22 annotated bibliographies AMA style

Systematic Review

Factors Influencing Efficacy of Nutrition Education Interventions: A Systematic Review

Mary W. Murimi, PhD, RDN; Michael Kanyi, PhD; Tatenda Mupfudze, PhD; Md. Ruhul Amin, MPH, MS; Teresia Mbogori, MS; Khalid Aldubayan, PhD

College Conflict o with this the JNE Editor-in Address Sciences TX 7940 �2016 S reserved http://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 from previous 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-analysis criteria. 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 initially considered. The number was screened and scaled down to 40 publications for the final analysis. Quality assessment was based on the Cochrane Handbook for Systematic Reviews of Intervention. Studies were rated as having low risk of bias, moderate risk, or high risk. Results: Efficacy of nutrition education interventions depended on major factors: interventions that lasted $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 of this 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 as any set of learning experiences designed to facilitate the voluntary adoption of eating and other nutrition-related be- haviors conducive to health and well- being.1 Efficacy describes the ability to 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 factors including the duration and frequency of intervention, the number and relat-

of Human Sciences, Texas Tech Unive f Interest Disclosure: The authors’ confli article on www.jneb.org. The first auth B staff as Associate Editor. Review of -Chief to minimize conflict of interest for correspondence: Mary W. Murim , College of Human Sciences, Texas T 9; Phone: (806) 834-1812; Fax: (806) 7 ociety for Nutrition Education and Beh . .doi.org/10.1016/j.jneb.2016.09.003

edness of the study objectives, study design and theory, and fidelity in inter- vention.

The specific characteristics of the de- terminants of success of interventions are still unclear.2 However, several studies have been conducted to ascer- tain determinants of efficacy of nutri- tional education interventions. For example, another systematic review3

concluded that educational interven- tions that are sustained for a longer time, >5 months, and offer personal- ized feedback on dietary behavior and related health risk factors, are more

rsity, Lubbock, TX ct of interest disclosures can be found online or of this article (M. W. Murimi) served on this article was handled, exclusively, by the . i, PhD, RDN, Department of Nutritional ech University, PO Box 41240, Lubbock, 42-3042; E-mail: [email protected] avior. Published by Elsevier, Inc. All rights

Journal of Nutrition Education and Beh

likelytobeeffectivethanthoseconduct- edforashortperiod,<5months,anddo not offer personalized feedback. Other studies concluded that expert-led inter- ventions as well as studies that used behavioral theories, social support, and an educational approach to guide die- tary interventions were more likely to be successful.4 Despite previous studies on the wider area of nutrition educa- tion, there is still inadequate literature on the efficacies of the various nutrition education interventions that were im- plemented in recent years. In a departure from previous reviews that concen- trated primarily on a single type of intervention and its related outcome, the current review investigated several factors that led to success of various types of interventions. The purpose of this review was to examine system- atically the factors that contribute to the efficacy of nutrition education in- terventionsinpromotingbehaviorchange for good health and well-being based on their stated objective. To achieve this purpose, the researchers used population, intervention, comparison, and outcomes criteria 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

METHODS Literature Search

This systematic review was conducted in accordance with recommendations and criteria outlined in the Preferred Re- porting Items for Systematic Reviews and Meta-analysis statement.6,7 Articles on studies that conducted nutrition education interventions on dietary be- haviors were identified by performing literature searches in: PubMed, Medline, Web of Science, Academic Search Comp- lete, Science Direct, Cochrane Reviews, ERIC and PsychLIT. The search was limited to articles published between 2009 and 2015. Key search words were nutrition education, nutrition education interventions, dietary behavior, food, and health living. References of all retrieved studies were used to determine the source of information, whether they were primary, secondary, or website based, and to understand better the basis for conclusions of the studies that were reviewed.

All 6 members of the research team were independently involved in re- viewing the references. Inclusion and quality measures were determined by the 3 senior researchers who conduct- ed an independent evaluation of each article; afterward, several discussions were held to reach a consensus, hence monitoring bias. A total of 246 original studies published since 2009 and tar- geting healthy individuals without preexisting medical conditions were reviewed. This initial number was screened and scaled down to 40 publi- cations for the final analysis. Screening criteria for inclusion and elimination are illustrated in the Figure.

Members of the Research Team

The research team was composed of 6 members, 3 of whom held doctoral de- gree; the others had a master's degree in nutrition. The lead researcher was a full professor of nutrition and a regis- tered dietitian. Two other researchers were faculty members in recognized in- ternational universities with wide expe- rience in nutrition, education, and research.Eachofthe3seniorresearchers paired with 1 junior researcher in each database for article search and retrieval. All 6 members were independently involved in reviewing the articles and initially screening them.

Inclusion/Exclusion Criteria

The authors included in the review research articles published in English that 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 the target population had special nutri- tional needs (eating disorders, dia- betic, hospitalized, etc). In addition, studies that failed to achieve any of their objectives were excluded. In the cases where multiple studies were con- ducted on the same data set, only the most recently published study was included. There were 2 reviewers per database. Trained reviewers evaluated whether articles met inclusion criteria and determined the quality of the study. All researchers except the lead researcher went through group training, conducted by a systematic review and meta-analysis expert, which also involved watching a webcast.

Assessment of Study Quality/ Risk of Bias

In the initial part of work of the current review, researchers worked in pairs in which data were extracted by 1 reviewer and verified by a second reviewer. The risk of bias in any re- ported evidence should be at mini- mum and evidence that is likely to have high risk of bias serves a negli- gible purpose and thus should not be included in a systematic review even if there is no better evidence.8 In this review, determination of the quality of studies was guided by the Grading of Recommendations Assessment, Development, and Evaluation system of rating quality of evidence.9 A thor- ough assessment of the study's fidelity, perceived conflict of interest regarding outcome owing to sponsorship, study design, imprecision, inconsistency, appropriate use of theories, reasonable duration of intervention, and whether a study achieved the stated objectives formed criteria for quality assessment. Rating scores ranged from 1 to 6. Any discrepancies were discussed until an agreement was reached. Based on these criteria of assessment of study quality, studies were rated as having a low risk of bias (5–6 scores), moderate risk (3–4 scores), or high risk (1–2

scores) (Tables 1 and 2). Fidelity as a factor in this systematic review was assessed from authors' declaration of limitation in their respective studies.

Reviewers completed a detailed data extraction form. Extracted data were transferred to a spreadsheet (Tables 1 and 2).

Analysis Approach

The primary analytic goal was to deter- mine the overall effectiveness of nutri- tion education interventions to modify dietary and exercise behaviors. To deter- mine whether an intervention was suc- cessful, the outcome of the study was compared with the stated purpose and/or objectives of the study. Once a study was classified as having achieved its intended purpose, the contributing factors were assessed. Assessed factors included: (1) the design of the study including randomness, (2) the type of intervention and activities imple- mented, (3) the duration and dosage of the interventions, (4) number of objec- tives in a study, (5) fidelity in interven- tion implementation, and (6) the use of theory in directing the studies. These factors were identified through a thor- ough review of published nutrition edu- cation interventions. They were found to be common in almost all published studies. The duration of intervention wascategorizedasshortifithadacumu- lative length of >5 months and long if a study lasted for an accumulated period of $5 months. This classification of duration was deemed appropriate based on the descriptions authors used of the respectiveoriginal studies.Thereviewed studies rarely reported the dosage and frequency of interventions. Therefore it was reasonable to report the total amount of time spent in intervention in months.

Another factor that emerged during the review was worksite environment interventions. Worksite environments differ from one site to another. There are various worksite environment in- terventions for health living. These include the provision of health mes- sages around cafeterias, the provision of healthy food in cafeterias, encour- aging and providing walking space as part of exercise for healthy living, and schedules and amounts regarding eating, among others. The analysis of worksite environment interventions

Figure. Flow diagram illustrating the article filtering process as part of the systematic review. 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 on work- site environment interventions were analyzed and reported along with other initially identified factors.

A semiquantitative approach was used to summarize findings from nutri- tion education interventions. Results fromnutritioneducationinterventions were dichotomized based on whether they reported a statistically significant (P < .05) improvement in diet intake, exercise, or other related risk factors for obesity and diet-related chronic diseases. The researchers used this approach to allow for the diverse range of reported statistics, outcomes, and measurement 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 of intervention, number of objectives, fi- delity in intervention, use of theories, and use of worksite environment were identified as determinants of the efficacy 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,orphysiologicoutcomes based on their primary objectives. Of the 21 successful studies, 14 used the- ories and 7 did not. Another 48% of the interventions (n ¼ 19) met some but not all of their primary objectives; 9 of them used theories and 10 did not.

Study Designs in the Reviewed Studies

Randomized control trials accounted for 70% of the studies (n ¼ 28), the pre-

test–posttest/quasi-experimental design accounted for 23% (n ¼ 9), and the non-experimental design constituted 8% (n ¼ 3) of the total articles included in the analysis. Over half of the total number of studies (58%; n ¼ 23) were based on theory.

Interventions in the Reviewed Studies

This review considered studies con- ducted in person or face-to-face inter- ventions at the individual or group level. A majority of the studies (68%; n ¼ 27) used a single type of interven- tion. For example, a study whose objective was to create education materials using the target audience's preferences and to implement a heart-healthy diet education program used a single intervention in which 2 registered dietitians led in-home edu- cation sessions.49 Table 1 lists more ex- amples of studies that employed a single intervention. Another 33% of studies (n ¼ 13) employed multiple in- terventions including cooking, watching videos, attending exercise classes, and gardening. An example is a study whose objective was to examine the extent to which participants in a com- bined physical activity and dietary intervention achieved changes in mul- tiple health behaviors.36 Multiple in- terventions involved the provision of opportunities for physical activity and healthy eating before, during, and af- ter church services. Table 2 provides more examples on studies that used multiple interventions.

Effect of Duration of Intervention

For the purposes of this systematic re- view, the duration of the intervention was categorized as short if it had a cu- mulative length of <5 months and long if a study lasted for an accumu- lated period of $5 months. This was informed by the way studies reported their time. Therefore, for convenience in reporting of the results, the terms short duration and long duration were adopted. The amount of time spent in intervention was reported in months. The reviewed studies largely left out dosage and frequency of intervention.

In this review, two thirds (n ¼ 12) of nutrition education interventions that

Table 1. Face-to-Face, Individual, Group, and Peer Counseling/Nutrition Education Studies (n ¼ 27)

Authors Study Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias

(Quality Measure) Major Findings

Arrebola et al, 201110

60 patients with grade II overweight and non-morbid grade I–II obesity (age range, 18–50 y)

To evaluate effects of lifestyle modification program focused on diet, exercise, and psychological support on health- related quality of life

Pre–post. Intervention: physical activity recommendations and psychological support. Group sessions led by doctor, nurse, or dietitian.

No control

11 sessions conducted every 2 wk for 6 mo

No theory All Low Intervention was associated with significant improvements in physical functioning (80.37 � 18.90 vs 89.40 � 13.95; P < .001) and role of physical (20.37 � 9.10 vs 23.14 � 6.67; P < .05), vitality (58.71 � 21.98 vs 70.91 � 26.56; P < .01), social functioning (79.62 � 27.76 vs 86.57 � 25.45; P < .03), and general health (61.03 � 19.13 vs 69.42 � 18.80; P < .001) factors.

Auld et al, 201511

723 adults To analyze impact of weight management intervention on physical activity/ exercise and body weight and composition

Pre–post. Intervention: ESBA curriculum delivered in group settings.

No control

9 lessons taught for 8–12 wk

SCT and Adult Learning Theory

Some Moderate ESBA elicited mean positive behavior change for food resource management (P < .01), food safety (P < .001), nutrition (P < .001), and physical activity level in participating states (P < .01) except New York. There was an increase in dairy, fruit, and vegetable intake in Arkansas and California (P < .05) but not in Colorado, New York, and Ohio.

Babatunde et al, 201112

110 African American adults aged 50–93 y

To assess effectiveness of osteoporosis education program to improve calcium

RCT. Sessions of #15 people; short presentations/ lectures, hands-on

6 sessions (30– 45 min/wk) for 6 wk

Health Belief Model

All Low Overall, an educational program developed with a theoretical background was associated with an improvement in calcium

(continued)

Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

M u rim

i et

al 1 4 5

Table 1. Continued

Authors Study Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias

(Quality Measure) Major Findings

intake, knowledge, and self-efficacy

activities, and demonstrations to help participants increase self- efficacy.

Control: Delayed NE

intake (mean increase, 556 mg dietary calcium; P < .001), knowledge (P < .001), and self- efficacy (P < .001).

Backman et al, 201113

327 participants (156 treatment; 171 control), 75% of whom were low- income African American women aged 18–54 y

To evaluate effectiveness of fruit, vegetable, and physical activity toolbox for community educators in changing knowledge, attitudes, and behavior among women of low- income

Quasi-experimental design with treatment and control groups. Control group did not receive NE.

Intervention was 1-h nutrition and physicalactivity education classes per week for 6 wk

SCT All Low Women in the treatment group reported significant changes in 9 measures of attitude, compared with 1 measure in the control group (P < .05). Compared with those in the control group, women in the treatment group were also more likely to make behavioral changes to meet recommendations for fruit and vegetable consumption (P < .001) and physical activity (P < .001).

Brennen and Williams, 201314

16 African American women aged 25– 63 y

To evaluate effects of culturally sensitive lifestyle intervention on blood pressure and weight

Quasi-experiment. Intervention: Counseling and education on increasing physical activity and dietary intake of fruits and vegetables while decreasing dietary intake of salt and fat.

No control

10 individual sessions, 30 min each, and 11 group sessions, 60 min each, for 12 wk

No theory All Moderate Both systolic and diastolic blood pressure decreased from a mean of 151/90 pre-intervention to 131/76 post- intervention. There was a significant decrease from pre- to post-intervention systolic (P ¼ .03) and diastolic (P ¼ .001) blood pressures. There was a statistically significant improvement in self- efficacy to exercise if bored (P ¼ .02) or if busy with other activities

1 4 6

M u rim

i et

al Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

(P ¼ .008). There was a statistically significant improvement in self- perception as valuable/ worthless (P ¼ .02) and choice as superficial/ profound (P ¼ .04).

Clifford et al, 200915

101 college students To determine whether a series of 4 15- min, theory-driven (SCT) cooking programs aimed at college students living off campus improved cooking self-efficacy, knowledge, attitudes, and behaviors regarding fruit and vegetable intake

RCT. Subjects in intervention group viewed 4 15-min cooking programs over 4 wk. Subjects in control group viewed 4 5-min programs on sleep disorders.

4 weekly 15-min episodes

SCT Some Moderate There were significant improvements in knowledge of fruit and vegetable recommendations in the intervention group compared with the control group post¼intervention and at 4-mo follow-up (P < .05). There were no significant changes in fruit and vegetable motivators, barriers, self-efficacy, or intake.

Craigie et al, 201116

75 adults aged $40 y To assess the feasibility of a lifestyle intervention focusing on diet and activity, participating in cardiovascular screening

RCT. Lifestyle intervention composed of 3 personalized counseling sessions plus telephone contact.

12 wk No theory reported

Some Low 82% successfully maintained or lost weight (mean loss 1.1 kg, and 2.6 cm waist circumference) and 85% reported eating 5 portions of fruits and vegetables compared with 56% at baseline. No behavior changes were detected in control group.

Davis et al, 200917

46 individuals aged >16 y

Assessment of peer- led approach to improving diet of South Asians in Southampton

Quasi-experiment. Intervention: 10 taster sessions and 28 cookery club sessions.

No control

Length not specified but follow up was conducted after 1 y

No theory All Moderate There was increased intake of low-fat dairy products and reduced fat and salt intake. 80% and 75% made positive changes to cooking practices and eating patterns, respectively.

Duncan et al, 201318

286 adults enrolled in English as a Second Language, aged 18–73 y

To conduct pre–post feasibility trial of Healthy Eating for Life, which

Pre–post design. Intervention: Healthy Eating for Life curriculum for at

At least 2 h/wk for 12 wk

Social Learning Theories

All Low There was a significant increase in fruit, vegetable intake, nutrition knowledge, action

(continued)

Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

M u rim

i et

al 1 4 7

Table 1. Continued

Authors Study Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias

(Quality Measure) Major Findings

integrates content about healthy nutrition to decrease cancer health disparities

least 2 h/wk of classroom instruction for 12 wk.

No control group

planning, and coping planning among participants (P < .05 for all).

Endevelt et al, 201119

127 older adults aged $75 y

To determine impact of intensive nutritional intervention program on health and nutritional status of malnourished community

Partial RCT. DIT: Intervention led by dietitian or medical treatment. A physician led a standard care group with an educational booklet.

Nonrandomized ‘‘untreated nutrition’’ group.

5 visits for 6 mo No theory All Low DIT group showed significant improvement in cognitive function and depression score compared with the change in the other 2 groups. DIT group showed a significant improvement in intake of carbohydrates, protein, vitamin B6, and vitamin B1 and had a significantly lower cost of physician visits than did the other 2 groups (P < .05 for all).

Francis and Taylor, 200949

58 women aged 54–83 y

To create education materials using target audience’s preferences and implement heart- healthy diet education program designed using needs and preferences

RCT. Intervention: 2 individual registered dietitian– led in-home education sessions

Control: 2 education material mailings

3 mo Social Marketing Theory

All Low Intervention and control mini- nutritional assessment scores improved (P < .001). Intervention subjects consumed more fiber than did control subjects (P ¼ .01) and reduced sodium intake (P ¼ .02). Controls reduced energy (P ¼ .01) and cholesterol intakes (P ¼ .03), likely because of decreased food intake.

Ha and Caine-Bish, 201120

80 college students aged 18–24 y

To determine whether there would be an increase in whole- grain consumption after students

Pre–post. Intervention: General nutrition class.

No control

3 times/wk for 50 min for semester

SCT Some Moderate At baseline, total mean grain consumption was 3.07oz and mean whole-grain consumption was 0.37 oz. After the study, mean

1 4 8

M u rim

i et

al Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

completed an interactive introductory nutrition course focusing on disease prevention

consumption of whole- grain products significantly increased to 1.16 oz (P < .001) but total mean grain intake remained the same (3.06 oz).

Hsu et al, 201321

25 adults aged >18 y To examine feasibility, acceptability, and preliminary results of exercise intervention with a Healthy at Every Size orientation

RCT. Intervention: Exercise training and weekly behavioral intervention.

Control: Exercise only

Project CHANGE was 8-wk randomized, controlled trial with follow-up at 4 wk

SDT All Low Both interventions showed large effect sizes on changes in weekly energy expenditure, moderate PA, and brisk walking. Both interventions showed small effect sizes for all fitness variables, including body mass index, waist– hip ratio, predicted VO2max, 1RM machine chest press, and leg press. Adherence to PA goal was better for the intervention group at follow-up. The Self-Determination Theory based exercise intervention with a Healthy at Every Size resulted in larger effect sizes for changes in key motivational variables, including self- determination, autonomy, and goal-setting, planning and scheduling self-efficacy (P not listed).

Ireland, et al, 201022

43 healthy adults aged 20–75 y

To investigate whether dietary education enabled reduction in salt consumption

RCT. 2 different education methods using either Australia’s National Heart Foundation Tick symbol or Food Standards Australia and New Zealand’s low-salt guideline of 120 mg sodium/ 100 g food.

8 wk No theory Some Low After 8 wk, urinary sodium excretion decreased from 121 � 50 to 106 � 47 mmol/24 h (7.3 � 3.0 to 6.4 � 2.8 g salt/24 h) in the Tick group and from 132 � 44 to 98 � 50 mmol/24 h (7.9 � 2.6 to 6.0 � 3.0 g salt/24 h) in the Food Standards Australia New Zealand group (P < .05,

(continued)

Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

M u rim

i et

al 1 4 9

Table 1. Continued

Authors Study Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias

(Quality Measure) Major Findings

with no between-group difference).

Kannan et al, 201023

102 low-income African American women aged 18–45 y

To reduce nutrition risks and enhance protective nutrition and analyze changes in self- efficacy

Pre–post. Intervention: peer-led nutrition curriculum.

No control

13 lessons taught at 1-wk intervals for 13 wk

PEN-3 model and TTM

All Low 77% reported adopting at least 1 healthy eating behavior (moderating sodium or serving more fruits and vegetables to their families), 23% adopted at least 2 such behaviors, and 45% adopted both dietary and biomedical behaviors (self-monitoring blood pressure, and exercising).

Kontogianni et al, 201224

126 individuals aged 45–67 y

To evaluate impact on dietary and activity habits of non- intensive, community-based lifestyle intervention for type 2 diabetes prevention in high- risk Greek individuals

Pre–post. Intervention: NE with dietician.

No control

6 bimonthly sessions for 1 y

No theory Some Moderate There was decreased consumption of whole-fat dairy and processed meat (P ¼ .02 and .02, respectively), sugar (P ¼ .006), and refined cereals (P ¼ .05). There was improved diet, decreased body weight (P ¼ .04), plasma triglycerides (P ¼ .02), and 2-h post-load plasma glucose (P ¼ .05) compared with those who had worsened dietary habits. Total time spent daily on physical activity remained unchanged throughout the intervention.

Kreausukon et al, 201225

114 full-time undergraduate students (aged 18–

To improve fruit and vegetable consumption

RCT. Intervention group received psychological

Not specified but follow-up at 6 wk

SCT All Low A social-cognitive intervention to improve fruit and vegetable consumption was

1 5 0

M u rim

i et

al Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

25 y) at Chiang Mai University, Thailand

program that addressed self- efficacy and strategic planning.

Control group received handouts about general nutrition guidelines

superior to a knowledge- based education session with significantly greater intention, planning, and self-efficacy for fruit and vegetable consumption (P < .05 for all). Both the intervention and control groups demonstrated an increase in fruit and vegetable consumption from baseline to 1 wk after the intervention (P < .001 for both) but the increase was greater in the intervention group (P ¼ .01).

Mendonca Rde and Lopes, 201226

167 adults aged 40–65 y

To determine effects of health interventions on dietary habits and physical measurements

Quasi-experimental/ pre–post

Intervention: Guided physical exercise, nutrition intervention, nutritional education groups, individual nutritional care.

No control

4 sessions, 60 min each for 7 mo

No theory All Low There was a reduction in systolic blood pressure (P ¼ .02) and use of animal fats (P < .01) as well as an increase in the percentage of individuals with a normal waist circumference and daily consumption of greens/ vegetables and milk/dairy products (P < .01)

Milliron et al, 201227

153 adults aged 20–65 y

To promote healthy eating behavior and weight management

RCT. Intervention: 10 min face-to-face.

Control: Received no education on Eat Smart shelf tags posted in store.

10 min No theory Some Moderate No significant differences between the 2 groups on purchased total, saturated, or trans-fat and servings of total vegetables. However, the intervention group purchased significantly more servings (per 1,000 kcal) of whole fruit and dark green/bright yellow vegetables compared with the control group.

(continued)

Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

M u rim

i et

al 1 5 1

Table 1. Continued

Authors Study Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias

(Quality Measure) Major Findings

Nakade et al, 201228

226 overweight/ obese adults in Japan, aged 40–65 y

To evaluate effects of behavioral approach that emphasized tailored behavior counseling, diet, weight loss, and weight maintenance

RCT. Intervention: 30 min individual counseling and 20 min group sessions about effective exercise provided by registered dietitians and exercise instructors.

Control: No NE

5 counseling sessions for 1 y and follow-up 1 y after intervention

No theory All Low The intervention group lost significantly more weight than the control group (– 5.0 kg vs 0.1 kg for men and –3.9 kg vs –0.2 kg for women). Dietary intake and number of walking steps improved in the intervention group. After 1-y follow-up, the intervention group maintained significantly lower weight, lower energy intake, and improvement in irregular eating habits (P < .05 for all variables).

Pimentel et al, 201029

67 Brazilian adults aged 50–69 y, with impaired glucose tolerance and at least 1 other risk factor for diabetes mellitus 2

To evaluate effectiveness of nutritional education program on anthropometric, dietary, and metabolic parameters with impaired glucose tolerance

RCT. Intervention: individual and group counseling once and twice per month, respectively with team of nutritionists

Control: No NE

36 sessions over 12 mo

No theory Some Low The intervention group showed a significant decline in body weight (�3.4%), body mass index (�5.7%), cholesterol intake (�49.5%), fasting glycemia (�14.0%), fasting insulin (�9.0%), postprandial glycemia (�21.0%), postprandial insulin (�71.0%), total serum cholesterol �23.0%), and glycated hemoglobin (�24.0%). A decrease in energy intake (5%; P ¼ .06) and low density lipoprotein cholesterol (25%; P ¼ .07) was observed in the interventional group, although it did not reach statistical significance.

1 5 2

M u rim

i et

al Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

Plawecki and Chapman- Novakofski, 201330

69 adults aged 55–75 y

To enhance physical activity and nutritional behaviors

RCT. Intervention: lectures and hands-on active learning

Control: intervention was delayed

3 sessions (duration not mentioned) for 8 wk

Health Belief Model and Theory of Reasoned Action

Some Low Comparison of week 1 and week 8 data indicated significant improvement for the treatment but not the control group for calcium, and vitamin D (P < .05). There was limited response to the exercise outcome variables, with many not participating in that section of the program.

Ritchie et al, 201031

3,015 and 3,004 before and after NE; pregnant or postpartum women/caregivers of children enrolled in WIC, aged 22–36 y

To explore impact of WIC on family behavior regarding fruits and vegetables, whole grains, and lower- fat milk

Pre–post cross- sectional design. Caregivers received education intervention in a group (class) or in individual (one-on- one counseling) format.

No control

3 sessions over 6 mo

TTM All Low After nutrition education, women and caregivers reported increased recognition of education messages, positive movement in stage of change for target food items, increased family consumption of fruits and whole grains, and replacement of whole milk with lower-fat milk.

Shahnazari et al, 201332

84 US veterans aged 25–80 y

To determine effectiveness of nutrition-based wellness coaching using multiple contacts and simple educational tips on health eating and weight management

RCT. Intervention: 9 individualized NE sessions. Coaching consisted of 15-min sessions with final 60-min session at end of 6 mo. Total of 3.75 h educational contact for intervention and 1 h for control group by same nutrition coach for each veteran throughout study

Control: 1 h individualized NE session.

9 NE sessions for 6 mo

Stage of Change Model

All Low Multiple coaching contacts decreased intake of energy, fat, and carbohydrates by 31% (P < .001). Weight loss of 5% from baseline (92.8 to 88.2 kg; P < .01) was observed in the intervention group with mean body mass index decreasing from 30.4 to 28.9 (P < .05). The control group showed a decrease in fat intake by 20% (P < .01) but no statistically significant changes in intake of other nutrients or body weight (88.7 to 87.4 kg). Veterans’ readiness to change eating behavior

(continued)

Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

M u rim

i et

al 1 5 3

Table 1. Continued

Authors Study Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias

(Quality Measure) Major Findings

for weight loss improved with nutrition coaching.

Silva et al, 201033

239 women aged 30–45 y

To analyze impact of weight management intervention on theory-based psychosocial mediators, physical activity/exercise, and body weight and composition

RCT. Intervention: NE sessions covering: physical activity, eating/nutrition, body image, and other cognitive and behavioral contents

Control: General health education

30 sessions, 2 h each weekly or bimonthly for 1 y

SDT All Low At 12 mo, the intervention group showed increased weight loss (–7.29%,) and higher levels of physical activity/exercise (þ138 � 26 min/d of moderate plus vigorous exercise; þ2,049 � 571 steps/d) compared with control subjects (P < .001).

Sorensen et al, 201134

56 individuals aged 22–55 y

To compare effect of behavior modification consisting of either a gourmet cooking course or NLP therapy on weight regain

RCT. The first step was 12- wk weight loss program. Participants achieving at least 8% weight loss were randomized to 5 mo of either NLP therapy or course in gourmet cooking.

8 mo No theory Some Low The NLP therapy group lost 1.8 kg and the cooking group lost 0.2 kg during the 5 mo of weight maintenance. The dropout rate was lower during the active cooking treatment compared with the NLP group. There was no difference in weight maintenance after 2 and 3 y of follow-up.

Wieland et al, 201235

34 women (Hispanic, Somali, and Cambodian) aged 22–68 y

To evaluate socio- culturally appropriate physical activity and nutrition intervention in community-based participatory research approach

Pre–post. Intervention. 6-wk program with 2 90- min classes per week.

No control

6 wk No theory Some Moderate After the intervention, participants were more likely to exercise regularly (P < .001). They reported higher health-related quality of life (P < .001). Self-efficacy for diet and exercise; weight loss, waist circumference, and blood pressure were not significantly different after the 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 Program for Women, Infants, and Children.

1 5 4

M u rim

i et

al Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

Table 2. Multicomponent Nutrition Education Intervention Studies (n ¼ 13)

Authors Study

Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias (Quality Measure) Major Findings

Baruth and Wilcox, 201336

360 African Americans aged $18 y

To examine extent to which participants in combined physical activity and dietary intervention achieved changes in multiple health behaviors

RCT. Intervention: PA and

healthy eating before, during and after church services

Control: delayed intervention

15 mo SEM All Low Up to 19% indicated no change in health behavior, 31% changed 1 health behavior, 31% changed 2 health behaviors, 13% changed 3 health behaviors, and 5% changed all 4 targeted health behaviors. Combinations of multiple behavior change included PA and dietary behaviors, which suggests that both behaviors can be changed simultaneously.

Cullen et al, 200937

1,004 Texas EFNEP clients in 100 classes, mean age 35 y

To evaluate modified curriculum for 6- session Texas EFNEP promoting healthful home food environments and parenting skills related to obesity prevention

RCT. Intervention: 6 short

videos with goal setting, problem solving, guided discussion, and handouts. Then a weekly goal sheet for recording was issued Participants monitored daily goal attainment and return goal sheet the following week

Control: Traditional EFNEP class included brief discussion and food preparation

6 classes for 6 wk

No theory Some Low There was a significant BMI decrease at postintervention compared with baseline only for the intervention group. This change was not maintained at follow-up.

(continued)

Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

M u rim

i et

al 1 5 5

Table 2. Continued

Authors Study

Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias (Quality Measure) Major Findings

Dirige et al, 201338

673 Filipino- American adults aged $18

To evaluate 18-mo nutrition and physical activity intervention (active life) conducted through culturally specific organizations

RCT. Intervention:

Workshops and activities (cooking demonstrations, recipe contests, supermarket tours, group aerobic classes, gardening)

Control: Cancer screening, alternative medicine, and stress management

18 mo TTM Some Low Intervention participants showed significant increases in PA (P < .05), adoption of a low-fat diet (P < .05), and stage of change for fruits and vegetables (P < .05), dietary fat intake (P < .01), and PA (P < .01). Intervention did not lead to increases in number of participants eating $5 servings/d of fruits and vegetables.

French et al, 201039

160 metropolitan transit workers aged 20–79 y

To describe and report results from worksite obesity prevention intervention that targeted transit employees

Group RCT (4 garages)

Intervention: Enhancement of PA facilities, increased availability of and lower prices for healthy vending machine choices, etc

Control: No intervention

18 mo worksite intervention

No theory Some Low Energy intake decreased significantly and fruit and vegetable intake increased significantly in intervention garages compared with control garages. However, BMI and PA changes were not significant.

Iriyama and Murayama, 201440

57 male workers in Japan

To evaluate effects of new worksite weight-control program using combination of nutrition education environmental interventions

RCT crossover. Intervention:

Intervention group received 6-mo program consisting of nutrition education and provision of healthy cafeteria meals and nutritional information. 1-y

6 mo TTM and Precede- Proceed model

All Low Mean BMI was significantly reduced from baseline value of 25.6 kg/m2 to 25.3 kg/ m2 at month 6 and to 24.8 kg/m2 at year 1 (P ¼ .008) and was significantly lower at year 1 than at baseline and month 6 in multiple comparison

1 5 6

M u rim

i et

al Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

follow-up Control received

same intervention as counterparts in crossover design from 6 mo after study entry

tests in intervention group; BMI increased over time in control group (P < .001).

Johansen et al, 201041

198 adults aged 25–63 y

To present effect of intervention study intentions to change dietary behavior and changes made in dietary intake

RCT. Intervention:

Combination of group sessions, individual counseling, and organized exercise groups

Control: General advice participants would receive from general practitioner

6 sessions, 2 h each, for 7 mo

TTM Some Low Differences between intervention and control after intervention were significant for sugar-rich drinks and rapeseed oil (P < .05). Intention to reduce dietary intake of fat, sugar, and white flour, and to change type of fat and increase intake of vegetables and legumes shifted for intervention group (P < .05) from pre- action (pre- contemplation, contemplation, and preparation) stages to action stage in intervention group but not in control group. Difference between groups at follow-up was significant (P < .05). No significant differences were found for intention to increase fruit intake.

Linde et al, 201242

1,672 participants aged 18–75 y

To influence weight gain positively among employees over 2 y

RCT (worksite level). Six worksites in US metropolitan area were recruited and randomized in pairs at worksite level to 2-y intervention or no-contact control

Intervention: Posters at worksites, link to Web sites with

2 y, frequency varied with center

No theory reported

None Moderate No differences between sites in key outcome of weight change over 2- y study period.

(continued)

Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

M u rim

i et

al 1 5 7

Table 2. Continued

Authors Study

Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias (Quality Measure) Major Findings

useful information on food availability and price, PA promotion, scale access, and media enhancements

Control: No contact

Racette et al, 200943

123 participants aged 36–54 y with BMI 32.9 � 8.8 kg/ m2, employed at 1 of 2 selected worksites within a large medical center

To evaluate effectiveness of worksite health promotion program on improving cardiovascular disease risk factors

Cohort randomized trials. Intervention included pedometers, healthy snack cart, Weight Watchers meetings, group exercise classes, seminars, team competitions, and participation rewards

Control: Personal health reports containing assessment results

Weekly for 1 y

TTM Some Low Improvements (P < .05) were observed at both worksites for fitness, blood pressure, and total, high-density lipoprotein, and LDL-C. Additional improvements occurred in intervention group in BMI, fat mass, Framingham risk score, and prevalence of metabolic syndrome; only changes in BMI and fat mass were different between control and intervention worksites.

Rustad and Smith, 201344

118 ethnically diverse, low- income women aged 23–45 y

To assess impact of short-term nutrition intervention using education on comprehensive array of nutrition and health topics

Pre–post intervention. Experiential and interactive lectures, activities, and demonstrations. Educational sessions on shopping and budgeting, healthy cooking, improving food security by growing foods.

No control

3 sessions in 6 wk.

Each class lasted 75–90 min

No theory All Moderate Postintervention increased nutrition knowledge and favorable nutrition behavior. Responses to 7 of 11 questions in knowledge and 9 of 11 variable set changed significantly at postintervention (P < .01 for both). Women also decreased consumption of fast

1 5 8

M u rim

i et

al Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

foods and processed snacks high in sugar, salt, and fat, and fatty cuts of meat, and decreased addition of sugar, salt, and butter to foods (P < .01).

Sarrafzadegan et al, 201345

12,514 participated at baseline and 9,570 in post- intervention survey, aged 24–54 y

To evaluate component of Healthy Heart Programs to assess feasibility and outcomes of program on lifestyle behaviors and risk factors for chronic non-communicable diseases

RCT. Intervention: Public

education through mass media, intersectoral cooperation and collaboration, professional education and involvement, marketing, organizational development, legislation and policy development, as well as research and evaluation

Duration of intervention activities varied: 3–4 y.

Precede– Proceed model, Social Learning Theory, and innovation diffusion approach

Some Moderate Prevalence of abdominal obesity, hypertension, hypercholesterolemia, hypertriglyceridemia, and high LDL-C decreased significantly in intervention area vs control area in both sexes. However, reduction in overweight and obesity was significant only in females (P < .05 for all). There were no significant changes in prevalence of diabetes mellitus.

Savoie et al, 201546

203 participants aged $18 y

To determine whether participation in selected Supplemental Nutrition Assistance Program–Education

lessons had an impact on intent to improve nutrition- related behaviors of participants

Retrospective post- then-pre design

Intervention: lecture, cooking demonstration, sample tasting of food prepared in class, handout.

No control

Not indicated TPB All Low Mean responses of individual questions and mean lesson scores increased significantly from pretest to posttest in the menu planning, shopping lesson, and the My Plate lesson (P < .001).

Wilcox et al, 201347

74 African Methodist Episcopal churches and 1,257 adult members within them

To report results of intervention targeting physical activity and healthy eating

RCT. Intervention: Churches

implemented; sharing messages from pulpit; passing out educational materials (provided), create Faith, Activity, and

15 mo SEM Some Low There was a significant effect favoring intervention group in self-reported leisure time (d ¼ 0.18; P ¼ .02). No group differences were found for self-reported fruit and vegetable

(continued)

Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

M u rim

i et

al 1 5 9

Table 2. Continued

Authors Study

Population Study Objective Study Design/ Intervention

Length and Frequency

Behavioral Theory/

Construct Achieved Objectives

Risk of Bias (Quality Measure) Major Findings

Nutrition Program Control: delayed intervention components at end of 15 mo

consumption, measured blood pressure, and self- reported fat- and fiber- related behaviors.

Lowe et al, 201048

96 adult participants (BMI 29.7 � 6.0 kg/ m2) hospital or university employees aged 21–65y

To evaluate nutritional and weight changes in program that used worksite cafeterias to reduce employees’ calorie content of purchased foods and improve their macronutrient intake

RCT. 3 mo of baseline data

collection, then 3- mo intervention; 6- and 12-mo postintervention follow-ups. Participants were randomly assigned to 1 of 2 intervention groups: Environmental Change or Environmental Change Plus Energy Density Education and Incentives. Randomization of participants occurred within each worksite

3 mo No theory Some Low There was no difference between groups in total energy intake over study period.

Across groups, energy and percentage of energy from fat decreased and percent of energy from carbohydrates increased from baseline tointerventionperiod(all P < .01). Follow-up analyses, conducted by averaging baseline months 1 and 2 and comparing them with intervention month 3 as a conservative estimate of overall impact of intervention, indicated that change in energy, carbohydrate, and fat intake remained significant (P < .001).

Providing nutrition labels and reducing energy- density of selected foods was associated with improved dietary intake.

BMI indicates body mass 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.

1 6 0

M u rim

i et

al Jo u rn al o f N u tritio

n E d u catio

n an d B eh av io r �

V o lu m e 4 9 , N u m b er

2 , 2 0 1 7

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 study whose objective was to evaluate the effectiveness of a nutritional education program that involved bimonthly group discussions for 12 months, and that included written and oral didactic instructions on anthropometric, dietary, and metabolic parameters with impa- ired glucose tolerance, reported a decrease in 2 risk factors related to diabetes mellitus.29

Similarly, a study that took 1 year and whose objective was to study the impact of a weight management inter- vention on theory-based psychosocial mediators reported weight loss (–7.29%,) and increased levels of physical activ- ity/exercise (þ138 � 26 min/d of mod- erateplusvigorousexercise;þ2,049� 571 steps/d) compared with control sub- jects (P < .001).33 Finally, a study that involved a physical activity and healthy eating intervention over 15 months reported significant results and an 18% effect on the intervention group in self-reported leisure time (P ¼ .02; d ¼ 0.18).47

On the contrary, this review found that nutrition education interventions that lasted for a short duration were less likely to meet their stated objec- tives. A relevant example here is a study whose purpose was to promote healthy eating behavior and weight management.27 In this study, inter- vention involved 10 minutes of face- to-face education on the printed EatSmart shelf tags posted in the store. The control group received no educa- tion about the EatSmart shelf tags posted in the store. Outcome measures included purchases of total saturated and 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 analysis of participants' shopping baskets. Re- sults showed no significant differences between the control and intervention groups on total purchases as well as in the choice of health eating grocery products except in fruit and dark green or yellow vegetables, thus indicating minimal attainment of the study ob- jectives.

Similarly, a nutrition education intervention on physical activity and nutrition35 revealed that self-efficacy for diet and exercise, weight loss,

waist circumference, and blood pres- sure were not significantly different from baseline after 2-hour classes each week for 6 weeks.

Effect of Number of Study Objectives/Focus

In addition to the long duration, this systematic review found that studies with few or focused objectives were more successful in meeting all of their stated objectives than were interven- tions that had several unrelated objec- tives. For example, a 4-month nutrition education intervention whose objec- tive was to increase whole-grain con- sumption among students who completed an interactive introductory nutrition course focusing on disease preven- tion20 indicated that student partici- pants increased whole-grain consumption from 0.37 to 1.16 oz (P < .001). Simi- larly, another study22 with only 1 objective, to investigate whether die- tary education enabled a reduction in salt consumption, indicated that after 8 weeks of dietary education interven- tion there was a reduction in salt con- sumption and urinary sodium excretion.

In contrast, interventions with >3 unrelated objectives were not success- ful in meeting all of their objectives. For instance, a study that had >3 ob- jectives, conducted in 3 phases with different feeding regimes, reported inconsistent results at the 3 phases.34

In phase 1, 88% of participants (n ¼ 49) completed 12 weeks of calorie re- striction and achieved 8% weight loss. Phase 2 involved 5 months of weight maintenance; participants were divided into 2 groups each with different feeding regimes and hence different objectives. The 2 groups experienced different numbers of participant dropout, which affected the final results. There was no differ- ence in weight maintenance after 2 and 3 years of follow-up. In general, it was observed that follow-up studies did not yield many results. A lot of follow-up studies did not yield a significant change from the initial re- sults. Therefore, it can be implied that a majority of researchers could have paid less attention to the follow-ups. The researchers in this re- view recommend stringent measures in follow-up studies, just as in the

initial phase of the study, to enhance reliability of the results of follow-ups.

Lack of Fidelity in Delivery

Fidelity in intervention ensures that all intervention activities are executed as planned in the methods. There were few reported cases of lack of intervention fidelity that could have compromised the findings. A peer- led nutrition education intervention that addressed maternal and infant health through dietary patterns re- ported that some facilitators neglected to follow the complete lesson plans by omitting parts of a lesson or failing to use the facilitator's guide, or they did not show enthusiasm in promoting the desired behavior among the peer- led cohorts. Although the interven- tion was the same by design and content information, results of the study varied among cohorts. As a result, the success of the interventions was affected by the human factors of the presenter.23

Theory-Based Studies

Slightly over half of the studies (57.5%; n ¼ 23) reported being theory based and used at least 1 theory. The most common theories used to design and implement nutrition education interventions in studies selected in this review were the Trans-theoretical Model and Social Cognitive Theory. The majority of the theory-based studies (61%; n ¼ 14 of 23) were successful in achieving their stated ob- jectives, whereas the remaining theory-based studies (39%; n ¼ 9) achieved some but not all of their pri- mary objectives.

This review considered studies that provided information about how they used theories in the design of the study as a best practice, rather than just mentioning the theory casually in the introduction or methods: for instance, a study conducted by Savoie et al46 to determine whether partici- pation in a selected Supplemental Nutrition Assistance Program–Education (SNAP-Ed) that clearly showed how the constructs of the Theory of Planned Behavior used in the design and 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 the study objective. Results of this study showed that posttest scores were significantly higher than pretest scores related to menu planning, shopping les- sons, and My Plate lessons (P < .001).

However, this review observed that although some studies indicated that they were theory based, they failed to describe explicitlyhow thetheories guided the studies. For example, some re- searchers15 reported using Social Cogni- tive Theory; others45 reported using several theories including the Precede– Proceed Model, Social Learning Theory, and the Innovation Diffusion Approach. However, the theory constructs and how they were used or measured were not described in either article.

Nevertheless, although 45% of studies (n ¼ 18) were not informed by a theory, they were equally as suc- cessful as those that were. For instance, a study to describe and report the re- sults from a worksite obesity preven- tion intervention that targeted transit employees was not guided by a theory, and yet it indicated success.39 The re- sults indicated that energy intake decreased significantly and fruit and vegetable intake increased significantly in intervention garages compared with control garages. A summary of the studies that used theories and those that did not is provided in Tables 1 and 2.

Environmental Interventions at the Worksite

Worksite environmental interventions have an integral part in modifying die- tary and weight management behav- iors when executed appropriately. For example, a study to evaluate the effects of a new worksite weight control pro- gram using nutrition education envi- ronmental interventions among male adults in Japan reported significant re- sults. At the 1-year follow-up, the inter- vention group hadsignificantly greater reductions in body weight, body mass index, and alanine aminotransferase than the control group did (P ¼ .02, .02, and .86, respectively).40

For worksite environmental inter- ventions to be successful, sufficient appropriate changes must be imple- mented at the right places. This re- view discovered that some worksite

environmental interventions did not ensure changes sufficient for the inter- vention, which affected the results. Such a failure occurred when the worksite management and collabora- tors resisted making sufficient envi- ronmental changes to modify dietary and exercise behavior in employees at work despite promoting the behavior.39,42 For example, a worksite environmental intervention study used posters at worksites, links to Web sites with useful information about food availability and prices, physical activity promotion, access to scales, and media enhancements to promote weight gain control.42

Intervention components were food selection, promotion of walking and stair use, weight self-monitoring, and health information at work. However, the provision of a variety of healthy options and extra time for exercise were not offered to employees. The re- sults of this study indicated that there were no differences between the inter- vention and control sites in the key outcome of weight change over the 2-year study.

DISCUSSION

The purpose of this review was to sys- tematically examine factors that contribute to the efficacy of nutrition education interventions in promoting behavior change for good health and well-being based on their stated objec- tive. The main findings of this review indicated that the efficacy of nutrition education interventions depends on the duration of the intervention, having few focused objectives, the appropriate use of theories, fidelity in interventions, and support from policy makers and management for the environmental interventions. These findings are largely congruent with the results of another review conducted by Baird et al.4

Therefore, factors that were identi- fied as determinants of efficacy and which form the discussion of this study are: (1) the types and use of de- signs, (2) the type of intervention that characterized the studies, (3) the dura- tion and dosage of the interventions, (4) the number of objectives in a study, (5) fidelity in intervention, and (6) the use of theories in nutrition education interventions. Worksite environmental interventions are also

featured in this discussion. Critical analysis of these factors provides the outcome of the current review.

The randomized control trial (RCT) design has a reputation of being robust. This design is therefore appropriate for baseline studies intended to inform a larger intervention project.51 This re- view indicated that a majority (70%; n ¼ 28) of the nutrition education in- terventions used an RCT design. The RCT design is robust and may be attrib- uted to the success of the interventions in achieving their stated objectives.

Interventions that lasted for >5 months reported a higher level of success. They were mainly multiple- component interventions. Thisfinding supported the results of a previous re- view that reported that remarkably more studies on nutrition education with long-term follow-up were associ- ated with success.51 Another study noted that behavior change takes time and practice.52 Therefore, it may be argued that the length of time taken for intervention and the frequency of exposure are important factors for the success of a nutrition education inter- vention. However, another study noted that interventions with long du- rations are associated with a higher cost of implementation and participants' attrition, which constrained some in- terventions.24

This systematic review revealed that studies with few and succinct ob- jectives were more successful than were those with numerous and at times unrelated objectives. Other re- views acknowledged the effectiveness of few objectives in nutrition educa- tion interventions.51 The current re- view noted that studies with #3 objectives that were related were suc- cessful even when the duration of intervention was <6 months.

It is important to report fidelity in interventions because it allows readers and other researchers to judge the quality of the intervention and how various factors may have influenced the outcome.53 Lack of fidelity in the delivery of a program has a counter- productive effect on the results of an intervention. Fidelity in intervention is a critical element that is rarely re- ported in many studies.54 This review found that the few studies that re- ported fidelity discovered that it nega- tively affected the results: Some sites achieved 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 educators might have experienced some profes- sional challenges in the delivery of educationalinterventions,otherstudies noted that they could provide a good form of social support associated with successful behavior change.4 It is there- fore important for the training of peer- led interventions to emphasize fidelity in the implementation in an effort to achieve desired results.

This systematic review revealed that although the majority of the studies that were theory based were successful in achieving all of their primary objec- tives (53.8%), a good number (45%) of well-designed, non-theory interventions were equally successful in achieving their primary objectives. The current findings support a review by Baba- tunde et al,12 who noted that overall, an educational program developed with a theoretical background was associ- ated with an improvement in calcium intake (mean increase, 556 mg dietary calcium;P< .001),knowledge(P< .001), and self-efficacy (P < .001). This indi- cates that well-used theory is likely to make an intervention successful. This review observed that some of the studies that claimed to use a theory failed to describe explicitly how the theory was used in the study. This finding agrees with the results of a re- view on experimentally based evi- dence of the theoretical mechanisms of dietary behavior change,55 which concluded that future intervention tri- als need to focus on identifying effec- tive procedures for mediator change and adopting a more rigorous and sys- tematic approach to theory testing.

Finally, in worksite environmental interventions, interventions without appropriate support from collabora- tors to support the desired behavior were less likely to meet their objec- tives. For instance, an intervention did not meet the primary objective, which was to affect weight gain posi- tively over 2 years, owing to weak and inconsistent implementation of environmental changes in the work- place.42 In contrast, an intervention by Iriyama and Murayama40 on work- site weight control was able to imple- ment changes in the workplace cafeteria and introduce healthy me- nus. The results of this intervention indicated a significant decrease in

mean body mass index from a base- line value of 25.6 kg/m2 to 25.3 and 24.8 kg/m2 at 6 months and 1 year, respectively (P < .05). These worksite environmental interventions high- light the need for effective collabora- tion among nutritionists, policy makers and stakeholders within institutions including schools and workplaces where the food environment has a big role in the success of a nutrition education intervention. These find- ings are in agreement with results of another study regarding the effects of environmental, policy, and social marketing interventions on physical activity and fat intake of middle school students. The study noted that envi- ronmental and policy interventions were effective in increasing physical activity at school. Appropriate changes are 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, there is a possibility that some recent and important findings published in lan- guages other than English were left out. Another limitation is that there was a potential that studies that did notfindasignificanteffectintheirinter- vention were not published, and there- fore were not included in this review.

Finally, the review was limited by articles that did not include adequate information in their methods and re- sults. This posed a challenge to eval- uate the contributions of specific components of nutrition education interventions and their effectiveness properly, including the use of theory and the dosage of intervention (fre- quency and duration). Despite the limitations, the current review drew its strength from the fact that the re- searchers investigated several factors that led to the success of various types of interventions. This was a departure from previous reviews that concen- trated primarily on a single type of intervention and the related outcome.

IMPLICATIONS FOR RESEARCH AND PRACTICE

The results of this review suggest that nutrition education interventions with longer duration have few and focused objectives, and that those that are guided

by a theory have a higher chance of achieving their purpose. Lack of fidel- ity in peer-led interventions, lack of behavior support for environmental interventions, and too-short dura- tions were factors that contributed to a lack of success in some of the inter- ventions. It was observed that various human factors affected the effective- ness of an intervention in cases where peers and paraprofessionals imple- mented the nutrition interventions. Although the lack of fidelity during interventions was not rampant, the few reported incidences had profound ramifications on the results. The studies with control groups had better inter- pretation of the results, which enhanced the validity of the outcome. This im- plies that studies that used RCTs had a better chance of replicability, followed by those that employed a quasi- experimental design. The eligible studies that were reviewed and whose inter- ventions were considered successful were considered largely sustainable and could easily be replicated.

The researchers concluded that the use of theories in designing nutri- tional education interventions was a common practice in which 55% of the analyzed studies (n ¼ 22) were informed by at least 1 theory. Many studies that used theories indicated success in achieving their objectives. However, a lack of details regarding how the behavior theories guided the studies made it difficult to assess the effect of the theories mentioned in some studies. The researchers concluded that the use of theories is a good practice in interventions and that worksite environmental inter- ventions provide an important oppor- tunity for behavioral adjustments for better health, but that they need the cooperation of the policy makers.

The results of this study suggest that more focused, clearly defined, measur- able objectives are associated with behavior change, whereas the more ambitious use of many objectives may limit the effectiveness of nutrition education by taking away from the main message and confusing partici- pants. The objective should have a clear targeted behavior, followed by adequate dosage or exposure to facili- tate the desired behavior change. A purposeful selection of behavior the- ory that will guide the intervention based 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 designing and implementing the intervention should be reflected in the results. For studies that use multiple sites with different implementers, ensuring fidel- ity is critical to the success of the inter- vention. Training should emphasize the important message and targeted behavior. Researchers should also consider the Guide for Effective Nutri- tion Interventions and Education as a checklist to help design or improve their research methods for effective in- terventions.57

For worksite and other environ- mental interventions, it is important for policy makers to make healthy choices the easy ones by allowing time for the exercise. This can be im- plemented in many ways depending on the worksite environment: for example, including markings indi- cating the distance covered along a path 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 nutrition education policy, programs, and research: a review of research. J Nutr Educ. 1995;27:312-328.

2. Ammerman AS, Lindquist CH, Lohr KN, Hersey J. The efficacy of behavioral interventions to modify die- tary fat and fruit and vegetable intake: a review of the evidence. Prev Med. 2002; 35:25-41.

3. Pomerleau J, Lock K, Knai C, McKee M. Interventions designed to increase adult fruit 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, Barker M, Inskip HM. Changing health behaviour of young women from disadvantaged backgrounds: evidence from systematic reviews. Proc Nutr Soc. 2009;68:195-204.

5. Guyatt G, Drummond R, Meade M, Cook D. The Evidence Based-Medicine Working 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 reporting systematic reviews and meta-analyses of studies that evaluate health care inter-

ventions: explanation and elaboration. J Clin Epidemiol. 2009;62:1-34.

7. Moher D, Tetzlaff J, Tricco AC, SampsonM, Altman DG. Epidemiology and 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. Cochrane Handbook for Systematic Reviews of Inter- ventions. Version 5.1.0. http://www. cochrane-handbook.org. Accessed October 1, 2016.

9. GRADE Working Group. Grading quality 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 program for treatment of overweight and non- morbid obesity in primary healthcare and itsinfluenceonhealth-relatedquality of 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 widely adopted EFNEP curriculum Eating Smart-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 Nutr Educ Behav. 2011;43:434-440.

13. BackmanD,ScruggsAA, Atiedu S,etal. Using a toolbox of tailored educational lessons to improve fruit, vegetable, and physical activity behaviors among Afri- can American women in California. J Nutr Educ Behav. 2011;43:S75-S85.

14. Brennen M, Williams CL. Lifestyle management 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 a TV cooking show for college students living 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 in cardiovascular risk screening. Prev Med. 2011;52:387-389.

17. Davies JA, Damani P, Margetts BM. Intervening to change the diets of low-income women. Proc Nutr Soc. 2009;68:210-215.

18. Duncan LR, Martinez JL, Rivers SE, et al. Healthy eating for life English as a second language curriculum: primary outcomes from a nutrition education intervention targeting cancer risk reduc- tion. J Health Psychol. 2013;18:950-961.

19. Endevelt R, Lemberger J, Bregman J, et al. Intensive dietary intervention by a dietitian as a case manager among community dwelling older adults: the EDIT study. J Nutr Health Aging. 2011;15:624-630.

20. Ha EJ, Caine-Bish N. Interactive intro- ductory nutrition course focusing on disease 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 exercise intervention promoting healthy at every size with sedentary overweight women: Project CHANGE. Psychol Sport 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. Healthy Eating and Harambee: curriculum development for a culturally-centered bio-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 a non-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 vegetable consumption: a randomized controlled trial. J Behav Med. 2012;35:443-451.

26. Mendonca Rde D, Lopes AC. The ef- fects of health interventions on dietary habits and physical measurements. Rev Esc Enferm USP. 2012;46:573-579.

27. Milliron BJ, Woolf K, Appelhans BM. A point-of-purchase intervention featuring in-person supermarket educa- tion affects healthful food purchases. J Nutr 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 loss program 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 nutrition education reduces several risk factors for type 2 diabetes mellitus in Brazilians with impaired glucose tolerance. Nutr Res. 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 Gerontol Geriatr. 2013;32:145-160.

31. Ritchie LD, Whaley SE, Spector P, Gomez J, Crawford PB. Favorable impact 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 in overweight US veterans. J Acad Nutr Diet. 2013;113:928-935.

33. Silva MN, Vieira PN, Coutinho SR, et al. Using self-determination theory to promote physical activity and weight control: a randomized controlled trial in women. J Behav Med. 2010;33: 110-122.

34. Sorensen LB, Greve T, Kreutzer M, et al. Weight maintenance through behaviour modification with a cooking course 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 among immigrant and refugee women: a community-based participatory research approach. Womens Health Is- sues. 2012;22:225-232.

36. Baruth M, Wilcox S. Multiple behavior change among church members taking part in the faith, activity, and nutrition program. J Nutr Educ Behav. 2013;45: 428-434.

37. Cullen KW, Lara Smalling A, Thompson D, Watson KB, Reed D, Konzelmann K. Creating healthful home food environments: results of a study with participants in the expanded food 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 physical activity 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 to prevent obesity among metropolitan transit workers. Prev Med. 2010;50: 180-185.

40. Iriyama Y, Murayama N. Effects of a worksite weight-control programme in obese male workers: a randomized controlled crossover trial. Health Educ J. 2014;73:247-261.

41. Johansen KS, Bjorge B, Hjellset VT, Holmboe-Ottesen G, Raberg M, Wandel M. Changes in food habits and motivation for healthy eating among 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-randomized worksite environmental intervention trial for weight gain prevention. Int J Behav Nutr Phys Act. 2012;9:14.

43. Racette SB, Deusinger SS, Inman CL, et al. Worksite Opportunities for Well- ness (WOW): effects on cardiovascular disease risk factors after 1 year. Prev Med. 2009;49:108-114.

44. Rustad C, Smith C. Nutrition knowl- edge and associated behavior changes in a holistic, short-term nutrition edu- cation intervention with low-income women. J Nutr Educ Behav. 2013;45: 490-498.

45. Sarrafzadegan N, Kelishadi R, Sadri G, et al. Outcomes of a comprehensive healthy lifestyle pro- gram on cardiometabolic risk factors in a developing country: the isfahan healthy heart program. Arch Iran Med. 2013;16:4-11.

46. Savoie MR, Mispireta M, Rankin LL, Neill K, LeBlanc H, Christofferson D. Intention to change nutrition-related behaviors 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 in African-American churches. Am J Prev Med. 2013;44:122-131.

48. Lowe MR, Tappe KA, Butryn ML, et al. An intervention study targeting nutritional intake in worksite cafeterias. Eat Behav. 2010;11:144-151.

49. Francis SL, Taylor ML. A social mar- keting theory-based diet-education program for women ages 54 to 83 years improved 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 mineral density: the Postmenopausal Health Study. Br J Nutr. 2010;104:100-107.

51. Broekhuizen K, Kroeze W, Poppel MN, Oenema A, Brug J. A systematic review of randomized controlled trials on the effectiveness of computer-tailored phys- ical activityanddietary behaviorpromo- tion programs: an update. Ann Behav Med. 2012;44:259-286.

52. Olander C: Office of analysis, nutrition and evaluationfood and nutrition service. Nutrition Education and the Role of Dosage. US Department of Agriculture; 2007.

53. Gutman SA, Murphy SL. From the desk of the editor and associate edi- tor—information commonly unre- ported in intervention effectiveness studies. Am J Occup Ther. 2012;66:7-10.

54. Gearing RE, El-Bassel N, Ghesquiere A, Baldwin S, Gillies J, Ngeow E. Major ingredients of fidelity: a review and sci- entific guide to improving quality of intervention research implementation. Clin Psychol Rev. 2011;31:79-88.

55. Cerin E, Barnett A, Baranowski T. Testing theories of dietary behavior change in youth using the mediating variable model with intervention pro- grams. J Nutr Educ Behav. 2009;41: 309-318.

56. Sallis JF, McKenzie TL, Conway TL, et al. Environmental interventions for eating and physical activity: a random- ized controlled trial in middle schools. Am J Prev Med. 2003;24:209-217.

57. Hand R, Abram J, Brown K, Ziegler P, Scott Parrott J, Steiber A. Development and Validation of the Guide for Effec- tive Nutrition Interventions and Educa- tion (GENIE): a tool for assessing the quality of proposed nutrition education programs. 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.