Quantifying parental preferences for interventions designed to improve home food preparation and home food environments during early childhood

Accepted Manuscript Quantifying parental preferences for interventions designed to improve home food preparation and home food environments during early childhood Senbagam Virudachalam, Paul J. Chung, Jennifer A. Faerber, Timothy M. Pian, Karen Thomas, Chris Feudtner PII:

S0195-6663(15)30084-2

DOI:

10.1016/j.appet.2015.11.007

Reference:

APPET 2759

To appear in:

Appetite

Received Date: 22 June 2015 Revised Date:

13 October 2015

Accepted Date: 8 November 2015

Please cite this article as: Virudachalam S., Chung P.J., Faerber J.A., Pian T.M., Thomas K. & Feudtner C., Quantifying parental preferences for interventions designed to improve home food preparation and home food environments during early childhood, Appetite (2015), doi: 10.1016/j.appet.2015.11.007. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1 Quantifying parental preferences for interventions designed to improve home food preparation and home food environments during early childhood

RI PT

Authors and e-mail addresses: Senbagam Virudachalam, MD, MSHPa,b ([email protected]), Paul J. Chung, MD, MSc,d,e ([email protected]), Jennifer A. Faerber, PhDa ([email protected]), Timothy M. Pian, BAf ([email protected]), Karen Thomas, MPHa ([email protected]), Chris Feudtner, MD, PhD, MPHa,b ([email protected])

AC C

EP

TE D

Corresponding author: Senbagam Virudachalam, MD, MSHP Division of General Pediatrics The Children’s Hospital of Philadelphia 34th St. & Civic Center Blvd. CHOP North Room 1545 Philadelphia, PA, 19104 E-mail: [email protected] Phone: 215-590-6753 Fax: 267-426-0380

M AN U

SC

Author affiliations: a Division of General Pediatrics, The Children’s Hospital of Philadelphia, Philadelphia, PA, USA b Department of Pediatrics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA c Departments of Pediatrics and Health Policy & Management, University of California Los Angeles, Los Angeles, CA, USA d RAND Corporation, Santa Monica, CA, USA e Children's Discovery & Innovation Institute, Mattel Children's Hospital, University of California Los Angeles, Los Angeles, CA, USA f Warren Alpert Medical School, Brown University, Providence, RI, USA

ACCEPTED MANUSCRIPT 2 ABSTRACT

TE D

M AN U

SC

RI PT

Though preparing healthy food at home is a critical health promotion habit, few interventions have aimed to improve parental cooking skills and behaviors. We sought to understand parents’ preferences and priorities regarding interventions to improve home food preparation practices and home food environments during early childhood. We administered a discrete choice experiment using maximum difference scaling. Eighty English-speaking parents of healthy 1-4 year-old children rated the relative importance of potential attributes of interventions to improve home food preparation practices and home food environments. We performed latent class analysis to identify subgroups of parents with similar preferences and tested for differences between the subgroups. Participants were mostly white or black 21-45 year-old women whose prevalence of overweight/obesity mirrored the general population. Latent class analysis revealed three distinct groups of parental preferences for intervention content: a healthy cooking group, focused on nutrition and cooking healthier food; a child persuasion group, focused on convincing toddlers to eat home-cooked food; and a creative cooking group, focused on cooking without recipes, meal planning, and time-saving strategies. Younger, lower income, 1-parent households comprised the healthy cooking group, while older, higher income, 2-parent households comprised the creative cooking group (p<0.05). The child persuasion group was more varied with regard to age, income, and household structure but cooked dinner regularly, unlike the other two groups (p<0.05). Discrete choice experiments using maximum difference scaling can be employed to design and tailor interventions to change health behaviors. Segmenting a diverse target population by needs and preferences enables the tailoring and optimization of future interventions to improve parental home food preparation practices. Such interventions are important for creating healthier home food environments and preventing obesity starting from early childhood.

AC C

EP

Keywords: cooking, home food preparation, home food environment, early childhood, nutrition, and obesity

ACCEPTED MANUSCRIPT 3 INTRODUCTION The American Academy of Pediatrics’ Bright Futures: Guidelines for the Health Supervision of

2

Infants, Children, and Adolescents describes the need for “well-designed studies that examine a

3

range of interventions” focused on promoting healthy eating behaviors and “helping families

4

initiate lifestyle changes” during early childhood (Hagan JF, 2008). Inherent in this approach is

5

improving a family’s ability to create a healthy home food environment by preparing food at

6

home (Couch, Glanz, Zhou, Sallis, & Saelens, 2014). Many studies have shown that home-

7

prepared food is healthier than food prepared outside the home (Briefel, Wilson, & Gleason,

8

2009; Larson, Perry, Story, & Neumark-Sztainer, 2006; Lin B, 1999; Stephens, McNaughton,

9

Crawford, & Ball, 2014), and more than a decade of research has established the beneficial

M AN U

SC

RI PT

1

effects of family meals (Gillman et al., 2000; Martin-Biggers et al., 2014; Taveras et al., 2005).

11

A recent study among school-age children showed that the frequency of family meals was only

12

moderately correlated with the frequency of home-prepared dinner consumption, but a higher

13

frequency of either family meals or home-prepared food was associated with a healthier diet

14

(Appelhans, Waring, Schneider, & Pagoto, 2014). Further, many prevailing strategies to promote

15

healthy eating, such as recommending decreased fast and processed food consumption

16

(U.S.D.A., 2013) and improved access to fresh produce (Kohan, 2011), assume that most

17

Americans possess the skillset needed to translate nutrition knowledge and fresh food access into

18

healthier diets by regularly preparing food at home. Yet only fifty percent of Americans cook

19

dinner 6 or 7 nights per week (Virudachalam, Long, Harhay, Polsky, & Feudtner, 2014), and

20

little is known about how to reintroduce such practices into family life. While many families

21

with unhealthy eating patterns are motivated to change, pediatricians have few resources to offer

22

them.

AC C

EP

TE D

10

ACCEPTED MANUSCRIPT 4 23 There are two key reasons to focus on families with young children. First, early childhood is the

25

time when children begin acculturating into the household’s food preparation and eating patterns

26

(Benton, 2004; Birch, 1999; Institute of Medicine (U.S.). Committee on Obesity Prevention

27

Policies for Young Children., Birch, Burns, & Parker, 2011). When young children transition to

28

solid food from breast milk or formula, there is a window of opportunity to teach them healthy

29

food-related routines they will carry forward into later childhood, adolescence, and adulthood.

30

Second, 2-5 year-old children consume more than 75% of their meals at home and depend more

31

on their parents and home food environments for nourishment than school-age children and

32

adolescents (Anzman, Rollins, & Birch, 2010; Lin B, 1999). Young children depend on their

33

parents for both physical sustenance and modeling of food-related behaviors, including planning,

34

preparing, and eating healthy food (Savage, Fisher, & Birch, 2007; Sweetman, McGowan,

35

Croker, & Cooke, 2011). Parents are essential for providing children with healthy home food

36

environments during early childhood (Ogata & Hayes, 2014). Creating a household culture that

37

includes home-prepared foods and healthy food-related routines starting from early childhood

38

should help to prevent the struggle of having to break unhealthy habits and introduce new

39

routines during later childhood and adolescence.

SC

M AN U

TE D

EP

AC C

40

RI PT

24

41

To our knowledge, no one has focused on improving parental food preparation practices to

42

enhance the home food environments of families with young children. Previous studies have

43

shown that many factors influence home food preparation practices and home food

44

environments, including time, skills, culture, food availability, resources, and nutrition-related

45

knowledge (Couch et al., 2014; Jabs & Devine, 2006; Malhotra et al., 2013). Yet the literature is

46

silent regarding parents’ preferences for future interventions focused on home food preparation.

ACCEPTED MANUSCRIPT 5 Studies among parents regarding young children’s diets and family meals do, however, provide

48

relevant information. One study found healthier diets among young children whose mothers

49

provided breakfast daily, cooked from scratch, and facilitated family meals, and maternal

50

motivations and attitudes regarding these behaviors significantly predicted their frequency

51

(Swanson et al., 2011). Another study found that young children’s vegetable consumption was

52

positively correlated with cooking from scratch, children and parents eating similar foods, and

53

the use of ready-made sauces (Sweetman et al., 2011). Two series of focus groups among parents

54

of young children explored barriers and facilitators for family meals. In one study, highly

55

educated parents reported several perceived benefits to regular family meals; they also identified

56

barriers, including child behavior, time, and cooking ability (Quick, Fiese, Anderson, Koester, &

57

Marlin, 2011). In another study, low-income mothers’ reported that while they are motivated to

58

have regular family meals, they require support for both meal preparation and managing social

59

interactions during mealtime (Malhotra et al., 2013). While all these studies support the need for

60

interventions focused on home food preparation for parents of young children, they do not

61

examine which intervention strategies parents prefer. The default is a “one-size-fits-all”

62

approach, offering the same intervention for all potential participants. Much work in the fields of

63

intervention design and implementation science suggests, however, that interventions to change

64

complex health-related behaviors, such as home food preparation practices, must take into

65

account varied individual needs and preferences for intervention content, even within specific

66

target populations (Campbell et al., 2007; Glanz, Rimer, & Viswanath, 2008).

SC

M AN U

TE D

EP

AC C

67

RI PT

47

68

We sought to lay a systematic foundation for designing and tailoring interventions to improve

69

home food preparation practices and home food environments during early childhood by asking

ACCEPTED MANUSCRIPT 6 what parents of young children want from such interventions. We administered a discrete choice

71

experiment to ascertain parents’ views on the relative importance of characteristics they would

72

consider when deciding whether to participate in interventions to improve home food preparation

73

practices (Cohen, 2003; Paulhus, 1991). We then identified subgroups of parents with similar

74

preferences for intervention content and examined whether demographic and behavioral

75

characteristics differed across subgroups.

AC C

EP

TE D

M AN U

SC

RI PT

70

ACCEPTED MANUSCRIPT 7 METHODS

77

Study population

78

We recruited caregivers during their child’s outpatient visit at one of three urban or suburban

79

pediatric primary care clinics in 2013. Eligible caregivers were ≥21 years and English-speaking,

80

with at least one child age 1-4 years who was born at ≥36 weeks gestational age without chronic

81

medical conditions. Using electronic health record (EHR) rosters of scheduled pediatric visits,

82

we identified potentially eligible caregivers based on the child’s age, gestational age, and health

83

status. Study staff also reviewed medical records to purposefully sample at least 20 African

84

American, 10 Latino, and 40 Medicaid-enrolled families.

M AN U

SC

RI PT

76

85

159 potentially eligible families were approached, 122 were eligible, 83 enrolled, and 80 families

87

participated in the study. The final study sample was comprised of 80 adult caregivers and 87

88

children. Seven families had more than one child in the eligible age range; all eligible children in

89

each family were enrolled in the study. Each adult caregiver completed only one discrete choice

90

experiment, regardless of the number of children they enrolled in the study. Thirty-seven

91

families were ineligible because the caregiver did not speak English. None of the 159 families

92

we approached were excluded based on the other exclusion criteria (caregiver age, child

93

gestational age, or child health status). Thirty-nine eligible English-speaking caregivers chose

94

not to enroll in the study; 31 of these caregivers completed a short survey to determine if non-

95

participants differed from participants with regard to gender, age, BMI, and cooking habits.

EP

AC C

96

TE D

86

97

Informed consent was obtained under a study protocol that was reviewed and deemed exempt by

98

The Children’s Hospital of Philadelphia Institutional Review Board.

ACCEPTED MANUSCRIPT 8 99 Measures

101

Drawing on the literature (Evans et al., 2011; Malhotra et al., 2013; Rosenkranz &

102

Dzewaltowski, 2008), as well as feedback from expert colleagues and parents from the target

103

population, we developed and refined a list of 16 content and 13 logistical items of importance

104

for potential interventions to improve home food preparation practices and home food

105

environments among families with young children (Text Box). The final discrete choice

106

experiment was field tested among a convenience sample of parents and other adults to ascertain

107

comprehension and ease of completion. The experiment was revised accordingly prior to being

108

administered to study participants. Parental opinions regarding the relative importance of these

109

items were ascertained using maximum difference scaling, described below.

M AN U

SC

RI PT

100

110

We collected socio-demographic and behavioral characteristics of interest using a written

112

questionnaire. All parental characteristics were self-reported, including gender, age, race,

113

ethnicity, country of birth, education level, employment status, income, household structure, and

114

height and weight. Child characteristics obtained from the EHR included gender, age, insurance

115

payer (as a marker of income), and measured height and weight; location of daytime childcare

116

was reported by caregivers. Current home food preparation practices and factors affecting the

117

home food environment were reported by caregivers in answer to questions from the National

118

Health and Nutrition Examination Survey (CDC, 2007).

120

EP

AC C

119

TE D

111

Statistical Analysis

ACCEPTED MANUSCRIPT 9 Parents rated the relative importance of 16 content and 13 logistical items that would, in their

122

view, comprise effective interventions to improve home food preparation practices and home

123

food environments. This exercise, referred to as best/worst or maximum difference scaling, was

124

administered and analyzed using MaxDiff/Web v.6.0 (Sawtooth Software, Inc., Sequim, WA)

125

(The MaxDiff/Web v6.0 Technical Paper, 2007). The 16 content items were arrayed into 12 sets

126

of 4 items each; the 12 sets provided equal representation of all 16 items. For each set of 4 items,

127

participants selected the most and least important items. We used a similar exercise to ascertain

128

parental preferences regarding the 13 logistical items. MaxDiff employs hierarchical Bayes

129

estimation to predict the probability of choosing each item (as best or worst) from the set of

130

items shown; this provides both individual and group-level estimates of the relative importance

131

of each item. The raw logit scores are transformed to ratio-scaled scores on a scale of 0-100, with

132

higher scores representing greater importance. The final output provided the rank order and

133

strength of parental preferences regarding content and logistical considerations for a potential

134

intervention to improve home food preparation practices.

TE D

M AN U

SC

RI PT

121

135

We performed latent class analysis on the preference data to identify subgroups of respondents

137

with similar intervention preferences using Sawtooth Software’s Latent Class v.4.6.5 (Sawtooth

138

Software, Inc., Sequim, WA). We examined solutions with 2-5 distinct classes and replicated

139

each latent class solution 10 times using different random seeds. We chose the solution that had

140

classes with conceptual meaning, as well as the best fit using the Bayesian information criterion

141

(Schwarz, 1978).

142

AC C

EP

136

ACCEPTED MANUSCRIPT 10 We described the socio-demographic and behavioral characteristics of the total sample and each

144

latent class subgroup, and tested for differences between the latent class subgroups using Fisher’s

145

exact test. All p-values represent two-sided hypothesis tests; we set a significance level of 0.05

146

for all tests.

AC C

EP

TE D

M AN U

SC

RI PT

143

ACCEPTED MANUSCRIPT 11 RESULTS

148

Eighty of 122 eligible adult primary caregivers participated, a 66% response rate. Participating

149

caregivers (Table 1) were mostly women who were 21-45 years old and self-identified as either

150

white or black. Participants had varied education level, employment status, income, and

151

household structure. Participants mirrored the US population with regard to overweight and

152

obesity prevalence and cooking habits (Virudachalam et al., 2014), with 8% never, 50%

153

sometimes, and 42% always cooking dinner. At least 40% of study participants were receiving

154

food assistance (WIC or SNAP). More than half of participating children were cared for at home

155

during the day. Eligible adult caregivers who declined to participate did not differ significantly

156

from study participants with regard to socio-demographic characteristics and cooking habits

157

(determined using Chi-square and t-tests; data not shown).

M AN U

SC

RI PT

147

158

Parental preferences for intervention content are shown in Figure 1a. Rating scores are relative,

160

on a scale from 0-100; for instance, an item with a score of 10 is preferred twice as much as an

161

item with a score of 5. Parents most highly prioritized learning how to cook usual, staple meals

162

in a healthier way (#1, score 12.7) and how to cook healthy meals (#2, score 12.1). How to

163

convince toddlers to eat home-cooked food was third (#3, score 10.5). The two lowest ranked

164

items were how to stock the kitchen and pantry (#15, score 1.8) and make grocery lists (#16,

165

score 0.6); parents expressed a comparatively stronger interest in learning about meal planning

166

(#6, score 8.0).

EP

AC C

167

TE D

159

168

The most preferred logistical consideration (Figure 1b) was for professionally taught classes

169

(#1, score 18.5), and having free, on-site childcare during class was second (#2, score 15.6).

ACCEPTED MANUSCRIPT 12 Parental preferences for logistical considerations were uniform, with little variability across the

171

sample. The least important logistical considerations were proximity to public transportation

172

(#11, score 3.4), having classes taught by community members (#12, score 3.0), and holding

173

classes at participants’ homes (#13, score 2.7).

174

RI PT

170

We performed latent class analysis on the content preferences to identify subgroups of parents

176

with similar preferences. We identified the best solution as a 3-class solution on the grounds of

177

parsimony and strong conceptual meaning of each class. The healthy cooking group (n=21) was

178

named based on their three top preferences for intervention content: nutrition (#1, score 16.1),

179

how to cook usual, staple meals in a healthier way (#2, score 14.0), and how to cook healthy

180

meals (#3, score 13.7). In the child persuasion group (n=34), the most preferred content item

181

was how to convince toddlers to eat home-cooked food (#1, score 15.6), with the second and

182

third most preferred items focused on healthy cooking: how to cook usual, staple meals in a

183

healthier way (#2, score 14.3) and how to cook healthy meals (#3, score 13.0). Parents in the

184

creative cooking group (n=25) were less focused on health and nutrition than the other two

185

groups; their most preferred items instead related to creativity in the kitchen and meal planning:

186

cooking without recipes and being creative (#1, score 11.3), how to plan meals and make meals

187

out of leftovers (tied for #2, score 9.7 for both).

M AN U

TE D

EP

AC C

188

SC

175

189

Similarities and differences between parental preferences in the three subgroups are shown in

190

Figure 2. Learning about nutrition was more important to the healthy cooking group than the

191

creative cooking and child persuasion groups (score of 16.1 vs. 4.3 and 10.0, respectively,

192

p<0.0001 for both t-tests). On the other hand, learning about toddler feeding strategies was 2-3

ACCEPTED MANUSCRIPT 13 times more important to parents in the child persuasion group than for the creative cooking group

194

(15.6 vs. 7.5, p<0.0001) and the healthy cooking group (15.6 vs. 5.4, p<0.0001). The creative

195

cooking group wanted to learn about cooking creatively without recipes nearly twice as much as

196

parents in the child persuasion group (11.3 vs. 6.0, p=0.001), and they wanted to learn about

197

cooking quickly >4 times as much as the healthy cooking group (8.8 vs. 2.0, p<0.0001).

198

Learning about making meals out of leftovers was much more important to the creative cooking

199

group than the healthy cooking and child persuasion groups (9.7 vs. 1.2 and 3.9, respectively,

200

p<0.0001). All the groups felt that learning about making time to cook at home was a very low

201

priority (rank ≤ 11 and rating scores ≤ 4 for all groups). Parents in the healthy cooking and

202

creative cooking groups wanted to learn cooking skills 2-4 times more than parents in the child

203

persuasion group (scores of 4.3, 7.9, and 1.8, respectively, p<0.05). Learning to choose and buy

204

fresh ingredients, as well as how to find and use recipes, was much more important to parents in

205

the healthy cooking group than the other two groups (fresh ingredients: 3.7 and 2.2 times more

206

important than for the creative cooking and child persuasion groups, respectively; recipes: >4

207

times as important as for the other two groups; p<0.0001 for all).

SC

M AN U

TE D

EP

208

RI PT

193

The three latent class subgroups differed significantly from one another by caregiver age,

210

insurance payer, receipt of food assistance, income, and household structure (Table 2). The

211

healthy cooking group was largely comprised of younger, single, lower income parents whose

212

children receive Medicaid. This was in contrast to the creative cooking group, which was largely

213

comprised of older, 2-parent, higher income households whose children receive private

214

insurance. The child persuasion group was more varied with regard to age (nearly equal

215

distribution of parents across age categories), household structure (mix of 1- and 2-parent

AC C

209

ACCEPTED MANUSCRIPT 14 households), and income (high proportion of both low- and high-income households). The child

217

persuasion group also reported cooking dinner more regularly (62% always cooked) than either

218

the healthy cooking or creative cooking groups (33% and 24%, respectively). The groups did not

219

defer significantly by race, ethnicity, education level, employment, weight status, or number of

220

children in the household. There were no differences between the children of caregivers in each

221

of the three latent class subgroups with regard to gender, age, weight status, or location of

222

daytime childcare.

AC C

EP

TE D

M AN U

SC

RI PT

216

ACCEPTED MANUSCRIPT 15 DISCUSSION

224

In our diverse sample of 80 parents of young children, respondents reported cooking an average

225

of 5 dinners per week, reflective of the national average (Virudachalam et al., 2014). In a discrete

226

choice experiment, parents expressed one of three different sets of preferences for the content of

227

potential interventions to improve home food preparation practices and home food environments.

228

The healthy cooking and child persuasion groups both strongly preferred learning how to cook

229

healthy and nutritious meals. The child persuasion group, however, highly prioritized learning

230

how to convince toddlers to eat home-cooked food, while the other groups did not prioritize this

231

topic. A more disparate third group, the creative cooking group, was focused on culinary

232

creativity and time-saving strategies – learning how to cook without recipes, plan meals, and use

233

leftovers. The groups were demographically different and exhibited different cooking habits. The

234

creative cooking group had a higher proportion of older, higher income, partnered parents

235

compared to the healthy cooking group, which had a higher proportion of younger, lower

236

income, single parents. The child persuasion group was more varied with regard to age, income,

237

and household structure, but they reported cooking dinner more regularly than either of the other

238

two groups. Parents across the sample had similar preferences regarding logistical considerations

239

for a potential intervention, preferring professionally taught classes and free, on-site childcare.

SC

M AN U

TE D

EP

AC C

240

RI PT

223

241

This study has two key implications. First, this study begins answering the question of how to

242

reintroduce the kitchen back into the family structures and routines of the twenty-first century,

243

especially during early childhood. Americans spend 40% less time cooking now than in the

244

1960s (Jabs & Devine, 2006). There are many potential reasons for this shift. Today, creating a

245

healthy home food environment is a complex endeavor, from planning, to shopping for

ACCEPTED MANUSCRIPT 16 ingredients, to preparing and eating family meals, to budgeting time and money for the entire

247

process. Our findings can be applied to design tailored interventions to improve home food

248

preparation practices and home food environments for our target population, as well as similar

249

populations of families with young children.

250

RI PT

246

Second, this study offers a rigorous, underutilized methodology that can be applied to design and

252

appropriately tailor either similar interventions for different populations or interventions focused

253

on changing other health behaviors. Discrete choice experiments and maximum difference

254

scaling offer an effective and efficient way to gather a target population’s preferences regarding

255

various aspects of a potential intervention. Knowing both the rank order and the strength of

256

respondent preferences allows findings to be easily applied to real-world contexts, where trade-

257

offs are inevitable. For instance, in this study parents uniformly felt that compared to having on-

258

site childcare, holding classes in a location easily accessible by public transportation was an

259

unimportant logistical consideration. This was unexpected, but helpful for future planning.

M AN U

TE D

260

SC

251

Prior studies have suggested tailoring health behavior interventions to address the varying needs

262

of particular groups within a population (Campbell et al., 2007; Deal, 2013; Grisolia, Longo,

263

Boeri, Hutchinson, & Kee, 2013). This study supports that assertion by showing that parents

264

within a specific target population have substantially different preferences for intervention

265

content. The three latent class groups identified imply specific lifestyle issues within each group.

266

Many prior studies have identified time as a significant barrier for preparing food at home

267

(Jones, Walter, Soliah, & Phifer, 2014; Monsivais, Aggarwal, & Drewnowski, 2014; Pelletier &

268

Laska, 2012; Quick et al., 2011; Smith, Ng, & Popkin, 2013; Storfer-Isser & Musher-Eizenman,

AC C

EP

261

ACCEPTED MANUSCRIPT 17 2013; Cathleen D. Zick & Stevens, 2010; C. D. Zick, Stevens, & Bryant, 2011). In this study,

270

time was the most significant issue for parents in the creative cooking group. The child

271

persuasion group rated some time-related items (cooking quick meals and meal planning) as

272

moderately important, while the healthy cooking group generally did not give importance to

273

time-related topics. The creative cooking group’s most preferred items (cooking without recipes

274

and culinary creativity, planning meals, and making meals out of leftovers) are all strategies that

275

will allow them to cook despite limited time. Interestingly though, all groups felt that learning to

276

make time to cook at home was a very low priority, suggesting that the pressures parents face are

277

unlikely to be malleable, and they must learn to prepare food efficiently within existing time

278

constraints. The creative cooking group had the highest socioeconomic status (SES), while the

279

healthy cooking group had the lowest SES. Based on these findings, we suggest teaching busy,

280

high SES parents like those in the creative cooking group how to prepare quick, healthy meals,

281

as well as other time-saving strategies such as preparing meals from leftovers. Interventions for

282

low SES parents like those in the healthy cooking group should focus on healthy cooking, while

283

bearing in mind that these parents also have constraints on their time. Unlike the other two

284

groups, the child persuasion group cooks regularly and has a bimodal SES distribution. This

285

suggests that once a family has established a regular cooking routine, convincing toddlers to eat

286

home-cooked food is challenging, regardless of SES. Parents in the child persuasion group

287

prioritized learning feeding strategies, consistent with research showing that parental feeding

288

style is a key aspect of the home food environment during early childhood (Savage et al., 2007).

SC

M AN U

TE D

EP

AC C

289

RI PT

269

290

The results also have implications for how to most effectively deliver key content. Many parents

291

prioritized learning about meal planning but felt that important components of meal planning,

ACCEPTED MANUSCRIPT 18 such as learning what to buy and how to shop, were very low priorities. Parents also highly

293

prioritized having professionally taught classes, rather than classes taught by community

294

members. This is inconsistent with prior studies showing the efficacy of peer mentors for

295

improving health behaviors (Dorgo, King, Bader, & Limon, 2012; Funnell, 2010; Long, Jahnle,

296

Richardson, Loewenstein, & Volpp, 2012). Though parental preferences may not always align

297

with evidence-based approaches, these approaches should not necessarily be excluded from

298

future interventions. Parental preferences should instead be taken into account during

299

intervention delivery. For instance, mundane tasks such as grocery shopping should be framed

300

within the overarching theme of meal planning, which parents find to be important. If

301

community members or peer mentors are delivering the intervention, then their on-the-ground

302

expertise should be presented in a professional light to maximize their acceptance by

303

participating parents.

M AN U

SC

RI PT

292

TE D

304

This study has three key limitations. First, we utilized a convenience sample of families with

306

young children, limiting generalizability of our findings. We deliberately focused on this age

307

group, however, because interventions to improve home food environments during early

308

childhood offer the greatest potential impact (Anzman et al., 2010; Institute of Medicine (U.S.).

309

Committee on Obesity Prevention Policies for Young Children. et al., 2011). Second, while

310

discrete choice methodology using maximum difference scaling offers many advantages, the

311

findings should still be considered in the context of prior work. Third, we included English-

312

speaking participants living in urban and suburban areas, which limits the generalizability of our

313

findings to non-English-speaking populations or those living in rural areas.

314

AC C

EP

305

ACCEPTED MANUSCRIPT 19 Conclusions

316

Discrete choice experiments using maximum difference scaling can be employed to efficiently

317

design and tailor interventions to change health behaviors. Regarding potential interventions to

318

improve home food preparation practices among English-speaking urban and suburban families

319

with young children, parents expressed relatively uniform views with regard to logistical

320

considerations. Parents diverged into three distinct groups with regard to content preferences;

321

they prioritized learning about health and nutrition, convincing their toddlers to eat home-cooked

322

food, or culinary creativity and time-saving strategies. These findings have implications for

323

designing and appropriately tailoring interventions to improve home food preparation practices

324

and home food environments among families with young children. Such interventions are

325

important for promoting healthier diets and preventing obesity starting from early childhood.

AC C

EP

TE D

M AN U

SC

RI PT

315

ACCEPTED MANUSCRIPT 20 ACKNOWLEDGEMENTS We would like to acknowledge the network of primary care clinicians as well as their patients and families for their contribution to this project and clinical research, facilitated through the Pediatric Research Consortium (PeRC) at The Children's Hospital of Philadelphia.

SC

RI PT

All phases of this study were supported by the Academic Pediatric Association Bright Futures Young Investigator Award, supported by the Maternal and Child Health Bureau in partnership with the American Academy of Pediatrics. SV was supported by a National Research Service Award institutional training grant for primary medical care, #T32-HP10026. The study sponsors had no role in study design, collection, analysis, and interpretation of data, writing the manuscript, or the decision to submit the manuscript for publication.

M AN U

SV and CF conceptualized and designed the study. TMP and SV acquired data. JAF and TMP analyzed the data, with guidance from SV, CF, and PJC. SV, PJC, JAF, TMP, and CF interpreted the data. SV and JAF drafted the initial manuscript. PJC, JAF, TMP, KT, and CF critically reviewed and revised the manuscript. All authors approved the final manuscript as submitted. None of the authors have any conflicts of interest to disclose.

AC C

EP

TE D

The contents of this article were presented at the Pediatric Academic Societies Annual Meeting on May 3, 2014 in Vancouver, BC.

ACCEPTED MANUSCRIPT 21 REFERENCES

AC C

EP

TE D

M AN U

SC

RI PT

Anzman, S. L., Rollins, B. Y., & Birch, L. L. (2010). Parental influence on children's early eating environments and obesity risk: implications for prevention. International Journal of Obesity, 34(7), 1116-1124. doi:10.1038/ijo.2010.43 Appelhans, B. M., Waring, M. E., Schneider, K. L., & Pagoto, S. L. (2014). Food preparation supplies predict children's family meal and home-prepared dinner consumption in lowincome households. Appetite, 76, 1-8. doi:10.1016/j.appet.2014.01.008 Benton, D. (2004). Role of parents in the determination of the food preferences of children and the development of obesity. International Journal of Obesity and Related Metabolic Disorders : Journal of the International Association for the Study of Obesity, 28(7), 858869. doi:10.1038/sj.ijo.0802532 Birch, L. L. (1999). Development of food preferences. Annual Review of Nutrition, 19, 41-62. doi:10.1146/annurev.nutr.19.1.41 Briefel, R. R., Wilson, A., & Gleason, P. M. (2009). Consumption of low-nutrient, energy-dense foods and beverages at school, home, and other locations among school lunch participants and nonparticipants. Journal of the American Dietetic Association, 109(2 Suppl), S79-90. doi:10.1016/j.jada.2008.10.064 Campbell, N. C., Murray, E., Darbyshire, J., Emery, J., Farmer, A., Griffiths, F., . . . Kinmonth, A. L. (2007). Designing and evaluating complex interventions to improve health care. British Medical Journal, 334(7591), 455-459. doi:10.1136/bmj.39108.379965.BE CDC. (2007). Consumer Behavior Questionnaire. Retrieved from http://www.cdc.gov/nchs/nhanes/nhanes2007-2008/questexam07_08.htm questionnaire Cohen, S. H. (2003). Maximum Difference Scaling: Improved Measures of Importance and Preference for Segmentation. Sawtooth Software: Research Paper Series. Sawtooth Software, Inc. Sequim, WA. Retrieved from http://www.sawtoothsoftware.com/download/techpap/maxdiff.pdf Couch, S. C., Glanz, K., Zhou, C., Sallis, J. F., & Saelens, B. E. (2014). Home Food Environment in Relation to Children's Diet Quality and Weight Status. Journal of the Academy of Nutrition and Dietetics. doi:10.1016/j.jand.2014.05.015 Deal, K. (2013). Segmenting Patients and Physicians Using Preferences from Discrete Choice Experiments. The Patient. doi:10.1007/s40271-013-0037-9 Dorgo, S., King, G. A., Bader, J. O., & Limon, J. S. (2012). Outcomes of a peer mentor implemented fitness program in older adults: A quasi-randomized controlled trial. International Journal of Nursing Studies. doi:10.1016/j.ijnurstu.2012.12.006 Evans, A., Chow, S., Jennings, R., Dave, J., Scoblick, K., Sterba, K. R., & Loyo, J. (2011). Traditional foods and practices of Spanish-speaking Latina mothers influence the home food environment: implications for future interventions. Journal of the American Dietetic Association, 111(7), 1031-1038. doi:10.1016/j.jada.2011.04.007 Funnell, M. M. (2010). Peer-based behavioural strategies to improve chronic disease selfmanagement and clinical outcomes: evidence, logistics, evaluation considerations and needs for future research. Family Practice, 27 Suppl 1, i17-22. doi:10.1093/fampra/cmp027 Gillman, M. W., Rifas-Shiman, S. L., Frazier, A. L., Rockett, H. R., Camargo, C. A., Jr., Field, A. E., . . . Colditz, G. A. (2000). Family dinner and diet quality among older children and

ACCEPTED MANUSCRIPT 22

AC C

EP

TE D

M AN U

SC

RI PT

adolescents. Archives of Family Medicine, 9(3), 235-240. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10728109 Glanz, K., Rimer, B. K., & Viswanath, K. (2008). Health Behavior and Health Education : Theory, Research, and Practice (4th ed.). San Francisco, CA: Jossey-Bass. Grisolia, J. M., Longo, A., Boeri, M., Hutchinson, G., & Kee, F. (2013). Trading off dietary choices, physical exercise and cardiovascular disease risks. Social Science & Medicine, 93, 130-138. doi:10.1016/j.socscimed.2013.05.031 Hagan JF, S. J., Duncan PM, eds. (2008). Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents (Third ed.). Elk Grove Village, IL: American Academy of Pediatrics. Institute of Medicine (U.S.). Committee on Obesity Prevention Policies for Young Children., Birch, L. L., Burns, A. C., & Parker, L. (2011). Early Childhood Obesity Prevention Policies. Washington, D.C.: National Academies Press. Jabs, J., & Devine, C. M. (2006). Time scarcity and food choices: an overview. Appetite, 47(2), 196-204. doi:S0195-6663(06)00381-3 [pii] 10.1016/j.appet.2006.02.014 Jones, S. A., Walter, J., Soliah, L., & Phifer, J. T. (2014). Perceived Motivators to Home Food Preparation: Focus Group Findings. Journal of the Academy of Nutrition and Dietetics. doi:10.1016/j.jand.2014.05.003 Kohan, E. G. (2011, October 25, 2011). Sam Kass: First Lady will acheive goal of eradicating all US food deserts by 2017. Obama Foodorama: The Blog Of Record About White House Food Initiatives, From Policy To Pie. Retrieved from http://obamafoodorama.blogspot.com/2011/10/sam-kass-first-lady-will-meetgoal-of.html Larson, N. I., Perry, C. L., Story, M., & Neumark-Sztainer, D. (2006). Food preparation by young adults is associated with better diet quality. Journal of the American Dietetic Association, 106(12), 2001-2007. doi:S0002-8223(06)02090-6 [pii] 10.1016/j.jada.2006.09.008 Lin B, G. J., Frazao E. (1999). Quality of children’s diets at and away from home- 1994-96. Food Review, 22, 2-10. Long, J. A., Jahnle, E. C., Richardson, D. M., Loewenstein, G., & Volpp, K. G. (2012). Peer mentoring and financial incentives to improve glucose control in African American veterans: a randomized trial. Annals of Internal Medicine, 156(6), 416-424. doi:10.1059/0003-4819-156-6-201203200-00004 Malhotra, K., Herman, A. N., Wright, G., Bruton, Y., Fisher, J. O., & Whitaker, R. C. (2013). Perceived benefits and challenges for low-income mothers of having family meals with preschool-aged children: childhood memories matter. Journal of the Academy of Nutrition and Dietetics, 113(11), 1484-1493. doi:10.1016/j.jand.2013.07.028 Martin-Biggers, J., Spaccarotella, K., Berhaupt-Glickstein, A., Hongu, N., Worobey, J., & ByrdBredbenner, C. (2014). Come and get it! A discussion of family mealtime literature and factors affecting obesity risk. Advances in Nutrition, 5(3), 235-247. doi:10.3945/an.113.005116 The MaxDiff/Web v6.0 Technical Paper. (2007). Sawtooth Software Inc. Sequim, WA. Retrieved from http://www.sawtoothsoftware.com/download/techpap/maxdifftech.pdf Monsivais, P., Aggarwal, A., & Drewnowski, A. (2014). Time Spent on Home Food Preparation and Indicators of Healthy Eating. American Journal of Preventive Medicine. doi:10.1016/j.amepre.2014.07.033

ACCEPTED MANUSCRIPT 23

AC C

EP

TE D

M AN U

SC

RI PT

Ogata, B. N., & Hayes, D. (2014). Position of the academy of nutrition and dietetics: nutrition guidance for healthy children ages 2 to 11 years. Journal of the Academy of Nutrition and Dietetics, 114(8), 1257-1276. doi:10.1016/j.jand.2014.06.001 Paulhus, D. L. (1991). Measurement and Control of Response Bias. San Diego: Academic Press. Pelletier, J. E., & Laska, M. N. (2012). Balancing healthy meals and busy lives: associations between work, school, and family responsibilities and perceived time constraints among young adults. Journal of Nutrition Education and Behavior, 44(6), 481-489. doi:10.1016/j.jneb.2012.04.001 Quick, B. L., Fiese, B. H., Anderson, B., Koester, B. D., & Marlin, D. W. (2011). A formative evaluation of shared family mealtime for parents of toddlers and young children. Health Communication, 26(7), 656-666. doi:10.1080/10410236.2011.561920 Rosenkranz, R. R., & Dzewaltowski, D. A. (2008). Model of the home food environment pertaining to childhood obesity. Nutrition Reviews, 66(3), 123-140. doi:10.1111/j.17534887.2008.00017.x Savage, J. S., Fisher, J. O., & Birch, L. L. (2007). Parental influence on eating behavior: conception to adolescence. The Journal of Law, Medicine & Ethics : A Journal of the American Society of Law, Medicine & Ethics, 35(1), 22-34. doi:10.1111/j.1748720X.2007.00111.x Schwarz, G. (1978). Estimating the Dimension of a Model. Annals of Statistics, 6(2), 461-464. Retrieved from http://projecteuclid.org/DPubS?service=UI&version=1.0&verb=Display&handle=eu clid.aos/1176344136 Smith, L. P., Ng, S. W., & Popkin, B. M. (2013). Trends in US home food preparation and consumption: analysis of national nutrition surveys and time use studies from 1965-1966 to 2007-2008. Nutrition Journal, 12, 45. doi:10.1186/1475-2891-12-45 Stephens, L. D., McNaughton, S. A., Crawford, D., & Ball, K. (2014). Predictors of high-energy foods and beverages: a longitudinal study among socio-economically disadvantaged adolescents. Public Health Nutrition, 17(2), 324-337. doi:10.1017/S136898001200482X Storfer-Isser, A., & Musher-Eizenman, D. (2013). Measuring parent time scarcity and fatigue as barriers to meal planning and preparation: quantitative scale development. Journal of Nutrition Education and Behavior, 45(2), 176-182. doi:10.1016/j.jneb.2012.08.007 Swanson, V., Power, K. G., Crombie, I. K., Irvine, L., Kiezebrink, K., Wrieden, W., & Slane, P. W. (2011). Maternal feeding behaviour and young children's dietary quality: a crosssectional study of socially disadvantaged mothers of two-year old children using the Theory of Planned Behaviour. The International Journal of Behavioral Nutrition and Physical Activity, 8, 65. doi:10.1186/1479-5868-8-65 Sweetman, C., McGowan, L., Croker, H., & Cooke, L. (2011). Characteristics of family mealtimes affecting children's vegetable consumption and liking. Journal of the American Dietetic Association, 111(2), 269-273. doi:10.1016/j.jada.2010.10.050 Taveras, E. M., Rifas-Shiman, S. L., Berkey, C. S., Rockett, H. R., Field, A. E., Frazier, A. L., . . . Gillman, M. W. (2005). Family dinner and adolescent overweight. Obesity Research, 13(5), 900-906. doi:10.1038/oby.2005.104 U.S.D.A. (2013). Build a Healthy Plate With Less Salt and Sodium. Retrieved from http://www.fns.usda.gov/sites/default/files/salt_sodium.pdf. Virudachalam, S., Long, J. A., Harhay, M. O., Polsky, D. E., & Feudtner, C. (2014). Prevalence and patterns of cooking dinner at home in the USA: National Health and Nutrition

ACCEPTED MANUSCRIPT 24

AC C

EP

TE D

M AN U

SC

RI PT

Examination Survey (NHANES) 2007-2008. Public Health Nutrition, 17(5), 1022-1030. doi:10.1017/S1368980013002589 Zick, C. D., & Stevens, R. B. (2010). Trends in Americans' food-related time use: 1975-2006. Public Health Nutrition, 13(7), 1064-1072. Retrieved from http://ovidsp.ovid.com/ovidweb.cgi?T=JS&CSC=Y&NEWS=N&PAGE=fulltext&D=me dl&AN=19943999 Zick, C. D., Stevens, R. B., & Bryant, W. K. (2011). Time use choices and healthy body weight: a multivariate analysis of data from the American Time Use Survey. The International Journal of Behavioral Nutrition and Physical Activity, 8, 84. doi:10.1186/1479-5868-884

ACCEPTED MANUSCRIPT 25 Table 1: Study participant and household characteristics n (%) ADULTS Total Gender

80 (100%) 68 (85%) 12 (15%)

RI PT

Female Male Age (years) 21-25 26-30 31-35 36-45

20 (25%) 17 (21%) 21 (26%) 22 (28%)

Country of birth USA Other Education level

M AN U

White Black Latino Other

SC

Race/ethnicity

AC C

EP

TE D

Some high school High school grad or GED Some college College graduate Professional degree Employment status Employed Unemployed Insurance payer Public Private Weight category (BMI) Underweight (BMI < 18.5) Normal (18.5 ≤ BMI < 25) Overweight (25 ≤ BMI < 30) Obese (BMI ≥ 30) Person who plans and prepares most family meals Yes No Currently receiving WIC benefits Yes No Currently receiving SNAP benefits Yes

37 (46%) 30 (38%) 9 (11%) 4 (5%)

70 (87%) 10 (13%) 7 (9%) 24 (30%) 22 (28%) 10 (12%) 17 (21%) 55 (69%) 25 (31%) 44 (55%) 36 (45%) 1 (1%) 35 (44%) 26 (32%) 18 (23%) 74 (93%) 6 (7%) 23 (29%) 57 (71%) 32 (40%)

ACCEPTED MANUSCRIPT 26 No

48 (60%)

RI PT

19 (24%) 14 (17%) 7 (9%) 11 (14%) 29 (36%)

EP

CHILDREN Total Gender

TE D

Female Male

Age (months)

AC C

12-23 24-35 36-47 48-59 Weight category (WHO BMI percentile) Underweight (BMI < 5%) Normal (5% ≤ BMI < 85%) Overweight (85% ≤ BMI < 95%) Obese (BMI ≥ 95%) Location of daytime childcare Home Daycare or preschool

21 (26%) 27 (34%) 32 (40%)

SC

M AN U

HOUSEHOLD Annual household income ≤ $22,000 $22,000 - $32,999 $33,000 - $54,999 $55,000 - $74,999 ≥ $75,000 Household size 2-3 4 ≥5 Other adults in household No Yes, romantic partner only Yes, romantic partner and other adults Yes, other adults only Number of children in household 1 2 ≥3 Frequency of cooking dinner at home Never (0-1 night per week) Sometimes (2-5 nights per week) Always (6-7 nights per week)

11 (13%) 41 (51%) 14 (18%) 14 (18%) 26 (32%) 32 (40%) 22 (28%)

6 (8%) 40 (50%) 34 (42%)

87 (100%) 51 (59%) 36 (41%) 28 (32%) 21 (24%) 17 (20%) 21 (24%) 1 (1%) 53 (61%) 18 (21%) 15 (17%) 49 (56%) 38 (44%)

ACCEPTED MANUSCRIPT 27

AC C

EP

TE D

M AN U

SC

RI PT

Table 2: Adult, household, and child characteristics for intervention content preference subgroups Healthy Child Creative p-value Cooking Persuasion Cooking n (%) n (%) n (%) ADULTS Total 21 (26%) 34 (43%) 25 (31%) Gender 0.16 Female 15 (71%) 30 (88%) 23 (92%) Male 6 (29%) 4 (12%) 2 (8%) Age (years) 0.04 21-25 11 (52%) 7 (21%) 2 (8%) 26-30 3 (14%) 7 (21%) 7 (28%) 31-35 2 (10%) 10 (29%) 9 (36%) 36-45 5 (24%) 10 (29%) 7 (28%) Race/ethnicity 0.17 White 5 (24%) 18 (53%) 14 (56%) Black 13 (62%) 9 (27%) 8 (32%) Latino 2 (9%) 5 (15%) 2 (8%) Other 1 (5%) 2 (5%) 1 (4%) Country of birth 0.63 USA 19 (90%) 28 (82%) 23 (92%) Other 2 (10%) 6 (18%) 2 (8%) Education level 0.19 Some high school 3 (13%) 2 (6%) 2 (8%) High school grad or GED 10 (48%) 9 (26%) 5 (20%) Some college 6 (29%) 10 (29%) 6 (24%) College graduate 0 6 (18%) 4 (16%) Professional degree 2 (10%) 7 (21%) 8 (32%) Employment status 0.66 Unemployed 8 (38%) 9 (26%) 8 (32%) Employed 13 (62%) 25 (74%) 17 (68%) Insurance payer 0.004 Public 18 (86%) 15 (44%) 11 (44%) Private 3 (14%) 19 (56%) 14 (56%) Weight category (BMI) 0.54 Underweight (BMI < 18.5) 0 1 (3%) 0 Normal (18.5 ≤ BMI < 25) 8 (38%) 15 (44%) 12 (48%) Overweight (25 ≤ BMI < 30) 5 (24%) 12 (35%) 9 (36%) Obese (BMI ≥ 30) 8 (38%) 6 (18%) 4 (16%) Person who plans and prepares most meals in the family 0.27 Yes 18 (86%) 33 (97%) 23 (92%) No 3 (14%) 1 (3%) 2 (8%) Currently receiving WIC benefits 0.02 Yes 11 (52%) 8 (24%) 4 (16%)

ACCEPTED MANUSCRIPT 28 26 (76%)

21 (84%) 0.01

14 (67%) 7 (33%)

12 (35%) 22 (65%)

9 (43%) 3 (14%) 2 (10%) 1 (5%) 3 (14%) 3 (14%)

7 (21%) 8 (24%) 1 (3%) 0 4 (12%) 14 (40%)

6 (24%) 19 (76%)

5 (24%) 8 (38%) 8 (38%)

8 (24%) 11 (32%) 15 (44%)

8 (32%) 8 (32%) 9 (36%)

5 (24%) 5 (24%) 5 (24%) 6 (28%)

5 (15%) 20 (59%) 6 (18%) 3 (8%)

1 (4%) 16 (64%) 3 (12%) 5 (20%)

RI PT

0.05

10 (48%) 7 (33%) 4 (19%)

TE D

3 (12%) 3 (12%) 0 3 (12%) 4 (16%) 12 (48%)

SC

HOUSEHOLD Annual household income ≤ $22,000 per year $22,000 - $32,999 per year $33,000 - $43,999 per year $44,000 - $54,999 per year $55,000 - $74,999 per year ≥ $75,000 per year Household size 2-3 4 ≥5 Other adults in household No Yes, romantic partner only Yes, romantic partner and other adults Yes, other adults only Number of children in household 1 2 ≥3 Frequency of cooking dinner at home Never (0-1 night per week) Sometimes (2-5 nights per week) Always (6-7 nights per week)

10 (48%)

M AN U

No Currently receiving SNAP benefits Yes No

0.05

0.45 8 (32%) 9 (36%) 8 (32%) 0.02 2 (8%) 17 (68%) 6 (24%)

23 (100%) 38 (100%)

26 (100%)

13 (57%) 10 (43%)

19 (50%) 19 (50%)

19 (73%) 7 (27%)

12-23 8 (35%) 24-35 4 (17%) 36-47 5 (22%) 48-59 6 (26%) Weight Category (WHO BMI percentile) Underweight (BMI < 5%) 0 Normal (5% ≤ BMI < 85%) 12 (52%) Overweight (85% ≤ BMI < 95%) 8 (35%)

11 (29%) 8 (21%) 7 (18%) 12 (32%)

9 (35%) 9 (35%) 5 (19%) 3 (11%)

EP

3 (9%) 10 (29%) 21 (62%)

CHILDREN

Total

AC C

Gender

Female Male

1 (5%) 13 (62%) 7 (33%)

8 (24%) 16 (46%) 10 (30%)

0.93

0.18

Age (months)

0.59

0.43 0 24 (63%) 6 (16%)

1 (4%) 17 (66%) 4 (15%)

ACCEPTED MANUSCRIPT 29

AC C

EP

TE D

M AN U

SC

RI PT

Obese (BMI ≥ 95%) 3 (13%) 8 (21%) 4 (15%) Location of daytime childcare 0.54 Home 12 (52%) 24 (63%) 13 (50%) Daycare or preschool 11 (48%) 14 (37%) 13 (50%) Bold text indicates that differences between subgroups are statistically significant (p≤0.05) by Fisher’s exact test.

ACCEPTED MANUSCRIPT 30 FIGURE CAPTIONS

RI PT

Text Box: Discrete choice experiment items rated by adult caregivers Two separate maximum difference scaling experiments were conducted to assess parental preferences for intervention content and logistical considerations. The questions and answer choices presented to parents are shown here.

SC

Figure 1: Parental preferences for intervention content and logistical considerations Maximum difference scaling results for parental preferences regarding intervention content (1a) and logistical considerations (1b) are shown, including the rank, mean rating score, and 95% CI for each item. Scores are relative, on a scale from 1 to 100. For instance, a score of 10 indicates that the item is preferred twice as much as an item with a score of 5.

AC C

EP

TE D

M AN U

Figure 2: Intervention content preferences for each latent class subgroup The rating scores for several intervention content items are shown for each latent class subgroup. The subgroups had similar preferences for some items, but diverging preferences for other items.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

Highlights

AC C

EP

TE D

M AN U

SC

• • • •

Parents of young children have varied preferences for cooking-related interventions Parents prefer interventions that focus on one of the following themes: Healthy cooking – preferred by lower-income, single, younger parents Culinary creativity – preferred by higher-income, partnered, older parents Convincing toddlers to eat homemade food – preferred by parents who cook regularly

RI PT

•