Correlates of participation in a pediatric primary care-based obesity prevention intervention

Abstract

The purpose of this study was to examine correlates of participation in a childhood obesity prevention trial. We sampled parents of children recruited to participate in a randomized controlled trial. Eligible children were 2.0 - 6.9 years with BMI ≥ 95^th percentile or 85^th-<95^th percentile if at least one parent was overweight. We attempted contact with parents of children who were potentially eligible. We recruited 475 parents via telephone following an introductory letter. We also interviewed 329 parents who refused participation. Parents who refused participation (n=329) did not differ from those who participated (n=475) by number of children at home (OR 0.94 per child; 95% CI: 0.77, 1.15) or by child age (OR 1.07 per year; 95% CI: 0.95, 1.20) or sex (OR 1.06 for females v. males; 95% CI: 0.80, 1.41). After multivariate adjustment, parents who were college graduates v. < college graduates were less likely to participate (OR 0.62; 95% CI: 0.46, 0.83). In addition, parents were less likely (OR 0.41; 95% CI: 0.31, 0.56) to participate if their child was overweight v. obese. Among the 115 refusers with obese children, 21% cited as a reason for refusal that their children did not have a weight problem, v. 30% among the 214 refusers with overweight children. In conclusion, parents of preschool-age children with a BMI 85-95^th%ile are less likely to have their children participate in an obesity prevention trial than parents of children with BMI >95^th%ile. One reason appears to be they less frequently consider their children to have a weight problem.

INTRODUCTION

In the U.S., approximately 21.2% of children ages 2 - 5 years are overweight (age- and sex-specific BMI 85^th to 94^th percentile) and 10.4% are obese (BMI ≥ 95^th percentile).¹ Obesity prevention interventions in primary care settings targeting preschool age children are likely to have benefit because compared with older children, children this age more often visit their primary care clinicians for routine medical care; parents have more control over their children’s health-related behaviors; and these behaviors may be more malleable. Thus, obesity prevention and management, starting in the preschool age group, could help stem the rising tide of obesity.

Few interventions to prevent childhood obesity among preschool age children have been conducted ^2-4 and no study has examined parent-child correlates of participation in obesity interventions. Identification of the factors that influence parents and children’s participation in obesity trials is essential to the development of effective intervention strategies aimed at the prevention of obesity in this age group.

The purpose of this study was to examine correlates of participation in a primary care-based intervention to reduce body mass index and improve specified obesity-related behaviors among children age 2 through 6 years at elevated risk of obesity.

METHODS

We sampled parents of children recruited to participate in High Five for Kids, a cluster-randomized controlled trial in 10 primary care pediatric offices of Harvard Vanguard Medical Associates (HVMA), a multi-site group practice in Massachusetts. Details of the intervention have been described elsewhere.⁵ The primary aims of the trials were to assess the extent to which a multi-factorial, primary care-based intervention blunted the age-associated increase in BMI and improved specified obesity-related behaviors. Eligible children were age 2.0 - 6.9 years at baseline with BMI ≥ 95^th percentile or 85^th-<95^th percentile if at least one parent was overweight (BMI ≥ 25 kg/m²), who received their pediatric care at any of the 10 primary care offices of HVMA, between August 2006 and October 2008.

Using the electronic medical records of HVMA, we identified a total of 3253 children as being “pre-eligible” for the trial based on having a BMI ≥ 85^th percentile sometime within the year prior to their next scheduled well child care visit. After each pediatric provider offered medical clearance, and approximately 1 month prior to their scheduled well child care visit, we mailed a letter to each parent introducing the study and encouraging participation. We informed parents that their child could be assigned to usual care or to a practice that was offered an intervention consisting of six 20- to 40- minute meetings and two 15- to 20- minute phone calls with a nurse practitioner from their practice over a 2-year period. The letter included an opt-out telephone number to call if the family did not want to participate. Twenty three parents opted out by calling our toll free telephone number and we did not attempt contact with 737 because we had reached our anticipated sample size. Within 7 days of mailing the letter, we attempted contact with the 2493 parents. A total of 533 (21% of 2493) parents actively declined to participate in the intervention of whom 329 (62%) agreed to answer 3 questions regarding their reason for refusal to participate. An additional 475 parents of children agreed to have their child participate and were enrolled in the study. The remainder of children were either ineligible or passively declined participation. All study procedures were in accordance with ethical standards for human experimentation by the Declaration of Helsinki and approved by the human subjects committee of Harvard Pilgrim Health Care.

Measures

We asked parents who refused participation, 1) “why did you decide not to participate in the study” (responses were recorded verbatim and later coded according to categories), 2) the number of children under the age of 18 living in the household, and 3) the highest grade or degree that the parent completed in school. From the electronic medical records we obtained the child’s age, sex, and most recent height and weight from which we calculated their BMI.

Data Analysis

We first tested for differences between parents and children who participated in the intervention and those who refused using unadjusted logistic regression models, corrected for clustering by site, using proc GLIMMIX (general linear mixed models). Next, we used multivariate adjusted logistic regression models to examine differences between participation and refusal to participate. The multivariate models included child’s age, sex, BMI category (BMI ≥ 95^th percentile or 85^th-<95^th percentile), and parental educational attainment, and the number of children < 18 years living in the household. We also ran descriptive analyses of reasons for refusal, overall, and according to each baseline characteristic of parent-child pairs who refused to participate. We performed all analyses using SAS version 9.2.

RESULTS

Figure 1 shows the participant flow of the study enrollment. Among all of the participants who actively or passively declined, we collected the gender and BMI of the child from the electronic records. There were no substantial differences in gender or BMI among the 329 who actively declined and completed the refusal interview v. all participants who actively or passively declined. For example, there were an equal proportion of males in each group (52% v. 54%) and there was only a small difference in children having a BMI > 95^th percentile among those who completed the refusal survey e.g. 35% among those who completed the refusal survey v. 42% among the total who actively or passively declined to participate.

Figure 1.

Participant flow of the High Five for Kids study enrollment

Table 1 shows characteristics of parent-child pairs who agreed to participate in the study and of those who refused. Overall, the children were a mean age of 4.8 (SD=1.2) years; 47% were female; 48% of the children had a BMI ≥ 95^th percentile at baseline; 43% of parents had less than a college degree and there were an average of 2.1 (SD=0.7) children under the age of 18 years in the homes.

Table 1. Unadjusted and Multivariate Adjusted Correlates of Participation in the High Five for Kids Study.

Child Characteristics	Total (N=804)	Agreed to participate (N=475)	Refused to participate (N=329)	Unadjusted	Adjusted*
		N (%) or Mean (SD)		OR (95% Confidence Interval)
Age, years (SD)	4.8 (1.2)	4.9 (1.2)	4.8 (1.2)	1.07 (0.95, 1.20)	1.10 (0.97, 1.24)
Gender
Female	381 (47)	228 (48)	153 (47)	1.06 (0.80, 1.41)	1.14 (0.85, 1.53)
Male	423 (53)	247 (52)	176 (54)	1.00 (reference)	1.00 (reference)
Body mass index category
85th to 94th percentile	419 (52)	205 (43)	214 (65)	0.41 (0.31, 0.55)	0.41 (0.31, 0.56)
≥ 95th percentile	385 (48)	270 (57)	115 (35)	1.00 (reference)	1.00 (reference)
Parent and Household Characteristics
Number of children < 18 years living in the home	2.1 (0.7)	2.1 (0.7)	2.1 (0.7)	0.94 (0.77, 1.15)	1.01 (0.81, 1.25)
Parent educational attainment
< Some college or below	342 (43)	228 (48)	114 (35)	1.00 (reference)	1.00 (reference)
≥ College graduate	455 (57)	247 (52)	208 (65)	0.59 (0.44, 0.79)	0.62 (0.46, 0.83)

^*Adjusted for child’s age, sex, BMI category (BMI ≥ 95^th percentile or 85^th-<95^th percentile) at baseline, and parental educational attainment, and the number of children < 18 years living in the household; and corrected for clustering by site.

Parents who refused participation (n=329) did not differ from those who participated (n=475) by number of children at home (OR 0.94 per child; 95% CI: 0.77, 1.15) or by child age (OR 1.07 per year; 95% CI: 0.95, 1.20) or sex (OR 1.06 for females v. males; 95% CI: 0.80, 1.41). After multivariate adjustment, we found that parents who were college graduates v. < college graduates were less likely to participate (OR 0.62; 95% CI: 0.46, 0.83). In addition, parents were much less likely (OR 0.41; 95% CI: 0.31, 0.56) to participate if their child was overweight v. obese.

Table 2 shows parental report of reasons for refusing to participate in the study. Parents reported several reasons for declining to participate including “study will take up too much time” (60%), “things [being] too difficult in the family right now – illness, divorce, etc” (9%), and “clinical site too far away” (5%). Among the 115 refusers with obese children, 21% cited as a reason for refusal that their children did not have a weight problem, v. 30% among the 214 refusers with overweight children (p=0.06). We did not find significant differences in reasons for refusing to participate among college graduate v. non-college graduate families.

Table 2. Parental report of reasons for refusing to participate in the High Five for Kids Study^*.

Reasons for Refusal to Participate	N (%)
Study will take too much time	196 (60)
Believe child does not have a weight problem	89 (27)
Things too difficult in the family at the moment (e.g. illness,
divorce, new baby, etc.)	30 (9)
Clinical site is too far	15 (5)
Concerns about harm or negative effects on child	11 (3)
Already involved in another research study	8 (2)
Pediatrician did not say child’s weight was a problem	7 (2)
Spouse’s decision	5 (2)
Lack of trust in research	3 (1)
Other or don’t know	38 (12)

^*Sample size = 329 participants who actively declined but completed a refusal interview.

DISCUSSION

Parents of preschool-age children with a BMI between the 85^th and 95^th percentile are less likely to have their children participate in an obesity prevention trial than parents of children with BMI ≥95^th percentile. One reason appears to be they less frequently perceive their children to have a weight problem. Neither child age nor sex affected parental willingness to participate but parental educational attainment did. The main reason parents cited for refusing to participate was concern with the amount of time the intervention would involve.

Parental misperception of weight status among overweight and obese children is well documented even among parents of preschool-age children.^{6, 7} Studies of parents of preschool-age children also show that parents do not recognize the physical and mental health risks of childhood overweight.⁸ Our findings are consistent with these previous studies and also suggest that parents of children with a BMI between the 85^th to 95^th percentiles, who may be likely to benefit from an obesity prevention trial, are less likely to participate largely due to their misperception of their child’s weight.

Our findings have relevant implications for interventions to prevent and manage obesity among preschool-age children. First, increasing parental awareness of their child’s weight status and potential health risks could increase their willingness to participate in obesity prevention programs. Recent recommendations for pediatric providers to regularly monitor and track the body mass index of their patients and offer appropriate evidence-based advice to families to promote healthy lifestyles, ^9-11 could increase parental awareness of their child’s weight status and help support families in making changes in their obesogenic behaviors. Second, efforts to enhance parental participation in childhood obesity interventions will need to consider parental concerns about the time commitment of participating in interventions. Recruitment strategies that address potential benefits of interventions for the amount of time to be committed may reduce this barrier to parental participation. Finally, it is possible that unmeasured characteristics or perceptions could explain the differences in participation rates we observed by parental educational attainment. This difference merits further study.