Accuracy of Perceptions of Overweight and Relation to Self-Care Behaviors Among Adolescents With Type 2 Diabetes and Their Parents

Our sample consisted of adolescents aged 12–20 years with type 2 diabetes who received care at the Vanderbilt Eskind Pediatric Diabetes Clinic and their parents/primary caregivers. Via telephone interview, parents provided information on demographics, the adolescent's diabetes regimen, perceptions of the adolescent's weight, the adolescent's self-care behaviors, and perceptions of barriers to self-care. Adolescent telephone interviews included the same questions, with additional questions about perceived barriers to healthy diet and exercise behaviors.

Weight perceptions were based on parent and adolescent responses to the following question: “Do you think [your child's/your] weight is very overweight, slightly overweight, about right, slightly thin, or very thin?” We calculated adolescents’ BMI using actual weight from medical records and categorized it using current Centers for Disease Control (CDC) recommendations (2,3).

We assessed both parents’ and adolescents’ reports of the adolescents’ diet and exercise behaviors. For diet, adolescents reported frequency of consuming sugary drinks, eating fast food, having unplanned snacks, and overeating; parents reported only the first two. For exercise, adolescents reported exercise frequency, hours spent watching television, days taking physical education, and time spent in physical education; parents reported only exercise frequency. Responses were categorized by frequency (0 = never, 1 = once per week, 2 = two to three times per week). Adolescents reported perceived barriers to diet and exercise (0 = never, 1 = sometimes, 2 = always) using statements previously validated among adolescents and adults with diabetes (4–6). A complete description of the methods is published elsewhere (7).

For bivariate analyses, we used t (continuous variables) and χ² (categorical variables) tests. We categorized weight perceptions into four categories (very overweight, slightly overweight, about right, and slightly thin/thin) that were consistent with the four CDC BMI categories for adolescents (overweight, at-risk for overweight, normal weight, and underweight). We created summary variables of reported diet and exercise behaviors by scoring each diet or exercise frequency variable as above and summing items to create separate diet and activity scores for both parents and adolescents. The highest quartile on these summary measures was considered “good” diet or activity. Similarly, we summed each barrier to a diet and exercise subscale and considered the lowest quartile to perceive fewer barriers.

Of 139 adolescent-parent pairs contacted, 104 (75%) participated. Parents were 85% mothers, 8.5% fathers, and 6.5% other guardians. There were no differences in responses for these groups and they are therefore reported together as “parents.” The adolescents’ mean weight was 100.3 kg (220.7 lbs). Most (69%) were female, and 47% were African American. Based on CDC guidelines, 87% of adolescents were classified as overweight, and an additional 5.9% were at-risk for overweight. Mean A1C levels were 7.7 ± 2.6%. Most adolescents were taking insulin, oral agents (typically metformin), or both.

While 87% of children were overweight by CDC standards, only 41% of parents and 35% of adolescents considered the adolescent “very overweight.” Among parents who reported their child's weight as “about right,” 40% had children whose BMI was ≥95th percentile; 55% of adolescents who reported their weight as “about right” had BMI ≥95th percentile. Adolescents were more likely to underestimate their weight when their parents also underestimated weight than when parents accurately perceived weight (66.2vs. 34.2%, P < 0.001).

Girls were more likely than boys to underestimate the severity of their weight (42.9 vs. 22.0%, P < 0.05), although parents’ accuracy did not differ by their child's sex. There were also no differences in the accuracy of weight perceptions by race or insulin use for either parents or adolescents. Parents and adolescents were both more accurate in their perceptions for adolescents <13 and >16 years; adolescents aged 13–16 years and their parents were the most inaccurate.

Compared with parents who either correctly estimated or overestimated the adolescents’ weight, those who underestimated their adolescents’ weight were less likely to report good dietary behaviors (Table 1). Similarly, adolescents who underestimated their weight were significantly less likely to report good diet behaviors. The pattern was similar, but weaker, for physical activity behaviors. Finally, adolescents who better estimated the severity of their weight, and whose parents better estimated the severity of the adolescent's weight, reported fewer barriers to healthy diet and exercise behaviors.

We found that the poor recognition of overweight seen among overweight adolescents (8–14) is also seen in adolescents with type 2 diabetes. What makes our findings particularly troublesome and important is that we studied adolescents who should be more cognizant of their weight status because they 1) have a diagnosis of type 2 diabetes, 2) are under physicians’ care, and 3) are severely overweight (mean BMI = 36.4 kg/m²).

There are some limitations to this study are that. First, it was conducted at a single academic medical center, and, second, although we have a relatively large sample, we are limited in our ability to analyze subgroup differences in the relationships between weight perceptions and health behaviors.

Overweight adolescents under physician care for type 2 diabetes, as well as their parents, failed to recognize the adolescents’ overweight status. Consistent with health behavior models, failing to recognize overweight was associated with poorer diet and exercise behaviors and more perceived barriers to performing diet and exercise. Clinicians should recognize that even extremely overweight children and their parents may not accurately perceive the presence of weight problems, let alone the negative consequences of failing to make difficult lifestyle changes that result in weight loss.

R.R. was supported by the Vanderbilt Physician Scientist Development Award and a National Institutes of Health K23 Career Development Award (DK-065294). A.C.S. was funded by an Agency for Healthcare Research and Quality National Research Service Award (H-T32-HS00032–14). M.W. is supported by a Department of Veterans Affairs Health Services Research and Development Service Senior Career Scientist Award.