Research Design and Methods
Adolescents with type 1 diabetes who were followed at a Diabetes Clinic at Children's Hospital in St. Paul, MN, were invited to participate in the AHEAD study (Assessing Health and Eating among Adolescents with Diabetes). The AHEAD study comprised two phases: 1) a paper-pencil survey sent to all eligible youth being followed at the clinic and 2) in-depth interviews with a subsample of adolescents who completed the survey and one of their parents. The present study describes the first phase of the AHEAD study.
Participants were between the ages of 12 and 21 years and had been diagnosed with type 1 diabetes at least 1 year before study participation. Eligible participants under age 18 years and their parents were invited to participate in the study by the adolescents' physicians, either through a mailed letter or during a clinic visit. After receipt of signed informed parental consent and adolescent assent, adolescent participants were mailed a survey to complete. For eligible participants older than 18 years, similar procedures were used, but only adolescent informed consent (sent together with the survey) was required. To improve response rate and minimize bias, study staff mailed invitation packets to nonresponders on two additional occasions. Concurrently, clinic staff distributed invitation packets to potential participants during their clinic visits. All participants received a small monetary incentive for completing the survey. All study procedures were approved by both the University of Minnesota and the Children's Hospital/Clinic Institutional Review Boards.
Of a potential 246 adolescents who were being seen in the clinic, surveys were completed by 143 youth, yielding a response rate of 58%. Responders and nonresponders were compared for BMI, age, and HbA1c levels. Mean BMI values were similar for responders (23.8 ± 6.0 kg/m2) and nonresponders (23.8 ± 4.2 kg/m2) (P = 0.989). However, responders tended to be older (15.2 ± 2.3 years) than nonresponders (14.3 ± 2.4 years) (P < 0.001). In addition, responders had lower HbA1c levels (8.8 ± 1.6%) than nonresponders (9.5 ± 2.0%) (P = 0.001).
The study population was equally split on gender (females: n = 70, 49%; males: n = 73, 51%). The mean age of the study population was 15.3 years (SD = 2.3). The mean BMI of the population was 23.8 kg/m2 (SD = 4.2). It is noteworthy that a large proportion of participants were overweight; 41% (n = 58) had a BMI greater than the 85th percentile; of these, 8% (n = 12) had a BMI greater than the 95th percentile. The majority of the participants (84%) lived in two-parent families, and more than half of participants (59%) reported that at least one parent had attained a level of education equivalent to a college degree or higher. While the majority of the population identified their race/ethnicity as white, the sample included 10% (n = 14) minority youth. This proportion is representative of the number of minority youth living in the geographic area of the study.
Items assessing weight perceptions and weight control behaviors were drawn from the Project EAT (Eating Among Teens) Survey. The Diabetes Eating Problems Survey (DEPS) was used to further assess disordered eating attitudes and behaviors and manipulation of insulin for weight control purposes among the respondents. Adolescent-perceived family functioning was assessed using three subscales of the Family Environment Scale-Revised (FES-R): cohesion, independence, and control. These subscales were chosen because they assess constructs of interest to the research questions. Responsibility for diabetes management was assessed with the Diabetes and Family Responsibility Questionnaire (DFRQ). The DFRQ assesses the adolescent's perceptions of who in the family (parent most of the time, adolescent most of the time, or both about equally) is responsible for 17 different diabetes-management tasks. Sociodemographic variables were based on self-report and included age, grade level, ethnicity, family structure, and parental educational level. BMI was based on the respondent's height and weight recorded during the medical visit that was closest to the date the adolescent filled out the survey. Metabolic control was assessed using the HbA1c value obtained at the medical visit that was closest to the date the adolescent filled out the survey. A more detailed description of each of these measures is included in Appendix A.
Demographic and weight control behavior frequencies were examined with cross-tabulations using the 2 as the test for significance. Frequency of weight-control categories (none, healthy, unhealthy, very unhealthy) was assessed separately by gender. Continuous scores on dependent variables were examined with means and were compared using ANOVA across demographic categories and categories of weight control behaviors. For all group comparisons, post hoc Tukey multiple comparison tests were conducted to assess differences between pairs of groups at P < 0.05. Two-tailed correlation coefficients between continuous variables were calculated using Pearson's r. Analyses controlling for age were also run. Patterns tended to be similar in age-adjusted analyses; therefore, these data are not presented. All analyses were completed using SPSS for the Macintosh, Version 6.1.
Diabetes Care. 2002;25(8) © 2002 American Diabetes Association, Inc.
Cite this: Weight Control Practices and Disordered Eating Behaviors Among Adolescent Females and Males With Type 1 Diabetes - Medscape - Aug 01, 2002.