Impact of a Home‐Based Obesity Prevention Intervention on Children's and Parents' BMI: Findings From the Guelph Family Health Study Randomised Controlled Trial

ABSTRACT

Background

To prevent the development of obesity and related comorbidities, interventions must begin early in life. This study aimed to assess the impact of a home‐based obesity prevention intervention on children's and parents' body mass index (BMI).

Methods

The Guelph Family Health Study is a randomised controlled trial of a 6‐month obesity prevention intervention aimed at establishing healthful household routines regarding family meals, children's screen time, physical activity and sleep among 285 families with preschool‐aged children. The control group received 6 emails on child health behaviours. The analytic sample included 376 children (3.6 ± 1.3 years old, 20.2% at risk for overweight, 8.2% with overweight/obesity) and 481 parents (58.2% with overweight/obesity).

Results

Compared to controls, the intervention had no significant impact on children's or parents' BMI at post‐intervention and 1 year after the intervention. Children in both the intervention and control groups had a non‐significant decrease in BMI from baseline to post‐intervention [intervention; −0.23 kg/m² (95% CI, −0.46, 0.002); control; −0.17 kg/m² (95% CI, −0.40, 0.06)] and from baseline to 1‐year follow‐up [intervention, −0.29 kg/m² (95% CI, −0.67, 0.08); control, −0.06 kg/m² (95% CI, −0.42, 0.30)].

Conclusions

Compared to controls, this home‐based obesity prevention intervention had no significant impact on children's or parents' BMI immediately after the intervention or after 1 year of follow‐up, indicating that the intervention had no additional effect compared to the control group on children's or parents' BMI over time. Future data collection phases will allow for the examination of the intervention effect on BMI as children age.

Trial Registration

This study is registered at clinicaltrial.gov identifier: NCT02939261

Abbreviations

BMI

body mass index

GFHS

Guelph Family Health Study

SD

standard deviation

1. Background

Obesity represents an important public health issue [1, 2]. Since 1990, the prevalence of adult obesity has more than doubled, reaching 16% worldwide in 2022 [2, 3]. Children are not spared this upward trend in obesity [2, 3, 4]. From 1990 to 2022, the worldwide prevalence of overweight and obesity among children aged 5–19 years increased from 6% to 12% and from 2% to 8%, respectively [2, 3]. Among Canadian adults, 61% were living with overweight or obesity in 2015 [5], and this prevalence is projected to reach 66% by 2030 [6]. Thirty‐four percent of Canadian children aged 2–5 years were at risk for overweight or were living with overweight or obesity, and 31% of children aged 5–17 years were living with overweight or obesity in 2015 [7, 8]. In addition to this high and rising prevalence of obesity among both children and adults, it has been shown that obesity developed during childhood and adolescence tends to persist into adulthood [1, 9, 10]. For instance, two studies showed that almost 90% of children living with obesity at 3 years of age were living with overweight or obesity during adolescence [11], and that children having obesity at 2 and 19 years of age have a relative risk of 1.30 and 1.99 of living with obesity at age 35, respectively [12]. Since childhood obesity is a predisposing factor for many comorbidities, including type 2 diabetes, hypertension, cardiovascular diseases, some cancers, musculoskeletal problems, lower psychological health and is associated with excess and premature morbidity and mortality which can manifest in young adulthood [1, 10, 13, 14], childhood obesity prevention interventions are needed [15].

Behavioural modifications are the cornerstone of childhood obesity treatment and prevention [10]. While preventive efforts are required at every stage of life to help reduce the obesity epidemic, the preschool‐age period is important for primary prevention as many health behaviours are established early in life [16]. Moreover, studies have shown that health behaviours including dietary habits and physical activity acquired during childhood tend to track over time [16, 17, 18], highlighting the relevance of early prevention interventions. Parental involvement is critical for childhood obesity prevention as parents exert a strong influence on children's health behaviours by being the primary role models of health behaviours for their children and the providers and regulators of a healthy home environment [19]. Family‐based interventions are thus considered the gold standard of childhood weight management and prevention interventions, especially among preschool‐aged children [15, 19]. However, the literature on childhood obesity treatment shows that most family‐based interventions primarily focus on a single parent–child dyad, typically involving the mother [19]. Including all family members in interventions allows healthcare providers to better understand the family dynamics, including relationships, interactions, roles and the environment within a family, which may facilitate the tailoring of interventions to the unique needs of the family and may lead to improvement in the health and body weight of all family members [19].

Intervening within the home setting may reduce challenges associated with whole family interventions by increasing accessibility and convenience, enhancing the applicability of interventions and helping families overcome barriers that limit participation in long‐term follow‐ups [20, 21]. Despite high parental acceptability of interventions delivered in the home setting [22], only seven obesity prevention interventions with home visits conducted among families—mostly parent–child dyads—of preschool‐aged children have been identified from three recent systematic reviews [15, 23, 24]. To the best of our knowledge, the only family‐based obesity prevention intervention targeting preschool‐aged children that includes home visits and all family members identified from these three systematic reviews was the external pilot study of the Guelph Family Health Study (GFHS), which was conducted by our team among 44 Canadian families [25]. This pilot randomised controlled trial showed a high level of acceptability and feasibility of the intervention and study design, and preliminary data suggest a favourable impact of the GFHS intervention on children's fruit and fibre intake and on children's and parents' body weight outcomes [25, 26, 27, 28]. A full‐scale trial of this study with long‐term assessment of body weight outcomes in both children and parents is needed to test the effectiveness of this intervention.

This study aimed to assess the impact of the GFHS intervention, a 6‐month home‐based obesity prevention intervention among families with preschool‐aged children, on children's BMI immediately after the intervention and after 1 year of follow‐up (18 months from baseline) as a primary outcome and on parents' BMI at these same time points as a secondary outcome. The main hypothesis was that children randomised to the intervention group would show a slower body weight gain over time compared to those in the control group, which would result in a lower BMI among children in the intervention group compared to children in the control group at post‐intervention and at 1‐year follow‐up. Our secondary hypothesis was that the GFHS intervention would result in a lower BMI among parents immediately after the intervention and after 1 year of follow‐up as compared to parents in the control group.

2. Methods

2.1. Study Design and Participants

The GFHS is a randomised controlled trial of a 6‐month home‐based obesity prevention intervention conducted among families with preschool‐aged children. Participants of the current study were enrolled in the GFHS Full Study or the GFHS Pilot 2, which is an internal pilot study of the main study funded through bridge funding. Participants of the GFHS Pilot 2 and Full Study were recruited from December 2015 to December 2016 and from April 2017 to March 2020, respectively. Families were recruited through advertisements on the University of Guelph and GFHS websites, community centres, newspapers, social media (e.g., Facebook and Instagram) and through in‐person community events. The GFHS was advertised as a long‐term study testing ways to help children learn healthy habits early in life to reduce their risk of disease both now and in the future. Families were eligible to participate in the study if they had at least one child aged between 1.5 and 5 years, were able to speak English and complete English surveys, were living in Guelph, Ontario, Canada, or the surrounding area, and were not planning to move out of the region within the following year. After baseline assessment, families were randomised to the study groups by the study coordinator using a random number generator to assign each family to the intervention or control groups in a 1:1 ratio. Participants were aware of the study group to which they were assigned, but measurement staff were blinded to families' randomisation. Families received financial compensation in the form of grocery gift cards for completing baseline and follow‐up assessments. The GFHS aims to follow families over the long term, with yearly assessments over a period of up to 20 years. The current study included data from baseline to 1‐year after the end of the intervention (18 months from baseline). The 1‐year post‐intervention assessment was completed in March 2022. A flowchart diagram of the study is provided in Figure 1. The study was approved by the Research Ethics Board of the University of Guelph (REB14AP009 and REB17‐07‐003). Parents provided a written informed consent for themselves and their children. When possible, children also gave a verbal assent to participate in the study. This study is registered at clinicaltrial.gov (NCT02939261).

FIGURE 1.

Flowchart diagram of the Guelph Family Health Study until 1 year of follow‐up (18 months from baseline).

2.2. Intervention Group

The GFHS intervention is adapted from the Healthy Habits, Happy Homes intervention [29, 30]. The Healthy Habits, Happy Homes intervention was a 6‐month home‐based obesity prevention intervention conducted among low‐income, ethnic minority parent–child dyads from Boston, MA, USA. The Healthy Habits, Happy Homes intervention promoted the adoption of household routines associated with a lower prevalence of obesity among preschool‐aged children, namely regularly eating family meals, obtaining adequate sleep duration, limiting screen time and encouraging families to remove TV from the child's bedroom [29, 30, 31]. This intervention was effective in reducing BMI among preschool‐aged children [30]. Before being implemented in Canada, the Healthy Habits, Happy Homes intervention was adapted based on feedback on intervention content and delivery mode provided by 28 Canadian parents of preschool‐aged children [22].

The GFHS intervention focused on establishing healthful household routines regarding family meals, screen time, sleep, physical activity, fruit and vegetable intake and sugar‐sweetened beverages. The intervention is based on the Family Systems Theory and the Self‐Determination Theory, and uses motivational interviewing delivered by trained health educators [32, 33, 34, 35, 36, 37]. According to the Family Systems Theory, families are systems of interconnected and interdependent individuals who cannot be understood in isolation from the system [36, 37]. Interventions informed by the Family Systems Theory foster changes at the family level to improve family functioning and promote long‐term behaviour changes by establishing new routines, rules and communication dynamics [36, 37]. Consistent with this theory, the GFHS intervention uses a home‐based, whole family approach to promote sustainable behaviour changes. The Self‐Determination Theory is a predicting behaviour and behaviour change theory that conceptualises motivation as being regulated by different motives, both internal and external to the individual and organised along a continuum of self‐determination or autonomy [32, 33, 34]. Numerous studies showed that internally regulated, self‐determined behaviours are more likely to be maintained than those engaged in for more controlled, external reasons [32]. Motivational interviewing is defined as a collaborative, goal‐oriented style of communication with particular attention to the language of change, designed to strengthen personal motivation for and commitment to a specific goal by eliciting and exploring the person's own reasons for change within an atmosphere of acceptance and compassion [35]. Before the intervention, health educators received a 2‐day intensive advanced training in motivational interviewing by experts (J.D.I. and D.M.) from the Monarch System^TM [38]. As part of the training, health educators performed two evaluated practice sessions of motivational interviewing before leading the home visits. Thereafter, health educators received a 1‐day motivational interviewing refresher training once a year. All health educators (n = 8) were graduate students and registered dietitians with at least 1 year of counselling experience. The health educators' fidelity to motivational interviewing at the home visits was 96% (range: 88%–100%) based on assessment using the Motivational Interviewing Treatment Integrity (MITI) version 4.2.1 [39].

2.2.1. Home Visits

During a 6‐month period, families randomised to the intervention group received four home visits with a health educator, biweekly emails and monthly mailed behavioural supports.

At the first home visit, the health educator presented the structure of home visits and an overview of the goals of the intervention to the families. Families were invited to describe their current daily family routines relating to the goals of the intervention and were invited to reflect on which parts of the day, if any, could be more ideal for them in terms of family routines. At all home visits, families were presented the specific health behaviour goals targeted by the intervention and were invited to rate their current routines regarding these behaviours using a scale of 1 (very unsatisfied) to 10 (very satisfied). The specific goals of the study were (1) to increase the number of meals eaten together as a family with the TV off, (2) to be more physically active so that children meet a recommendation of at least 1 h of active play per day, (3) to limit TV and screen time to a maximum of one to 2 h per day, (4) to set a bedtime routine to help children get between 10 and 13 h of sleep per day, (5) to increase fruit and vegetables intake so that children eat four to five servings of fruit and vegetables per day and (6) to choose water to drink to reduce the amount of sugar‐sweetened beverages that children drink and limit their fruit juice intake to 125 mL or less per day. To support participants' autonomy, families were given the choice to set one or more behaviour change goals, to not choose any goals, or to set a different health behaviour goal than those targeted by the intervention at each home visit. Though the primary aim of the intervention was to improve children's health behaviours, changes were encouraged to be made at a family level by establishing new family routines involving all family members. For each goal set, families worked with the health educator to establish an action plan to meet their goal and to overcome possible barriers according to motivational interviewing. If families did not set a goal at a visit, the health educator acknowledged they were not ready to make behaviour change and asked them what, if any, additional information or support they would find helpful regarding their family routines. More details about the use of motivational interviewing during the home visits of the GFHS intervention have been previously published [26].

Follow‐up home visits were scheduled every 4–6 weeks. During these visits, families were invited to discuss their progress, successes and challenges in implementing their behavioural goal(s) with the health educator. They worked with the health educators to find solutions to overcome identified barriers to their behaviour change. During the last follow‐up visit, families were invited to reflect on progress made towards family routines, to reflect on the importance they place on maintaining their newly established routines and to set an action plan for maintaining them. The typical duration of home visits was between 40 and 70 min for the initial visit, and between 25 and 55 min for the follow‐up visits.

Due to the COVID‐19 pandemic physical distancing requirements, 5 families received the four sessions with the health educator remotely, and 16 families received between one and three remote sessions with the health educator after having completed at least one in‐person home visit before the start of social distancing. Among the 21 families who received remote delivery of the intervention, 10 families chose to receive phone calls, 8 families chose online video calls through the Microsoft Teams platform and 3 families did a combination of phone and video calls [40].

2.2.2. Health Behaviour Emails and Mailed Behavioural Supports

In addition to the home visits, families in the intervention group received biweekly emails (n = 14) providing tips, strategies and information to support their behaviour change. Every month, families were also mailed incentives for each child in the family to support behaviour change for each goal of the study. These incentives included a placemat to encourage family meals, a dice game to increase physical activity, a book to support their bedtime routine, a colouring book with crayons to reduce screen time, a reusable straw to make drinking water more fun and a planter pot and a pack of string bean or cherry tomato seeds to support fruit and vegetable intake.

2.3. Control Group

Families randomised to the control group received 6 monthly emails providing publicly available documents on child health behaviours related to healthy eating, physical activity, sleep habits, sedentary behaviours and screen time. These documents are from Canadian organisations, such as Health Canada (e.g., Easy ways to eating healthy meals when you're busy at home) [41] and the Canadian Society for Exercise Physiology (CSEP, e.g., Canadian Physical Activity Guidelines for the Early Years—0–4 years) [42].

2.4. Measures

2.4.1. Anthropometric Measurements

Anthropometric measurements were performed before the intervention, immediately after the intervention (6 months from baseline) and after 1 year of follow‐up (18 months from baseline). At baseline, all families came to the Body Composition Laboratory at the University of Guelph for a health assessment visit during which height and body weight were measured by research staff. At the 6‐month assessment, 176 families came to the Body Composition Laboratory, while 52 families performed at‐home self‐measurement due to the COVID‐19 pandemic. At the 1‐year post‐intervention follow‐up, 83 families had in‐person laboratory measurements and 113 families performed at‐home self‐measurements. During in‐person laboratory measurements, height was measured to the nearest 0.1 cm using a stadiometer for adults and children aged 2 years and over (Seca Model 222, Mount Pleasant, SC, USA) or a paediatric measuring board in a standing position for children under the age of two (ShorrBoard, Weigh and Measure LLC, Olney, MD, USA), following standardised procedures [43]. Body weight was measured with the participants wearing light clothing to the nearest 0.001 kg using a BODPOD digital scale. During the COVID‐19 pandemic, families were provided with a non‐elastic flexible tape measure and a digital scale (Seca 803 digital scale), together with instructions on how to measure their children's height and to use the digital scale according to standardised procedures that have been validated among families with young children [44, 45]. Children's height was measured to the nearest 0.1 cm. Body weight was measured in kg to two decimal places using the digital scale, with participants wearing light clothing. Body mass index was calculated as kg/m². If a parent reported being pregnant or breastfeeding at a time point, BMI was recorded as missing at the corresponding time point. Parents' baseline height was used to calculate BMI at each time point. Children's BMI z‐score was calculated based on the World Health Organization (WHO) Child Growth Standards [46] using the R package zscorer, version 0.3.1 [47, 48]. For descriptive purposes, children under the age of 5 years were classified as being at risk for overweight, with overweight, or with obesity, if their BMI z‐score was above 1, 2 and 3 standard deviations (SD), respectively [49]. Children aged 5 years were classified as living with overweight or obesity if their BMI z‐score was above 1 and 2 SD, respectively [49, 50]. For all children, underweight was defined as a BMI z‐score lower than −2 SD. Body mass index was used to assess changes in response to the intervention since it has been shown to better reflect changes in fat mass than BMI z‐score [51, 52, 53].

2.4.2. Assessment of Sociodemographic Variables

Information on parents' ethnicity, educational attainment and household income were collected through a questionnaire completed by each parent at baseline. Information on children's ethnicity was provided through a questionnaire completed at baseline by the first parent to enrol in the study. Parents' and children's age and sex were collected by the research staff at the baseline health assessment visit.

2.5. Statistical Analyses

A priori sample size calculation was performed using the R package pwr (R, version 3.2.1) [47, 54]. A conservative approach based on a two‐sample Student's t‐test with one observation per family was used given the lack of information available on the strength of correlation among participants of a same family to inform a more complex power analysis. Using a statistical power of 80%, a 2‐sided statistical significance level of 0.05, and a standard deviation of 1.25, it was determined that 300 families with one child would be required to detect a between‐group difference of 0.4 kg/m² in children's BMI immediately after the intervention. A similar magnitude of difference in children's BMI between groups has been found in previous obesity prevention trials, including in the Healthy Habits, Happy Homes randomised controlled trial on which the GFHS is based [30, 55, 56]. Using a conservative attrition rate of 15% after 18 months, given the GFHS Pilot 1 Study resulted in a 4.5% loss to follow‐up at post‐intervention (6 months from baseline) [25], we aimed to randomise 356 families with at least one child, in the intervention and control groups, with 178 families in each group.

Statistical analyses related to study outcomes were performed using SAS Studio (SAS Institute, Cary, NC, USA). Changes in BMI in children and parents in response to the intervention were investigated using linear mixed models for repeated measures with Tukey–Kramer's post hoc tests, using group, time and their interaction as fixed effects and participants and families as random effects. Group and time main effects were obtained from models that excluded the group by time interaction term. The dependence among repeated measurements was modelled using an unstructured matrix of covariance. Three different models were performed for parents and children, namely unadjusted and adjusted models and a model further accounting for the COVID‐19 pandemic. The adjusted model for children included age, sex, ethnicity (White vs. other) and parents' highest education level (some university, some college or technical school or less; college; university; postgraduate training or degree) as covariates. Parents' adjusted model accounted for age, sex, ethnicity and education level. Missing data on baseline parents' education level (n = 8) were handled by assigning the response provided at the 6‐month assessment (n = 1) or by performing a mode imputation (i.e., postgraduate training or degree, n = 7) [57, 58]. Parents' highest education level used in the children's model was calculated before performing mode imputation for those 7 parents and considered education level provided by the other parent as the highest parental education level (n = 12 children). As sensitivity analyses, the adjusted model for both children and parents was run excluding BMI data collected at home because of the COVID‐19 pandemic at the post‐intervention and 1‐year follow‐up time points.

A total of 285 families were randomised to the intervention and control groups. The analytic sample includes 481 parents and 376 children, from 284 families (Figure 1). A total of 14 parents and 8 children were excluded from the analyses since their BMI was missing at all three time points. Families included in the analyses had a mean of 1.4 ± 0.5 children and 1.7 ± 0.5 parents participating in the study. Each family included between 1 and 3 children and between 1 and 2 parents. Analyses were performed as intention to treat. Therefore, families who dropped out of the study but did not ask for their data to be removed from the database, and those who did not receive all four home visits (n = 7), were included in the analyses. Moreover, since the GFHS is also a cohort study with yearly assessments planned over a period of up to 20 years, families who declined to participate or were lost to follow‐up at a specific time point are not considered to have dropped out, as they may have participated or may participate at a subsequent follow‐up measurement time point.

3. Results

Among the 285 families randomised to the intervention or the control groups, a total of 15 families dropped out of the study; 8 in the intervention group and 7 in the control group (Figure 1). Follow‐up rates were 90.5% immediately after the intervention (6 months from baseline) (intervention group, 88.9%; control group, 92.2%) and 80.4% 1 year after the end of the intervention (18 months from baseline) (intervention group, 77.8%; control group, 83.0%). One hundred and thirty‐five families (93.8%) randomised to the intervention group completed all four home visits (Figure 1). At baseline, children included in the analyses had a mean age of 3.6 ± 1.3 years and about half of them were female (50.5%) (Table 1). Children were primarily identified as White (80.1%) and 71.6% had a BMI z‐score categorised as normal weight. Twenty percent of children were at risk for overweight, and 7.4% and 0.9% of children were living with overweight and obesity, respectively. Parents had a mean age of 36.0 ± 4.9 years at baseline and 58.2% identified as female. Fifty‐eight percent of parents were living with overweight (35.1%) or obesity (23.1%). Most parents identified as White (83.8%) and had a university degree (undergraduate degree, 32.7%; postgraduate training or degree, 35.4%). More than half of parents had an annual household income over $100 000 (53.2%) (Table 1).

TABLE 1.

Baseline characteristics of children and parents from the Guelph Family Health Study.

	All	Intervention	Control
Children	376	185	191
Sex
Female	190 (50.5)	87 (47.0)	103 (53.9)
Male	186 (49.5)	98 (53.0)	88 (46.1)
Age, years	3.6 ± 1.3	3.5 ± 1.3	3.6 ± 1.3
BMI z‐score a	0.56 ± 0.94	0.58 ± 0.95	0.55 ± 0.94
BMI category a
Normal weight b	252 (71.6)	125 (71.4)	127 (71.8)
At risk for overweight	71 (20.2)	37 (21.1)	34 (19.2)
Overweight	26 (7.4)	11 (6.3)	15 (8.5)
Obesity	3 (0.9)	2 (1.1)	1 (0.6)
Ethnicity
White	301 (80.1)	152 (82.2)	149 (78.0)
Other	75 (20.0)	33 (17.8)	42 (22.0)
Parents	481	243	238
Sex
Female	280 (58.2)	139 (57.2)	141 (59.2)
Male	201 (41.8)	104 (42.8)	97 (40.8)
Age, years	36.0 ± 4.9	35.7 ± 5.1	36.2 ± 4.7
BMI, c kg/m2	27.1 ± 5.9	27.0 ± 5.7	27.2 ± 6.2
BMI category c
Underweight (BMI < 18.5 kg/m2)	7 (1.5)	4 (1.7)	3 (1.3)
Normal weight (18.5 ≤ BMI < 25 kg/m2)	185 (40.3)	91 (39.2)	94 (41.4)
Overweight (25 ≤ BMI < 30 kg/m2)	161 (35.1)	86 (37.1)	75 (33.0)
Obesity (BMI ≥ 30 kg/m2)	106 (23.1)	51 (22.0)	55 (24.2)
Ethnicity
White	403 (83.8)	213 (87.7)	190 (79.8)
Other	78 (16.2)	30 (12.4)	48 (20.2)
Education level d
Some university, some college or technical school, or less	75 (15.8)	44 (18.5)	31 (13.1)
College	76 (16.0)	40 (16.8)	36 (15.3)
University	155 (32.7)	74 (31.1)	81 (34.3)
Postgraduate training or degree	168 (35.4)	80 (33.6)	88 (37.3)
Household income, $CA e
< $40 000	32 (6.7)	16 (6.6)	16 (6.8)
$40 000–$59 999	44 (9.2)	22 (9.1)	22 (9.4)
$60 000–$79 999	70 (14.7)	38 (15.7)	32 (13.6)
$80 000–$99 999	77 (16.1)	40 (16.5)	37 (15.7)
$100 000–$149 999	156 (32.7)	72 (29.8)	84 (35.7)
≥ $150 000	98 (20.5)	54 (22.3)	44 (18.7)

Note: Values are n (%) or mean ± SD.

Abbreviations: BMI, body mass index; SD, standard deviation.

^a Missing data, n = 24 (intervention group, n = 10; control group, n = 14). Normal weight, −2 SDs ≤ BMI z‐score ≤ 1 SD; at risk for overweight (children < 5 years: 1 SD < BMI z‐score ≤ 2 SDs; children aged 5 years: Not applicable); overweight (children < 5 years: 2 SDs < BMI z‐score ≤ 3 SDs; children aged 5 years: 1 SD < BMI z‐score ≤ 2 SDs); obesity (children < 5 years: BMI z‐score > 3 SDs; children aged 5 years: BMI z‐score > 2 SDs).

^b Includes 1 child categorised as being underweight (BMI z‐score < 2 SDs) randomised to the control group.

^c Missing data, n = 22 (intervention group, n = 11; control group, n = 11).

^d Missing data, n = 7 (intervention group, n = 5; control group, n = 2).

^e Missing data, n = 4 (intervention group, n = 1; control group, n = 3). Household income data provided by the first parent to enrol in the study.

Compared to the control group, no intervention effect on BMI over time was observed in children (adjusted and unadjusted models, group by time interaction p = 0.43, and p = 0.42, respectively) (Table 2). Accordingly, estimates showed no between‐group differences in BMI at 6 and 18 months. Based on the adjusted model, children in the intervention group had a non‐significant decrease in BMI of −0.23 kg/m² (95% CI, −0.46, 0.002) from baseline to 6 months and −0.29 kg/m² (95% CI, −0.67, 0.08) from baseline to 18 months. Relative to baseline, children in the control group had a non‐significant decrease in BMI of −0.17 kg/m² (95% CI, −0.40, 0.06) at 6 months and −0.06 kg/m² (95% CI, −0.42, 0.30) at 18 months. A time effect was observed (p = 0.002), indicating that irrespective of study groups, children reduced their BMI by −0.20 kg/m² (95% CI, −0.33, −0.07) between baseline and 6 months (adjusted model). Sensitivity analysis excluding data collected at home during the COVID‐19 pandemic showed similar results as no group by time interaction was observed (Table S1). However, the estimates from the time effect suggest a higher reduction in BMI at 6 and 18 months, irrespective of study groups, compared to the adjusted model that included data collected at home.

TABLE 2.

Children's and parents' mean BMI (95% CI) at baseline, immediately after the intervention (6 months from baseline) and after 1 year of follow‐up (18 months from baseline) in the intervention and control groups of the Guelph Family Health Study.

	n		BMI			p
	Intervention	Control	Intervention	Control	Difference vs. control	Group a	Time a	Group × time
Children
Unadjusted model						0.60	0.004	0.42
Baseline	175	177	16.36 (16.13, 16.60)	16.24 (16.01, 16.48)	0.12 (−0.37, 0.61)
6 months	145	148	16.15 (15.89, 16.41)	16.08 (15.83, 16.34)	0.07 (−0.46, 0.59)
18 months	121	130	16.09 (15.77, 16.40)	16.20 (15.89, 16.51)	−0.11 (−0.76, 0.53)
Adjusted model						0.85	0.002	0.43
Baseline	175	177	16.47 (16.18, 16.76)	16.43 (16.14, 16.72)	0.04 (−0.43, 0.52)
6 months	145	148	16.24 (15.92, 16.57)	16.26 (15.94, 16.58)	−0.02 (−0.57, 0.54)
18 months	121	130	16.18 (15.80, 16.56)	16.37 (16.00, 16.74)	−0.19 (−0.87, 0.49)
Parents
Unadjusted model						0.89	< 0.0001	0.10
Baseline	232	227	26.88 (26.03, 27.73)	27.10 (26.24, 27.96)	−0.22 (−1.98, 1.54)
6 months	184	174	27.12 (26.27, 27.97)	27.03 (26.18, 27.89)	0.08 (−1.68, 1.85)
18 months	151	157	27.50 (26.63, 28.38)	27.43 (26.55, 28.31)	0.07 (−1.74, 1.88)
Adjusted model						0.78	< 0.0001	0.09
Baseline	232	227	27.96 (26.98, 28.94)	28.26 (27.32, 29.20)	−0.30 (−1.95, 1.35)
6 months	184	174	28.20 (27.22, 29.18)	28.19 (27.25, 29.14)	0.01 (−1.65, 1.67)
18 months	151	157	28.59 (27.59, 29.59)	28.59 (27.63, 29.56)	−0.0049 (−1.71, 1.70)

Note: Adjusted models: Children's model is adjusted for age, sex, ethnicity and parents' highest education level. Parents' model is adjusted for age, sex, ethnicity and education level.

Abbreviations: BMI, body mass index; CI, confidence interval.

^a Group and time effects are from models that excluded the group by time interaction term.

Among parents, both the adjusted and unadjusted models showed no group by time interaction (p = 0.09, and p = 0.10, respectively), indicating that compared to the control group, no intervention effect on BMI over time was observed (Table 2). Accordingly, and as with children, estimates showed no between‐group differences in BMI at any time points. Parents in the intervention group had a non‐significant change in BMI of +0.24 kg/m² (95% CI, −0.05, 0.53) from baseline to 6 months and a significant increase in BMI of 0.63 kg/m² (95% CI, 0.24, 1.02) from baseline to 18 months (adjusted model). Parents in the control group had a non‐significant decrease in BMI of −0.07 kg/m² (95% CI, −0.36, 0.23) from baseline to 6 months and a non‐significant change of +0.33 kg/m² (95% CI, −0.05, 0.72) from baseline to 18 months (adjusted model). Despite the significant increase in BMI from baseline to 18 months in the intervention group, the non‐significant group by time interaction indicates that the overall changes in BMI over time are not different between groups. Moreover, a time effect was observed (p < 0.0001), indicating that irrespective of study groups, parents gained body weight from baseline to 18 months [0.48 kg/m² (95% CI, 0.26, 0.71)] and during the 1‐year follow‐up period [0.39 kg/m² (95% CI, 0.20, 0.58)] (adjusted model). Sensitivity analysis excluding data collected at home during the COVID‐19 pandemic also found no group by time interaction (Table S1). A significant time effect was also observed in this model. The increase in BMI from baseline to 18 months in the control and intervention groups combined (time effect) remained significant, but the estimate was slightly attenuated compared to the adjusted model that includes data collected at home.

4. Discussion

This study assessed the impact of a 6‐month home‐based multicomponent behavioural obesity prevention intervention focused on establishing healthful household routines among families with preschool‐aged children on children's and parents' BMI. Families in the intervention groups received four home visits with a health educator, biweekly emails and monthly‐mailed incentives over 6 months. Compared to a control group that received monthly emails on child health behaviours, the invention did not differentially impact children's or parents' BMI immediately after the intervention or after 1 year of follow‐up (18 months from baseline).

The lack of a significant intervention effect over time on children's BMI compared to the control group does not support our hypothesis. However, the absolute mean changes in BMI at post‐intervention and after 1 year of follow‐up in the intervention group and the effect size for group difference in children's BMI at the 18‐month time point from the adjusted model are slightly higher than the effect size observed [i.e., −0.11 kg/m², 95% CI (−0.21, 0.00)] in previous intervention studies that included diet and physical activity components among preschool‐aged children [15]. The results of this study are therefore consistent with the literature showing a null to small effect of obesity prevention interventions on body weight outcomes among preschool‐aged children [15, 23, 24]. In the current study, the control group also showed a non‐significant reduction in children's BMI after the intervention and after the 1‐year follow‐up period. The estimates for changes in BMI observed at 6 months in each group suggest that both groups had a similar effect on children's BMI immediately after the intervention, which is consistent with the time effect observed from baseline to 6 months. Given that children had a mean age of 3.6 ± 1.3 years at baseline, the reduction in BMI from baseline to 6 months among the two groups combined (time effect) could also be explained by BMI‐for‐age growth trajectories, or by a combination of group‐related effects (intervention or control) and expected developmental changes in BMI. Indeed, BMI generally shows a rapid increase during the first year of life, then decreases to reach a nadir around 6 years of age and thereafter increases again throughout growth [59, 60]. The next phases of GFHS data collection will provide us with more information on the effect of the intervention on children's weight trajectory as they age.

The non‐significant reduction in children's BMI at 6 months in the intervention group [i.e., −0.23 kg/m² (95% CI, −0.46, 0.002)] is also similar to that of the Healthy Habits, Happy Homes intervention [i.e., −0.18 ± 0.98 (SD) kg/m²] conducted in Boston, USA and on which the GFHS intervention is based [30]. However, the main difference in the results of these two studies is that children in the control group of the GFHS had a non‐significant decrease in BMI of −0.17 kg/m² (95% CI, −0.40, 0.06) at 6 months, while children in the control group of the Healthy Habits, Happy Homes Study had a change in BMI of +0.21 ± 1.07 kg/m² after the intervention period. This explains the discordant results between the two studies, where a significant effect of the intervention on children's BMI compared to the control group was observed in the Healthy Habits, Happy Homes Study but not in the GFHS. Several factors may explain why children in the control group of the GFHS better responded to the minimal‐attention intervention they received. First, the email content provided to families in the control group of the GFHS, which targeted children's healthy lifestyle habits, was more likely to influence body weight than the email content sent to the control group of the Healthy Habits, Happy Homes Study, which mainly included information on developmental milestones related to social, emotional, language, cognitive and motor domains, but also included some recommendations on screen time, sleep, physical activity and eating habits [30]. A recent scoping review on family systems approaches to paediatric obesity management found that the type of control group influences study outcomes [61]. It is also possible that families participating in the GFHS, who have a higher socioeconomic level, had more resources allowing them to be better able to engage with and benefit from the email content that the control group received compared to families of the Healthy Habits, Happy Homes Study, the majority of whom had a low socioeconomic status and belonged to US ethnic minorities. Indeed, such minimal‐attention interventions may be insufficient for low‐income ethnic minorities to benefit from due to their heightened psychosocial stress, less favourable home and neighbourhood environments and cultural differences in attitudes and social norms towards obesity [62, 63]. However, low‐income ethnic minorities seem to respond well to family‐based obesity treatment interventions [64].

Among parents, the lack of an intervention effect on BMI over time compared to the control group is also discordant with our hypothesis. Irrespective of study groups (time effect), parents showed an increase in BMI of 0.48 kg/m² (95% CI, 0.26, 0.71) from baseline to 18 months and this body weight gain occurred mainly during the 1‐year follow‐up period [i.e., 0.39 kg/m² (95% CI, 0.20, 0.58)]. The BMI change observed in both groups combined from baseline to 18 months represents a body weight gain of 1.44 kg (95% CI, 0.76, 2.11) (data not shown). In comparison, data from 1996/1997 to 2004/2005 showed that Canadian adults have an average body weight gain of 0.5–1 kg per 2‐year period [65], whereas data from 2013 to 2018 showed a weight gain of 0.69 kg per year, 95% CI (0.64–0.74), from young adulthood to midlife among Americans [66]. The results thus suggest that neither the family‐based intervention nor the minimal‐attention control intervention reduced the natural body weight gain observed during adulthood. The sensitivity analysis model that excluded data collected at home during the COVID‐19 pandemic showed a slight attenuation in the increase in BMI from baseline to 18 months in both groups combined and among the intervention group. Although this model should be interpreted with caution given the sample size at both post‐intervention time points is reduced compared to the main (adjusted) model, this result suggests that the COVID‐19 pandemic may have exacerbated body weight gain in parents, which is consistent with the literature showing that between 30% and 50% of the adult population gained body weight during the first year of the COVID‐19 pandemic [67, 68, 69]. A similar impact of the pandemic was observed in children, where the sensitivity analysis model showed slightly higher reductions in BMI in both groups combined (time effect) at 6 and 18 months from baseline compared to the adjusted model, suggesting that the pandemic attenuated BMI reduction in children over time in both groups. As with the adults, these results are in accordance with the literature showing that children gained body weight during the COVID‐19 pandemic [69, 70].

Very few studies on family‐based early childhood obesity prevention intervention assessed the ripple effect on parents' body weight outcome [24]. A recent systematic review identified three studies on that topic and found no impact of the interventions on parental body weight compared to the control group [24]. Although these results are in accordance with those of the current study, it is worth noting that these studies only included a parent–child dyad, which may limit the tailoring of the intervention to the family system and reduce improvement in health behaviours and body weight outcome of all family members [19]. Another recent systematic review on childhood obesity prevention interventions not limited to family‐based interventions in preschool‐aged children also found no spillover effect on parental body weight [71]. However, the external pilot study of the GFHS, which included both parents and their children aged 1.5–5 years, found a significant impact of the two home visit intervention compared to the control group on parental body weight [27]. The positive effect of the intervention on parental body weight in this pilot study suggests that family‐based obesity prevention interventions that include home visits may potentially have spillover effects on the entire family, which is in accordance with the literature on early childhood obesity treatment interventions [72]. Indeed, although the literature is scarce on the parental effect of early childhood obesity treatment, and that treatment interventions are typically more intensive than prevention interventions, spillover effect on parental body weight outcomes have been observed in family‐based obesity treatment interventions for preschool‐aged children that included a home‐based component [72]. More research is needed on the effect of family‐based childhood obesity prevention on parental body weight before drawing firm conclusions about their spillover effect, or lack thereof, among parents. In addition, future studies should investigate whether the GFHS intervention leads to behavioural modifications that may be a precursor to future changes in body weight or improvements in cardiometabolic health.

To the best of our knowledge, this study is the first full‐scale trial of a home‐based obesity prevention intervention targeting preschool‐aged children that includes all family members. Moreover, very few studies have assessed the spillover effect of childhood obesity prevention intervention among parents. Although the study did not find an impact of the intervention over time on BMI compared to the control group among children and parents, the study is grounded in a theoretical foundation with the potential to positively influence weight trajectories and lifestyle habits in a sustainable manner [19]. The 1‐year post‐intervention follow‐up included in this study represents an important strength as it enables the assessment of long‐term intervention outcomes. The laboratory measurements of height and body weight are also a strength of the study. However, these measurements were affected by the restrictions imposed by the COVID‐19 pandemic, which required 52 and 113 families to self‐report these measurements at post‐intervention and at the 1‐year follow‐up, respectively. To reduce measurement errors associated with self‐reported measures, material and a validated and standardised procedure were provided to families [44, 45]. Moreover, 21 families received at least one of their sessions with the health educator remotely. These modifications to the study protocol did not seem to have had a strong impact on the results, as sensitivity analysis excluding self‐reported measurements showed similar results to the models including all data. Lastly, given the study sample included socioeconomically advantaged and predominantly White families, the findings of this study may not be generalisable to other populations.

5. Conclusions

In conclusion, this study found no impact of a 6‐month home‐based obesity prevention intervention compared to the control group on preschool‐aged children's or parents' BMI immediately after the intervention or after 1 year of follow‐up. Irrespective of study groups, children slightly reduced their BMI from baseline to 6 months. Future studies should assess whether this intervention improved health behaviours in children and parents. Moreover, future data collection phases of the GFHS will allow for the examination of the intervention effect on BMI as children age.

Author Contributions

D.W.L.M., J.H., A.C.B., G.D., A.M.D., L.A.V., J.‐P.C., J.D.I. and D.M. designed the research. A.A. and M.N. coordinated the study. R.J. analysed data and wrote the first draft of the manuscript; all authors read, edited and approved the final version of the manuscript. J.H. had primary responsibility for final content.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: ijpo70062‐sup‐0001‐Supinfo.pdf.

Jacob R., Buchholz A. C., Darlington G., et al., “Impact of a Home‐Based Obesity Prevention Intervention on Children's and Parents' BMI: Findings From the Guelph Family Health Study Randomised Controlled Trial,” Pediatric Obesity 21, no. 1 (2026): e70062, 10.1111/ijpo.70062.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.