Differences Between the Perceived and Actual Age of Overweight Onset in Children and Adolescents

Abstract

Objective

Little is known about whether children or their parents can accurately recall the age at which they became overweight.

Design, Subjects and Main Outcome Measures

We interviewed 64 overweight children (7–18 years old) about their weight history and compared reported age of overweight onset with actual onset, as determined by the age at which the child's measured BMI first exceeded the 95th percentile.

Results

Only 28% of children reported overweight onset within 1 year of actual overweight onset. Reported overweight onset age (7.6 ± 2.5y) and actual onset age (5.3 ± 2.5y; P < .001) were not significantly correlated (r² = .03, P = .22). Children who became overweight before 8 years of age tended to report becoming overweight at a later age than actual onset, whereas children who became overweight after 8 years of age tended to report becoming overweight at an earlier age than actual onset (P < .001), with 27% of children either underreporting or overreporting their overweight onset by at least 5 years. Similar results were found when analyzing parent reports of their children's overweight onset.

Conclusions

Reported and actual overweight onset ages were uncorrelated in our sample, suggesting that families may not be cognizant of children's growth trajectories. Greater efforts should be made to help parents and children understand and track growth patterns with the aim of preventing excessive weight gain.

Introduction

Clinicians and researchers often query overweight individuals about the age at which they became overweight. In assessing the need for weight loss among obese patients, it is recommended that the duration of obesity be considered by practitioners.[1] Moreover, as part of an effort to understand the etiology of obesity and develop novel treatment and prevention programs, researchers often ask individuals about the age at which they become overweight.[2–5] However, few data exist regarding how well individuals recall the age at which they became overweight.

It might be presumed that overweight children, compared with obese adults, would better recall when they became overweight, given that they are closer in age to their own overweight onset. However, the only relevant prior study found that among 25 children, self-reported overweight onset age correlated poorly with actual overweight onset.[3] We, therefore, interviewed children about their age of overweight onset and compared child-reported age with the actual age of overweight onset, on the basis of growth records retrieved from their healthcare providers. Furthermore, since parent knowledge of child overweight is typically crucial for intervention, we also queried parents about the age at which their children became overweight. Given preliminary findings[3] and data suggesting that neither children[6–11] nor their parents[12–14] can accurately report whether a child is currently overweight, we hypothesized that both child- and parent-reported onset of overweight would be unrelated to actual onset.

Methods

Parents in a convenience sample of 338 children participating in observational or weight-loss treatment studies approved by the NICHD Institutional Review Board[15,16] were mailed notices requesting consent to contact their children's healthcare providers to obtain height and weight measurements from birth to present. One-hundred-sixty-one (48%) families provided permission, with healthcare providers supplying growth data for 123 children.

Obtained data were considered sufficient for use if yearly heights and weights were available beginning at age 5 years or earlier. The age at which BMI first equaled or exceeded the 95th percentile for age and sex[17] was determined. Current weight and height were also measured, as previously described.[15,16]

Participants were interviewed in person (n = 106) or by telephone (n = 10) regarding their weight history, as previously described.[3] If children reported that they had been overweight since “infancy,” or “as long as they could remember,” we used the age of 5 years in order to provide a conservative estimate. Parents of overweight children were also asked to report at what age their child first became overweight.

Simple regression and Bland-Altman comparisons[17] were used to assess agreement between actual and reported overweight onset. Partial regressions were used to control for the child's current age for analyses involving child reports. Because no differences in results were found when nontreatment and treatment-seeking children were analyzed separately, data for all children were combined for presentation.

Results

One-hundred-sixteen growth records, including those of 64 overweight children, contained sufficient data for analysis (Table). Among overweight participants, 87.5% correctly reported being currently overweight. Of the 56 parents of overweight children queried, 87.5% accurately reported that their child was overweight. The reported age of overweight onset (7.6 ± 2.5y) significantly differed from actual overweight onset (5.3 ± 2.5y; P < .001). Only 28% of children accurately reported overweight onset (ie, within 1 year of actual overweight onset according to their growth records); 22% reported overweight onset within 1–2 years of actual onset, 23% within 2–4 years, and 27% were inaccurate by >/= 5 years. Actual overweight onset was not significantly related to reported onset (r² = 0.03, P = .22; Figure, A). Parent-reported age of their child's overweight onset was significantly related to child reports (r² = 0.10, P = .04; Figure, B), and parent reports were unrelated to actual overweight onset (r² = 0.01, P = .44; Figure, C).

Table.

Subject Demographics for Overweight Participants

Variable	Mean ± SD	Range
Age (y)	12.0 ± 2.5	7.0 – 17.9
Sex	Male = 30 (46.9%)
	Female = 34 (53.1%)
Race	African American = 35.9 (35.9%)
	White = 40 (62.5%)
	Hispanic = 1 (1.6%)
Socioeconomic status (median)	3	1 – 5
Body mass index (kg/m2)	33.3 ± 8.4	17.9 – 54.6
Body mass index z-score	2.3 ± 0.4	1.2 – 2.9
Age of actual overweight onset according to growth chart (y)	5.3 ± 2.5	2.0 – 14.0
Age of overweight onset according to child report (y)	7.6 ± 2.5	2.0 – 12.0
Age of overweight onset according to parent report (y)†	5.9 ± 3.2	1.0 – 13.0

Note. N = 64

^†56 parents were queried. Socioeconomic status was determined as suggested by Hollingshead.[20]

Figure.

The relationship between actual age of overweight onset, on the basis of growth chart data, and child's self-reported age of overweight onset, accounting for child's current age (A); the relationship between parent-reported age of child's overweight onset and child's self-reported age of overweight onset, controlling for child's current age (B); the relationship between actual age of overweight onset, on the basis of growth chart data, and parent-reported age of child's overweight onset (C); Bland-Altman comparison between actual age of overweight onset, on the basis of growth chart data, and the difference between actual and child's self-reported age of onset, accounting for child's current age (D).

Bland-Altman comparisons revealed a significant magnitude bias (r² = 0.38, P < .001; Figure, D). Among children whose growth charts indicated overweight onset before 8 years of age, subjects' reported age of overweight was older than actual onset, while children with measured overweight onset after 8 years of age reported becoming overweight at a younger age than actual onset. A similar inverse relationship was found between actual onset ages and the difference between parent report and actual onset age (r² = 0.26, P < .001).

Discussion

In this study, actual age of overweight onset on the basis of growth chart data was unrelated to self-reported overweight onset age. Participants who became overweight at a younger age reported a later onset age, and those who became overweight at an older age reported a younger onset age. A similar inverse relationship was found for parent reports of their children's overweight onset.

Unlike studies comparing measured-to-reported current body weight, most children in the present investigation were able to report accurately whether or not they were currently overweight. This finding may be the result of characteristics specific to our sample. Sixteen percent of participants were seeking weight-loss treatment and thus were certainly aware of their excessive weight. Moreover, individuals who are knowledgeable or concerned about health might be more cognizant of their own weight and more willing to participate in health-related studies.

Most overweight participants were, however, unable to report with accuracy the actual age when they became overweight. Only 28% of the overweight children reported having become overweight within 1 year of their actual age of overweight onset. Most reported a later age of overweight onset. Children who become overweight at a very young age may not be aware of their body size at the time and, therefore, may base their reported onset age on a significant recollection (eg, being teased) that caused them to become cognizant of their actual body size. Alternatively, the children in our sample may have reported an age of overweight onset on the basis of their parents' perceptions. Indeed parent and child reports were significantly related. Given that the mean age of our sample was 12 years, most of the children would be presumed to have the cognitive ability to base their reports on meaningful information (eg, parent reports, doctor visits, or other significant experiences). As described previously,[3] in order to assist younger children in making the most accurate assessment of the age of overweight onset, we asked them to recall not only their age in years, but also their school grade, their teacher, and friends at the time, with the aim of anchoring their memory to a specific time period. If a response appeared questionable, we probed participants as to why they recalled each behavior occurring at the time period stated. In some cases, children reported that their parents had informed them.

The findings that neither children nor their parents can accurately recall the age at which children become overweight raise concern since a lack of awareness may deter families from seeking early intervention. Treating overweight children prior to adolescence has demonstrated promising results[18] and prevention has been suggested as the most important approach to decreasing the obesity epidemic.[19] As such, familial knowledge of excessive weight gain trajectories must be addressed early in order to engage both parents and children to participate in the necessary dietary and activity changes. At well-child visits, we recommend a routine discussion of growth charts with families of children who are at-risk for overweight or who are already overweight so that pediatricians have an opportunity to raise the potentially sensitive issue of overweight in a nonjudgmental and straightforward manner.

Study limitations include the relatively small sample size; indeed, our finding that older children reported an earlier overweight onset age should be interpreted cautiously due to the small number of subjects in this category (n = 5). A second limitation is the possibility that families who chose to participate may have been more health conscious than the general population, limiting external validity. However, such families might be expected to have greater awareness of overweight onset. Thus, the perception of overweight onset among children and their parents in the general population may demonstrate an even larger discrepancy with actual onset compared with our sample. Strengths include the use of measured heights and weights from growth charts and the racially diverse sample.

We conclude that because reported and actual overweight onset ages are not significantly related, child and parent reports of the age children first became overweight are unreliable. Pediatric healthcare providers should make greater efforts to educate families about growth trajectories, rather than just current body weight, so that prevention and early intervention may be implemented in those approaching the 95th percentile for BMI.

Funding Information

This research was supported by the Intramural Research Program of the NIH, grant ZO1-HD-00641 (NICHD, NIH).