factorsinfluencingefficacyofnutritioneducation · pdf...
TRANSCRIPT
Systematic Review
Factors Influencing Efficacy of Nutrition EducationInterventions: A Systematic Review Mary W. Murimi, PhD, RDN; Michael Kanyi, PhD; Tatenda Mupfudze, PhD;Md. Ruhul Amin, MPH, MS; Teresia Mbogori, MS; Khalid Aldubayan, PhDCollegeConflict owith thisthe JNEEditor-inAddressSciencesTX 7940�2016 Sreservedhttp://dx
142
ABSTRACT
Objective: To examine systematically factors that contribute to the efficacy of nutrition education inter-ventions in promoting behavior change for good health based on their stated objective. In a departure fromprevious reviews, the researchers investigated factors that lead to success of various types of interventions.Critical analysis of these factors constituted the outcome of this review.Methods: This study followed Preferred Reporting Items for Systematic Reviews and Meta-analysiscriteria. A total of 246 original articles published between 2009 and 2015 in PubMed, Medline, Web of Sci-ence, Academic Search Complete, Science Direct, Cochrane Reviews, ERIC, and PsychLIT were initiallyconsidered. The number was screened and scaled down to 40 publications for the final analysis. Qualityassessment was based on the Cochrane Handbook for Systematic Reviews of Intervention. Studies were ratedas having low risk of bias, moderate risk, or high risk.Results: Efficacy of nutrition education interventions depended on major factors: interventions thatlasted $5 months; having #3 focused objectives; appropriate design and use of theories; fidelity in inter-ventions; and support from policy makers and management for worksite environmental interventions.Conclusions and Implications: Intervention duration of $5 months, #3 focused objectives, random-ization, use of theories, and fidelity are factors that enhance success of interventions based on the results ofthis study.Key Words: efficacy, interventions, nutrition education, systematic review (J Nutr Educ Behav. 2017;49:142-165.)
Accepted September 8, 2016. Published online November 1, 2016.
INTRODUCTION
Nutrition education can be viewed asany set of learning experiences designedto facilitate the voluntary adoption ofeating and other nutrition-related be-haviors conducive to health and well-being.1 Efficacy describes the abilityto yield intended outcome; for the effi-cacy of an intervention to be evalu-ated, it must be adequately described.2
Efficacy of nutrition education inter-ventions depends on several factorsincluding the duration and frequencyof intervention, the number and relat-
of Human Sciences, Texas Tech Univef Interest Disclosure: The authors’ confliarticle on www.jneb.org. The first authB staff as Associate Editor. Review of-Chief to minimize conflict of interestfor correspondence: Mary W. Murim, College of Human Sciences, Texas T9; Phone: (806) 834-1812; Fax: (806) 7ociety for Nutrition Education and Beh..doi.org/10.1016/j.jneb.2016.09.003
edness of the study objectives, studydesign and theory, and fidelity in inter-vention.
The specific characteristics of the de-terminants of success of interventionsare still unclear.2 However, severalstudies have been conducted to ascer-tain determinants of efficacy of nutri-tional education interventions. Forexample, another systematic review3
concluded that educational interven-tions that are sustained for a longertime, >5 months, and offer personal-ized feedback on dietary behavior andrelated health risk factors, are more
rsity, Lubbock, TXct of interest disclosures can be found onlineor of this article (M.W. Murimi) served onthis article was handled, exclusively, by the.i, PhD, RDN, Department of Nutritionalech University, PO Box 41240, Lubbock,42-3042; E-mail: [email protected]. Published by Elsevier, Inc. All rights
Journal of Nutrition Education and Beh
likely tobeeffective thanthoseconduct-ed fora shortperiod,<5months, anddonot offer personalized feedback. Otherstudies concluded that expert-led inter-ventions as well as studies that usedbehavioral theories, social support, andan educational approach to guide die-tary interventions were more likely tobe successful.4 Despite previous studieson the wider area of nutrition educa-tion, there is still inadequate literatureon the efficacies of the various nutritioneducation interventions that were im-plemented in recent years. In a departurefrom previous reviews that concen-trated primarily on a single type ofintervention and its related outcome,the current review investigated severalfactors that led to success of varioustypes of interventions. The purposeof this review was to examine system-atically the factors that contribute tothe efficacy of nutrition education in-terventions inpromotingbehaviorchangefor goodhealth andwell-being basedontheir stated objective. To achieve thispurpose, the researchers used population,intervention, comparison, and outcomescriteria to frame the research questions.5
avior � Volume 49, Number 2, 2017
Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Murimi et al 143
METHODSLiterature Search
This systematic review was conductedin accordance with recommendationsand criteria outlined in thePreferredRe-porting Items for Systematic Reviewsand Meta-analysis statement.6,7 Articleson studies that conducted nutritioneducation interventions on dietary be-haviors were identified by performingliterature searches in: PubMed, Medline,Web of Science, Academic Search Comp-lete, Science Direct, Cochrane Reviews,ERIC and PsychLIT. The search waslimited to articles published between2009 and 2015. Key search words werenutrition education, nutrition educationinterventions, dietary behavior, food,and health living. References of allretrieved studies were used to determinethe source of information, whether theywere primary, secondary, or websitebased, and to understand better thebasis for conclusions of the studiesthat were reviewed.
All 6 members of the research teamwere independently involved in re-viewing the references. Inclusion andquality measures were determined bythe 3 senior researchers who conduct-ed an independent evaluation of eacharticle; afterward, several discussionswere held to reach a consensus, hencemonitoring bias. A total of 246 originalstudies published since 2009 and tar-geting healthy individuals withoutpreexisting medical conditions werereviewed. This initial number wasscreened and scaled down to 40 publi-cations for the final analysis. Screeningcriteria for inclusion and eliminationare illustrated in the Figure.
Members of the Research Team
The research team was composed of 6members, 3 of whom held doctoral de-gree; the others had a master's degreein nutrition. The lead researcher was afull professor of nutrition and a regis-tered dietitian. Two other researcherswere faculty members in recognized in-ternational universities with wide expe-rience in nutrition, education, andresearch.Eachof the3senior researcherspaired with 1 junior researcher in eachdatabase for article search and retrieval.All 6 members were independentlyinvolved in reviewing the articles andinitially screening them.
Inclusion/Exclusion Criteria
The authors included in the reviewresearch articles published in Englishthat examined nutrition education in-terventions in adults aged >18 years.Studies were excluded if they were re-view articles, poster abstracts, or quali-tative, cross-sectional studies, or if thetarget population had special nutri-tional needs (eating disorders, dia-betic, hospitalized, etc). In addition,studies that failed to achieve any oftheir objectives were excluded. In thecases wheremultiple studies were con-ducted on the same data set, only themost recently published study wasincluded. There were 2 reviewers perdatabase. Trained reviewers evaluatedwhether articles met inclusion criteriaand determined the quality of thestudy. All researchers except the leadresearcher went through group training,conducted by a systematic review andmeta-analysis expert, which also involvedwatching a webcast.
Assessment of Study Quality/Risk of Bias
In the initial part ofworkof the currentreview, researchers worked in pairs inwhich data were extracted by 1reviewer and verified by a secondreviewer. The risk of bias in any re-ported evidence should be at mini-mum and evidence that is likely tohave high risk of bias serves a negli-gible purpose and thus should not beincluded in a systematic review evenif there is no better evidence.8 In thisreview, determination of the qualityof studies was guided by the Gradingof Recommendations Assessment,Development, and Evaluation systemof rating quality of evidence.9 A thor-ough assessment of the study'sfidelity,perceived conflict of interest regardingoutcome owing to sponsorship, studydesign, imprecision, inconsistency,appropriate use of theories, reasonableduration of intervention, andwhethera study achieved the stated objectivesformed criteria for quality assessment.Rating scores ranged from 1 to 6. Anydiscrepancies were discussed until anagreement was reached. Based onthese criteria of assessment of studyquality, studies were rated as having alow risk of bias (5–6 scores), moderaterisk (3–4 scores), or high risk (1–2
scores) (Tables 1 and 2). Fidelity as afactor in this systematic review wasassessed from authors' declaration oflimitation in their respective studies.
Reviewers completed a detailed dataextraction form. Extracted data weretransferred to a spreadsheet (Tables 1and 2).
Analysis Approach
The primary analytic goal was to deter-mine the overall effectiveness of nutri-tion education interventions to modifydietary and exercise behaviors. To deter-mine whether an intervention was suc-cessful, the outcome of the study wascompared with the stated purposeand/or objectives of the study. Once astudy was classified as having achievedits intended purpose, the contributingfactors were assessed. Assessed factorsincluded: (1) the design of the studyincluding randomness, (2) the type ofintervention and activities imple-mented, (3) the duration and dosage ofthe interventions, (4) number of objec-tives in a study, (5) fidelity in interven-tion implementation, and (6) the useof theory in directing the studies. Thesefactors were identified through a thor-ough reviewof publishednutrition edu-cation interventions. They were foundto be common in almost all publishedstudies. The duration of interventionwascategorizedas short if ithadacumu-lative length of>5months and long if astudy lasted for an accumulated periodof $5 months. This classification ofdurationwas deemed appropriate basedon the descriptions authors used of therespectiveoriginal studies. The reviewedstudies rarely reported the dosage andfrequency of interventions. Thereforeit was reasonable to report the totalamount of time spent in interventionin months.
Another factor that emerged duringthe review was worksite environmentinterventions. Worksite environmentsdiffer from one site to another. Thereare various worksite environment in-terventions for health living. Theseinclude the provision of health mes-sages around cafeterias, the provisionof healthy food in cafeterias, encour-aging and providing walking space aspart of exercise for healthy living,and schedules and amounts regardingeating, among others. The analysis ofworksite environment interventions
Figure. Flow diagram illustrating the article filtering process as part of the systematicreview. GEMs indicates Great Educational Materials.
144 Murimi et al Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017
is therefore case specific. Data onwork-site environment interventions wereanalyzed and reported along with otherinitially identified factors.
A semiquantitative approach wasused to summarizefindings fromnutri-tion education interventions. Resultsfromnutritioneducation interventionswere dichotomized based on whetherthey reported a statistically significant(P < .05) improvement in diet intake,exercise, or other related risk factorsfor obesity and diet-related chronicdiseases. The researchers used thisapproach to allow for the diverse rangeof reported statistics, outcomes, andmeasurement units.2
RESULTS
A majority of studies (68%; n ¼ 27)were conducted in the US; the remain-ing studies (33%; n ¼ 13) were con-ducted in other parts of the world. In
the current review, the duration ofintervention, number of objectives, fi-delity in intervention, use of theories,and use of worksite environmentwere identified as determinants of theefficacy of nutrition education inter-ventions. This review provides the re-sults and a discussion of these factors.
Slightly over half of the nutrition ed-ucation interventions (53%; n ¼ 21)were successful in modifying knowl-edge,behavior,orphysiologicoutcomesbased on their primary objectives. Ofthe 21 successful studies, 14 used the-ories and 7 did not. Another 48% ofthe interventions (n ¼ 19) met somebut not all of their primary objectives;9 of them used theories and 10 did not.
Study Designs in the ReviewedStudies
Randomized control trials accountedfor 70%of the studies (n¼ 28), the pre-
test–posttest/quasi-experimental designaccounted for 23% (n ¼ 9), and thenon-experimental design constituted8% (n¼ 3) of the total articles includedin the analysis. Over half of the totalnumber of studies (58%; n ¼ 23) werebased on theory.
Interventions in the ReviewedStudies
This review considered studies con-ducted in person or face-to-face inter-ventions at the individual or grouplevel. A majority of the studies (68%;n ¼ 27) used a single type of interven-tion. For example, a study whoseobjective was to create educationmaterials using the target audience'spreferences and to implement aheart-healthy diet education programused a single intervention in which 2registered dietitians led in-home edu-cation sessions.49 Table 1 listsmore ex-amples of studies that employed asingle intervention. Another 33% ofstudies (n¼ 13) employedmultiple in-terventions includingcooking,watchingvideos, attending exercise classes, andgardening. An example is a studywhose objective was to examine theextent to which participants in a com-bined physical activity and dietaryintervention achieved changes inmul-tiple health behaviors.36 Multiple in-terventions involved the provision ofopportunities for physical activity andhealthy eating before, during, and af-ter church services. Table 2 providesmore examples on studies that usedmultiple interventions.
Effect of Duration ofIntervention
For the purposes of this systematic re-view, the duration of the interventionwas categorized as short if it had a cu-mulative length of <5 months andlong if a study lasted for an accumu-lated period of $5 months. This wasinformed by the way studies reportedtheir time. Therefore, for conveniencein reporting of the results, the termsshort duration and long duration wereadopted. The amount of time spent inintervention was reported in months.The reviewed studies largely left outdosage and frequency of intervention.
In this review, two thirds (n¼ 12) ofnutrition education interventions that
Table 1. Face-to-Face, Individual, Group, and Peer Counseling/Nutrition Education Studies (n ¼ 27)
Authors Study Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias
(QualityMeasure) Major Findings
Arrebolaet al, 201110
60 patients with gradeII overweight andnon-morbid gradeI–II obesity (agerange, 18–50 y)
To evaluate effects oflifestylemodificationprogram focusedon diet, exercise,and psychologicalsupport on health-related quality of life
Pre–post.Intervention:physical activityrecommendationsand psychologicalsupport. Groupsessions led bydoctor, nurse, ordietitian.
No control
11 sessionsconductedevery 2 wk for6 mo
No theory All Low Intervention was associatedwith significantimprovements in physicalfunctioning (80.37� 18.90vs 89.40 � 13.95;P < .001) and role ofphysical (20.37 � 9.10 vs23.14 � 6.67; P < .05),vitality (58.71 � 21.98 vs70.91 � 26.56; P < .01),social functioning(79.62 � 27.76 vs86.57 � 25.45; P < .03),and general health(61.03 � 19.13 vs69.42 � 18.80; P < .001)factors.
Auld et al,201511
723 adults To analyze impact ofweightmanagementintervention onphysical activity/exercise and bodyweight andcomposition
Pre–post.Intervention: ESBAcurriculumdelivered in groupsettings.
No control
9 lessons taughtfor 8–12 wk
SCT andAdultLearningTheory
Some Moderate ESBA elicited mean positivebehavior change for foodresource management(P < .01), food safety(P < .001), nutrition(P < .001), and physicalactivity level in participatingstates (P < .01) exceptNew York. There was anincrease in dairy, fruit, andvegetable intake inArkansas and California(P < .05) but not inColorado, New York, andOhio.
Babatundeet al, 201112
110 African Americanadults aged50–93 y
To assesseffectiveness ofosteoporosiseducation programto improve calcium
RCT.Sessions of #15people; shortpresentations/lectures, hands-on
6 sessions (30–45 min/wk) for6 wk
Health BeliefModel
All Low Overall, an educationalprogram developed with atheoretical backgroundwas associated with animprovement in calcium
(continued)
JournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Murim
iet
al145
Table 1. Continued
Authors Study Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias
(QualityMeasure) Major Findings
intake, knowledge,and self-efficacy
activities, anddemonstrations tohelp participantsincrease self-efficacy.
Control: Delayed NE
intake (mean increase,556 mg dietary calcium;P < .001), knowledge(P < .001), and self-efficacy (P < .001).
Backmanet al, 201113
327 participants (156treatment; 171control), 75% ofwhom were low-income AfricanAmerican womenaged 18–54 y
To evaluateeffectiveness offruit, vegetable, andphysical activitytoolbox forcommunityeducators inchangingknowledge,attitudes, andbehavior amongwomen of low-income
Quasi-experimentaldesign withtreatment andcontrol groups.Control group didnot receive NE.
Intervention was1-h nutrition andphysicalactivityeducationclasses perweek for 6 wk
SCT All Low Women in the treatmentgroup reported significantchanges in 9 measures ofattitude, compared with 1measure in the controlgroup (P < .05).Compared with those inthe control group, womenin the treatment groupwere also more likely tomake behavioral changestomeet recommendationsfor fruit and vegetableconsumption (P < .001)and physical activity(P < .001).
Brennen andWilliams,201314
16 African Americanwomen aged 25–63 y
To evaluate effects ofculturally sensitivelifestyle interventionon blood pressureand weight
Quasi-experiment.Intervention:Counseling andeducation onincreasing physicalactivity and dietaryintake of fruits andvegetables whiledecreasing dietaryintake of salt andfat.
No control
10 individualsessions,30 min each,and 11 groupsessions,60 min each, for12 wk
No theory All Moderate Both systolic and diastolicblood pressuredecreased from amean of151/90 pre-interventionto 131/76 post-intervention. There was asignificant decrease frompre- to post-interventionsystolic (P ¼ .03) anddiastolic (P ¼ .001) bloodpressures. There was astatistically significantimprovement in self-efficacy to exercise ifbored (P ¼ .02) or if busywith other activities
146Murim
iet
alJournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
(P ¼ .008). There was astatistically significantimprovement in self-perception as valuable/worthless (P ¼ .02) andchoice as superficial/profound (P ¼ .04).
Clifford et al,200915
101 college students To determine whethera series of 4 15-min, theory-driven(SCT) cookingprograms aimed atcollege studentsliving off campusimproved cookingself-efficacy,knowledge,attitudes, andbehaviorsregarding fruit andvegetable intake
RCT. Subjects inintervention groupviewed 4 15-mincooking programsover 4 wk. Subjectsin control groupviewed 4 5-minprograms on sleepdisorders.
4 weekly 15-minepisodes
SCT Some Moderate There were significantimprovements in knowledgeof fruit and vegetablerecommendations in theintervention groupcompared with the controlgroup post¼interventionand at 4-mo follow-up(P < .05). There were nosignificant changes in fruitand vegetable motivators,barriers, self-efficacy, orintake.
Craigie et al,201116
75 adults aged $40 y To assess thefeasibility of alifestyle interventionfocusing on dietand activity,participating incardiovascularscreening
RCT. Lifestyleinterventioncomposed of 3personalizedcounselingsessions plustelephone contact.
12 wk No theoryreported
Some Low 82% successfullymaintained or lost weight(mean loss 1.1 kg, and2.6 cm waistcircumference) and 85%reported eating 5 portionsof fruits and vegetablescompared with 56% atbaseline. No behaviorchanges were detected incontrol group.
Davis et al,200917
46 individuals aged>16 y
Assessment of peer-led approach toimproving diet ofSouth Asians inSouthampton
Quasi-experiment.Intervention: 10 tastersessions and 28cookery clubsessions.
No control
Length notspecified butfollow up wasconducted after1 y
No theory All Moderate There was increased intakeof low-fat dairy productsand reduced fat and saltintake. 80% and 75%made positive changes tocooking practices andeating patterns,respectively.
Duncanet al, 201318
286 adults enrolled inEnglish as aSecond Language,aged 18–73 y
To conduct pre–postfeasibility trial ofHealthy Eating forLife, which
Pre–post design.Intervention: HealthyEating for Lifecurriculum for at
At least 2 h/wk for12 wk
SocialLearningTheories
All Low There was a significantincrease in fruit, vegetableintake, nutritionknowledge, action
(continued)
JournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Murim
iet
al147
Table 1. Continued
Authors Study Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias
(QualityMeasure) Major Findings
integrates contentabout healthynutrition todecrease cancerhealth disparities
least 2 h/wk ofclassroominstruction for12 wk.
No control group
planning, and copingplanning amongparticipants (P< .05 for all).
Endeveltet al,201119
127 older adults aged$75 y
To determine impactof intensivenutritionalinterventionprogram on healthand nutritionalstatus ofmalnourishedcommunity
Partial RCT.DIT: Intervention ledby dietitian ormedical treatment.A physician led astandard caregroup with aneducationalbooklet.
Nonrandomized‘‘untreatednutrition’’ group.
5 visits for 6 mo No theory All Low DIT group showedsignificant improvement incognitive function anddepression scorecompared with thechange in the other 2groups. DIT groupshowed a significantimprovement in intake ofcarbohydrates, protein,vitamin B6, and vitaminB1 and had a significantlylower cost of physicianvisits than did the other 2groups (P < .05 for all).
Francisand Taylor,200949
58 women aged54–83 y
To create educationmaterials usingtarget audience’spreferences andimplement heart-healthy dieteducation programdesigned usingneeds andpreferences
RCT.Intervention: 2individualregistered dietitian–led in-homeeducation sessions
Control: 2 educationmaterial mailings
3 mo SocialMarketingTheory
All Low Intervention and control mini-nutritional assessmentscores improved(P < .001). Interventionsubjects consumed morefiber than did controlsubjects (P ¼ .01) andreduced sodium intake(P ¼ .02). Controlsreduced energy (P ¼ .01)and cholesterol intakes(P¼ .03), likely because ofdecreased food intake.
Ha andCaine-Bish,201120
80 college studentsaged 18–24 y
To determine whetherthere would be anincrease in whole-grain consumptionafter students
Pre–post.Intervention: Generalnutrition class.
No control
3 times/wk for50 min forsemester
SCT Some Moderate At baseline, total mean grainconsumption was 3.07ozand mean whole-grainconsumption was 0.37 oz.After the study, mean
148Murim
iet
alJournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
completed aninteractiveintroductorynutrition coursefocusing ondisease prevention
consumption of whole-grain products significantlyincreased to 1.16 oz(P < .001) but total meangrain intake remained thesame (3.06 oz).
Hsu et al,201321
25 adults aged >18 y To examine feasibility,acceptability, andpreliminary resultsof exerciseintervention with aHealthy at EverySize orientation
RCT.Intervention: Exercisetraining and weeklybehavioralintervention.
Control: Exercise only
Project CHANGEwas 8-wkrandomized,controlled trialwith follow-up at4 wk
SDT All Low Both interventions showedlarge effect sizes onchanges in weekly energyexpenditure,moderate PA,and brisk walking. Bothinterventions showedsmall effect sizes for allfitness variables, includingbody mass index, waist–hip ratio, predictedVO2max, 1RM machinechest press, and leg press.Adherence to PA goal wasbetter for the interventiongroup at follow-up. TheSelf-DeterminationTheory based exerciseintervention with a Healthyat Every Size resulted inlarger effect sizes forchanges in keymotivational variables,including self-determination, autonomy,and goal-setting,planning and schedulingself-efficacy (P not listed).
Ireland, et al,201022
43 healthy adultsaged 20–75 y
To investigatewhether dietaryeducation enabledreduction in saltconsumption
RCT.2 different educationmethods usingeither Australia’sNational HeartFoundation Ticksymbol or FoodStandards Australiaand New Zealand’slow-salt guideline of120 mg sodium/100 g food.
8 wk No theory Some Low After 8 wk, urinary sodiumexcretion decreased from121 � 50 to106 � 47 mmol/24 h(7.3 � 3.0 to 6.4 � 2.8 gsalt/24 h) in the Tick groupand from 132 � 44 to98 � 50 mmol/24 h(7.9 � 2.6 to 6.0 � 3.0 gsalt/24 h) in the FoodStandards Australia NewZealand group (P < .05,
(continued)
JournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Murim
iet
al149
Table 1. Continued
Authors Study Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias
(QualityMeasure) Major Findings
with no between-groupdifference).
Kannan et al,201023
102 low-incomeAfrican Americanwomen aged18–45 y
To reduce nutritionrisks and enhanceprotective nutritionand analyzechanges in self-efficacy
Pre–post.Intervention: peer-lednutritioncurriculum.
No control
13 lessons taughtat 1-wk intervalsfor 13 wk
PEN-3modeland TTM
All Low 77% reported adopting atleast 1 healthy eatingbehavior (moderatingsodium or serving morefruits and vegetables totheir families), 23%adopted at least 2 suchbehaviors, and 45%adopted both dietary andbiomedical behaviors(self-monitoring bloodpressure, and exercising).
Kontogianniet al,201224
126 individuals aged45–67 y
To evaluate impact ondietary and activityhabits of non-intensive,community-basedlifestyle interventionfor type 2 diabetesprevention in high-risk Greekindividuals
Pre–post.Intervention: NE withdietician.
No control
6 bimonthlysessions for 1 y
No theory Some Moderate There was decreasedconsumption of whole-fatdairy and processed meat(P ¼ .02 and .02,respectively), sugar(P ¼ .006), and refinedcereals (P ¼ .05). Therewas improved diet,decreased body weight(P ¼ .04), plasmatriglycerides (P¼ .02), and2-h post-load plasmaglucose (P ¼ .05)compared with those whohad worsened dietaryhabits. Total time spentdaily on physical activityremained unchangedthroughout theintervention.
Kreausukonet al,201225
114 full-timeundergraduatestudents (aged 18–
To improve fruit andvegetableconsumption
RCT.Intervention groupreceivedpsychological
Not specified butfollow-up at6 wk
SCT All Low A social-cognitiveintervention to improvefruit and vegetableconsumption was
150Murim
iet
alJournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
25 y) at Chiang MaiUniversity, Thailand
program thataddressed self-efficacy andstrategic planning.
Control groupreceived handoutsabout generalnutrition guidelines
superior to a knowledge-based education sessionwith significantly greaterintention, planning, andself-efficacy for fruit andvegetable consumption(P < .05 for all). Both theintervention and controlgroups demonstrated anincrease in fruit andvegetable consumptionfrom baseline to 1 wk afterthe intervention (P < .001for both) but the increasewas greater in theintervention group(P ¼ .01).
MendoncaRde andLopes,201226
167 adults aged40–65 y
To determine effectsof healthinterventions ondietary habits andphysicalmeasurements
Quasi-experimental/pre–post
Intervention: Guidedphysical exercise,nutritionintervention,nutritionaleducation groups,individual nutritionalcare.
No control
4 sessions, 60 mineach for 7 mo
No theory All Low There was a reduction insystolic blood pressure(P ¼ .02) and use ofanimal fats (P < .01) aswell as an increase in thepercentage of individualswith a normal waistcircumference and dailyconsumption of greens/vegetables and milk/dairyproducts (P < .01)
Milliron et al,201227
153 adults aged20–65 y
To promote healthyeating behavior andweightmanagement
RCT.Intervention: 10 minface-to-face.
Control: Received noeducation on EatSmart shelf tagsposted in store.
10 min No theory Some Moderate No significant differencesbetween the 2 groups onpurchased total,saturated, or trans-fat andservings of totalvegetables. However, theintervention grouppurchased significantlymore servings (per1,000 kcal) of whole fruitand dark green/brightyellow vegetablescompared with thecontrol group.
(continued)
JournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Murim
iet
al151
Table 1. Continued
Authors Study Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias
(QualityMeasure) Major Findings
Nakade et al,201228
226 overweight/obese adults inJapan, aged40–65 y
To evaluate effects ofbehavioralapproach thatemphasizedtailored behaviorcounseling, diet,weight loss, andweightmaintenance
RCT.Intervention: 30 minindividualcounseling and20 min groupsessions abouteffective exerciseprovided byregistered dietitiansand exerciseinstructors.
Control: No NE
5 counselingsessions for 1 yand follow-up1 y afterintervention
No theory All Low The intervention group lostsignificantly more weightthan the control group(– 5.0 kg vs 0.1 kg for menand –3.9 kg vs –0.2 kg forwomen). Dietary intakeand number of walkingsteps improved in theintervention group. After1-y follow-up, theintervention groupmaintained significantlylower weight, lower energyintake, and improvementin irregular eating habits(P < .05 for all variables).
Pimentel et al,201029
67 Brazilian adultsaged 50–69 y, withimpaired glucosetolerance and atleast 1 other riskfactor for diabetesmellitus 2
To evaluateeffectiveness ofnutritionaleducation programon anthropometric,dietary, andmetabolicparameters withimpaired glucosetolerance
RCT.Intervention:individual andgroup counselingonce and twice permonth, respectivelywith team ofnutritionists
Control: No NE
36 sessions over12 mo
No theory Some Low The intervention groupshowed a significantdecline in body weight(�3.4%), body mass index(�5.7%), cholesterolintake (�49.5%), fastingglycemia (�14.0%), fastinginsulin (�9.0%),postprandial glycemia(�21.0%), postprandialinsulin (�71.0%), totalserum cholesterol�23.0%), and glycatedhemoglobin (�24.0%). Adecrease in energy intake(5%; P ¼ .06) and lowdensity lipoproteincholesterol (25%; P ¼ .07)was observed in theinterventional group,although it did not reachstatistical significance.
152Murim
iet
alJournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Plawecki andChapman-Novakofski,201330
69 adults aged55–75 y
To enhance physicalactivity andnutritionalbehaviors
RCT.Intervention: lecturesand hands-onactive learning
Control: interventionwas delayed
3 sessions(duration notmentioned) for8 wk
Health BeliefModel andTheory ofReasonedAction
Some Low Comparison of week 1 andweek 8 data indicatedsignificant improvementfor the treatment but notthe control group forcalcium, and vitamin D(P < .05). There waslimited response to theexercise outcomevariables, with many notparticipating in thatsection of the program.
Ritchie et al,201031
3,015 and 3,004before and afterNE; pregnant orpostpartumwomen/caregiversof children enrolledin WIC, aged22–36 y
To explore impact ofWIC on familybehavior regardingfruits andvegetables, wholegrains, and lower-fat milk
Pre–post cross-sectional design.Caregiversreceived educationintervention in agroup (class) or inindividual (one-on-one counseling)format.
No control
3 sessions over6 mo
TTM All Low After nutrition education,women and caregiversreported increasedrecognition of educationmessages, positivemovement in stage ofchange for target fooditems, increased familyconsumption of fruits andwhole grains, andreplacement of whole milkwith lower-fat milk.
Shahnazariet al,201332
84 US veterans aged25–80 y
To determineeffectiveness ofnutrition-basedwellness coachingusing multiplecontacts andsimple educationaltips on healtheating and weightmanagement
RCT.Intervention: 9individualized NEsessions. Coachingconsisted of 15-minsessions with final60-min session atend of 6mo. Total of3.75 h educationalcontact forintervention and 1 hfor control group bysame nutritioncoach for eachveteran throughoutstudy
Control: 1 hindividualized NEsession.
9 NE sessions for6 mo
Stage ofChangeModel
All Low Multiple coaching contactsdecreased intake ofenergy, fat, andcarbohydrates by 31%(P < .001). Weight loss of5% from baseline (92.8 to88.2 kg; P < .01) wasobserved in theintervention group withmean body mass indexdecreasing from 30.4 to28.9 (P< .05). The controlgroup showed a decreasein fat intake by 20%(P < .01) but nostatistically significantchanges in intake of othernutrients or body weight(88.7 to 87.4 kg).Veterans’ readiness tochange eating behavior
(continued)
JournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Murim
iet
al153
Table 1. Continued
Authors Study Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias
(QualityMeasure) Major Findings
for weight loss improvedwith nutrition coaching.
Silva et al,201033
239 women aged30–45 y
To analyze impact ofweightmanagementintervention ontheory-basedpsychosocialmediators, physicalactivity/exercise,and body weightand composition
RCT.Intervention: NEsessions covering:physical activity,eating/nutrition,body image, andother cognitive andbehavioral contents
Control: Generalhealth education
30 sessions, 2 heach weekly orbimonthly for 1 y
SDT All Low At 12 mo, the interventiongroup showed increasedweight loss (–7.29%,) andhigher levels of physicalactivity/exercise(þ138 � 26 min/d ofmoderate plus vigorousexercise; þ2,049 � 571steps/d) compared withcontrol subjects (P< .001).
Sorensenet al,201134
56 individuals aged22–55 y
To compare effect ofbehaviormodificationconsisting of eithera gourmet cookingcourse or NLPtherapy on weightregain
RCT.The first step was 12-wk weight lossprogram.Participantsachieving at least8% weight losswere randomizedto 5 mo of eitherNLP therapy orcourse in gourmetcooking.
8 mo No theory Some Low The NLP therapy group lost1.8 kg and the cookinggroup lost 0.2 kg duringthe 5 mo of weightmaintenance. Thedropout rate was lowerduring the active cookingtreatment compared withthe NLP group. There wasno difference in weightmaintenance after 2 and3 y of follow-up.
Wielandet al,201235
34 women (Hispanic,Somali, andCambodian) aged22–68 y
To evaluate socio-culturallyappropriatephysical activityand nutritionintervention incommunity-basedparticipatoryresearch approach
Pre–post.Intervention. 6-wkprogram with 2 90-min classes perweek.
No control
6 wk No theory Some Moderate After the intervention,participants were morelikely to exercise regularly(P < .001). They reportedhigher health-relatedquality of life (P < .001).Self-efficacy for diet andexercise; weight loss,waist circumference, andblood pressure were notsignificantly different afterthe intervention.
DIT indicates Dietetic Intensive Treatment; ESBA, Eating Smart Being Active; NE, nutrition education; NLP, Neurolinguistic Programming; PA, Physical Activity; RCT, ran-domized control trials; SCM, Stage of Change Model; TRA, SCT, Social Cognitive Theory; SDT, Self-determination Theory; WIC, Special Supplemental Nutrition Programfor Women, Infants, and Children.
154Murim
iet
alJournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Table 2. Multicomponent Nutrition Education Intervention Studies (n ¼ 13)
AuthorsStudy
Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias (QualityMeasure) Major Findings
Baruth andWilcox,201336
360 AfricanAmericans aged$18 y
To examine extent towhich participantsin combinedphysical activityand dietaryinterventionachieved changesin multiple healthbehaviors
RCT.Intervention: PA and
healthy eatingbefore, during andafter churchservices
Control: delayedintervention
15 mo SEM All Low Up to 19% indicated nochange in healthbehavior, 31%changed 1 healthbehavior, 31%changed 2 healthbehaviors, 13%changed 3 healthbehaviors, and 5%changed all 4 targetedhealth behaviors.Combinations ofmultiple behaviorchange included PAand dietary behaviors,which suggests thatboth behaviors can bechangedsimultaneously.
Cullen et al,200937
1,004 TexasEFNEP clients in100 classes,mean age 35 y
To evaluate modifiedcurriculum for 6-session TexasEFNEP promotinghealthful homefood environmentsand parenting skillsrelated to obesityprevention
RCT.Intervention: 6 short
videos with goalsetting, problemsolving, guideddiscussion, andhandouts. Then aweekly goal sheetfor recording wasissued Participantsmonitored daily goalattainment andreturn goal sheetthe following week
Control: TraditionalEFNEP classincluded briefdiscussion and foodpreparation
6 classesfor 6 wk
No theory Some Low There was a significantBMI decrease atpostinterventioncompared withbaseline only for theintervention group.This change was notmaintained atfollow-up.
(continued)
JournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Murim
iet
al155
Table 2. Continued
AuthorsStudy
Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias (QualityMeasure) Major Findings
Dirige et al,201338
673 Filipino-American adultsaged $18
To evaluate 18-monutrition andphysical activityintervention (activelife) conductedthrough culturallyspecificorganizations
RCT.Intervention:
Workshops andactivities (cookingdemonstrations,recipe contests,supermarket tours,group aerobicclasses, gardening)
Control: Cancerscreening,alternativemedicine, andstress management
18 mo TTM Some Low Intervention participantsshowed significantincreases in PA(P< .05), adoption of alow-fat diet (P < .05),and stage of changefor fruits andvegetables (P < .05),dietary fat intake(P < .01), and PA(P < .01). Interventiondid not lead toincreases in number ofparticipants eating $5servings/d of fruits andvegetables.
French et al,201039
160 metropolitantransit workersaged 20–79 y
To describe andreport results fromworksite obesitypreventionintervention thattargeted transitemployees
Group RCT (4garages)
Intervention:Enhancement ofPA facilities,increasedavailability of andlower prices forhealthy vendingmachinechoices, etc
Control: Nointervention
18 mo worksiteintervention
No theory Some Low Energy intake decreasedsignificantly and fruitand vegetable intakeincreased significantlyin intervention garagescompared with controlgarages. However,BMI and PA changeswere not significant.
Iriyama andMurayama,201440
57 male workers inJapan
To evaluate effects ofnew worksiteweight-controlprogram usingcombination ofnutrition educationenvironmentalinterventions
RCT crossover.Intervention:
Intervention groupreceived 6-moprogram consistingof nutritioneducation andprovision of healthycafeteria meals andnutritionalinformation. 1-y
6 mo TTM andPrecede-Proceedmodel
All Low Mean BMI wassignificantly reducedfrom baseline value of25.6 kg/m2 to 25.3 kg/m2 at month 6 and to24.8 kg/m2 at year 1(P ¼ .008) and wassignificantly lower atyear 1 than at baselineand month 6 inmultiple comparison
156Murim
iet
alJournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
follow-upControl received
same interventionas counterparts incrossover designfrom 6 mo afterstudy entry
tests in interventiongroup; BMI increasedover time in controlgroup (P < .001).
Johansen et al,201041
198 adults aged25–63 y
To present effect ofintervention studyintentions tochange dietarybehavior andchanges made indietary intake
RCT.Intervention:
Combination ofgroup sessions,individualcounseling, andorganized exercisegroups
Control: Generaladvice participantswould receive fromgeneral practitioner
6 sessions,2 h each,for 7 mo
TTM Some Low Differences betweenintervention and controlafter intervention weresignificant for sugar-richdrinks and rapeseed oil(P < .05). Intention toreduce dietary intake offat, sugar, and whiteflour, and to changetype of fat and increaseintake of vegetables andlegumes shifted forintervention group(P < .05) from pre-action (pre-contemplation,contemplation, andpreparation) stages toaction stage inintervention group butnot in control group.Difference betweengroups at follow-up wassignificant (P < .05). Nosignificant differenceswere found for intentionto increase fruit intake.
Linde et al,201242
1,672 participantsaged 18–75 y
To influence weightgain positivelyamong employeesover 2 y
RCT (worksite level).Six worksites in USmetropolitan areawere recruited andrandomized in pairsat worksite level to2-y intervention orno-contact control
Intervention: Postersat worksites, link toWeb sites with
2 y, frequencyvariedwith center
No theoryreported
None Moderate No differences betweensites in key outcome ofweight change over 2-y study period.
(continued)
JournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Murim
iet
al157
Table 2. Continued
AuthorsStudy
Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias (QualityMeasure) Major Findings
useful informationon food availabilityand price, PApromotion, scaleaccess, and mediaenhancements
Control: No contact
Racette et al,200943
123 participantsaged 36–54 ywith BMI32.9 � 8.8 kg/m2, employed at1 of 2 selectedworksites withina large medicalcenter
To evaluateeffectiveness ofworksite healthpromotion programon improvingcardiovasculardisease risk factors
Cohort randomizedtrials. Interventionincludedpedometers,healthy snack cart,Weight Watchersmeetings, groupexercise classes,seminars, teamcompetitions, andparticipationrewards
Control: Personalhealth reportscontainingassessment results
Weeklyfor 1 y
TTM Some Low Improvements (P < .05)were observed at bothworksites for fitness,blood pressure, andtotal, high-densitylipoprotein, and LDL-C.Additionalimprovementsoccurred in interventiongroup in BMI, fat mass,Framingham risk score,and prevalence ofmetabolic syndrome;only changes in BMIand fat mass weredifferent betweencontrol and interventionworksites.
Rustad andSmith,201344
118 ethnicallydiverse, low-income womenaged 23–45 y
To assess impact ofshort-term nutritionintervention usingeducation oncomprehensivearray of nutritionand health topics
Pre–post intervention.Experiential andinteractive lectures,activities, anddemonstrations.Educationalsessions onshopping andbudgeting, healthycooking, improvingfood security bygrowing foods.
No control
3 sessionsin 6 wk.
Each classlasted75–90 min
No theory All Moderate Postinterventionincreased nutritionknowledge andfavorable nutritionbehavior. Responsesto 7 of 11 questions inknowledge and 9 of 11variable set changedsignificantly atpostintervention(P < .01 for both).Women alsodecreasedconsumption of fast
158Murim
iet
alJournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
foods and processedsnacks high in sugar,salt, and fat, and fattycuts of meat, anddecreased addition ofsugar, salt, and butterto foods (P < .01).
Sarrafzadeganet al, 201345
12,514participated atbaseline and9,570 in post-interventionsurvey, aged24–54 y
To evaluatecomponent ofHealthy HeartPrograms toassess feasibilityand outcomes ofprogram on lifestylebehaviors and riskfactors for chronicnon-communicablediseases
RCT.Intervention: Public
education throughmass media,intersectoralcooperation andcollaboration,professionaleducation andinvolvement,marketing,organizationaldevelopment,legislation andpolicy development,as well as researchand evaluation
Duration ofinterventionactivitiesvaried: 3–4 y.
Precede–Proceedmodel, SocialLearningTheory, andinnovationdiffusionapproach
Some Moderate Prevalence of abdominalobesity, hypertension,hypercholesterolemia,hypertriglyceridemia,and high LDL-Cdecreased significantlyin intervention area vscontrol area in bothsexes. However,reduction inoverweight andobesity was significantonly in females(P < .05 for all). Therewere no significantchanges in prevalenceof diabetes mellitus.
Savoie et al,201546
203 participantsaged $18 y
To determine whetherparticipation inselectedSupplementalNutrition AssistanceProgram–Education
lessons had animpact on intent toimprove nutrition-related behaviors ofparticipants
Retrospective post-then-pre design
Intervention: lecture,cookingdemonstration,sample tasting offood prepared inclass, handout.
No control
Not indicated TPB All Low Mean responses ofindividual questionsand mean lessonscores increasedsignificantly frompretest to posttest inthe menu planning,shopping lesson, andthe My Plate lesson(P < .001).
Wilcox et al,201347
74 AfricanMethodistEpiscopalchurches and1,257 adultmembers withinthem
To report results ofinterventiontargeting physicalactivity and healthyeating
RCT.Intervention: Churches
implemented;sharing messagesfrom pulpit; passingout educationalmaterials (provided),create Faith,Activity, and
15 mo SEM Some Low There was a significanteffect favoringintervention group inself-reported leisuretime (d ¼ 0.18;P ¼ .02). No groupdifferences were foundfor self-reported fruitand vegetable
(continued)
JournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Murim
iet
al159
Table 2. Continued
AuthorsStudy
Population Study ObjectiveStudy Design/Intervention
Length andFrequency
BehavioralTheory/
ConstructAchievedObjectives
Risk ofBias (QualityMeasure) Major Findings
Nutrition ProgramControl: delayedinterventioncomponents at endof 15 mo
consumption,measured bloodpressure, and self-reported fat- and fiber-related behaviors.
Lowe et al,201048
96 adultparticipants(BMI29.7 � 6.0 kg/m2) hospital oruniversityemployees aged21–65y
To evaluate nutritionaland weightchanges inprogram that usedworksite cafeteriasto reduceemployees’ caloriecontent ofpurchased foodsand improve theirmacronutrientintake
RCT.3 mo of baseline data
collection, then 3-mo intervention; 6-and 12-mopostinterventionfollow-ups.Participants wererandomly assignedto 1 of 2interventiongroups:EnvironmentalChange orEnvironmentalChange PlusEnergy DensityEducation andIncentives.Randomization ofparticipantsoccurred withineach worksite
3 mo No theory Some Low There was no differencebetween groups in totalenergy intake overstudy period.
Across groups, energyand percentage ofenergy from fatdecreased and percentof energy fromcarbohydratesincreased from baselineto interventionperiod (allP< .01). Follow-upanalyses, conducted byaveraging baselinemonths 1 and 2 andcomparing them withintervention month 3 asa conservative estimateof overall impact ofintervention, indicatedthat change in energy,carbohydrate, and fatintake remainedsignificant (P< .001).
Providing nutrition labelsand reducing energy-density of selected foodswas associated withimproved dietary intake.
BMI indicates bodymass index; EFNEP, Expanded Food and Nutrition Education Program; LDL-C, low-density lipoprotein cholesterol; PA, physical activity; RCT, random-ized control trials; SEM, Structural Ecological Model; TPB, Theory of Planned Behavior; TTM, Trans-theoretical Model.
160Murim
iet
alJournalofNutritio
nEducatio
nandBehavior�
Volume49,
Number
2,2017
Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Murimi et al 161
lasted longer than 5 months (n ¼ 18)met their primary objectives.17,19,26,28,29,31,32,36,40,43,50 For example, a studywhose objective was to evaluate theeffectiveness of a nutritional educationprogram that involved bimonthlygroup discussions for 12 months, andthat included written and oral didacticinstructions on anthropometric, dietary,and metabolic parameters with impa-ired glucose tolerance, reported adecrease in 2 risk factors related todiabetes mellitus.29
Similarly, a study that took 1 yearand whose objective was to study theimpact of a weight management inter-vention on theory-based psychosocialmediators reported weight loss (–7.29%,)and increased levels of physical activ-ity/exercise (þ138� 26min/d of mod-erateplusvigorousexercise;þ2,049�571steps/d) compared with control sub-jects (P < .001).33 Finally, a study thatinvolved a physical activity and healthyeating intervention over 15 monthsreported significant results and an 18%effect on the intervention group inself-reported leisure time (P ¼ .02;d ¼ 0.18).47
On the contrary, this review foundthat nutrition education interventionsthat lasted for a short duration wereless likely to meet their stated objec-tives. A relevant example here is astudy whose purpose was to promotehealthy eating behavior and weightmanagement.27 In this study, inter-vention involved 10 minutes of face-to-face education on the printedEatSmart shelf tags posted in the store.The control group received no educa-tion about the EatSmart shelf tagsposted in the store. Outcomemeasuresincluded purchases of total saturatedand trans fat (grams per 1,000 kcal),fruit, vegetables, and dark green or yel-low vegetables (servings per 1,000 kcal)derived through a nutritional analysisof participants' shopping baskets. Re-sults showed no significant differencesbetween the control and interventiongroups on total purchases as well asin the choice of health eating groceryproducts except in fruit and dark greenor yellow vegetables, thus indicatingminimal attainment of the study ob-jectives.
Similarly, a nutrition educationintervention on physical activity andnutrition35 revealed that self-efficacyfor diet and exercise, weight loss,
waist circumference, and blood pres-sure were not significantly differentfrom baseline after 2-hour classeseach week for 6 weeks.
Effect of Number of StudyObjectives/Focus
In addition to the long duration, thissystematic review found that studieswith few or focused objectives weremore successful in meeting all of theirstated objectives than were interven-tions that had several unrelated objec-tives. For example, a 4-month nutritioneducation intervention whose objec-tive was to increase whole-grain con-sumption among students who completedan interactive introductory nutritioncourse focusing on disease preven-tion20 indicated that student partici-pants increasedwhole-grain consumptionfrom 0.37 to 1.16 oz (P < .001). Simi-larly, another study22 with only 1objective, to investigate whether die-tary education enabled a reduction insalt consumption, indicated that after8 weeks of dietary education interven-tion there was a reduction in salt con-sumption andurinary sodiumexcretion.
In contrast, interventions with >3unrelated objectives were not success-ful in meeting all of their objectives.For instance, a study that had >3 ob-jectives, conducted in 3 phases withdifferent feeding regimes, reportedinconsistent results at the 3 phases.34
In phase 1, 88% of participants (n ¼49) completed 12 weeks of calorie re-striction and achieved 8% weightloss. Phase 2 involved 5 months ofweight maintenance; participantswere divided into 2 groups each withdifferent feeding regimes and hencedifferent objectives. The 2 groupsexperienced different numbers ofparticipant dropout, which affectedthe final results. There was no differ-ence in weight maintenance after 2and 3 years of follow-up. In general,it was observed that follow-up studiesdid not yield many results. A lot offollow-up studies did not yield asignificant change from the initial re-sults. Therefore, it can be impliedthat a majority of researchers couldhave paid less attention to thefollow-ups. The researchers in this re-view recommend stringent measuresin follow-up studies, just as in the
initial phase of the study, to enhancereliability of the results of follow-ups.
Lack of Fidelity in Delivery
Fidelity in intervention ensures thatall intervention activities are executedas planned in the methods. Therewere few reported cases of lack ofintervention fidelity that could havecompromised the findings. A peer-led nutrition education interventionthat addressed maternal and infanthealth through dietary patterns re-ported that some facilitators neglectedto follow the complete lesson plans byomitting parts of a lesson or failing touse the facilitator's guide, or they didnot show enthusiasm in promotingthe desired behavior among the peer-led cohorts. Although the interven-tion was the same by design andcontent information, results of thestudy varied among cohorts. As aresult, the success of the interventionswas affected by the human factors ofthe presenter.23
Theory-Based Studies
Slightly over half of the studies(57.5%; n ¼ 23) reported being theorybased and used at least 1 theory. Themost common theories used to designand implement nutrition educationinterventions in studies selected inthis review were the Trans-theoreticalModel and Social Cognitive Theory.The majority of the theory-basedstudies (61%; n ¼ 14 of 23) weresuccessful in achieving their stated ob-jectives, whereas the remainingtheory-based studies (39%; n ¼ 9)achieved some but not all of their pri-mary objectives.
This review considered studies thatprovided information about how theyused theories in the design of thestudy as a best practice, rather thanjust mentioning the theory casuallyin the introduction or methods: forinstance, a study conducted by Savoieet al46 to determine whether partici-pation in a selected SupplementalNutrition Assistance Program–Education(SNAP-Ed) that clearly showed howthe constructs of the Theory ofPlanned Behavior used in the designand implementation of the study les-sons showed an impact on the intent
162 Murimi et al Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017
to improve nutrition-related behav-iors of participants as stated in thestudy objective. Results of this studyshowed that posttest scores weresignificantly higher than pretest scoresrelated to menu planning, shopping les-sons, and My Plate lessons (P < .001).
However, this review observed thatalthough some studies indicated thatthey were theory based, they failed todescribe explicitlyhowthe theories guidedthe studies. For example, some re-searchers15 reported using Social Cogni-tive Theory; others45 reported usingseveral theories including thePrecede–Proceed Model, Social Learning Theory,and the InnovationDiffusionApproach.However, the theory constructs andhow they were used or measured werenot described in either article.
Nevertheless, although 45% ofstudies (n ¼ 18) were not informedby a theory, they were equally as suc-cessful as those thatwere. For instance,a study to describe and report the re-sults from a worksite obesity preven-tion intervention that targeted transitemployees was not guided by a theory,and yet it indicated success.39 The re-sults indicated that energy intakedecreased significantly and fruit andvegetable intake increased significantlyin intervention garages comparedwith control garages. A summary ofthe studies that used theories andthose that did not is provided inTables 1 and 2.
Environmental Interventions atthe Worksite
Worksite environmental interventionshave an integral part inmodifying die-tary and weight management behav-iors when executed appropriately. Forexample, a study to evaluate the effectsof a new worksite weight control pro-gram using nutrition education envi-ronmental interventions among maleadults in Japan reported significant re-sults. At the 1-year follow-up, the inter-vention grouphad significantly greaterreductions in body weight, body massindex, and alanine aminotransferasethan the control group did (P ¼ .02,.02, and .86, respectively).40
For worksite environmental inter-ventions to be successful, sufficientappropriate changes must be imple-mented at the right places. This re-view discovered that some worksite
environmental interventions did notensure changes sufficient for the inter-vention, which affected the results.Such a failure occurred when theworksite management and collabora-tors resisted making sufficient envi-ronmental changes to modify dietaryand exercise behavior in employeesat work despite promoting thebehavior.39,42 For example, a worksiteenvironmental intervention studyused posters at worksites, links toWeb sites with useful informationabout food availability and prices,physical activity promotion, accessto scales, and media enhancementsto promote weight gain control.42
Intervention components were foodselection, promotion of walking andstair use, weight self-monitoring, andhealth information at work. However,the provision of a variety of healthyoptions and extra time for exercisewere not offered to employees. The re-sults of this study indicated that therewere no differences between the inter-vention and control sites in the keyoutcome of weight change over the2-year study.
DISCUSSION
The purpose of this review was to sys-tematically examine factors thatcontribute to the efficacy of nutritioneducation interventions in promotingbehavior change for good health andwell-being based on their stated objec-tive. The main findings of this reviewindicated that the efficacy of nutritioneducation interventions depends on theduration of the intervention, having fewfocused objectives, the appropriateuse of theories, fidelity in interventions,and support from policy makers andmanagement for the environmentalinterventions. These findings are largelycongruent with the results of anotherreview conducted by Baird et al.4
Therefore, factors that were identi-fied as determinants of efficacy andwhich form the discussion of thisstudy are: (1) the types and use of de-signs, (2) the type of intervention thatcharacterized the studies, (3) the dura-tion and dosage of the interventions,(4) the number of objectives in astudy, (5) fidelity in intervention,and (6) the use of theories in nutritioneducation interventions. Worksiteenvironmental interventions are also
featured in this discussion. Criticalanalysis of these factors provides theoutcome of the current review.
The randomized control trial (RCT)designhas a reputation of being robust.This design is therefore appropriate forbaseline studies intended to inform alarger intervention project.51 This re-view indicated that a majority (70%;n ¼ 28) of the nutrition education in-terventions used an RCT design. TheRCT design is robust andmay be attrib-uted to the success of the interventionsin achieving their stated objectives.
Interventions that lasted for>5 months reported a higher level ofsuccess. They were mainly multiple-component interventions. Thisfindingsupported the results of a previous re-view that reported that remarkablymore studies on nutrition educationwith long-term follow-up were associ-ated with success.51 Another studynoted that behavior change takes timeand practice.52 Therefore, it may beargued that the length of time takenfor intervention and the frequency ofexposure are important factors for thesuccess of a nutrition education inter-vention. However, another studynoted that interventions with long du-rations are associatedwith ahigher costof implementation and participants'attrition, which constrained some in-terventions.24
This systematic review revealedthat studies with few and succinct ob-jectives were more successful thanwere those with numerous and attimes unrelated objectives. Other re-views acknowledged the effectivenessof few objectives in nutrition educa-tion interventions.51 The current re-view noted that studies with #3objectives that were related were suc-cessful even when the duration ofintervention was <6 months.
It is important to report fidelity ininterventions because it allows readersand other researchers to judge thequality of the intervention and howvarious factors may have influencedthe outcome.53 Lack of fidelity in thedelivery of a program has a counter-productive effect on the results of anintervention. Fidelity in interventionis a critical element that is rarely re-ported in many studies.54 This reviewfound that the few studies that re-ported fidelity discovered that it nega-tively affected the results: Some sitesachieved their objectives whereas
Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Murimi et al 163
others failed to achieve their objec-tives despite a similar program.
However, whereas peer educatorsmight have experienced some profes-sional challenges in the delivery ofeducational interventions,other studiesnoted that they could provide a goodform of social support associated withsuccessful behavior change.4 It is there-fore important for the training of peer-led interventions to emphasize fidelityin the implementation in an effort toachieve desired results.
This systematic review revealed thatalthough the majority of the studiesthat were theory based were successfulin achieving all of their primary objec-tives (53.8%), a good number (45%) ofwell-designed, non-theory interventionswere equally successful in achievingtheir primary objectives. The currentfindings support a review by Baba-tunde et al,12 who noted that overall,an educational program developed witha theoretical background was associ-ated with an improvement in calciumintake (mean increase, 556 mg dietarycalcium;P< .001), knowledge (P< .001),and self-efficacy (P < .001). This indi-cates that well-used theory is likely tomake an intervention successful. Thisreview observed that some of thestudies that claimed to use a theoryfailed to describe explicitly how thetheory was used in the study. Thisfinding agrees with the results of a re-view on experimentally based evi-dence of the theoretical mechanismsof dietary behavior change,55 whichconcluded that future intervention tri-als need to focus on identifying effec-tive procedures for mediator changeand adopting a more rigorous and sys-tematic approach to theory testing.
Finally, in worksite environmentalinterventions, interventions withoutappropriate support from collabora-tors to support the desired behaviorwere less likely to meet their objec-tives. For instance, an interventiondid not meet the primary objective,which was to affect weight gain posi-tively over 2 years, owing to weakand inconsistent implementation ofenvironmental changes in the work-place.42 In contrast, an interventionby Iriyama and Murayama40 on work-site weight control was able to imple-ment changes in the workplacecafeteria and introduce healthy me-nus. The results of this interventionindicated a significant decrease in
mean body mass index from a base-line value of 25.6 kg/m2 to 25.3 and24.8 kg/m2 at 6 months and 1 year,respectively (P < .05). These worksiteenvironmental interventions high-light the need for effective collabora-tion among nutritionists, policy makersand stakeholders within institutionsincluding schools and workplaceswhere the food environment has abig role in the success of a nutritioneducation intervention. These find-ings are in agreement with results ofanother study regarding the effectsof environmental, policy, and socialmarketing interventions on physicalactivity and fat intake ofmiddle schoolstudents. The study noted that envi-ronmental and policy interventionswere effective in increasing physicalactivity at school. Appropriate changesare necessary for the success of work-site environmental interventions.48,56
This review has limitations. First,only articles that were published in En-glish were considered. Therefore, thereis a possibility that some recent andimportant findings published in lan-guages other than English were leftout. Another limitation is that therewas a potential that studies that didnotfindasignificanteffect in their inter-vention were not published, and there-fore were not included in this review.
Finally, the review was limited byarticles that did not include adequateinformation in their methods and re-sults. This posed a challenge to eval-uate the contributions of specificcomponents of nutrition educationinterventions and their effectivenessproperly, including the use of theoryand the dosage of intervention (fre-quency and duration). Despite thelimitations, the current review drewits strength from the fact that the re-searchers investigated several factorsthat led to the success of various typesof interventions. This was a departurefrom previous reviews that concen-trated primarily on a single type ofintervention and the related outcome.
IMPLICATIONS FORRESEARCH ANDPRACTICE
The results of this review suggest thatnutrition education interventions withlonger duration have few and focusedobjectives, and that those that are guided
by a theory have a higher chance ofachieving their purpose. Lack of fidel-ity in peer-led interventions, lack ofbehavior support for environmentalinterventions, and too-short dura-tions were factors that contributed toa lack of success in some of the inter-ventions. It was observed that varioushuman factors affected the effective-ness of an intervention in cases wherepeers and paraprofessionals imple-mented the nutrition interventions.Although the lack of fidelity duringinterventions was not rampant, thefew reported incidences had profoundramifications on the results. The studieswith control groups had better inter-pretation of the results, which enhancedthe validity of the outcome. This im-plies that studies that used RCTs hada better chance of replicability, followedby those that employed a quasi-experimental design. The eligible studiesthat were reviewed and whose inter-ventions were considered successfulwere considered largely sustainableand could easily be replicated.
The researchers concluded that theuse of theories in designing nutri-tional education interventions was acommon practice in which 55% ofthe analyzed studies (n ¼ 22) wereinformed by at least 1 theory. Manystudies that used theories indicatedsuccess in achieving their objectives.However, a lack of details regardinghow the behavior theories guidedthe studies made it difficult to assessthe effect of the theories mentionedin some studies. The researchersconcluded that the use of theories isa good practice in interventions andthat worksite environmental inter-ventions provide an important oppor-tunity for behavioral adjustments forbetter health, but that they need thecooperation of the policy makers.
The results of this study suggest thatmore focused, clearly defined, measur-able objectives are associated withbehavior change, whereas the moreambitious use of many objectivesmay limit the effectiveness of nutritioneducation by taking away from themain message and confusing partici-pants. The objective should have aclear targeted behavior, followed byadequate dosage or exposure to facili-tate the desired behavior change. Apurposeful selection of behavior the-ory that will guide the interventionbased on the desired behavior change.
164 Murimi et al Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017
A clear use of the theory in designingand implementing the interventionshould be reflected in the results. Forstudies that use multiple sites withdifferent implementers, ensuring fidel-ity is critical to the success of the inter-vention. Training should emphasizethe important message and targetedbehavior. Researchers should alsoconsider the Guide for Effective Nutri-tion Interventions and Education as achecklist to help design or improvetheir research methods for effective in-terventions.57
For worksite and other environ-mental interventions, it is importantfor policy makers to make healthychoices the easy ones by allowingtime for the exercise. This can be im-plemented in many ways dependingon the worksite environment: forexample, including markings indi-cating the distance covered along apath between 2 points, such as build-ings, or by providing healthy alterna-tives or choices at the worksite.
REFERENCES
1. Contento IR, Balch GI, Bronner YL,et al. The effectiveness of nutrition edu-cation and implications for nutritioneducation policy, programs, andresearch: a review of research. J NutrEduc. 1995;27:312-328.
2. Ammerman AS, Lindquist CH,Lohr KN, Hersey J. The efficacy ofbehavioral interventions to modify die-tary fat and fruit and vegetable intake: areview of the evidence. Prev Med. 2002;35:25-41.
3. Pomerleau J, LockK,Knai C,McKeeM.Interventions designed to increase adultfruit and vegetable intake can be effec-tive: a systematic review of the literature.J Nutr. 2005;135:2486-2495.
4. Baird J, Cooper C, Margetts BM,BarkerM, Inskip HM. Changing healthbehaviour of young women fromdisadvantaged backgrounds: evidencefrom systematic reviews. Proc NutrSoc. 2009;68:195-204.
5. Guyatt G, Drummond R, Meade M,Cook D. The Evidence Based-MedicineWorking Group Users’ Guides to the Med-ical Literature. 2nd ed. Chicago, IL:McGraw Hill; 2008.
6. Liberati A, Altman DG, Tetzlaff J, et al.The PRISMA statement for reportingsystematic reviews and meta-analysesof studies that evaluate health care inter-
ventions: explanation and elaboration. JClin Epidemiol. 2009;62:1-34.
7. Moher D, Tetzlaff J, Tricco AC,SampsonM,AltmanDG. Epidemiologyand reporting characteristics of system-atic reviews. PLoS Med. 2007;4:78.
8. Higgins JPT, Altman DG, Sterne JAC.Assessing risk of bias in included studies.In: Higgins JPT, Green S, eds. CochraneHandbook for Systematic Reviews of Inter-ventions. Version 5.1.0. http://www.cochrane-handbook.org. AccessedOctober 1, 2016.
9. GRADE Working Group. Gradingquality of evidence and strength of rec-ommendations. BMJ. 2004;328:1490.
10. Arrebola E, Gomez-Candela C,Fernandez-Fernandez C, Loria V,Munoz-Perez E, Bermejo LM. Evalua-tion of a lifestyle modification programfor treatment of overweight and non-morbid obesity in primary healthcareand its influence onhealth-relatedqualityof life.Nutr Clin Pract. 2011;26:316-321.
11. Auld GS, Baker L, Conway J,Dollahite MC, Lambea MC, McGirr K.Outcome effectiveness of the widelyadopted EFNEP curriculum EatingSmart-Being Active. J Nutr Educ Behav.2015;47:19-27.
12. Babatunde OT, Himburg SP,Newman FL, Campa A, Dixon Z. The-ory-driven intervention improves cal-cium intake, osteoporosis knowledge,and self-efficacy in community-dwelling older Black adults. J NutrEduc Behav. 2011;43:434-440.
13. BackmanD,ScruggsAA,AtieduS, et al.Using a toolbox of tailored educationallessons to improve fruit, vegetable, andphysical activity behaviors among Afri-can American women in California. JNutr Educ Behav. 2011;43:S75-S85.
14. Brennen M, Williams CL. Lifestylemanagement of cardiovascular risk fac-tors in African American women.ABNF J. 2013;24:92-97.
15. Clifford D, Anderson J, Auld G,Champ J. Good Grubbin’: impact of aTV cooking show for college studentsliving off campus. J Nutr Educ Behav.2009;41:194-200.
16. Craigie AM, Barton KL, Macleod M,et al. A feasibility study of a personal-ised lifestyle programme (HealthForce)for individuals who have participated incardiovascular risk screening. Prev Med.2011;52:387-389.
17. Davies JA, Damani P, Margetts BM.Intervening to change the diets oflow-income women. Proc Nutr Soc.2009;68:210-215.
18. Duncan LR, Martinez JL, Rivers SE,et al. Healthy eating for life English asa second language curriculum: primaryoutcomes from a nutrition educationintervention targeting cancer risk reduc-tion. J Health Psychol. 2013;18:950-961.
19. Endevelt R, Lemberger J, Bregman J,et al. Intensive dietary intervention bya dietitian as a case manager amongcommunity dwelling older adults: theEDIT study. J Nutr Health Aging.2011;15:624-630.
20. Ha EJ, Caine-Bish N. Interactive intro-ductory nutrition course focusing ondisease prevention increased whole-grain consumption by college students.J Nutr Educ Behav. 2011;43:263-267.
21. Hsu YT, Buckworth J, Focht BC,O’Connell AA. Feasibility of a self-determination theory-based exerciseintervention promoting healthy atevery size with sedentary overweightwomen: Project CHANGE. PsycholSport Exerc. 2013;14:283-292.
22. Ireland DM, Clifton PM, Keogh JB.Achieving the salt intake target of 6 g/day in the current food supply in free-living adults using two dietary educa-tion strategies. J Am Diet Assoc. 2010;110:763-767.
23. Kannan S, Sparks AV, Webster JD,Krishnakumar A, Lumeng J. HealthyEating and Harambee: curriculumdevelopment for a culturally-centeredbio-medically oriented nutrition educa-tion program to reach African Amer-ican women of childbearing age.Matern Child Health J. 2010;14:535-547.
24. Kontogianni MD, Liatis S,Grammatikou S, Perrea D,Katsilambros N, Makrilakis K.Changes in dietary habits and their asso-ciation with metabolic markers after anon-intensive, community-based life-style intervention to prevent type 2 dia-betes, in Greece. The DEPLAN study.Diabetes Res Clin Pract. 2012;95:207-214.
25. Kreausukon PP, Gellert S, Lippke S,Schwarzer R. Planning and self-efficacy can increase fruit and vegetableconsumption: a randomized controlledtrial. J Behav Med. 2012;35:443-451.
26. Mendonca Rde D, Lopes AC. The ef-fects of health interventions on dietaryhabits and physical measurements. RevEsc Enferm USP. 2012;46:573-579.
27. Milliron BJ, Woolf K, Appelhans BM.A point-of-purchase interventionfeaturing in-person supermarket educa-tion affects healthful food purchases. JNutr Educ Behav. 2012;44:225-232.
Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Murimi et al 165
28. Nakade M, Aiba N, Suda N, et al.Behavioral change during weight lossprogram and one-year follow-up:Saku Control Obesity Program(SCOP) in Japan. Asia Pac J Clin Nutr.2012;21:22-34.
29. Pimentel GD, Portero-McLellan KC,Oliveira EP, et al. Long-term nutritioneducation reduces several risk factorsfor type 2 diabetes mellitus in Brazilianswith impaired glucose tolerance. NutrRes. 2010;30:186-190.
30. Plawecki K, Chapman-Novakofski K.Effectiveness of community interven-tion in improving bone health behav-iors in older adults. J Nutr GerontolGeriatr. 2013;32:145-160.
31. Ritchie LD, Whaley SE, Spector P,Gomez J, Crawford PB. Favorableimpact of nutrition education on Cali-fornia WIC families. J Nutr Educ Behav.2010;42:2-10.
32. Shahnazari M, Ceresa C, Foley S,Fong A, Zidaru E, Moody S. Nutri-tion-focused wellness coaching pro-motes a reduction in body weight inoverweight US veterans. J Acad NutrDiet. 2013;113:928-935.
33. Silva MN, Vieira PN, Coutinho SR,et al. Using self-determination theoryto promote physical activity andweightcontrol: a randomized controlled trialin women. J Behav Med. 2010;33:110-122.
34. Sorensen LB, Greve T, Kreutzer M,et al. Weight maintenance throughbehaviour modification with a cookingcourse or neurolinguistic program-ming. Can J Diet Pract Res. 2011;72:181-185.
35. Wieland ML, Weis JA, Palmer T, et al.Physical activity and nutrition amongimmigrant and refugee women: acommunity-based participatoryresearch approach. Womens Health Is-sues. 2012;22:225-232.
36. BaruthM,Wilcox S.Multiple behaviorchange among church members takingpart in the faith, activity, and nutritionprogram. J Nutr Educ Behav. 2013;45:428-434.
37. Cullen KW, Lara Smalling A,Thompson D, Watson KB, Reed D,Konzelmann K. Creating healthfulhome food environments: results of astudy with participants in the expandedfood and nutrition education program.J Nutr Educ Behav. 2009;41:380-388.
38. Dirige OV, Carlson JA, Alcaraz J, et al.Siglang Buhay: nutrition and physicalactivity promotion in Filipino-Americans through community orga-nizations. J Public Health Manag Pract.2013;19:162-168.
39. French SA, Harnack LJ, Hannan PJ,Mitchell NR, Gerlach AF, Toomey TL.Worksite environment intervention toprevent obesity among metropolitantransit workers. Prev Med. 2010;50:180-185.
40. Iriyama Y, Murayama N. Effects of aworksite weight-control programmein obese male workers: a randomizedcontrolled crossover trial. Health EducJ. 2014;73:247-261.
41. Johansen KS, Bjorge B, Hjellset VT,Holmboe-Ottesen G, Raberg M,Wandel M. Changes in food habitsand motivation for healthy eatingamong Pakistani women living in Nor-way: results from the InnvaDiab-DEPLAN study. Public Health Nutr.2010;13:858-867.
42. Linde JA, Nygaard KE, MacLehose RF,et al. HealthWorks: results of a multi-component group-randomizedworksiteenvironmental intervention trial forweight gain prevention. Int J BehavNutr Phys Act. 2012;9:14.
43. Racette SB, Deusinger SS, Inman CL,et al. Worksite Opportunities for Well-ness (WOW): effects on cardiovasculardisease risk factors after 1 year. PrevMed. 2009;49:108-114.
44. Rustad C, Smith C. Nutrition knowl-edge and associated behavior changesin a holistic, short-term nutrition edu-cation intervention with low-incomewomen. J Nutr Educ Behav. 2013;45:490-498.
45. Sarrafzadegan N, Kelishadi R,Sadri G, et al. Outcomes of acomprehensive healthy lifestyle pro-gram on cardiometabolic risk factorsin a developing country: the isfahanhealthy heart program. Arch IranMed. 2013;16:4-11.
46. Savoie MR, Mispireta M, Rankin LL,Neill K, LeBlanc H, Christofferson D.Intention to change nutrition-relatedbehaviors in adult participants of a sup-plemental nutrition assistance program-education. J Nutr Educ Behav. 2015;47:81-85.
47. Wilcox S, Parrott A, Baruth M, et al.The Faith, Activity, and Nutrition pro-
gram: a randomized controlled trial inAfrican-American churches. Am J PrevMed. 2013;44:122-131.
48. Lowe MR, Tappe KA, Butryn ML,et al. An intervention study targetingnutritional intake in worksite cafeterias.Eat Behav. 2010;11:144-151.
49. Francis SL, Taylor ML. A social mar-keting theory-based diet-educationprogram for women ages 54 to 83 yearsimproved dietary status. J Am Diet As-soc. 2009;109:2052-2056.
50. Moschonis G, Katsaroli I, Lyritis GP,Manios Y. The effects of a 30-month di-etary intervention on bone mineraldensity: the Postmenopausal HealthStudy. Br J Nutr. 2010;104:100-107.
51. BroekhuizenK,KroezeW, PoppelMN,Oenema A, Brug J. A systematic reviewof randomized controlled trials on theeffectiveness of computer-tailored phys-ical activity anddietary behavior promo-tion programs: an update. Ann BehavMed. 2012;44:259-286.
52. Olander C: Office of analysis, nutritionand evaluation food andnutrition service.Nutrition Education and the Role of Dosage.US Department of Agriculture; 2007.
53. Gutman SA, Murphy SL. From thedesk of the editor and associate edi-tor—information commonly unre-ported in intervention effectivenessstudies. Am J Occup Ther. 2012;66:7-10.
54. GearingRE, El-Bassel N, Ghesquiere A,Baldwin S, Gillies J, Ngeow E. Majoringredients of fidelity: a review and sci-entific guide to improving quality ofintervention research implementation.Clin Psychol Rev. 2011;31:79-88.
55. Cerin E, Barnett A, Baranowski T.Testing theories of dietary behaviorchange in youth using the mediatingvariable model with intervention pro-grams. J Nutr Educ Behav. 2009;41:309-318.
56. Sallis JF, McKenzie TL, Conway TL,et al. Environmental interventions foreating and physical activity: a random-ized controlled trial in middle schools.Am J Prev Med. 2003;24:209-217.
57. HandR, Abram J, Brown K, Ziegler P,Scott Parrott J, Steiber A. Developmentand Validation of the Guide for Effec-tiveNutrition Interventions and Educa-tion (GENIE): a tool for assessing thequality of proposed nutrition educationprograms. J Nutr Educ Behav. 2015;47:308-316.
165.e1 Murimi et al Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017
CONFLICT OF INTEREST
The authors have not stated any con-flicts of interest.