Energy Balance, a new paradigm and methodological issues

12 Pages·2015·0.5 MB·English

Nutr Hosp. 2015;31(Supl. 3):101-112
ISSN 0212-1611 • CODEN NUHOEQ
S.V.R. 318
Energy Balance, a new paradigm and methodological issues: the ANIBES
study in Spain
Gregorio Varela Moreiras1,2, José Manuel Ávila2 and Emma Ruiz2
1Department of Pharmaceutical and Health Sciences, Faculty of Pharmacy, CEU San Pablo University, Madrid (SPAIN).
2Spanish Nutrition Foundation (FEN). Madrid (SPAIN).
Abstract BALANCE ENERGÉTICO, UN NUEVO
PARADIGMA Y ASPECTOS METODOLÓGICOS:
Energy Balance (EB) is an important topic to unders- ESTUDIO ANIBES EN ESPAÑA
tand how an imbalance in its main determinants (ener-
gy intake and consumption) may lead to inappropriate
weight gain, considered to be “dynamic”and not “sta- Resumen
tic”. There are no studies to evaluate EB in Spain and
El balance energético (BE) o equilibrio energético se
new technologies reveal as key tools to solve the common
refiere “simplemente” a que debemos ingerir la misma
problems to precisely quantify energy consumption and
cantidad de energía que gastamos. Conocer el concepto
expenditure at population level. Within this context, the
de BE y aplicarlo a nuestras vidas es quizá el factor más
increasing complexity of the diet, but also the common
importante para mantener una buena salud y tratar de
problems of under and over reporting in nutrition sur-
prevenir la obesidad. Sin embargo, la teoría no es sen-
veys have to be taken into account.
cilla llevarla a la práctica ya que, por un lado, en este
The overall purpose of the ANIBES (“Anthropometry,
ya avanzado siglo XXI desconocemos todavía en gran
Intake and Energy Balance in Spain”) Study was to carry
medida lo que comemos, en definitiva, nuestra alimen-
out an accurate updating of foods and beverages intake,
tación. Y ésta es cada vez más compleja, lo que dificulta
dietary habits/ behaviour and anthropometric data of the
sin duda controlar adecuadamente este lado de la “ba-
Spanish population as well as the energy expenditure and
lanza”, la ingesta. Pero además, en el otro lado, el corres-
physical activity patterns, by the use of new tested ins-
pondiente al gasto energético, aún es peor conocido y hay
truments (i.e. tablet device to assess energy intake and
muy escasa información en la cuantificación adecuada
accelerometer to evaluate physical activity). This new
del mismo. Debe recordarse, además, que no debemos
ANIBES Study will contribute to a better knowledge of
estudiar aisladamente los componentes del BE, sino de
the different key factors contributing to EB in Spain.
manera integrada, y como interaccionan. Problemas
(Nutr Hosp 2015;31(Supl. 3):101-112) como la infravaloración de la ingesta de energía, y la so-
brevaloración del gasto, son frecuentes en la mayoría de
DOI:10.3305/nh.2015.31.sup3.8758 las encuestas nutricionales, impactando más en aquellos
Key words: Energy balance. Dietary intake. Physical ac- grupos de población en los que el control del BE resulta
tivity quantification. New technologies. Nutrition surveys. aún más necesario. El empleo de las nuevas tecnologías
abre numerosas posibilidades para las encuestas de ba-
lance energético. Precisamente, la innovadora metodolo-
gía (empleo de “tablet” para cuantificación de la ingesta,
y de acelerómetros para el nivel de actividad física) en
el reciente Estudio ANIBES (“Antropometría, Ingesta, y
Balance Energético en España”), representativo de la po-
blación española, y que hemos desarrollado, proporciona
una herramienta útil y actualizada para un mejor cono-
cimiento del balance energético de la población española,
como se pone de manifiesto en el presente artículo.
(Nutr Hosp 2015;31(Supl. 3):101-112)
DOI:10.3305/nh.2015.31.sup3.8758
Palabras clave: Balance energético. Encuestas nutricona-
les. Ingesta. Actividad física. Nuevas tecnologías.
Correspondence: Gregorio Varela Moreiras.
Catedrático de Nutrición y Bromatología.
Director Dpto. CC. Farmacéuticas y de la Salud.
Facultad de Farmacia Universidad CEU San Pablo.
Urb. Montepríncipe, crta. Boadilla km. 5,3.
28668 Boadilla del Monte, Madrid. España.
E-mail: [email protected] / [email protected]
101
013 Balance energetico_Gregorio Varela.indd 101 12/02/15 13:50
Introduction and goals tabolizing food consumed; and the energy expended
through physical activity 2,6. RMR is proportional to
Energy Balance (EB) is defined as the state achie- body mass, particularly the amount of fat-free mass.
ved when energy intake equals energy expenditure and There is also, however, a large variability in RMR (250
is considered to be “dynamic” and not “static”1. The kcal/d ;1000 kJ/d) that is not explained by differen-
concept of energy balance for regulating body weight ces in body composition7.The TEF is proportional to
is simple in principle. When energy expenditure ex- the total food consumed and, in a typical mixed diet,
ceeds energy intake, weight is lost. When energy in- makes up 8% to 10% of total energy ingested. Diet
take exceeds energy expenditure, weight is gained2. composition has a strong effect on TEF, and a hierar-
Human physiology complies with the first law of ther- chy of macronutrient effects on the magnitude of TEF,
modynamics, which states that energy can be transfor- the highest for protein and lowest for fat. Normally,
med from one form to another but cannot be created TEF is assumed to be a fixed percentage of EI, but
or destroyed. A fundamental principle of nutrition and variation between and within individuals also occurs
metabolism is that body weight change is associated 7,8. The energy for physical activity, the most varia-
with an imbalance between the energy content of food ble component of energy expenditure, consists of the
eaten and energy expended by the body to maintain amount of physical activity performed multiplied by
life and to perform the variety of physical work. Such the energy cost of that activity and can be further parti-
an EB framework is a potentially powerful tool for in- tioned into exercise energy expenditure and non exer-
vestigating the regulation of body weight3. cise activity thermogenesis 9,10.
Energy intake (EI) includes three major macronu- In summary, the key factors that may regulate the
trient groups—carbohydrates, proteins, and fats—and energy expenditure side of the balance are: resting me-
alcohol. Once ingested, the net absorption of the major tabolic rate; activities of daily living/level of sedentary
macronutrient groups is variable and incomplete,wi- behavior; metabolic cost of digestión, absorption and
th fecal losses accounting for 2–10% of gross intake metabolism; amount (min/d); type, timing and inten-
3,4. The net absorption of dietary energy components sity of physical activity; thermogenesis; and current
varies among individuals and is dependent on the spe- body weight and composition.
cific foods eaten, how they are prepared, and intestinal The different components of the energy balance
factors. The metabolizable energy of a diet represents equation continuously change over time11. Beginning
the difference between the absolute energy of ingested at conception, and remains positive, on average, throu-
substrates and the energy losses found in feces and uri- ghout growth and development12. This positive energy
ne. Commonly used energy densities for carbohydrate imbalance is reflected by increasing body weight. If
(4 kcal/g, 17 kJ/g), protein (4 kcal/g, 17 kJ/g), and fat adult weight is then maintained over the long term,
(9 kcal/g, 38 kJ/g) represent population averages for average EB approaches zero, and an approximate ave-
metabolizable energy available to cells for conducting rage state of energy balance should be present. Howe-
the different biological processes4. ver, most adults gain fat throughout their lives and
Digestibility depends on the composition of the food in later life lose skeletal muscle; although the energy
item and on its content of fiber and other indigestible content of body fat change is much higher than that of
components. Such components can mechanically limit lean tissue change. Thus, even with weight stability,
the access of digestive enzymes to food that would po- “perfect” energy balance over the long term does not
tentially be digestible. These effects can have a large occur in most older adults 2,13,14.
impact on the absorption of ingested macronutrients. In terms of a short term period, over a 24-h period,
The variability in absorptive efficiency depends on a typical person eats several meals during the day, and
any additional factors (eg, gut flora, food preparation, EB is strongly positive during and after each meal.
diet composition), which may explain the individual Energy output is continuous but with increases due to
differences in metabolizable energy. Other factors to episodic physical activity and reduction during sleep.
influence EI are: the increasing complexity of the food In consequence, EB is thus highly variable even over
market nowdays (i.e. over 30,000 available products just a 1-d period. In addition, most adults also vary their
for shopping); total intake; energy density; timing and daily eating and activity patterns; thus, EB should be
composition of diet relative to physical activity; food seen as a dynamic rather that a static state since it also
liking and disliking; social and economical context; varies from day to day, and EB may be achieved only
or current body weight and composition 3,4. Absorbed when averaged over longer time periods. The energy
carbohydrates, proteins, and fats are transformed in content of a given meal is highly variable between in-
vivo to substrates that can ultimately either be oxidi- dividuals and highly variable between meals in an in-
zed to produce metabolically useful energy that drives dividual. However, the variation in total caloric intake
biological processes or they may be stored 5. summed across all meals over a day is far less variable.
Humans expend energy through resting metabolic This suggests that there is meal-to-meal compensation
rate (RMR), which is the amount of energy neces- of intake13. In addition to variation in intake between
sary to fuel the body at rest; the thermic effect of food meals on a given day, we also vary the amount of food
(TEF), which is the energy cost of absorbing and me- eaten each day. On the contrary, energy expenditure
102 DIETARY SURVEYS AND ENERGY BALANCE: METHODOLOGY, BIAS, ADJUSTMENT AND OTHER ISSUES
013 Balance energetico_Gregorio Varela.indd 102 12/02/15 13:50
rarely shows the same degree of variation across days. 1. Although we know much from short-term studies
Hence, and importantly, we are almost perpetually in about the major components of energy balance, our
energy imbalance on the time scale of hours or days, knowledge is still deficient regarding their interaction
and even through the whole life cycle. over the long term. Therefore, long-term, longitudinal
Furthermore, we know that the components of ener- studies to learn the details of the relations between
gy balance can be influenced by changes in each other components of energy balance and changes in body
as a consequence of positive or negative EB, which composition and weight among children and adults are
act to defend body energy stores, to maintain energy needed.
balance, and to prevent shifts in body mass2. We know 2. It has been shown that biological and psycholo-
that if energy balance were not controlled by such a gical factors affect the components of energy balance.
system and were subject only to behavioral mechanis- But generally, these have been studied independently
ms controlling food intake and energy expenditure, of one another and an integrative approach is required.
most people would routinely experience wide swings We need to know the relative importance of preinges-
in body weight over short periods of time. Therefore, tive factors on energy intake, energy balance, and the
the relative stability of body weight from day to day is physiologic response to a meal.
consistent with the view that EB is subject to physio- 3. Although our knowledge of the broader implica-
logical control. tions of physical activity and exercise have been inves-
The energy balance literature and information about tigated, we need to understand the effects of different
how our modern lifestyle differs from decades ago, doses (volume, intensity, pattern, timing) and types
hypothesize that human physiology developed under (endurance, resistance) of exercise on: a) total daily
circumstances that conferred an advantage for achie- energy expenditure and its components, b) EI and food
ving energy balance at a relatively high level of energy preferences, and c) body composition and body weight
expenditure or high energy flux, which was first shown in children and adults.
by Mayer in the 1950´ by their observation that energy 4. The individual variation in weight-loss response
intake was better matched to energy expenditure when to energy balance interventions is striking, and there-
people were physically active15. It is well known that fore we need to know the mechanism or mechanisms
over the past century and continue, the physical acti- responsible for the underlying active compensatory di-
vity level of most of the population has declined subs- fferences in energy intake, food preferences, and body
tantially. Although it is theoretically possible to avoid weight in children and adults. In particular, we have
weight gain in this situation, the fact that few people almost no information from energy balance studies
have accomplished this suggests that it is difficult to subsequent to weight loss during the difficult period of
maintain EB at a low energy throughput 16,17. weight maintenance.
Food restriction is a common strategy for reducing 5. And very important, measurements of ener-
excess body weight and/or treating obesity18. Howe- gy input and output are neither precise nor accurate
ver, food restriction produces weight loss, but it also enough to allow the calculation of energy balance over
produces compensatory decreases in other compo- the appropriate timeframe needed to understand the
nents of energy balance, such as decreases in energy mechanisms responsible for excess weight gain. Ac-
expenditure and body energy stores, and even an in- cordingly, we need to develop new methods that can
crease in hunger4. Because energy requirements fall reliably measure energy balance over extended time
with weight loss, a dangerous and unhealthy strategy periods in free-living people.
for weight loss maintenance is trying to match a lower The EB is not well defined for the Spanish popu-
level of energy expenditure with a lower energy intake. lation and it is essential to approach it with the aim
It has to been remained that lowering energy intake is of being able to properly establish the energy require-
opposed by biology and the environment. An additio- ments for our population and the subsequent Referen-
nal problem could be that at a reduced energy intake ce Intakes. It seems essential to improve the tools for
level, nutrients (mainly vitamins and minerals) dietary studying the energy intakes and losses of “free living”
reference intakes may not be achieved, which may re- independent subjects13. In this regard, the tools such
present a key problem in some vulnerable groups (e.g. as databases of the composition of quality foods, es-
children and elderly). Finally, at high energy expendi- pecially regarding energy and serving sizes, should be
ture, it would be easier to comply with the principle improved, as clearly stated at the recent (2013) Con-
that an adequate nutrition it is health but also pleasure. sensus Document and Conclusions on “Obesity and
In consequence, it matters how EB is achieved, at high Sedentarism in the 21st Century: What can be done and
or low level1. what must be done?”19.
It is therefore important to recognize that the EB Different valuable dietary surveys have been con-
system is interactive and complex: a change in one ducted in Spain, although to the best of our knowled-
component can affect one or more other components. ge, no one has attempted to specifically approach EB.
Recently, a Panel of Experts identified the following Briefly, the first Food Consumption Survey was per-
important gaps in our knowledge, and asked for future formed in 1956 under the National Health Survey. Fur-
investigation13: ther, several Spanish Food Consumption and Nutrition
Energy Balance, a new paradigm and methodological issues: The ANIBES Study in Spain 103
013 Balance energetico_Gregorio Varela.indd 103 12/02/15 13:50
Surveys have been carried out (ENNAs; 1964-1965, (error +/-6.9%) to increase the statistical power. The-
1980-1981 y 1990-1991) mainly in collaboration with refore, the random sample plus booster was 2,285 par-
the National Statistics Office (INE, Spain)20-23. From ticipants (Table 1).
1987 onwards, the current Ministry of Agriculture,
Food and Environment (MAGRAMA) in Spain laun- Table I
ches the National Food Consumption Survey (Panel), Distribution of the sample for the ANIBES study.
in for which the Spanish Nutrition Foundation (FEN)
SAMPLE
is responsible for analyzing the dietary patterns and
(n)
energy/nutrients intake of the Spanish population
since the year 2000 onwards24. At national level, the Initial
Final Final +
current AECOSAN (Spanish Agency for Consumer targeted
sample Boost
Affairs, Food Safety and Nutrition) recently carried sample
out the ENIDE Survey (Encuesta Nacional de Inges- Base 2634 2009 2285
ta Dietética) (AECOSAN, 2012)25. At present, the
SEX Men 1309 1013 1160
so-called ENALIA (Encuesta Nacional de Alimenta-
ción en la población Infantil y Adolescente) Survey in Women 1325 996 1125
children and adolescents from Spain is being carried
AGE (yr) Infants 240 100 213
out also under the auspices of AECOSAN. The latter 9 – 12
updates the reference survey in Spain in children and
Adolescents 246 124 211
young people (2-24 years old) called EnKid26, and the
13 – 17
AVENA study, a multicenter nutrition survey in Spa-
nish adolescents27. However, when approaching the Adults
1911 1588 1655
other main EB determinant (“energy expenditure”) 18 – 64
studies are much less frequent even scarcer and lack Elderly
237 197 206
of accuracy. The most recent National Health Survey 65 – 75
in Spain (2013)28 revealed that four out of ten persons
(41.3%) of the adult Spanish population is conside- In the ANIBES sample, 50.4% of adults (aged 9-75
red as sedentary, higher for women (46.6%) than for years) were men and 49.6% were women. The sample
men (35.9%) (they do not perform any kind of phy- reflects the distribution of males and females in the
sical activity during their leisure time). Considering general population within Spain. A more detailed des-
both their main and their leisure time activity, 40.9% cription of the sample for the ANIBES study is shown
of the adults (49.4% males, 32.4% females, aged 15- in Table 1.
69 years) perform strenuous to moderate weekly phy- The sample quotas according to the following va-
sical activity. There is consensus at present that not riables were:
only physical activity level but also inactivity and/or • Age groups (in years): 9-12, 13-17, 18-64 and
sedentary behavior should be taken into account and 65-75.
quantified 3,6. • Gender: men and women.
The ANIBES (“Anthropometry, Intake, and Energy • Region: 7 Nielsen areas (Northeast, Levant,
Balance in Spain”) Study, aims at adding new scien- South, West, North Central, Barcelona, Ma-
tific-based evidence to describe the interplay among drid) and Canary Islands.
energy intake, energy expenditure, and body energy • Habitats size: 2,000 to 30,000 inhabitants (ru-
stores and how an understanding of energy EB must ral population); from 30,000 to 200,000 in-
be considered as an useful tool either at individual or habitants (semi-urban population) and over
population level. 200,000 inhabitants (city/town population).
• Additionally other factors for sample adjust-
ment were considered: rate of unemployment;
Anibes study design and sampling procedure % of foreigners (immigrant population), level
of physical activity, or, education and econo-
The design of the ANIBES Study aimed a sample mical level.
size which should be representative of all individuals Study was conducted through a stratified multistage
living in Spain (excluding the autonomous cities of sampling and for more coverage and representative-
Melilla and Ceuta) aged 9-75 years, living in munici- ness, 128 sampling points were used, with 90 inter-
palities >2,000 inhabitants. The universe considered viewers allocated in 11 areas and 12 coordinators, all
about 37 millions of inhabitants. The initial potential previously trained by the Spanish Nutrition Founda-
sample was 2,634 individuals, and the final sample tion (FEN) (Figure 1). No previous pre-recruitment
was 2,009 individuals (2.23% error and 95.5% con- was considered, which minimized the risk of bias in
fidence interval). In addition, for the youngest groups responses.
(9-12, 13-17 and 18-24 years old), a boost was consi- The final protocol was approved by the Ethical
dered in order to have at least a n=200 per age group Committee for Clinical Research of the Region of Ma-
104 DIETARY SURVEYS AND ENERGY BALANCE: METHODOLOGY, BIAS, ADJUSTMENT AND OTHER ISSUES
013 Balance energetico_Gregorio Varela.indd 104 12/02/15 13:50
Fig. 1.—General sche-
me of the ANIBES Study
in Spain.
drid (Spain). The final fieldwork was carried out from Final Fieldwork
mid-September to November (three months), 2013.
Two pilot studies were previously carried out, as According to the number of interviews to be potentia-
follows: lly targeted at the sampling point, one or more random
Once the methodology was developed, a first pilot initial routes for the sampling process were considered.
study was carried out in June 2013. For this purpose, The later criteria were not used for municipalities over
2,060 individuals were contacted: 162 (7.8%) agreed 100,000 inhabitants where a postcode proportional cri-
to have the first visit/interview, 142 participated at the terion was considered. In the initial route, the apartment
second visit/interview, but only 97 were able to make building or family housing was randomly selected, as
the three-days dietary record by using the tablet. Fi- well as the first household to be approached. Non-eligi-
nally, only 57 participants were considered as fully ble addresses include vacant or derelict properties and
eligible. Therefore, a high rate of non-responders was institutions. If the uptake was positive, limits to be con-
observed mainly in the older age groups and parents of sidered for a potential participant were:
children and adolescents. The first pilot study allowed Apartment building:
reviewing several issues, both software and question- • 1-10 units, only one potential participant.
naires. • 11-20 units, two potential participants as maxi-
Once the results from the pilot study were comple- mum.
ted, several working/discussion groups were created in • 21-50 units, three potential participants maxi-
order to improve the study design, protocols, softwa- mum.
re and manuals. Therefore, four groups (one of inter- • >50 units, four participants.
viewers; two mixed groups of young adult people from For family housing, one possible participant per 10
25-35 years old; one group of parents with children units was the rule used.
aged 9-17 years) worked in order to improve the defi- The survey was designed in order that no more than
ciencies observed during the fieldwork. one adult and one child were selected from a household.
A second pilot study was carried out in order to eva- This meant that adults living in households with one or
luate the improvements after the first pilot study and more adults, and children in households with one or
comments and recommendations from the working more children were less likely to be selected than were
groups. A total of 60 individuals (52 used tablet devi- adults or children in single adult/child households.
ce; 5 photo camera; 3 by phone interview) participa- All interviewers, call centre and dieticians-nutritio-
ted. The second pilot study demonstrated the efficacy nists working on the ANIBES study were briefed and
of the amendments made and validated the tools and trained before undertaking an assignment and were
questionnaires to be used later during the main field- monitored during their assignment. All interviewers at-
work of the ANIBES Study. tended a two-day training course designed by the FEN
Energy Balance, a new paradigm and methodological issues: The ANIBES Study in Spain 105
013 Balance energetico_Gregorio Varela.indd 105 12/02/15 13:50
where they were fully briefed on the protocols and admi- Those following dietary advice such as for pre-
nistration of the survey. Fieldworkers were also issued vention of hypertension, diabetes, hypercholes-
with comprehensive written instructions covering sur- terolemia or hyperuricemia.
vey procedures and measurement protocols. The brie- Pregnant and lactating women.
fing sessions covered background and content, doorstep With diagnosed allergy and/or food intolerance.
approach, questionnaire administration (including prac- Suffering a metabolic disease such as hyper-
tice sessions), placement and collection of self-com- thyroidism or hypothyroidism.
pletions and ActiGraphs and the placement, checking c) Anthropometric measurements: the trained
and collection of the three-day food tablet diaries and interviewer collected the different measures fo-
24h-dietary recall and training in anthropometric data llowing the procedures tested before at the two
collection. After the briefing, “early work” checks were pilot studies:
carried out at the two pilot studies that were carried out. Height: by triplicate using a Stadiometer model
In order to cover a broad range of dates and to optimi- Seca 206 (Medizinische Messsysteme und Waa-
ze the devices to be used during the study, several stages gen seit 1840, Hamburg, Germany).
were designed (Figure 1), and comprised of: Weight: one determination in a weighing sca-
le model Seca 804 (Medizinische Messsysteme
und Waagen seit 1840, Hamburg, Germany).
Stage 1: the interviewer visits This scale provide information about BMI,
percentage of body fat and percentage of body
A letter and leaflet describing the purpose of the sur- water.
vey were previously posted in potential targeted apart- Waist circumference: by triplicate using a tape
ment building/family housing at the sampling points. measure model Seca 201.
A few days later, interviewers visited the addresses to d) 24-h Dietary Recall: an adhoc questionnaire
determine whether the address was private, residential was designed and previously checked and modi-
and occupied. They then carried out the selection pro- fied at the pilot studies. During the interview the
cess as already explained. participant recalled the food intake for the past
Interviewers carried out two main visits to house- 24 hours (working days). Food quantities were
holds who agreed to participate: assessed by using of household measures, food
The first visit (“face-to-face”) with an approximate models, pictures, or the brands. No prior notifi-
duration of 60 minutes comprised the following items: cation was given to the subjects about whether
a) Identification of the trained interviewer, as a or when they would be interviewed about their
collaborator of the FEN. The interviewer explai- food intake. Information on food consumption
ned the main goals of the study, the design and was recalled per day, per meal and in between
stages, the novelty of the tools to be used for co- meals. The 24-h dietary recall was designed to
llecting food intake and recording physical acti- further verification of the information collected
vity, as well as offered to have a feedback report at the Tablet, but also to make the participant
at the end of the study that included main results, more familiar with the type of information to be
dietary and physical activity advice, etc. The po- recorded in the diet record during the three-days
tential participant also received a letter from the period.
principal investigator of ANIBES Study, and was e) Tablet device for collection of dietary data:
informed for participation in the study. At this the three-day food diary:
point, the potential participant was asked to sign All the participants were provided with a ta-
the letter of consent for participation in the study. blet device (Samsung Galaxy Tab27.0.) and ins-
b) Inclusion/exclusion questionnaire: the inter- tructed on how to record by taking pictures of
viewer verified through a filter questionnaire that all foods and drinks consumed during these three
the participant was eligible for the ANIBES Study days, both at home and outside. Pictures had to
after reviewing the inclusion/exclusion criteria. be taken before and after finishing. Additionally,
Several exclusion criteria were applied: a brief description of the meals, recipes, brands,
Those individuals living in an institution (e.g. etc. had to be also recorded with the device. A
colleges; nursing homes, hospitals, etc.) specific software (“IPSOS Mobile for ANI-
Individuals following a therapeutic diet due to a BES”) was designed which allowed to save the
recent surgery or any medical prescription. uncompleted information before sending it for
If they were suffering a transitory pathology codification. The tablet was designed only to be
(i.e. flu, gastroenteritis, chicken pox, etc.) at the used to collect information related to ANIBES
time of the fieldwork. Study and no other uses were allowed.
Individuals employed in areas related to consu- A toll-free telephone number attended by call
mer science, marketing or the media. center operators of ipsos, trained by dieticians-nu-
However, individuals under the following condi- tritionists, was also available for the participants
tions were considered eligible to be included: in order to answer any questions about the softwa-
106 DIETARY SURVEYS AND ENERGY BALANCE: METHODOLOGY, BIAS, ADJUSTMENT AND OTHER ISSUES
013 Balance energetico_Gregorio Varela.indd 106 12/02/15 13:50
re, device working, food and beverage record, etc. three days. The average daily cpm for each parti-
Interviewers carried out a food diary check visit cipant was calculated as a weighted average based
with participants on the second visit or third day on the probability of wear/non-wear (for a mini-
of recording, either in person or over the telepho- mum wearing time of at least eight hours per day).
ne, with the aim of collecting missing details for The participants were also provided with a sheet
foods recorded after initial codification process to be filled in with the periods (hours/minutes)
was completed and lack of information was detec- of non-wear (shower, swimming, etc.). For those
ted. A manual of procedures to facilitate food co- participants that agreed to wear the accelerometer
llection was also given to the participants. Partici- (n=206) to quantify the physical activity, the devi-
pants were also informed that an insurance would ce was activated coincident with the Tablet based
cover any accident or incidence with the devices, three-day food diary. The subsample was selected
although they were asked for correct and watchful following the same criteria for representativeness
functioning and maintenance. as for the total sample included in the study. After
The different days of the week would (as far collecting the accelerometer devices at the second
as possible) be equally represented as each cy- interviewer visit, they were sent to IPSOS in order
cle always included two working days (Monday to download the recorded information from the par-
and Tuesday or Thursday and Friday) and one
ticipant (physical activity, but also additional data
weekend day (Saturday for Thursday and Friday
such as sex, date of birth, height, and weight) and
cycle or Sunday for Monday and Tuesday cycle).
to recharge the battery for the next participant. The
f) As an alternative to the tablet device, if the parti-
recorded information by the accelerometer in the
cipant declared or demonstrated that he/she was
subsample (167 adults and 39 children) was further
unable to use it, other possibilities were offered:
used to validate the physical activity questionnaire
photo camera and paper or telephone inter-
administered to the whole sample, and to build a
view. In both cases, the information to be collec-
mathematical model to quantify energy expenditu-
ted was the same as with the tablet.
re in combination with different standard formulas.
g) At the end of the first visit, the date for the se-
The second visit (“face-to-face”) with an approxi-
cond interviewer visit was agreed, as well as the
mate duration of 60 minutes comprised the following
telephone calls to be made for check up at the
items: detailed interview about physical activity (IPAQ
end of the collection of the data, if necessary.
questionnaire for children and adolescents modified
h) Accelerometer device to quantify physical ac-
according to the HELENA study29), and a detailed in-
tivity level: The objective physical activity mea-
terview by using validated questionnaires previously
surements were obtained with an accelerometer
tested at the pilot studies, designed to gain insights
ActiGraph (model GT3x y GT3x+; ActiGraph,
from the participants on important food safety, nutri-
Pensacola, FL, USA). This provides a measure of
tion and health-related topics, was also scheduled. In
the frequency, intensity, and duration of physical
addition, the tablet device and the accelerometer were
activity and allows classification of activity levels
collected. Participants who were considered “fully
as sedentary, light, moderate and vigorous. Indivi-
productive” according to the completion of the stages,
duals were asked to wear the ActiGraph on a belt
were asked whether they would like to receive feed-
above the right hip, during three consecutive full
back on the analysis of their diary and how was this
days including its cycle of the three days food &
compared to nutrient intake recommendations. The
beverages dairy record by the tablet. Objective me-
feedback also included general information related to
asurements of physical activity were taken using
healthy eating advice.
the ActiGraph, which recorded vertical movement,
where the number of movements (‘counts’) increa- • A summary of the ANIBES fieldwork follows:
sed with the intensity of activity. For any individual, • Fieldwork dates: September 19th through No-
the accelerometer recorded different periods during vember 16th, 2013, structured in 15 different
the day spent at different levels of activity, i.e. diffe- cycles/stages.
ring levels of ‘counts per minute’ (cpm), while they • 90 interviewers and 12 coordinators.
were being sedentary or engaging in light, mode- • Equipments:
rate, or vigorous activity. This provided a measure – 426 tablets devices
of the frequency, intensity, and duration of physi- – 90 devices for anthropometric measurements
cal activity and allowed classification of activity (weighing scale, stadiometers, tape measure).
levels as sedentary, light, moderate and vigorous. – 87 accelerometers.
Individuals were asked to wear the accelerometer • Devices employed
on a belt above the right hip, during waking hours – 79% of the sample used Tablet
for three consecutive full days in parallel with the – 12% used photo camera
three days of food & beverages dairy record by the – 9% used telephone interview
tablet. For the present study, the minimum wear A more detailed distribution of the devices used by
time criterion for inclusion in analysis was set at sex and age group is shown in Table 2.
Energy Balance, a new paradigm and methodological issues: The ANIBES Study in Spain 107
013 Balance energetico_Gregorio Varela.indd 107 12/02/15 13:50
E
N
O
EPH 202 100% 95 47% 107 53% 202 100% 8 4% 3 1% 114 56% 77 38%
L
E
T
A
st CAMAR 279 100% 143 51% 136 49% 279 100% 27 10% 18 6% 180 65% 54 19%
o
o
B T
ple + ABLE 1804 100% 922 51% 882 49% 1804 100% 178 10% 190 11% 1361 75% 75 4%
m T
a E
s C
Final DEVI Base 2285 100% 1160 51% 1125 49% 2285 100% 213 9% 211 9% 1655 72% 206 9%
E
N
O
study. LEFH 188 100% 89 47% 99 53% 188 100% 3 2% 3 2% 112 60% 70 37%
S E
E T
B A
he ANI AMAR 253 100% 124 49% 129 51% 253 100% 15 6% 8 3% 176 70% 54 21%
r t C
o
I e groups f ample E TABLET 1568 100% 800 51% 768 49% 1568 100% 82 5% 113 7% 1300 83% 73 5%
Table Ix and ag Final S DEVIC Base 2009 100% 1013 50% 996 50% 2009 100% 100 5% 124 6% 1588 79% 197 10%
e
s
o
g t NE
n O
accordi LEPH 237 100% 115 49% 122 51% 237 100% 10 4% 4 2% 133 56% 90 38%
d TE
e
s
Devices u ple PHOTO CAMERA 320 100% 156 49% 164 51% 320 100% 29 9% 21 7% 207 65% 63 20%
m
E (n) argeted sa E TABLET 2077 100% 1038 50% 1039 50% 2077 100% 201 10% 221 11% 1571 76% 84 4%
SAMPL Initial t DEVIC Base 2634 100% 1309 50% 1325 50% 2634 100% 240 9% 246 9% 1911 73% 237 9%
Base Men Women Base Children 9 – 12 Adolescents 13 – 17 Adults 18 – 64 Elderly 65 – 75
yr)
X E (
E G
S A
108 DIETARY SURVEYS AND ENERGY BALANCE: METHODOLOGY, BIAS, ADJUSTMENT AND OTHER ISSUES
013 Balance energetico_Gregorio Varela.indd 108 12/02/15 13:50
Fig. 2.—General sche-
me of the protocol
followed for data co-
llection, processing,
cleaning, and codifica-
tion during the ANIBES
Study in Spain.
Data processing seconds, and updated every 30 minutes. The Central
Server contained different modules to verify the in-
Figure 2 shows an scheme for a better understan- formation at individual level but also according to the
ding of the data processing, quality control,cleaning of specific cycle; food weight and intakes; food codifi-
the data, and codification and verification stages. cation and the assigning weight in grams. If for any
The innovative technology used in this study allowed reason, the terminal was unable to be connected to the
that all the collected information could be verified and network, the recorded information by the participant
codified in nearly real time. For that purpose, partici- was saved, and resubmitted once the problem was sol-
pants were asked to record everything eaten or drunk ved. The innovative Central Server system allowed
over three consecutive days. As already explained, checking the different cycles, the participants records,
when placing the tablet device, interviewers followed which meant that possible incidents were able to be
a protocol to explain the method, taking participants fixed in a very short period of time. Finally, 189,600
through the different sections including the instruction inputs (ingredients) were managed from the 2,009
page, how to describe details of food and drink and por- participants, about 73 items per participant, and 24.3
tion sizes and several examples. In addition to the detai- food/beverages items per person/day as mean.
led information about what and how much was eaten, Coders attempted to match each food or drink item
participants recorded for each eating occasion: where recorded in the tablet device with a food/portion code.
they were, who they were eating with and what they For composite items which could be split into their com-
were doing (for instance, of they were watching TV or ponent parts, each individual component was assigned.
not). After each day, participants recorded if their intake If an item had been recorded and there was no suitable
was typical for that day (and if not, the reason why) and code or there was insufficient detail to code the food,
details of any dietary supplements taken. The software the entry was flagged as a query. Each food code is
also contained a serie of questions about usual eating linked to appropriate portion size descriptors, such as a
habits (for example, type of milk or fat spread usually tablespoon for rice or pasta, which are then linked to the
consumed) to facilitate coding. correct weight for that descriptor. So if a participant re-
Food and beverage records were returned from the corded/described their food using household measures,
field in real time to be coded by trained coders and coders under dieticians-nutritionists checking would be
editors. For this purpose, an ad hoc Central Server sof- able to select the appropriate portion size. If the portion
tware/database was developed by IPSOS (Java/IE10 size was described as a weight, the weight would be en-
compatible) in order to be able to work in parallel with tered directly into the system. Where the coder could
the verification process followed by the codification. not resolve the food or portion consumed, the entry was
Within this context, the software was prepared to re- flagged as a query for action by a researcher who had
ceive the information from the field tablets every two greater nutrition knowledge and experience. The die-
Energy Balance, a new paradigm and methodological issues: The ANIBES Study in Spain 109
013 Balance energetico_Gregorio Varela.indd 109 12/02/15 13:50
ticians-nutritionists assigned appropriate codes for all Where errors were found they were corrected. These
flagged food and portion codes and checked any other checks ensured that error rates were monitored for all
queries raised by the coders. In general, where details the coders working on the project and helped identi-
for the coding of foods were missing, formally agreed fying any coding issues. All of the entries flagged as
default codes were used. Where portion sizes were mis- a query by the coders were categorized into different
sing, an estimate was made using the same weight if the query types, such as food code or portion code not
food was consumed on another dietary day, or a portion available in the VD-FEN Software, recipes, missing
size consistent with the participant’s usual consumption or insufficient detail to code food or portion. Initial
(e.g. small, medium or large), or an age-appropriate checks were carried out to highlight any missing data
average portion. fields, such as incomplete eating context or nutrient
For new products still not included in the ANIBES variables, followed by a feasibility check of the maxi-
program, supermarkets or retail markets were visited mum and minimum portion sizes entered within each
or the manufacturer contacted to obtain information on subsidiary food group. Final quality checking was per-
nutrient content in order to decide whether a new food formed using each participant’s mean energy and nu-
code was needed. This decision was based on nutritio- trient intake (all reported nutrients) over the food and
nal composition compared to that of existing codes, as beverages diary record period (three days). Extreme
well as the frequency of consumption. If a new food intakes were considered from the mean and all entries
code was required, the nutrient content was entered in this region were checked against the diary. All errors
into the database. In the case of school meals, school found were corrected to their appropriate entry as re-
caterers information about the nutrient content and flected by the diary entry. Intakes were also calculated.
portion size of dishes was considered. For homemade Intakes of energy and nutrients were calculated
dishes where a recipe had been recorded, the ingre- from the food consumption records using a special
dients were entered individually using the appropriate adapted VD-FEN 2.1 (Programa de Valoración Dieté-
cooked food codes, and all the codes for the dish were tica de la FEN, Dietary Evaluation Program from the
allocated to a recipe food group according to the type Spanish Nutrition Foundation) software for ANIBES
of dish. The weight of each cooked ingredient was cal- Study. Data obtained from food manufacturers were
culated using the raw weights recorded by the partici- also used, as was nutritional information provided on
pant, a weight loss factor for the whole dish if possible, food labels, and food photographic atlas to assign wei-
and the weight of the portion consumed. ghing grams for portion sizes. All data were carefully
evaluated before being incorporated into the VD-FEN
ANIBES Database, that is briefly stratified as:
Quality control • Level 1 – 16 food and beverage groups
• Level 2 – 29 food and beverage subgroups
The quality control of the collected information was • Level 3 – 761 food and beverages entries
supervised by trained dieticians-nutritionists, accor- • Company and brand
ding to the following protocol: • Culinary treatment
a) The same dietician-nutritionist was responsible • Household measure (tablespoons, glasses,
for checking the food records included by the parti- cups, plates); typical/most used portion sizes
cipant during the three-day dietary food record study. and recipes from Spain; or conventional units/
b) The initial quality control was based on the pho- measures (e.g. 1 yogurt, 1 apple piece, half to-
tographs and descriptions sent by the participants, but mato, 1 slice of bread, 1 soda can, 1 biscuit,
also the brief description that was asked before/after butter portion, etc.)
each meal and/or intake. Special care was given to va-
lidate some variables such as ingredients, brands of the
processed and ready-to-eat foods, portion size or culi- Data cleaning
nary technique in order to obtain accurate information
for further codification. Once the data from the Tablet devices were coded
c) The final approval of the received information and transferred into the ANIBES Database, a data cle-
was given by a dietician-nutritionist and supervisor. aning process was necessary as follows:
It is also of importance that the used software had First data cleaning stage: Participants were conside-
an alarm-system where no records from the different red as fully eligible if after a cautiously review of the
three main meals were available. information, it was verified that the three days were
At the start of the coding process, dieticians-nutri- recorded. Where registers were above or below the
tionists worked together with the coders checking the three-days established period, the following criteria
information from the coders and giving them indivi- was adopted:
dual feedback on their work (food and portion code en-
tries). Portion code errors (selecting the wrong portion – If a participant only had records from two or less
size descriptor or entering an incorrect weight) were days, he/she was considered as not valid and eli-
more common than selecting the wrong food code. minated from the final sample.
110 DIETARY SURVEYS AND ENERGY BALANCE: METHODOLOGY, BIAS, ADJUSTMENT AND OTHER ISSUES
013 Balance energetico_Gregorio Varela.indd 110 12/02/15 13:50