Knowledge about food classification systems and value attributes provides insight for understanding complementary food choices in Mexican working mothers

Research report

Knowledge about food classification systems and value attributesprovides insight for understanding complementary food choices inMexican working mothers

Maria Guadalupe Rodriguez-Oliveros a Carole A Bisogni b Edward A Frongillo ca Centro de Investigacioacuten en Salud Poblacional Instituto Nacional de Salud Puacuteblica Avenida Universidad 655 Colonia Santa Maria AhuacatitlanCuernavaca Morelos CP 62100 Mexicob Division of Nutritional Sciences Cornell University 1303B MVR Hall Ithaca NY 14853-4401 USAc Department of Health Promotion Education and Behavior Arnold School of Public Health University of South Carolina 915 Greene Street Room 529Columbia SC 29208 USA

A R T I C L E I N F O

Article historyReceived 1 April 2014Received in revised form 18 July 2014Accepted 17 August 2014Available online 19 August 2014

KeywordsWorking mothersFood value attributesFood classificationComplementary foodsYoung childrenInfant feeding

A B S T R A C T

Knowledge about mothersrsquo perceptions of food classification and values about complementary feedingis necessary for designing educational and food supply interventions targeted to young children To de-termine classification attributes and consumptionpreparation routines of key complementary foods44 mothers of children lt 2 y of age in 14 manufacturing businesses were studied Using 31 key foodswe conducted free-listings pile-sort and food attributes exercises Hierarchical clustering showed thatmothers identified nine classes of key foods including milk derivatives complements junk food infantproducts chicken parts and other meats From multidimensional scaling mothers used three primaryclassification systems food groups food introduction stages and food processing Secondary classifica-tion systems were healthyndashjunk heavyndashlight hotndashcold goodndashbad fat and main dishndashcomplement Childhealth and nutrition particularly vitamin content were salient attributes Fruits and vegetables were pre-ferred for initiating complementary feeding on the second month of age Consumption of guava mangoand legumes however was associated with digestive problems (empacho) Red meats were viewed ascold-type heavy and hard not suitable for young children but right for toddlers Chicken liver was con-sidered nutritious but dirty and bitter Egg and fish were viewed as a vitamin source but potentiallyallergenic Mothers valued vitamin content flavor and convenience of processed foods but some weresuspicious about expiration date chemical and excessive sugar content and overall safety of these foodsMothersrsquo perceptions and values may differ from those of nutritionists and program designers and shouldbe addressed when promoting opportune introduction of complementary foods in social programs

Introduction

Complementary food choices are important determinants ofinfant feeding practices and future eating habits relevant to child

health development and growth (Black et al 2013) Life-coursemodels explain how food choices develop in changing temporalsocial and historical contexts (Devine 2005) Peoplersquos life-courseexperiences affect food choice through ideals personal factors re-sources and the social and food context (Sobal amp Bisogni 2009)Value attributes play a role in attaching positive or negative va-lences to the potential physiological psychological and sociologicalconsequences when making food choices These valences are modi-fied by specific situations such as illness that eventually influenceconsumer behavior (Bisogni Jastran Seligson amp Thompson 2012Blake Bisogni Sobal Devine amp Jastran 2007 Myung McCool ampFeinstein 2008)

Food-related value attributes are health and nutrition conve-nience quality sensory perceptions and monetary considerationsamong others (Dutta Sywulka Frongillo amp Lutter 2006 Sobal ampBisogni 2009 Winter-Falk Sobal Bisogni Connors amp Devine 2001)People cope with the diversity of foods that are potential satisfiers

Acknowledgements The authors gratefully acknowledge study informants busi-nesses managers child care center staff and IMSS authorities for their collaborationin this study Special thanks to Lawrence Williams PhD Kathleen OrsquoConnor PhDand Pamela Tolbert PhD Professors of Cornell University for contributing to projectconception All authors designed the research and approved the final manuscriptGRO and EAF wrote the paper GRO conducted the research analyzed the dataand had primary responsibility for the final content Funding Supported in part bythe National Council of Science and Technology Mexico (grant number S 49934)the Mario Einaudi Center Cornell University US and the First Presbyterian Churchof Ithaca NY US Conflict of interest G Rodriguez-Oliveros C Bisogni EA Frongillohave no conflicts of interest to disclose

Corresponding authorE-mail address efrongilloscedu (EA Frongillo)

Appetite 83 (2014) 144ndash152

Contents lists available at ScienceDirect

Appetite

journal homepage wwwelseviercom locate appet

of relevant values by grouping them into sets or categories so asto reduce the complexity of their choice (Blake et al 2007 Siegler1991) Complementary food classifications are the developmentalstages of food introduction food groups based in the nutrient contentof foods food color shape texture effects on the child availabili-ty and accessibility among others (Bisogni et al 2012 Blake et al2007 Bovet Vauclair amp Blaye 2005 Dutta et al 2006 GerberSiegler 1991 US Department of Agriculture Food and NutritionService Winter-Falk et al 2001)

People construct and use food classifications in multiple dimen-sions including those culturally recognized socially significant andpersonally operational (Bisogni et al 2012 Siegler 1991) such astaxonomic categorizations (Bovet et al 2005)Understanding folkcategorizations is important because they may differ from scien-tific categorizations Consistently the social-sciences literature stressesthe importance of understanding emic perspectives by studying andanalyzing a setting or a behavior from the authorrsquos perspective (WorldBank 2014) Therefore understanding caregiversrsquo perceptions andvalue attributes about foods is needed if health providersrsquo and childfood programs are to assure proper communication and sustainability

Developing countries are experiencing increases in maternal em-ployment (Gonzaacutelez-Cossiacuteo Rivera-Dommarco Moreno-MaciacuteasMonterrubio amp Sepuacutelveda 2006) The mechanisms through whichchildrenrsquos eating patterns relate with maternal employment are mul-tifaceted Time limitation may constrain parental food choices byoffering low nutrient and energy-dense complementary food(Chaterji amp Frick 2005) Studies about maternal work and infantfeeding have mostly focused on breastfeeding behaviors (BranSkinner amp Carruth 2001 Chaterji amp Frick 2005 Hawkins GriffithsDezateux Law amp the Millennium Cohort Study Child Health Group2007 Hirani amp Karmaliani 2013 Lakati Binns amp Stevenson 2002)Few studies have been conducted in developing countries to un-derstand working mothersrsquo complementary feeding choices (Heiniget al 2006 Pan-American Health Organization 2013)

The objective of this study was to determine classificationsystems value attributes and food preparationconsumption rou-tines of key complementary foods among Mexican mothers ofchildren less than two years of age working in manufacturing busi-nesses This study aimed to expand knowledge about the barriersand facilitators for promoting desirable feeding choices of home andprocessed complementary foods in Hispanic working mothers usingdifferent caregiving arrangements

Methods

Study design and data collection

This cross-sectional qualitative study was conducted in CentralMexico in 14 manufacturing businesses located in the Cuern-avaca City industrial zone Businesses ranged from 115 to 920employees and had at least 25 of working women Data collec-tion followed an interpretative approach building upon theparticipantrsquos real-life experiences and understandings regardinginfant feeding (Hough amp Ferraris 2010) Three different purposivesamples were chosen for data collection which included free listingsurveys focus groups screening surveys and pile-sort and food-attributes exercises as described below In preparation for datacollection a one-week training workshop was conducted with localSpanish-speaking interviewers Data collection instruments werepilot tested through personal interviews with mothers working ina manufacturing business located in the study site

Free listing surveyInitially a purposive sample of 20 key local informants knowl-

edgeable of young child feeding was chosen for conducting a freelisting survey (Hough amp Ferraris 2010) A total of 15 working mothers

of young children five day-caregivers and five pediatric nurses wereselected from a National Institute of Social Security (IMSS) region-al hospital and a daycare in Cuernavaca City To identify the culturaldomain of foods given to young children informants were askedto freely list all foods that are usually given to children less thanone year of age in the region They were interviewed individuallyin a private room by a trained nutritionist who registered the listof foods mentioned by each informant This survey was con-ducted in the winter season when certain seasonal fruits such asmango tangerine and guava were available

Selection of key complementary foodsFrom the set of 112 foods given to children less than one year

of age that were mentioned at least one time in the free listingsurvey a list of 31 key foods was composed (Pan-American HealthOrganization 2013) This list included the 21 most frequently men-tioned foods five processed foods that were less mentioned butrelevant in infant feeding and five foods that are potentially richsources of iron zinc fat and simple carbohydrates Plastic color cardswith the photograph and name of each of the key foods weredeveloped to be used in the last phase of the data collection

Focus groupA purposive sample of 12 working mothers of children lt 2 y of

age attending an IMSS daycare in Cuernavaca City was selected forconducting a focus group to pilot test the food cards The sessionwas carried out by a trained nutritionist at the daycare An inter-view guide was developed including as inquiry topics a) thepertinenceclarity of the color size and shape of the images andb) the adequacy of the font size color and words used in the textThe session lasted 28 minutes it was audio-recorded andtranscribed verbatim Afterwards some cards were redesigned

Screening surveyTo identify informants for the pile-sort and food-attributes ex-

ercises a screening survey was conducted with all blue-collar womenworking in the 14 selected businesses This survey asked for socio-demographic characteristics including the number of children lt 2 yof age commuting time working conditions and childcare arrange-ments Six mothers did not participate in this survey because ofeither a vacation or illness leave

Single pile-sort and food-attributes exercisesA stratified purposive sample of 44 blue-collar working mothers

relying on either family care (Group F = 22) or institutional daycare(Group D = 22) and using public transportation to get to and fromwork was chosen from the screening survey group to conduct botha single pile-sort and food-attributes exercises (Pan-American HealthOrganization 2013) Mothers were selected looking for diversity indemographic (age born place schooling and family structure) andworking characteristics (income years employed extra time ab-senteeism commuting from home to work) All mothers selectedagreed to participate

A single pile-sort exercise was conducted first to identify cul-turally defined food groups Trained interviewers asked the mothersto form as many piles as they wanted with at least two cards butplacing one item in only one pile (Weller amp Rommney 1988)Themothers provided a name and a description of the classification cri-teria for each food group Afterwards a food-attributes exercise wasconducted with these mothers Interviewers showed one food cardat a time to the mothers while asking about their perceptions offood attributes and food preparationconsumption routines Inter-viewers clarified the connotation (positive or negative) of an attributewhen necessary The sessions ranged from 20 to 32 minutes theywere audio-recorded and transcribed verbatim

145MG Rodriguez-Oliveros et alAppetite 83 (2014) 144ndash152

Ethical considerations

The mothers health providers and caregivers participated vol-untarily and provided written informed consent Informants did notreceive economic incentives for attending data collection ses-sions As recognition for their time invested a childrenrsquos plastic plate-spoon kit was given to them The study objectives recruitmenttechniques and data collection procedures were informed througha letter to the businesses chief executive officer The Cornell Uni-versity Committee on Human Subjects and the Ethics Review Boardof the IMSS approved the research protocol

Data analysis

Data from the screening and free-listing surveys were summa-rized by descriptive statistical methods For the pile-sort exercisethe number of piles for each person ranged from two to eight pileseach consisting of an individualized set of two to twelve cards Foodclasses were generated by using two approaches for analyzingproximities Johnsonrsquos hierarchical clustering and non-metric mul-tidimensional scaling (MDS) Hierarchical clustering attempts to findgroups that are nested within each other The ldquoagglomerativerdquo al-gorithm developed by Johnson used in the hierarchical clusteringallows the identification of food groups while it starts with joiningmany small clusters (ie pair of similar foods) and gradually mergesinto fewer bigger clusters (ie food groups) The purpose of MDSis to provide a visual representation of the pattern of similaritiesamong a set of items such as foods (Cox amp Cox 2000) MDS plotsitems perceived to be similar to each other in doing so it finds aset of vectors in p-dimensional space by ordering the items in themap along a continuum The pile-sort survey analysis was run usingAnthropacreg software version 4 (Bogartti 1996) One participant wasleft out from this analysis because a food card was placed twice intotwo different groups

Qualitative information from the pile-sort exercise provided insightto interpret the dimensions of the MDS maps and to analyze thereasons for classifying foods a coding catalog was developed froma sample of the transcripts of the pile-sort exercise and then ex-panded as new concepts emerged from the analysis Coding was peer-reviewed for accuracy and dimensions were defined until consensuswas obtained by project co-investigators and assistants

For the food-attributes exercise the researchers reviewed thetranscriptions to develop a coding catalog that compiled the food-attributes mentioned by the mothers This catalog was organizedinto five clusters physical wellbeing child nutrition food prepa-ration and consumption routines food quality and age of foodintroduction The positive and negative attributes expressed by themothers about each key food were registered in two thematic con-ceptual matrixes respectively (Miles amp Huberman 1994) Thefrequency of mention per attribute was counted separately formothers relying on family childcare daycare or both

Results

Characteristics of study informants

The average age of the 44 study working mothers was 279 yearsMost of the informants were born in the study city (641) and com-pleted middle school (714) Families had from one to four childrenAlmost one of every two informants (461) did not have a spouseor partner at the time of the study Women earned the minimalwage and reported being employed by the current business for 52years on average 146 worked extra time in the past monthMothersrsquo extra time was carried out generally on weekends and waseither paid or considered a replacement for the time they took offto care for their ill children or attend their childrenrsquos school Almost

one of every three mothers missed at least one day at their job inthe three months previous to the study The mean transportationtime from home to work was 377 minutes ranging from 102 to755 minutes

Food classification

Hierarchical clustering revealed that mothers classified key foodsinto nine classes More than half of the mothers agreed on averageon the foods selected to compose these four food classes vegeta-ble (086) fruit (078) junk food (chatarra 067) and complements(053) Lower degree of agreement on the foods included in a spe-cific class was observed however when composing the classes ofmilk derivatives (044) soups (043) chicken parts (038) other meats(038) and infant products (030)

When plotting mothersrsquo food-classification similarities in mul-tidimensional scaling the plot using two dimensions showedsatisfactory stress (015) reflecting a reasonable fit A reduction ofthe stress was observed by using a three-dimensional model (009)reflecting a better fit (Fig 1) This model indicates that mothers usedat least three different dimensions or criteria when classifying thekey foods The identified dimensions shown in the MDS map werefood groups food introduction stages and food processing Figure 1shows the first dimension regarding food groups as indicated by thefood clusters Overall this classification was consistent with the oneobtained by hierarchical clustering The second dimension refers tostages of introduction of complementary foods as represented bythe dotted lines The third dimension refers to food processing rep-resented by the southeastndashnorthwest axis where the key foods wereclassified as natural and processed foods Fruits and vegetables weredescribed as natural foods while Danoninoreg yogurt and Gerberregamong other foods were classified as ldquopackagedrdquo or processedproducts

Stages of introduction of foodsMothers in both groups recognized at least four different stages

of food introduction over the first year of age The first stage from2 to 4 months of age refers to the initiation of complementaryfeeding by trying some fruits and vegetables with soft consistencyand lack of strings and seeds Both groups of mothers but mainlymothers relying on institutional child-care mentioned the issue ofldquopreparing their childrenrdquo during the second month of age by in-troducing formula as well as some solid foods so the children willbe ldquoreadyrdquo when they come back to work after 42 days of deliv-ery Even mothers who considered it appropriate to initiatecomplementary feeding at later stages recognized that they had in-troduced complementary foods earlier to facilitate child-care foralternate caregivers when the mothers were at work as in thefollowing statement

ldquoMy baby didnrsquot really have to get used to my milk because I hadplanned to go back to work So before she was born I had alreadybought formula to have it on hand at home that way she gotused to her bottle we also began to feed her porridges so I hadno problems and she didnrsquot suffer on my absencerdquo

The second stage from 5 to 6 months according to the mothersis when the baby is training ldquothe stomachrdquo to receive solids but stillis at high risk of allergies stomachache and choke Foods intro-duced in this stage are legumes broth and some cereals The thirdstage from 7 to 10 months coincides with teeth eruption This timeis associated not only with childrsquos capacity to chew foods with harddry or sticky texture but also with development of the stomachand an intestine that allows the child to tolerate ldquoheavy foodsrdquo betterThis is a landmark period in complementary feeding because of theincrease of mothersrsquo confidence about tolerability of foods such ascereals legume (solids) and certain meats such as chicken

146 MG Rodriguez-Oliveros et alAppetite 83 (2014) 144ndash152

During the fourth stage from 11 months and over children areoffered a diverse diet and are included in family meals Red meatfish and whole egg are offered Finger foods such as rolled tortillaldquotacordquo banana and other fruits are considered practical because chil-dren can eat these foods by themselves while the ldquobusyrdquo caregiveris doing home chores or taking care of other children Mothers gen-erally would not recommend soda and potato chips for children atany age although some of them acknowledged giving it to theirpreschool child once heshe was demanding them

Food attributes

Mothers recognized at least 82 positive and 94 negative attri-butes of key complementary foods (Tables 1 and 2) Overall foodswith high proportion of positive attributes and few negative attri-butes were apple banana carrots squash chayote oatmeal pastaand chicken (meat) In contrast pork potato chips and soda amongother processed products were described mainly by negative at-tributes The most frequently mentioned attributes were related tochild health and nutrition These attributes were linked togetherwhen reported

Physical well-being attributesPositive attributes related to child well-being focused on three

main topics child health child illnesses and growthdevelopmentBoth groups of mothers considered vegetables and a few animalfoods such as chicken as important for preventing illnesses and main-taining health of less than one-year-old children Some mothershowever were worried about chicken liver egg and beef causingdigestive problems in young infants Mothers brought up a condi-tion called ldquoempachordquo characterized by stomachache abdominaldistention and finally diarrhea or constipation This condition wasoften associated with consumption of certain fruits such as guavaand mango as well as with consumption of legumes describedbelow

ldquoI was told that my child can eat brown beans at six months butjust the broth otherwise my baby will get sick [se empacha]because beansrsquo peel sticks on my babyrsquos stomach rdquo

Another salient concept relates with the ldquoheavy foodsrdquo such aspork and fish Mothers explained that the stomach of a very youngchild does not tolerate foods that are too strong or heavy and con-sequently develops diarrhea Mothers associated allergic events withconsumption of animal food source such as whole egg red meatand fish Moreover a few mothers pointed out red meat ascarcinogenic

Child nutrition attributesMothersrsquo perceptions about the nutritional value of the key foods

mainly focused on their vitamin content Most of the studied fruitsand vegetables as well as chicken and its derivatives were recog-nized as vitamin-rich foods A general belief was that bean andchicken broth contain a lot of vitamins and are highly nutritiousbecause solids left most of ldquotheir substancerdquo in the broth Anothercommon belief was that black beans are more nutritious and suit-able for the child than brown beans Mothers expressed contrastingconcepts about the nutritious value of processed products Forexample some of them considered Danoninoreg as a vitamin sourcewhile others thought it has a low nutritious value or might makethe children overly fat A few informants declared not knowing thenutrient value of processed infant products and consequentlyquestioned their role in child feeding

Regarding mineral sources respondents in both groups charac-terized legumes Danoninoreg and vegetables as sources of iron Bothchicken liver and infant cereals were also recognized as iron-richfoods Mothers acknowledged the importance of iron in prevent-ing anemia but paradoxically bean broth unlike meat wasconsidered an iron food source Moreover three mothers de-scribed meat as low nutritious Some mothers extrapolated themineral content of certain foods such as milk to its derivatives Forexample they stated that cream contains the same nutrients thanmilk as in the following example

ldquoCream contains milk so it is calcium this is one of the foodsthat children must eat frequentlyrdquo

Concerning the main nutrients mothers made a distinctionbetween ldquogood and badrdquo sources of flour (starchy carbohydrates)

guava mango tangerine apple

banana

broccoli carrot

chayote

brown beans broth lentils pasta

oatmeal rice

yogurt

Danonino reg

pork beef fish

Gerber reg(meat) Boxed cereal

potato chips soda

infant cereal

tortilla bread

chicken liver

chicken

cheese

Stage One

2-4 mo of age

Stage Two

5-6 mo of age

Stage Three

7-10 mo of age

Stage Four

11 mo and over

Natural Processed

Fig 1 Three-dimensional model of mothersrsquo food-classification similarities obtained from multidimensional scaling Notes The circles refer to complementary food groupsThe dotted lines refer to food groups according to developmental stages The solid line refers to processed (left side) and natural (right side) food groups

Table 1Positive attributes of key complementary foods recognized by Mexican working mothers

Positive attributes Key complementary foodsa

Mothers F and D Mothers FFamily child-care

Mothers DDaycare

Child well-beingIt doesnrsquot harm the baby promotes

child well-being healthy foodsFruitsb vegetablesc brown bean chicken brothchicken liver pasta mango lentils oatmeal fishtangerine

Tortilla infant cereal pastacream rice

Chicken

Good for stomachache and a gooddigestion

Oatmeal Rice tortilla

Good for teeth itching Oatmeal TortillaGood in winter time for colds Tangerine OatmealGood for preventing cancer Green vegetablesHelps to grow-up well it is good

for child developmentChicken yogurt oatmeal apple banana Danoninoreg beef infant cereal

pastaCarrot mango

Helps to strengthen bones Tortilla Beans brothChild nutrition

Nutritious Chicken broth and meat infant cereal oatmeal blackbeans fish cheese rice beef vegetables

Apple banana yogurt breadpasta

Chicken liver eggstortilla fruits

Contains vitaminsa) General Fruits vegetables chicken (liver meat eggs and

broth) black bean broth Danoninoreg infant cereallentils fish tortilla boxed cereal rice cream oatmealbeef

Pasta yogurt breadcheese

b) Vitamin C Tangerine guava apple banana MangoContains minerals

a) Iron (prevents anemia) Brown bean broth lentils Danoninoreg Chicken liver and vegetables Infant cereal greenvegetable cream

b) Calcium Tortilla cream eggs Danoninoreg Infant cereal oatmeal brownbeans broth banana yogurtvegetables

Boxed cereal fishcarrot guava cheese

Contains proteins Chicken (liver and broth) beef fish Chicken (meat and eggs) ricemango

Pasta lentils

Contains good fat productive forthe body

Fish beef Danoninoreg Cheese

It is good flour [carbohydrates] Rice and breadContains fiber for good digestion Tortilla infant cereal rice boxed cereal Brown beans bread oatmeal

cheese meatMango vegetableslentils

Food preparation and consumptionroutinesI use it often at home Tortilla cream pasta Danoninoreg brown beans

oatmeal eggs lentils chicken (liver and meat)bread vegetables mango

Rice Chicken brothbanana cheese yogurtbeef tangerine

I prepare soups with these foods Chicken and legume broths with tortilla rice pastavegetables chicken (meat or liver)

I use it as a reward Soda potato chipsPractical ready to eat or easy to

prepareVegetables apple banana Infant cereal Gerberreg carrot

eggsBoxed cereal

Dinner food Tortilla Pasta oatmeal boxed cereal Bread banana withoatmeal

Breakfast food Apple banana Yogurt BreadSnacks Guava GerberregGood dessert Apple banana YogurtIt is a complement Bread tortilla Oatmeal and cream Soda

Food qualityChildren like its flavor Soda potato chips tortilla Danoninoreg fruits

chicken broth fish boxed cereal pasta rice brownbean broth

Lentils Chicken liver cream

It is creamy soft tender or easy tochew

Infant cereal Egg tuna fish pasta Chicken

It has an attractive color DanoninoregIt is clean meat and less aggressive

than red meatChicken Fish

It is a refreshing beverage SodaIt is natural so it is good Fruits Guava mango rice

breadSeasonal food (cheep) Tangerine Apple Guava

a Foods in bold were the most frequently mentionedb Fruits apple banana guava and tangerinec Vegetables squash chayote carrot and broccoli

protein and fat For example some mothers considered rice as ldquobadflourrdquo because it makes the child overly fat Some mothers de-scribed red meats as ldquobad proteinsrdquo for young children because oftheir content of fat cholesterol and hard texture A few mothersdid not consider egg yolk appropriate for young children becauseof its content of cholesterol Nevertheless some mothers pre-ferred egg yolk to white because the latter is considered allergenicfor young infants

Fiber content was associated with a good digestion and ldquothestrength of childrsquos intestinerdquo Mothers often pointed out cereals andtheir derivatives as fiber sources A few mothers included cheese

and meats in this category Furthermore a few mothers indicatedthat meat fiber is similar to vegetable fiber so both fibers are equallybeneficial for children

Food quality attributesFoods appreciated for their sensory characteristics were creamy

low viscosity soft and easy to chew foods which are consideredparticularly suitable for children lacking teeth In contrast hard thickdry and rough foods were disapproved because of mothersrsquo concernof choking Informants also had strong concerns about foods con-taining strings seeds or bones such as mango guava and fish not

Table 2Negative attributes of key complementary foods recognized by Mexican working mothers

Negative attributes Key complementary foodsa

Mothers DDaycare

Child well-beingHeavy foods Pork fish beef boxed cereal soda potato

chipsPasta (solids) Danoninoregmango guava

It causes digestive problems such asstomachache (empacho) flatulencedistention (se esponjan) andorcramps

Brown beans lentils Danoninoreg guavastrawberry mango

Chicken liver egg rice beefand tortilla

It causes constipation Guava TortillaIt causes diarrhea andor vomit Mango cream tangerineIt might cause allergies Egg and fish Gerberreg (meat) and other

animal productsIt contains cancer Red meatChoking hazard because of its texture Boxed cereal potato chips tangerine rice

mango oatmealBrown beans and lentils grainpork and tortilla

Child nutritionIt has low nutritious value not

beneficial for childrenBoxed cereal Danoninoreg soda potato chips Guava Pork sausage cream Gerberreg

(meat)It is junk food (chatarra gusqueria

chucheria fritanga)Soda potato chips Beef

It is just a treat (golosina) DanoninoregContains a lot of cholesterol andor fat Egg chicken skin cream red meat Fish chicken liver PorkMakes the child overly fat Rice Soda and potato chips Infant cereal DanoninoregContains bad carbohydratesbad flour Bread rice Soda potato chips Pasta oatmealIt contains a lot of sugar Soda Boxed cereal bread No commentsI do not know if it is good or bad for

infantsEgg lentils Gerberreg boxed cereal Chicken liver Danoninoreg fish

mango guavaChicken broth cream pastainfant cereal

Food preparation and consumptionroutinesI barely use it at homethe daycare

centerPork fish and lentils soda potato chips Chicken liver oatmeal Gerberreg

I have not offered it to my baby yet Chicken liver boxed cereal Gerberreg (meat)lentils mango tangerine cream fish

Oatmeal Danoninoreg guavayogurt

Infant cereal pork potatochips

It must be given infrequentlyin lowquantity

Pork egg Danoninoreg cheese soda beefpotato chips

Cream mango tortilla pastacombined with other ldquofloursrdquo

Children might get used to processedfood

Infant cereal

Food qualityIt is hard tick dry or rough Boxed cereal beef Oatmeal Pork guavaI dislike its flavor so I do not give it to

my childPork and Gerberreg (meat)

My child dislikes its flavor Guava chicken liver infant cereal fish lentilsoatmeal

Boxed cereal Danoninoregcream brown beans cheeseyogurt

Gerberreg brown beans lentils

It just calms down the childrsquos appetitefor a while

Potato chips

It is a cold-type of food Tangerine beef Cream PorkIt is a hot-type of food MangoIt contains chemicals not good for

babiesGerberreg tuna fish soda potato chips infantcereal

Pork sausage Danoninoregyogurt boxed cereal

It is a canned packaged syntheticartificial or processed food

Gerberreg Pork sausage Potato chips yogurt andcheese

It has been on storage for a long timeand might be expired so I do nottrust it

Gerberreg Danoninoreg

It is dirty food Pork chicken liverIt can be contaminated or spoiled Fish Chicken

a Foods in bold were the most frequently mentioned

only because of the choking hazard but also because they mightldquostick to the babyrsquos stomachrdquo

Food taste was cited as relevant in mothersrsquo child feeding choicesSome mothers recognized not buying foods they personally dislikeby assuming their child would dislike these foods as well On theother hand they stopped offering nutritious foods when their childrejected its flavor For example some mothers described that theirchildren refused to eat chicken liver because of its bitter flavor sothey did not offer it despite the fact that that they viewed it asnutritious and overall good food for children A number of mothersreported that they stopped offering chicken liver to theirgrown-up child because over time the child rejected this food

The notion of ldquocold and hot foodsrdquo came out in both groups ofmothers Informants stated that especially cold-type foods such astangerine might be strong for young children and there is a riskthat milk gets spoiled if consumed together A mother mentionedthat she soaks cold-type fruits in water before consumption to reducethat risk

Processed foods were considered less suitable for very youngbabies than ldquonatural foodsrdquo despite the former being viewed as easyto prepare and practical for busy caregivers Furthermore a fewmothers affirmed that only non-processed foods are suitable for thefirst year of age so they did not buy any packaged food for theirinfants Reasons for rejecting processed foods relate mainly withtheir ldquochemicalrdquo content but also with their sugar content and fresh-ness Mothers believe that chemicals particularly preservatives andartificial colors are harmful for young babies as a participant stated

ldquoI would rather feed my child with natural foods packaged foodsare for lazy mothers These foods [processed food] contain a lotof preservatives and just look at me I am overweight becauseof the preservatives Preservatives might also harm my babyrsquosstomachrdquo

A few mothers described yogurt and cheese as foods with sig-nificant content of preservatives Mothers also complained aboutthe excessive sugar content of certain processed foods such as boxedcereal and soda Additionally some mothers were suspicious aboutthe freshness and overall quality of processed products as in thefollowing testimony

ldquoBecause it [a processed food] has been on storage for many daysit might be spoiled I just do not trust it no matter what theexpiration date saysrdquo

Safety concerns were also expressed about animal-source foodsSome mothers believed that pork is fed with excrement fish mightbe sold spoiled and chicken liver contains ldquothe bad emotions of theanimalrdquo so these foods are not suitable for infants

Mealtime attributesMothers stated attributes describing the role of foods either in

different mealtimes such as breakfast and dinner or in a main mealsuch as complements and snacks Dinner foods defined as ldquoveryfillingrdquo were particularly appreciated when causing the child to sleepfor a longer period of time These foods included cereals and bananaBrown beans were not considered as dinner foods because theyldquocause abdominal blow-up and flatulence at nightrdquo Guava and pro-cessed foods were considered snacks described as convenience foodsto eat outside home or when home cooked meals are not ready

Food preparation and consumption routines

Mothers provided a comprehensive description of the ways care-givers prepared the key foods for infants Boiling was the preferredpreparation technique during the first year of age Some easy-to-crush fruits such as apple and banana were also offered raw Seasonalfoods and ldquoatolesrdquo (ie starchy food beverages prepared with cereals)

were highly valued as well as chicken legume and pasta brothsMothers said that they usually combine broths with rice tortillavegetables or chicken liver indicating that broths might be aldquovehiclerdquo for nutritious foods

Mothers in Group F said that grandmothers usually prepare mainmeals for children given that they are the primary alternate care-givers A few mothers explained that grandmothers occasionally offerfoods and preparations they had not planned to give to their chil-dren at a certain age or time of the day as this mother explains

ldquoAt the beginning I did not want to give tortilla to my childbecause I was afraid of choking However one day after comingfrom work I saw my mom giving it to Juanito and I kept mymouth shutrdquo

Similarly a few mothers using daycare mentioned their chil-dren were offered some of the key foods that they do not prepareat home This was generally perceived as an opportunity for theirchildren to try new foods and preparations A few mothers indi-cated that they do not know how to prepare some key foods suchas lentils and fish so they do not offer these at home They ac-knowledge however that their children might consume these foodseither in their mother-in-lawrsquos home or in the daycare center

Discussion

Food groups stages of food introduction and food processingwere primary food classification systems used by Mexican urbanworking mothers with young children Secondary but relevant foodclasses were healthyndashjunk heavyndashlight hotndashcold goodndashbad fat andmain dishndashcomplement These findings are consistent with thoseof previous studies conducted in Latin America (Monterrosa PeltoFrongillo amp Rasmussen 2012) Positive and negative complemen-tary food attributes relate with child physical wellbeing nutritionfood quality and mealtime but barely with price

There were some similarities between the two groups of mothersin terms of food misconceptions Both groups associated mango withnegative attributes considering it a choking hazard ldquoheavyrdquo and acause of diarrhea Both groups reported offering cream to the youngchild because it prevents anemia (Group D) or promotes child well-being (Group F) Both groups agreed that chicken and beans brothsare highly nutritious and strengthens bones These food miscon-ceptions may drive food choices therefore nutrition educationactions targeted to caregivers must be implemented by employersand health personnel to support working mothers

We found two main differences between groups F and D relatedwith food convenience and use as a reward In terms of food con-venience group F mentioned using a greater number of foodsconsidered ldquopracticalrdquo than group D including fruits vegetables andprocessed infant products Mothers with time constraints might beparticularly sensitive to this issue so healthy food choices must beencouraged when selecting foods from this class When it comesto rewarding the child group F declares giving local fruits while groupD acknowledges that they give soda and potato chips even thoughthey classified these products as ldquojunk foodrdquo Using foods as a rewardis a practice that may have undesirable implications for comple-mentary feeding particularly when caregivers use energy densefoods promoting child overweight

The most salient classification system in both groups of motherswas food grouping consistent with international classifications thatinclude vegetable fruit milk derivatives and meats (Dixon Croninamp Krebs-Smith 2001 Kaufer-Horwitz Valdeacutes-Ramos WilletAnderson amp Solomons 2005 Painter Rah amp Lee 2002) A fat andsugar class present in several of these classifications however wasnot found in our study Mothers possibly were influenced by theMexican Food Guideline that lacks a fat and sugar class (Secretariacuteade Salud 2006) also they may consider this class unsuitable for

infants The classification based on stages of food introduction foundin our study emerged as promising for designing and marketinginfant products in nutrition social programs Moreover this systemis currently used by the baby food industry (Gerber) perhaps in-fluencing mothersrsquo belief system The secondary food classificationsystems found in our study have been reported in previous studiesconducted in Latin America (Behrens amp Shipibo 1986 Dutta et al2006 Kuhnlein amp Pelto 1997 Winter-Falk et al 2001) indicatingtheir relevance for understanding mothersrsquo food choices in local socialcontexts Our study supports that culturally sensitive complemen-tary food initiatives have to build on both primary and secondaryclassification systems to assure sustainability

Regarding complementary food attributes mothers bring out dis-trust concerns about expiration date nutrient value and safety ofprocessed products preferring natural or ldquohomerdquo foods Theygrouped together all processed foods except dairy products that wereoverall well-accepted perhaps by assuming that they share the nu-trition value of milk This can be advantageous for marketing milk-based fortified processed foods used in social initiatives such as theMexican ldquoOportunidadesrdquo Program (Mardones-Santander et al 1988Martiacutenez Campero Rodriacuteguez amp Rivera 1999 Rosado et al 1999)

Food-price attributes of animal-source foods were barely invokedin this study This suggests that economic factors appear to be weakerbarriers for the opportune introduction of beef fish and whole eggthan negative perceptions and misconceptions about these foodsAcceptance of animal source foods must be tested in food-based pro-grams oriented to increase the consumption of iron and zincSuccessful examples are reported in community-based studies con-ducting recipe trials (Creed-Kanashiro et al 1991 2003 Villanueva2001)

Our study provides limited information about dark-green leafyfoods used to increase vitamin A consumption in food-based pro-grams (Creed-Kanashiro et al 1991 Jones Specio Shrestha Brownamp Allen 2005) If the study had been conducted in the rainy seasonthese foods would have been more on informantsrsquo minds and men-tioned in the free-list We assumed that the single pile-sort exercisecaptured the most salient food arrangement Consecutive pile-sorts however might have provided additional insight aboutalternative food arrangements as those that emerged in the food-attributes exercise We acknowledge discretion inherent to themethod when selecting the MDS dimensions To address this po-tential limitation the researchers supported MDS findings byincorporating qualitative information about food grouping provid-ed by the mothers following an interpretative approach Cross-sectional data do not inform about permanent and temporaryproscriptions of complementary foods documented here Furtherstudies are necessary to have a better understanding of these issues

To our knowledge this is one of the few studies conducted indeveloping countries aimed to understand working mothersrsquo com-plementary feeding choices when using different caregivingarrangements We used robust methods and analytic techniques tobuild the food classification systems Our definition of key foods wasbroader than used previously (Pan-American Health Organization2013) by adding low-nutrient and energy-dense foods and bever-ages relevant for child overweight Our results sheds light on thecomplexity involved in decision-making for complementary feedingand provide insight to understand similar urban ethnoculturalgroups guiding inquiry for nutrition programming and educationThese findings though cannot be generalized to the larger popu-lation or working women having different socio-demographiccharacteristics or be extrapolated to ill children Finally our studydemonstrates that low-cost and feasible data-collection methodsprovide valuable qualitative information that can be used as for-mative research for developing behavior-change programs andfurther quantitative studies in the region targeted to working mothersof young children

Nutrition content linked to health outcomes should continuebeing stressed by nutrition educators when promoting comple-mentary foods Emic terms might be used for designing culturallyinsightful nutrition messages targeted to working mothers The ev-idence reported here will help to promote mission-based researchand policy development that is oriented to sustain and encourageexclusive breastfeeding and opportune introduction of nutritiouscomplementary foods to improve health and adequate develop-ment of young children of working mothers (Ruel Brown amp Caulfield2003)

References

Behrens C A amp Shipibo I (1986) Food categorization and preference Relationshipsbetween indigenous and western dietary concepts American Anthropologist 88(3)647ndash658

Bisogni C A Jastran M Seligson M amp Thompson A (2012) How people interprethealthy eating Contributions of qualitative research Journal of Nutrition Educationand Behavior 44 282ndash301

Black R E Victora C G Walker S P Bhutta Z A Christian P de Onis M et al(2013) Maternal and child undernutrition and overweight in low-income andmiddle-income countries Lancet 382(9890) 427ndash451

Blake C E Bisogni C A Sobal J Devine C M amp Jastran M (2007) Classifyingfoods in contexts How adults categorize foods for different eating settingsAppetite 49 500ndash510

Bogartti S P (1996) Anthropac four methods guide Natick MA Analytic TechnologiesBovet D Vauclair J amp Blaye A (2005) Categorization and abstraction abilities in

3-year-old children A comparison with monkey data Animal Cognition 8 53ndash59Bran L S Skinner J D amp Carruth B R (2001) Maternal employment and dietary

quality of children aged 42 to 60 months Topics in Clinical Nutrition 17(1) 27ndash35Chaterji P amp Frick K D (2005) Does returning to work after childbirth affect

breastfeeding Review of Economics of the Household 3(3) 315ndash335Cox T F amp Cox M A (2000) Multidimensional scaling (Vol 32 pp 241ndash254)

Washington DC Chapman and HallCreed-Kanashiro H C Fukumoto M Bentley M E Jacoby E Veroza C amp Brown

K H (1991) Use of recipe trials and anthropological techniques for thedevelopment of a home-prepared weaning food in the Central Highlands of PeruJournal of Nutrition Education 23(1) 30ndash35

Creed-Kanashiro H M Bartolini R M Fukumoto M N Uribe T G Robert R Camp Bentley M (2003) Formative research to develop a nutrition educationintervention to improve dietary iron intake among women and adolescent girlsthrough community kitchens in Lima Peru The Journal of Nutrition 133(11Sndash2)3987ndash3991

Devine C M (2005) A life course perspective Understanding food choices in timesocial location and history Journal of Nutrition Education and Behavior 37121ndash128

Dixon L B Cronin F J amp Krebs-Smith S M (2001) Let the pyramid guide yourfood choices Capturing the total diet concept The Journal of Nutrition 131(2S)461Sndash472S

Dutta T Sywulka S M Frongillo E A amp Lutter C K (2006) Caregiversrsquo attributesof complementary foods in four countries in Latin America and the CaribbeanFood and Nutrition Bulletin 27 316ndash326

Gerber (2014) Start healthy stay healthy nutrition system lthttpwwwgerbercomgtLast accessed 140825

Gonzaacutelez-Cossiacuteo T Rivera-Dommarco J Moreno-Maciacuteas H Monterrubio E ampSepuacutelveda J (2006) Poor compliance with appropriate feeding practices inchildren under 2 y in Mexico Journal of Nutrition 136 2928ndash2933

Hawkins S S Griffiths L J Dezateux C Law C amp the Millennium Cohort StudyChild Health Group (2007) The impact of maternal employment on breast-feeding duration in the UK Millennium Cohort Study Public Health Nutrition 10(9)891ndash896

Heinig M J Follett J R Ishii K D Kavanagh-Prochaska K Cohen R amp PanchulaJ (2006) Barriers to compliance with infant-feeding recommendations amonglow-income women Journal of Human Lactation 22(1) 27ndash38

Hirani S A amp Karmaliani R (2013) The experiences of urban professional womenwhen combining breastfeeding with paid employment in Karachi Pakistan Aqualitative study Women and Birth Journal of the Australian College of Midwives26(2) 147ndash151

Hough G amp Ferraris D (2010) Free listing A method to gain initial insight of a foodcategory Food Quality and Preference 21(3) 295ndash301

Johnson SC (1967) Hierarchical clustering schemes Psychometrika 2 241ndash254Jones K M Specio S E Shrestha P Brown K H amp Allen L (2005) Nutrition

knowledge and practices and consumption of vitamin A-rich plants by ruralNepali participants and nonparticipants in a kitchen-garden program Food andNutrition Bulletin 26(2) 198ndash207

Kaufer-Horwitz M Valdeacutes-Ramos R Willet W C Anderson A amp Solomons NW (2005) A comparative analysis of the scientific basis and visual appeal of sevendietary guideline graphics Nutrition Research 25 335ndash347

Kuhnlein H V amp Pelto G H (1997) Culture environment and food to prevent vitaminA deficiency Quebec Canada Centre for Nutrition and the Environment ofIndigenous People McGill University

Lakati A Binns C amp Stevenson M (2002) Breast-feeding and the working motherin Nairobi Public Health Nutrition 5(6) 715ndash718

Mardones-Santander F Rosso P Stekel A Ahumada E Llaguno S Pizarro F et al(1988) Effect of a milk-based food supplement on maternal nutrition status andfetal growth in underweight Chilean women The American Journal of ClinicalNutrition 47(3) 413ndash419

Martiacutenez H Campero L Rodriacuteguez G amp Rivera J (1999) Aceptabilidad asuplementos nutricios en mujeres embarazadas o lactando y nintildeos menores decinco antildeos Salud Puacuteblica de Meacutexico 41(3) 163ndash169

Miles M B amp Huberman A M (1994) Qualitative data analysis An expandedsourcebook (2nd ed) Thousand Oaks CA Sage

Monterrosa E C Pelto G H Frongillo E A amp Rasmussen K M (2012) Constructingmaternal knowledge frameworks How mothers conceptualize complementaryfeeding Appetite 59(2) 377ndash384

Myung E McCool A C amp Feinstein A H (2008) Understanding attributes affectingmeal choice decisions in a bundling context International Journal of HospitalityManagement 27(1) 119ndash125

Painter J Rah J amp Lee Y (2002) Comparison of international food guide pictorialrepresentations Journal of the American Dietetic Association 102(4) 483ndash489

Pan-American Health Organization (2013) ProPAN Process for the promotion of childfeeding Washington DC lthttpwwwpahoorghqindexphpoption=com_docmanamptask=doc_viewampgid=22301ampItemid=gt Last accessed 140210

Rosado J L Rivera J Loacutepez G Solano L Rodriacuteguez G Casanueva E et al (1999)Desarrollo y evaluacioacuten de suplementos alimenticios para el programa deeducacioacuten salud y alimentacioacuten Salud Puacuteblica de Meacutexico 41(3) 153ndash162

Ruel M T Brown K H amp Caulfield L E (2003) Moving forward with complementaryfeeding Indicators and research priorities Washington DC International FoodPolicy Research Institute Discussion paper 146

Secretariacutea de Salud (2006) Oficial Mexicana NOM-043-SSA2-2005 servicios baacutesicosde salud Promocioacuten y educacioacuten para la salud en materia alimentaria Criterios parabrindar orientacioacuten Mexico DF SSA

Siegler R S (1991) Childrenrsquos thinking Englewood Cliffs NJ Prentice HallSobal J amp Bisogni C A (2009) Constructing food decision making Annals Behavioral

Medicine 38(1) 37ndash46US Department of Agriculture Food and Nutrition Service (2014) Develop healthy

eating habits lthttpwwwchoosemyplategovpreschoolershealthy-habitshtmlgtLast accessed 140826

Villanueva M A (2001) Desarrollo y prueba de recomendaciones sobre alimentacioacutencomplementaria para un programa de educacioacuten en nutricioacuten dirigido a madres denintildeos menores de dos antildeos de edad en comunidades semirurales del estado deMorelos Mexico DF Universidad Iberoamericana

Weller S amp Rommney A (1988) Systematic data collection (Vol 10) Qualitativemethods series New York Newbury Park Sage Publications

Winter-Falk L Sobal J Bisogni C Connors M amp Devine C M (2001) Managinghealthy eating Definitions classifications and strategies Health Education andBehavior 28(4) 425ndash439

World Bank (2014) World development indicators Washington DC World Banklthttpsearchworldbankorgdataqterm=maternal+employmentamplanguage=ENampformat=gt Last accessed140324

Knowledge about food classification systems and value attributes provides insight for understanding complementary food choices in Mexican working mothers

Introduction

Methods

Free listing survey

Selection of key complementary foods

Focus group

Screening survey

Single pile-sort and food-attributes exercises

Data analysis

Results

Food classification

Stages of introduction of foods

Food attributes

Physical well-being attributes

Child nutrition attributes

Food quality attributes

Mealtime attributes