The SEARCH study is an ongoing multi-center study based on ascertainment of cases of physician-diagnosed diabetes mellitus in youth younger than age 20 years beginning in 2001 and continuing through the present. SEARCH has six centers, located in Ohio, Colorado, Washington, South Carolina, Hawaii, and California. Diabetes cases were also identified among three American Indian populations in Arizona and Colorado and among members of the Gila River Pima Indian community participating in the National Institute of Diabetes and Digestive and Kidney Diseases Pima Indian Diabetes Study. A detailed description of the SEARCH study was published elsewhere.6
The study was reviewed and approved by the local Institutional Review Boards at each center. Parents of participants under age 18 years at the time of data collection provided informed consent and minor participants assented; all participants aged 18 years or older provided signed informed consent.
The SEARCH food frequency questionnaire (FFQ) has been described in detail previously.4
The FFQ consisted of 85 food lines for which the participant indicated if the item(s) was/were consumed in the past week (“yes/no”) and if yes, how many days, and the average portion size. The food groups were created by either collapsing food lines based on their major components, or by disaggregating composite foods into constituent foods. The FFQ was self-administered by study participants age 10 years or older after careful instruction by study staff. A small percentage (7%) of FFQs were interviewer-administered because of participant difficulties in form completion.
The nutrient and portion-size databases for this instrument were modified from the respective Diabetes Prevention Program databases, using the Nutrition Data System for Research (database 3 version 4.05/33, 2002, Nutrition Coordinating Center, University of Minnesota, Minneapolis) and industry sources. This analysis focused on dairy, fruits, vegetables, sweetened soda, total fiber, calcium, and percent of calories from saturated fat. This selection was based on the components’ likely influence on metabolic status and on the development of complications associated with diabetes.
Race and ethnicity were obtained through self-report using the standard census questions.7
Highest parental education was based on the parent with the highest education as reported on a questionnaire. Household income was assessed through self-report in predetermined categories. Type of diabetes was based on the clinical diagnosis by the physician. Diabetes treatment mode, diabetes duration, diabetes-related self-care, weight management and parental education on healthy food choices were included in the analysis as probable correlates of dietary adherence. Diabetes-related self-care was assessed by the question “How much of your own diabetes care do you do for yourself?” with answers categorized as “none”, “less than 25%”, “25–75%”, “more than 75%” and “all”. Diabetes education was assessed by the question to the parent “Have you been taught about how to make healthy food choices?” Weight management and physical activity were assessed by questions developed for the Youth Risk Behavior Surveillance System (YRBSS).8
Weight management goal was assessed with a question “Which of the following are you trying to do about your weight?” with possible answers “not trying to do anything”, “trying to lose”, “trying to gain”, and “trying to stay the same weight”. Physical activity was assessed by the question: “On how many of the past 7 days did you exercise or participate in a physical activity for at least 20 minutes that made you sweat and breathe hard?” Sedentary behavior was assessed by a slightly modified YRBSS question “On each weekday, about how much do you usually spend watching TV?”
This analysis included youth whose diabetes was prevalent in 2001 or incident in 2002–2005 who participated in the SEARCH examination. Of the 11,437 registered, valid subjects, 8,338 youth (73%) were contacted by phone and had an initial patient survey; of these, 5,293 youth (63%) attended the SEARCH clinic visit; 3,074 of whom were age-eligible (10 years and older at the study visit) and completed the dietary assessment. Among these youth, we excluded those with provider-defined diabetes type other than type 1 and type 2 or with type missing (n=27); diabetes duration less than 6 months (n=369); race/ethnicity other than non-Hispanic white, African American, Hispanic, Asian/Pacific Islander, or Native American (n=12); and those who reported eating much more or much less in the week when food intake was assessed than their typical week (n=154). A total of 533 observations were excluded from the analysis due to one or more criteria listed above; data from 2,541 youth were included in these analyses.
Analyses were conducted using the SAS (version 9.1, 2003, SAS Institute Inc, Cary, NC). Across subgroups of demographic, socio-economic, behavioral and diabetes-related variables, the medians of intake were compared by means of quantile regression.9
We also explored the overall significance of the association for covariates with more than 2 levels or categories (i.e. race/ethnicity, income, diabetes-related self-management and weight management) with Wald test. Due to the number of statistical tests conducted, alpha=.01 was used to determine statistical significance.