public Health

Table 3 Mean food group intakes from EDR for categories based on FFQ tertiles with agreement of tertiles for both methods

Agreement of tertiles

Mean intake EDR (g) Same Adjacent Opposite Weighted kappa

tertile tertile (%) tertile (%)
Food groupa T1 T2 T3 pb (%)  (95% CI)
c
Beverages 814.0 1017.5 1211.6 0.002 50.0 38.3 11.7 0.30 (0.15–0.44)
Bread and cereals 120.8 163.9 151.9 0.064 39.2 39.2 21.6 0.13 (0.01–0.27)
Potatoes and grainsd 105.3 105.4 105.3 0.894 32.0 42.3 25.8 0.06 (0.20–0.08)
Vegetablese 90.2 107.2 132.1 0.024 46.9 38.8 14.3 0.23 (0.09–0.37)
Fruit 48.8 115.0 149.3 <0.001 45.5 42.4 12.1 0.26 (0.12–0.40)
Milk and soyf 71.6 117.5 227.1 <0.001 57.4 36.2 6.4 0.45 (0.31–0.60)
Cheese 22.1 25.4 37.2 0.001 50.0 37.5 12.5 0.30 (0.16–0.44)
Fishg 14.0 25.7 33.3 0.003 44.8 44.8 10.4 0.23 (0.09–0.38)
Meat and eggsh 108.4 122.8 163.5 0.002 44.7 42.6 12.8 0.23 (0.09–0.37)

Downloaded from https://academic.oup.com/eurpub/article/23/5/737/444280 by CNR user on 30 November 2023

Alcoholic beveragesi 18.7 120.3 437.5 <0.001 76.0 21.9 2.1 0.71 (0.56–0.85)
Restgroup foodsj 48.9 56.6 74.2 0.076 43.8 41.7 14.6 0.20 (0.05–0.34)
Fried restgroup foodsk 26.9 35.7 72.9 <0.001 52.0 39.8 8.2 0.37 (0.23–0.51)
Saucesl 6.3 6.2 11.0 0.137 38.1 42.3 19.6 0.10 (0.04–0.24)
Fatsm 4.1 15.8 22.4 <0.001 57.7 39.2 3.1 0.50 (0.36–0.64)

a: For the food group restgroup drinks, no tertiles could be calculated because >33.3% of the participants did not consume any food from
this food group during the 7-day EDR period.
b: Kruskal–Wallis one-way ANOVA.
c: All drinks (including fruit and vegetable juices and non-sugared soft drinks, excluding milk, soy drinks and drinks from restgroup).
d: Potatoes (excluding fried potato products), rice and pasta.
e: Raw and cooked vegetables including legumes.
f: Milk, buttermilk, chocolate milk, milk added to coffee or tea, yoghurt and soy drinks and desserts.
g: Fish, shellfish and fish products.
h: Meat, meat products, poultry, game, offal, eggs and vegetarian products (tofu, Quorn, tempeh).
i: Wine, beer and spirits.
j: Sweets and candy bars, chocolate, biscuits and pastry.
k: Fries, baked potatoes and crisps.
l: Cold sauces like mayonnaise and ketchup.
m: Butter, margarine, low-fat margarine and lard.

Table 4 Test statistics of mean differences and slopes of Bland and Altman data after Box-Cox transformation

Mean difference Regression

Food group FFQ-EDR Pdifference Intercept Slope Pslope n

a
Beverages 24.3 0.046 185.4 0.695 <0.001 94
Bread and cereals 7.7 <0.001 2.5 0.296 0.08 97
Potatoes and grainsb 4.4 <0.001 10.7 0.281 0.14 95
Vegetablesc 1.5 0.01 9.6 0.636 <0.001 97
Fruit 0.56 0.21 1.3 0.179 0.19 87
Milk and soy productsd 0.3 0.16 3.04 0.371 <0.001 87
Cheese 1.43 <0.001 1.36 0.018 0.89 81
Fishe 0.02 0.88 1.58 0.366 0.06 77
Meat and eggsf 5.1 <0.001 4.95 0.314 0.04 94
Alcoholic beveragesg 1.2 <0.001 1.20 0.004 0.98 68
Restgroup foodsh 0.8 0.03 3.04 0.276 0.02 96
Restgroup drinksi 0.47 0.046 1.38 0.167 0.25 37
Fried restgroup foodsj 0.043 0.75 1.06 0.196 0.15 83
Fatsk 0.095 0.70 0.43 0.101 0.30 97

Note: values for mean difference, intercept and slopes are in transformed scale.
n: number of participants with positive consumptions of food groups during both collections (FFQ and EDR).
a: All drinks (including fruit and vegetable juices and non-sugared soft drinks, excluding milk, soy drinks and
drinks from restgroup).
b: Potatoes (excluding fried potatoes and fries), rice and pasta.
c: Raw and cooked vegetables including legumes.
d: Milk, buttermilk, chocolate milk, milk added to coffee or tea, yoghurt, soy drinks and desserts.
e: Fish, shellfish and fish products.
f: Meat, meat products, poultry, game, offal, eggs and vegetarian products (tofu, Quorn, tempeh).
g: Wine, beer and spirits.
h: Sweets and candy bars, chocolate, biscuits and pastry.
i: Sugared soft drinks, sports drinks and energy drinks.
j: Fries, baked potatoes and crisps.
k: Butter, margarine, low-fat margarine and lard.
 Relative validity of a short qualitative FFQ 741

population, a higher correlation was found for vegetables compared amounts of intake is reduced. On the other hand, it was
with fruit by Fernandez-Ballart et al.17 demonstrated by Noethlings et al.19 that portion size adds only
Food groups for which relative validity turned out rather low in the limited information on variance of food intake in a large
current study were typical carbohydrate-containing food groups like European sample, suggesting that assignment of standard portions
bread and cereals, and potatoes and grains. A possible explanation for to frequencies of intake seems to be adequate. This finding was also
bread and cereals might be that bread is likely to be consumed more documented earlier in an American sample where it was concluded
than once a day with large differences in portion sizes between par- that due to a smaller contribution of between-person variance to the
ticipants, which is not reflected by the FFQ, and, especially in the total variance in portion size, specification of a standard portion size
older age category, breakfast cereals are consumed less frequently. may not introduce a large error in the estimation of food intake.20
Also, for the food group potatoes and grains, it was found that in
men, the FFQ largely overestimated potato consumption. For food
groups with low validity, more detailed questionnaires containing Conclusion
more food items may be needed to accurately assess actual food In general, the FFQ tends to underestimate food intake compared
consumption. On the other hand, the trade-off between adding with EDR. For fruit, fish, fried restgroup foods and fats, no systematic
items for improvement of validity and longer questionnaires, which

Downloaded from https://academic.oup.com/eurpub/article/23/5/737/444280 by CNR user on 30 November 2023

bias was present. Considering the short character and the absence
in turn can affect participation rate, should be kept in mind. of portion size questions, the FFQ appears to be reasonably valid
The correlation coefficients for cheese (0.33) and restgroup drinks in both genders and across different age categories for assessment
(0.18) in the oldest age category were lower than those in the other of group level intakes. However, for the food groups bread and
age categories, whereas for fish the correlation coefficient was higher cereals, potatoes and grains, and sauces, estimates should be inter-
(0.46). In elderly, data from both 7-day EDR and FFQ indicate very preted with caution because of poor ranking agreement.
low consumption of restgroup drinks (data not shown).
We are aware that the participants included in the different strata
are a selected group of the population and not a population-based Supplementary data
random sample due to the convenience sampling approach. No
upper age limitation was set for inclusion into the study; Supplementary data are available at EURPUB online.
therefore, some participants were older than 80 years (15% of the
oldest age category, data not shown). It was suggested by Acknowledgements
Rothenberg18 that elderly up to the age of 80 perform well in
reporting their food habits retrospectively and that from the age I.H. and H.V.O. were responsible for the study concept and design.
of 80, elderly tend to report food habits earlier in life. I.H. was responsible for the collection of data. W.D.K. was respon-
Although extensively used in other validation studies, the current sible for the statistical analysis and writing of the manuscript. A.D. is
methodology has some drawbacks. When designing a validation acknowledged for his statistical advice and programming of the
protocol, a key step lies in identifying a method that will serve as Bland and Altman plots and associated tests. All authors were
a reference for the instrument to be validated. Given that no method involved in the interpretation of data and reviewing of the
is perfect to serve as gold standard, it is of vital importance that manuscript. The authors acknowledge the students who conducted
errors of both methods are as independent as possible.12 Failing to the fieldwork of this study in the frame of their thesis: Karen
select a reference method of which errors are uncorrelated with those Bresseleers, Josefien Bisschop, Kim Den Blauwen-Van Driessche
from the instrument under study will inevitably lead to higher and Ann Devos. We also would like to thank the respondents who
estimates of validity. Major sources of error associated with FFQs voluntary participated. The University College Ghent is also
are caused by restrictions resulting from a fixed list of foods, acknowledged for the Doctoral Research Support Grant (W.D.K.).
memory and interpretation of questions. Among the available and This work was orally presented at the 11th European Nutrition
feasible comparison methods for validating an FFQ, diet records, Conference—FENS Madrid, Spain, October 2011.
with their open-ended format and independency towards memory,
are likely to have the least correlated errors.12 In addition, the Conflicts of interest: None declared.
reference method used in the present study (7-day EDR) is the
best obtainable instrument that takes into account within-person
variation in food intake, and therefore better reflects usual food Key points
intake. Also, replicate measures by EDR allow to correct
for within-person variability in food intake by calculating ratios of  Large-scale dietary intake assessments should preferably have
within- over between-person variances suitable for estimating low burden for participants. Short FFQs meet this condition
de-attenuated coefficients of correlation. and are therefore suitable instruments for nutrition research
A limitation of the current study is the different time frame over in the context of public health monitoring.
which food intake was assessed by both methods. Typically, FFQs are  To correctly interpret results resulting from the use of these
designed to measure long-term dietary intake, whereas EDRs instruments, it is necessary to validate them and investigate
measure short-term intake when not repeated over time. Ideally, their measurement error structure.
replicate 7-day periods, in which food intake is recorded over a  The present study shows that for some, but not all, food
time span of one year, is advised to include all seasonal variations. groups, the short FFQ is able to correctly rank participants
For instance, higher consumption of soups during winter months as according to their level of intake compared with a 7-day
opposed to summertime can have considerable impact on total EDR. For the food groups fruit, fish, fried foods and fats,
beverages intake. The same will count for consumption of raw no systematic group level bias was found.
vegetables during summer.
The current study compared a quantitative EDR with a qualitative References
FFQ. To do so, standard portions sizes were used to calculate food
intakes assessed by the FFQ. Therefore, it is very likely that two 1 De Vriese S, Huybrechts I, Moreau M, Van Oyen H. The Belgian Food
persons with identical frequencies of consumption have a different Consumption Survey 1—2004: Report (Enquête de consommation alimentaire
true consumption of a particular food due to differences in portion Belge 1—2004: Rapport) 2006.
sizes consumed. This loss of detail, inherent to qualitative FFQs, will 2 De Vriese S, De Backer G, De Henauw S, et al. The Belgian food consumption
certainly attenuate ranking of individuals because variability in survey: aims, design and methods. Arch Pub Health 2005;63:1–16.
742 European Journal of Public Health

3 Willett W. Food-frequency methods. In: Willett W, editor. Nutritional Epidemiology, 12 Willett W, Lenart E. Reproducibility and validity of food-frequency questionnaires.
2nd edn. New York / Oxford: Oxford University Press, 1998:74–100. In: Willett W, editor. Nutritional Epidemiology, 2nd edn. New York / Oxford:
4 Andersen LF, Johansson L, Solvoll K. Usefulness of a short food frequency ques- Oxford University Press, 1998.
tionnaire for screening of low intake of fruit and vegetable and for intake of fat. Eur J 13 Altman DG. Practical Statistics for Medical Research. London: Chapman & Hall,
Public Health 2002;12:208–13. 1991.
5 Osler M, Heitmann BL. The validity of a short food frequency questionnaire and its 14 Bland JM, Altman DG. Statistical methods for assessing agreement between two
ability to measure changes in food intake: a longitudinal study. Int J Epidemiol 1996; methods of clinical measurement. Int J Nurs Stud 2010;47:931–6.
25:1023–9. 15 Mikkelsen TB, Olsen SF, Rasmussen SE, Osler M. Relative validity of fruit and
6 Thompson FE, Subar AF, Smith AF, et al. Fruit and vegetable assessment: per- vegetable intake estimated by the food frequency questionnaire used in the Danish
formance of 2 new short instruments and a food frequency questionnaire. J Am Diet National Birth Cohort. Scand J Public Health 2007;35:172–9.
Assoc 2002;102:1764–72. 16 Wakai K. A review of food frequency questionnaires developed and validated in
7 Therese L, Lillegaard I. Evaluation of a short food frequency questionnaire used Japan. J Epidemiol 2009;19:1–11.
among Norwegian children. Food Nutr Res 2012;56, doi:10.3402/fnr.v56i0.6399. 17 Fernandez-Ballart JD, Pinol JL, Zazpe I, et al. Relative validity of a semi-quantitative
8 Van Oyen H, Tafforeau J, Hermans H, et al. The Belgian Health Interview Survey. food-frequency questionnaire in an elderly Mediterranean population of Spain. Br J

Downloaded from https://academic.oup.com/eurpub/article/23/5/737/444280 by CNR user on 30 November 2023

Arch Public Health 1997;55:1–13. Nutr 2010;103:1808–16.
9 De Vriese S, Huybrechts I, Moreau M, Van Oyen H. The Belgian national food 18 Rothenberg EM. Experience of dietary assessment and validation from three
consumption survey 1. Scientific Institute of Public Health 2004. [cited 2012 Swedish studies in the elderly. Eur J Clin Nutr 2009;63(Suppl 1):S64–8.
February]; available from: http://www.iph.fgov.be/epidemio/epien/index5.htm. 19 Noethlings U, Hoffmann K, Bergmann MM, Boeing H.European Investigation into
10 Health Counsil Belgium. Household Weights and Measures. A Manual for a Cancer and Nutrition. Portion size adds limited information on variance in food
Standardised Quantification of Food Items in Belgium 1997:280. intake of participants in the EPIC-Potsdam study. J Nutr 2003;133:510–5.
11 Beaton GH, Milner J, Corey P, et al. Sources of variance in 24-hour dietary recall 20 Hunter DL, Sampson L, Stampfer MJ, et al. Variability in portion sizes of commonly
data: implications for nutrition study design and interpretation. Am J Clin Nutr consumed foods among a population of women in the United States. Am J
1979;32:2546–59. Epidemiol 1988;127:1240–9.

.........................................................................................................
European Journal of Public Health, Vol. 23, No. 5, 742–746
ß The Author 2012. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.
doi:10.1093/eurpub/cks109 Advance Access published on 20 August 2012
.........................................................................................................
The mediating effect of Mediterranean diet on the
relation between smoking and colorectal cancer: a
case–control study
Niki Kontou1,2, Theodora Psaltopoulou3, Nick Soupos3, Evangelos Polychronopoulos1,
Dimitrios Xinopoulos2, Athena Linos3, Demosthenes B. Panagiotakos1

1 Department of Nutrition and Dietetics, Harokopio University, Athens, Greece

2 Saint Savvas Cancer Hospital, Athens, Greece
3 School of Medicine, University of Athens, Athens, Greece

Correspondence: Demosthenes B. Panagiotakos, 46 Paleon Polemiston Street, 166 74 Glyfada, Greece,

tel: +30210-9603116, fax: +30210-9600719, e-mail: d.b.panagiotakos@usa.net

Background: The protective role of Mediterranean diet (MD) and the detrimental effect of smoking on colorectal
cancer (CRC) have already been shown. The aim of this work was to evaluate the potential mediating effect of MD
on the association between the aforementioned factor (smoking) and CRC. Methods: It is a case–control study.
Two hundred fifty consecutive patients with CRC (63  12 years, 59% males) and 250 age–sex group-matched
controls, both from the area of Attica, were studied. Various socio-demographic, clinical, lifestyle (including
detailed smoking habits) and dietary characteristics were measured. Adherence to the MD was evaluated using
the MedDietScore (theoretical range 0–55). Results: Each unit increase in the MedDietScore was associated with
13% lower likelihood of CRC (P < 0.001). Smoking habits were associated with 2.9-fold the likelihood of CRC
among participants who were away from the MD (i.e. MedDietScore < 29) and with 2.1-fold the likelihood of
CRC among those who were close to the MD (P < 0.05). Conclusions: Adherence to the MD was associated with a
less detrimental association of smoking habits with CRC, suggesting indirect benefits of adherence to this dietary
pattern with regards to CRC morbidity and mortality.
.........................................................................................................

Introduction second place for women and at the fourth place for men), repre-
senting a major cause of cancer morbidity.1 Among several factors,
ccording to the International Agency for Research on Cancer smoking seems to play an important role. A recent meta-analysis
A(IARC) GLOBOCAN 2008 data, the colorectal cancer (CRC) on smoking and CRC revealed that ever smokers had 18% higher
incidence for both sexes was ranked at the third place worldwide, risk as compared with never smokers, and this association was
at the first place in Europe and at the third place in Greece (at the dose-dependent regarding pack-years.2 Several mechanisms have