The Midlife Health Study is a cross-sectional study that was conducted during 2000–2004 in the Baltimore metropolitan region. Detailed methods of this study have been published (Gallicchio et al., 2008
; Visvanathan et al., 2005
). Participants gave written informed consent according to procedures approved by the University of Maryland School of Medicine, University of Illinois, and Johns Hopkins University Institutional Review Boards.
Briefly, women aged 45 to 54 years of age were recruited by mass mailing an invitation to area households in the Baltimore metropolitan region. Women who were interested in participating were invited to call the clinic for more information. To be eligible for the study, women had to be between 45 and 54 years of age, have intact uterus and ovaries, and report at least three menstrual periods in the last 12 months. Women were excluded if they were pregnant, taking hormonal therapy or contraception, or had history of cancer of the reproductive organs.
At the clinic visit, women completed a 26-page questionnaire and had their height and weight measured. In total, 639 women completed the questionnaire and clinic visit.
Depressive symptoms were assessed using the Center for Epidemiologic Studies – Depression scale (CES-D), a 20-item, validated, self-report scale (Radloff, 1977
). Participants scoring 16 or greater were categorized as “depressive symptoms present” and those scoring 15 or less were categorized as “depressive symptoms absent” (Radloff, 1977
). The frequency of fast-food intake was measured by self-report using the question “How often did you eat foods from the following restaurants during the past year?” Participants were asked to score the frequency of food consumption for the following types of fast-food restaurants: “fried chicken”, “burgers”, and “fried fish” on a six-point scale (0, “never in past year”; 1, “1–4 times in past year”; 2, “5–11 times in pass year”; 3, “1–3 times a month”; 4, “once a week”; 5, “2–4 times a week”; or 6, “almost every day”). The responses were converted to median number of restaurant visits per month, and then added to yield total number of fast-food restaurant visits per month. Based on the distribution of the data, this variable were categorized as “at least weekly”, “at least monthly, but less than weekly” and “less than monthly.”
The following factors were considered as possible confounders in the analysis: age, race, marital status, education, annual household income, body mass index (BMI), current smoking, leisure physical activity, current alcohol use, and antidepressant use. The collection of data pertaining to these covariates has been reported previously (Gallicchio et al., 2008
; Visvanathan et al., 2005
Participants missing data on either depressive symptoms (n=6) or fast-food intake (n=7) were excluded, leaving 626 participants in the analytic dataset. Bivariate analyses were conducted using chi-square tests. Variables that were associated either with depressive symptoms or with fast-food intake in bivariate analyses at p<0.1 were considered potential confounders and inserted into the regression model. Ordinal logistic regression analysis was used to examine both the unadjusted and confounder-adjusted associations between depressive symptoms and the frequency of fast-food intake. Statistical analyses were performed using SAS version 9.2 (Cary, North Carolina).