Healthy Home Offerings via the Mealtime Environment (HOME): Feasibility, acceptability, and outcomes of a pilot study

Abstract

The primary objective was to develop and test the feasibility and acceptability of the Healthy Home Offerings via the Mealtime Environment (HOME) program, a pilot childhood obesity prevention intervention aimed at increasing the quality of foods in the home and at family meals. Forty-four child/parent dyads participated in a randomized controlled trial (n=22 in intervention and n=22 in control conditions). The intervention program, held at neighborhood facilities, included five, 90-minute sessions consisting of interactive nutrition education, taste-testing, cooking skill building, parent discussion groups, and hands-on meal preparation. Children (8–10 year olds) and parents (89% mothers), completed assessments at their home at baseline, post-intervention, and 6-month follow-up, including psychosocial surveys, anthropometry, 24-hour dietary recalls, and home food availability and meal offering inventories. Feasibility/acceptability was assessed with participant surveys and process data. All families completed all three home-based assessments. Most intervention families (86%) attended at least 4 of 5 sessions. Nearly all parents (95%) and 71% of children rated all sessions very positively. General linear models indicated that at post-intervention, compared to control children, intervention children were significantly more likely to report greater food preparation skill development (p<.001). There were trends suggesting that intervention children had higher consumption of fruits and vegetables (p <.08), and higher intakes of key nutrients (all p-values <.05) than control children. Obesity changes did not differ by condition. Not all findings were sustained at 6-month follow-up. Obesity prevention programming with families in community settings is feasible and well accepted. Results demonstrate the potential of the HOME program.

Introduction

Few successful childhood obesity prevention strategies exist (1), particularly those focused on the home or families. Relatively few randomized trials have been developed with the primary purpose of preventing childhood obesity (2–5), and most have not significantly affected weight gain (6,7). Although the most effective childhood obesity prevention programs generally include a family component, many programs focused on changing the children’s behavior and limited family involvement to the delivery of psycho-educational materials (6). In contrast, obesity treatment programs have shown that involving families is associated with improved child weight outcomes (8,9). Given the importance of the family environment in regard to healthy eating and activity, there is a dearth of childhood obesity prevention research that focuses on family units.

Parents are influential role models for healthy eating, and are gatekeepers for food and beverage availability and accessibility within the home (10). Almost 70% of calories and 80% of snacks consumed by 6–11 year old children are eaten in the home (11). Consumption of healthful (12,13) and unhealthful (14) foods and beverages is significantly predicted by their home availability. Furthermore, family purchases of convenient, unhealthful, away-from-home foods have increased dramatically over time (15,16) and may negatively affect the nutritional quality of family members’ diets (17), making it difficult for children to meet current dietary and weight recommendations (18,19). Thus, encouraging families to eat at home more often while positively changing the home food environment appears to be one way to decrease children’s consumption of unhealthful foods.

A review of the natural history of childhood obesity, and a study that examined the contribution of energy intake in the evening meal to longitudinal change in weight among children, indicate that the family mealtime has the potential to impact children’s dietary intake and obesity rates (1,20). Other research has shown significant positive associations between family meal frequency and nutritional intake and fruit and vegetable intake among children (21,22) and adolescents (23,24), and inverse associations with consumption of soft drinks and high-fat foods among children (21). In spite of these findings, associations between family meal frequency and obesity have been inconsistent across studies (21,22,25–28). More research is needed to fully understand these relationships; for example, associations may be mediated by the quality of the foods served at meals. Interventions to increase the frequency of family meals and the quality of the foods served are needed and may provide a fruitful avenue for the prevention of childhood obesity. Furthermore, although community-based obesity prevention programs may be difficult to implement, they are needed (29,30) and may be most conducive for families. The Healthy Home Offerings via the Mealtime Environment (HOME) pilot study was undertaken to assess the feasibility and acceptability of a community-based obesity prevention program that aimed at increasing the quality of foods in the home and at family meals among families of school-age children.

Methods and Procedures

Study Design

The HOME pilot study was designed as a two-arm randomized trial. The primary aim of the study was to develop, implement, and test the feasibility and acceptability of the HOME program. The secondary aims were to acquire preliminary data that would support the design of a larger trial. Thus, our secondary aims were that by the end of the 3-month intervention, relative to those in the control condition, households in the intervention condition would have more fruits and vegetables, and fewer high-fat foods and high-sugar foods and beverages available in the home and served at family meals; and target children in the intervention families would have higher fruit and vegetable consumption, lower fat and sugar intake, and lower age- and gender-specific body mass index (BMI) percentile values.

Participants

Parent/child dyads were recruited from two elementary schools/after-school programs via flyers, school newsletters, and small group presentations. After-school program staff were hired on a limited basis to aid recruitment efforts and provide childcare services during the intervention sessions. The parent/guardian that prepared most of the household meals and one 8–10 year old child were recruited per household. Targeting the primary meal-preparing parent was expected to increase the chance of making changes in the home food environment. The 8–10 year old developmental stage for target children was chosen to intervene on the dietary habits at a crucial developmental stage when children are involved in many activities and begin to assert some independence in managing their weight, but are at risk for obesity (4,31); moreover, the validity of assessment tools improves with this age group.

Interested parents (n=50) were directed to contact the project director by phone, email, or in-person for eligibility screening. Children were ineligible if they were very underweight, had conditions that would affect intervention program participation, and did not speak English. Six families were excluded (1 did not meet criteria, 5 declined after screening), leaving 44 families to be randomized to an intervention (n=22) or control (n=22) condition.

Assessments were conducted with all families at their homes at three time points (baseline, post-intervention, and 6-month follow-up). Assessments included psychosocial surveys, anthropometry, 24-hour dietary recalls, home food inventory, and a family meal inventory. Following standardized protocols, the project director scheduled home visits and collected study consent/assent and all survey and anthropometric data from target parents and children. Nutrition Data System for Research (NDS-R) certified staff conducted three dietary recall interviews with each target child. Baseline assessments at the participants’ homes took 1–2 hours, with shorter durations at follow-up visits. Families received $50 gift cards for each assessment (total of $150/family).

Randomization occurred after baseline assessments. None of the baseline demographic or weight-related characteristics differed significantly by condition. Families randomized to the intervention condition participated in HOME program sessions delivered in community settings (see below). Families randomized to the control condition participated in home assessments only and were sent written intervention materials at the end of the study. The research was approved by the Institutional Review Board of the University of Minnesota and the Research, Evaluation, and Assessment Board of the Minneapolis Public School District.

Intervention Program

Social Cognitive Theory (SCT) (32) provided the conceptual framework for the intervention program. In SCT, behavior is explained as dynamic, reciprocal interactions between personal factors, environmental factors, and behavior. Based on previous research (31,33) and formative qualitative work with parents prior to the development of the HOME program, we proposed that home environmental factors (e.g., home food availability) and family behavioral factors (e.g., cooking skills) impact a child’s eating behaviors (e.g., consumption of fruits and vegetables) and that these eating behaviors influence weight. Efforts to change the modifiable behaviors that are under parents’ control such as meal preparation and the types of foods served and accessible in the home meet the needs of parents (34) and are logical targets for prevention of obesity in the home (1). In our formative work, parents indicated that their preference was for evening programming in a form other than a lecture format where they could be with their children.

The intervention program was implemented by the study authors and trained students after baseline assessments were completed. Sessions were held at rented space in a church and community center (with kitchen and dining facilities) within close proximity to participants’ homes in the early evening (6–7:30pm). Families participated in five 90-minute intervention sessions in a multiple family-group format (3–8 families at one time). All family members (other adults and siblings) were encouraged to attend the program. Babysitting was available for children (< 8 years). Each session was offered to families twice at each location within a 2-week period to allow for scheduling flexibility.

The intervention session components are summarized on Table 1. Each session included a healthy snack, separate parent and child group time, family meal preparation, interactive nutrition education activities, a group meal, homework assignment, take home materials, and session evaluations. Activities were hands-on and interactive. Parent group time enabled parents to learn from each other in regards to dealing with picky eaters, meal planning, etc. Child group time included taste-testing, along with learning meal planning and cooking skills. The intervention components at each session focused on a specific topic (e.g., increasing fruits and vegetables). In choosing entrée recipes for the program, we aimed for <30% fat contribution, at least one serving of fruits/vegetables, and <30 minutes preparation time. To increase study retention among the intervention participants, a trained interventionist conducted an encouragement call to parents after the second session to identify barriers to meeting goals and to provide support. Door prize drawings were held at several sessions. Families who missed a session were sent an abbreviated written version of the intervention materials. Process data were collected at each session by program staff.

Table 1.

Overview of HOME Intervention Session Goals and Targets

Session	Session Name	Intervention Goals	Intervention Targets
1	Thinking Outside the Box	↑frequency of family meals and create healthful environment ↑child cooking skills ↓high-fat foods	Overcoming obstacles to family meals Child cooking skills (hygiene and vegetable peeling/chopping) Reducing fat (lower fat pizzas)
2	What’s for Dinner 2nights?	↑ fruits and vegetables ↑child cooking skills	Parent self-efficacy for meal planning and cooking Child cooking skills (meal preparation, measuring) Increasing fruits and vegetables in meals and home environment
3	Keep it Under Wraps	Resolving family conflict	Parental strategies for dealing with picky eaters Food and beverage advertising awareness Child taste-testing (fruits and vegetables) Child cooking skills (making wraps) Quick, healthy meals (wrap sandwiches)
4	AGreenable Dinners	↓high-fat foods ↑ fruits and vegetables	Parental skills for infusing meals with vegetables and reducing sweetened beverages Reducing high fat foods in home Child cooking skills (vegetable preparation and vegetarian lasagna) Child taste-testing (fruits and vegetables)
5	SOS Spells Soups or Sauce	↓ high-sugar foods and beverages	Reducing portion sizes Child cooking skills (salad preparation)

Process Measures

Recommended process data (35) were collected to examine the primary aims of program feasibility and acceptability. At each session, HOME staff collected attendance data at the family- and individual-level as well as homework completion (five assignments) to assess program dose, and implementation data of each component to assess intervention fidelity. A post-intervention survey assessed child and parent satisfaction of each component (on a 5-point satisfaction scale from very satisfied to very dissatisfied).

Outcome Measures

All outcome measures were collected in participants’ homes by trained staff to assess the study’s secondary aims.

Family Dinner Frequency and Source of Foods

The number of family dinners per week (0–7) was assessed with items from previous research (24,36).

Parental Self-Efficacy

A 9-item scale assessed parental self-efficacy regarding making healthful changes in the home (e.g., serving vegetables at every dinner, reducing soda in home). Scale scores ranged from 23–36 and had high internal consistency (α=.80) (3), with slightly lower reliability in the present sample (α=.67).

Child Food Preparation Skill

Child food preparation skills were assessed with seven items (e.g., child is able to …peel fruits/vegetables, …make a green salad) on both parent and child surveys. Responses were on a 5-point agree/disagree scale (parents) or a 3-point scale (“not true for me, “sort of true for me,” “very true for me”; children). A summative Child Food Preparation Skill scale was calculated (scores ranged from 13–35 for parents and from 8–21 for children; and had high internal consistency reliability, α=.75 for parents and α=.80 for children). Parents were also asked if their child helps make family dinner (1=never/rarely, 2=some days, 3=most days, 4=every day).

Anthropometry

Children’s objective height and weight measurements were collected in duplicate using a standardized protocol (37). Body mass index (BMI; weight (kg)/height (cm)²) and age- and gender-adjusted BMI percentiles were calculated (ANTHRO 1.01 software-CDC).

Home Food Availability

The target parent completed a Home Food Inventory (HFI) that has demonstrated substantial construct and criterion validity (38) to assess the number of types of fruits, vegetables, sweetened beverages, and high-fat foods available in the home.

Nutrition Quality of Foods Served at Family Meals

Meal quality was assessed with the Brief Mealtime Screener (BMS), developed specifically for this pilot intervention. The screener assessed types of foods/beverages served at dinner and has substantial construct and criterion validity (kappa=0.79–1.0 between participant and staff report) and was used to evaluate mealtime fruit and vegetable offerings. Parents completed the BMS for the seven days following each of the home assessments and returned them by mail. We calculated the percentage of dinners where fruits and vegetables were served.

Dietary Assessment

Three 24-hour recall interviews, the gold standard for dietary intake, were conducted using the multiple pass method with each target child at each of the assessments to provide an accurate measure of individual intake (39,40). At each assessment period, serving and nutrient data from three nonconsecutive days (two weekdays and one weekend day) were averaged. The first recall was conducted in-person at the child’s home, and the other two were conducted by telephone within 2 weeks. Parental assistance was permitted for clarification during the interviews (3). All data were collected with NDS-R software using a laptop computer.

Demographics

Parents completed demographic items regarding family SES, including race/ethnicity and date of birth of parents and children, age of all children living in the home, education of parent(s), number of adults in household, marital status of parent(s), employment status of parent(s) and hours worked, and biological/other relationship with child.

Statistical Analysis

As purported by Baranowski and colleagues (34), pilot study evaluations should focus on participation bias, feasibility and process evaluation to assess whether an adequate dose of a high quality intervention was delivered to an acceptable number of participants. To test the study’s primary feasibility aim, process measures of dose and fidelity were evaluated with frequencies. Frequencies of parent and child satisfaction were calculated to assess program acceptability.

The intention of the proposed pilot study was not to conclusively test for an intervention effect on child BMI because a larger randomized trial with more power is necessary for such an analysis. However, we proceeded with preliminary analysis of our pilot data by conducting statistical comparisons of our secondary outcomes by condition. General linear models were used to assess between-group differences in outcomes (adjusted for baseline levels of the outcome).

Results

Participant Characteristics

Child participants were 8–10 years old; 84% were Caucasian, followed by 11% mixed race/ethnicity, 5% American Indian, and 2% African American; 52% were female; and 30% were overweight (>85^th BMI percentile). Target mothers (89%) and fathers (11%) participated in the assessments; 86% of parents were Caucasian, followed by 7% mixed, 5% American Indian, and 2% African American. Many parents were college educated (75%), 52% were working full-time, and the average age was 41.5 years (SD=5.3).

Primary aims

All 44 families completed all three home assessments (100% retention over one calendar year). All families were successfully randomized to conditions. Eighty-six percent of intervention families attended at least four of five sessions (session completion was as follows: All 5 sessions (n=15 families), 4 sessions (n=4 families), 3 sessions (n=1 family), 2 sessions (n=1 family), and one family received only written materials). At each session, 5–10 adult partners (35–50% of partners from two-parent households) and 2–7 siblings participated in the program sessions, and 10–14 children attended childcare. Moreover, parents and children were very satisfied with the program, with 95% of parents and 71% of children rating sessions a “4” or “5” on a scale of 1–5 (1=didn’t like, 5=loved it). Family homework completion was as follows: all five (n=6), four (n=8), 1–3 (n=4), none (n=4).

Secondary aims

As shown in Table 2, families in both conditions reported an average of five family dinners per week at post-intervention. At post-intervention, children in the intervention group were significantly more likely than children in the control group to report gaining food preparation skills. In addition, analyses indicated trends for greater parental report of child food preparation skills, and reports of their own self-efficacy regarding meal preparation were in the expected direction. Intervention parents also reported that their children helped make dinner significantly more often than control parents (M=1.9 vs. M=1.6, p<.01).

Table 2.

Baseline to post-intervention psychosocial outcomes by condition

Survey scale scores	Intervention (n=22) Mean (SD)	Control (n=22) Mean (SD)	GLM1
Family dinner frequency (days per week)	5.0 (1.0)	5.0 (1.9)	p=.87
Parental self-efficacy for making healthful changes scale (parent report)	27.3 (3.2)	26.9 (4.0)	p=.55
Child food preparation skill scale (parent report)	28.6 (3.3)	26.9 (4.3)	p=.11
Child food preparation skill scale (child report)	17.5 (3.2)	15.2 (3.2)	p<.01

¹General Linear Model analysis by condition (adjusted for baseline level of outcome)

Note: SD=Standard deviation

As shown in Table 3, analyses indicated trends of higher fruit and vegetable availability (p=.12), and lower availability of quick, high-fat microwaveable foods (p=.11), and processed meats (p=.11) in intervention condition homes than control homes. Other positive food availability outcomes (sweetened beverages and high-sugar cereal) were in the expected direction. Similarly, analyses indicate that the percentage of family dinners in which vegetables and salads were served were in the expected direction. A trend of higher intakes of fruits and vegetables (p=.08) among children in the intervention condition compared to those in the control condition was also apparent. Positive nutrient-level outcomes were evident and in the expected direction, including reductions in percent of calories from fat, and increases in fiber, calcium, and Vitamins B6 and B12. At post-intervention, children in the intervention condition did not have significantly lower BMI percentiles or BMI z-scores (M=58.7, M=0.30) compared to children in the control condition (M=56.6, M=0.26).

Table 3.

Baseline to post-intervention mean changes in home food availability and child dietary outcomes by condition.

	Intervention (n=22)	Control (n=22)	GLM1
Home Food Availability (number of available distinct types of items)	Mean (SD)	Mean (SD)
Fruits & vegetables	23.7 (5.4)	21.7 (6.0)	p=0.12
Sweetened beverages	2.2 (1.4)	2.5 (1.1)	p=0.36
High sugar cereal	1.7 (1.0)	1.8 (0.9)	p=0.67
Quick, high-fat microwaveable foods	2.3 (1.4)	3.0 (1.6)	p=0.11
Processed meats	1.8 (1.3)	2.3 (1.3)	p=0.11
High-fat snacks	5.4 (1.5)	4.9 (1.7)	p=0.24
Family Meal Offerings	%	%
Vegetables (percent of times served at family dinner)	76.5	69.1	p=0.39
Fruit (percent of times served at family dinner)	59.1	68.5	p=0.43
Salad (percent of times served at family dinner)	34.2	29.4	p=0.61
Child Dietary Intake	Mean (SD)	Mean (SD)
Servings of fruits & vegetables	3.5 (1.7)	2.6 (1.3)	p=0.08
Servings of sweetened beverages	0.4 (0.4)	0.7 (0.5)	p=0.09
Energy (kcals)	1780.9 (352)	1740.2 (308)	p=0.45
% carbohydrate	57.4 (4.1)	54.4 (3.8)	p=0.02
% fat	30.4 (4.1)	32.1 (3.6)	p=0.34
% protein	14.0 (1.9)	14.8 (2.5)	p=0.14
Dietary fiber (g)	15.8 (3.3)	12.7 (3.5)	p<0.01
Vitamin B6 (mg)	1.6 (0.5)	1.4 (0.3)	p=0.10
Vit B12 (mcg)	4.8 (1.5)	4.4 (1.3)	p=0.28
Calcium (mg)	1144 (336)	961 (303)	p=0.05

¹General Linear Model analysis by condition (adjusted for baseline level of outcome)

Note: SD=Standard deviation

Discussion

The primary aims of the present study were to assess the feasibility and acceptability of the HOME pilot program. We met these aims by successfully recruiting families (not just parents OR children), retaining all families in three home assessments over one year, and engaging families in a group intervention format with a high participation rate. In addition, although the pilot study was not designed to power analyses of behavioral outcomes by condition, our post-intervention findings regarding psychosocial attitudes and behavioral outcomes by condition indicate that the HOME program may be a promising program; and thus support the need for a sufficiently powered, larger trial.

Our finding that intervention children were significantly more likely to report food preparation skill development than control children and that there was a trend for the same finding from parental report means that children were learning some of the key cooking skills taught in the program. Furthermore, they were extending these skills to mealtimes as intervention parents were significantly more likely than control parents to report that their children helped make dinner.

The increase of more than three-quarters of a serving of fruits/vegetables by the intervention group suggests that the HOME program may promote fruit and vegetable consumption. The positive nutrient-level outcomes support this interpretation.

Our null findings in regard to child BMI percentiles or z-scores by condition highlight the difficultly impacting excess weight with relatively brief intervention programs. The HOME pilot intervention dose of five sessions and relatively short duration may not have been long enough to get an impact on major outcomes such as weight. In addition, pilot studies such as HOME are usually small and underpowered by design. However, the promising findings regarding changes in home food availability and child dietary intake indicate that a trial with a larger sample may produce statistically significant findings. Perhaps a more intensive, longer intervention is needed to maintain or produce changes in body weight.

The present study had several strengths. To the best of our knowledge, although community organizations and researchers have been interested in promoting family meals, an evaluation of such a program has not been previously described in the scientific literature. Moreover, the present study findings show that even with a small sample, the HOME program was successful in showing trends in increasing a healthful home food environment. These findings were very encouraging given the pilot nature of the study and our small sample size. Furthermore, the measures used were of high quality with many demonstrating validity and reliability.

The study also had limitations in addition to those inherent in a pilot study (e.g., limited power). The study sample was predominantly Caucasian and highly educated, and included families who were already eating frequent meals together, limiting generalizability of the HOME program to other cultural groups, less educated populations, and those who eat few family meals. An intervention program like HOME that recruited more diverse families may need to focus more on making nutritious meals on a budget and increasing family meal frequency. In addition, a larger budget would have allowed for separate evaluation and intervention staff to lessen the likelihood of demand characteristics among participants.

In summary, the HOME pilot study was successful in meeting its aims, and provided many lessons learned to help inform community-based obesity prevention trials and the implementation of a full-scale trial. For example, although recruitment was successful, in future work, it would be important to expand recruitment sites to increase racial/ethnic and socioeconomic diversity. Thus, adaptation of the HOME program may be warranted (as described above). Given the 100% retention for the home assessments over the course of the study, it appears that the in-home assessment methodology likely contributed to the retention rate and is a fruitful method for community-based trials. The home visits were equally valued by both control and intervention families. In addition, the high number of non-target family members attending the intervention sessions and the number of families taking advantage of childcare reinforced expectations that inviting the entire family increased retention in the program. Similarly, the implementation of the intervention in convenient neighborhood locations likely contributed to the high session attendance rates. Thus, obesity prevention programming can be successfully implemented in the community by making it convenient for families. Furthermore, Park and Recreation Boards promote health by providing recreational services; thus, childhood obesity research partnerships may be mutually beneficial to the collaborating organizations and participants.