Study design and setting
The target population of the MOPO study consisted of all conscription-aged young men during 2009–2013 (n = 5,864) in the city of Oulu in Northern Finland, which has approximately 193,000 inhabitants. Data collection took place at the mandatory call-ups for military service arranged annually by the Finnish Defense Forces. Exclusion criteria to participate in the call-ups for military service included having a severe physical or mental illness that did not allow for independent living. The study included a medical examination before the call-ups and questionnaires and physiological measurements conducted during the call-up day. The aim of the study was to provide knowledge on measured physical activity and young men’s attitudes towards it, as well as health, fitness, nutrition, health information-seeking behavior, and other life habits and cultures among young men.
In this study, data collected during 2010, 2011, and 2013 that included information on disordered eating behavior. In total, 3,542 young men participated, of which 65.6 % (n = 2,322) took part in the physiological measurements and 70.3 % (n = 2,490) completed the questionnaire. Thus, the present study consisted of all those 2,096 (59.2 %) men (mean age of 17.9, SD 0.7 years) who responded to the questions on disordered eating behavior.
The procedures of the study were in accordance with the Declaration of Helsinki. The Ethics Committee of the Northern Ostrobothnia Hospital District approved the study design (ETTM123/2009), and written informed consent was obtained after a complete description of the study was given to the participants. Participation in the study did not affect the participants’ military service or their future health care.
Study questionnaire
The study questionnaire included items on age, socio-economic status, disordered eating behavior, weight, health, physical activity, sedentary behavior, fitness, and motives to exercise.
Disordered eating behavior
The participants were asked to respond to the two subscales of the Eating Disorder Inventory-3 questionnaire (EDI-3) as an indicator for DEB. One subscale described drive for thinness (DT), and the other described bulimic behavior (BB) [25]. Each subscale consisted of seven questions with a 6-point Likert scale from never to always, but the response alternatives never and rarely were combined for the analyses. Thus, the scale was scored as 0 = never or rarely; 1 = sometimes; 2 = often; 3 = usually; and 4 = always [25, 26]. The first question has an inverse scoring. The questionnaire has been validated for Danish women [26]. Cronbach’s alphas as a measure for internal consistencies were 0.74 for DT and 0.85 for BB.
Weight
Perceived weight was determined from the responses to the following question: “What do you think about your weight?” A 5-point Likert scale was used for the response (1 = very underweight through 5 = very overweight). The response alternatives very underweight and somewhat underweight were combined, as was also done for very overweight and somewhat overweight. Weight history was investigated with the following questions: “Have you been overweight at some point of your life?” (1 = yes, slightly overweight; 2 = yes, severely overweight; 3 = no; and 4 = do not know, which was interpreted as never having been overweight for the analyses) and “Have you sometimes lost several kilos of weight?” (1 = no; 2 = yes, weight loss was planned and controlled; 3 = yes, weight loss turned uncontrolled).
Health
Self-rated health was investigated with the following question: “How is the state of your health?” (1 = good through 5 = poor). The response alternatives good and quite good were combined, as was also done for poor and quite poor.
Physical activity, sedentary behavior, and fitness
Daily physical activity was assessed with the following question: “How much do you approximately move during the day (e.g., while working, biking or walking to school or work, during the breaks at school, in household chores, or in hobbies and leisure time)?” (1 = <1 h; 2 = 1–2 h; 3 = >2 h). Sedentary behavior was identified as daily sitting hours during leisure time with the following question: “How much do you sit per day outside school or work (e.g., watching TV, reading, using the computer, and driving a car)?” Self-rated physical fitness was assessed with the following question: “How fit do you think you are compared to your peers?” The response options used a 5-point scale (1 = distinctly poorer through 5 = significantly better). The response alternatives distinctly poorer and somewhat poorer were combined, as was also done for significantly better and somewhat better.
Motives to exercise
The motives to exercise were investigated by a modified version of Nigg’s question [27]: “Are the following issues important for your exercising?” The response choices given were the binary options yes or no. The motives to exercise were classified according to Markland and Ingledew [28] and Ingledew and Markland [29] into four categories:
-
a)
body (body acceptance, enhancing appearance, increasing sex appeal, weight loss);
-
b)
health/fitness (enhancing muscle mass, improving or maintaining muscular strength and/or physical performance, promoting health, reducing stress);
-
c)
social engagement (competing and succeeding in athletics or sports, suggestion of a friend or family member, helping relationships, improving one’s respect from peers); and
-
d)
enjoyment (enjoying the feelings of euphoria gained by exercise, enjoying heavy exertion, increasing energy, improving spirits).
Measurements
The physiological measurements included anthropometry (height, body weight, body mass index, body composition), maximal isometric hand grip strength (as a measure of upper body strength) and aerobic fitness assessments.
Height was measured with a wall-mounted measuring tape and was recorded at a 0.5-cm precision. Body composition (weight with 0.1 kg accuracy, body mass index (BMI), fat free mass in kilograms at 0.1 kg accuracy, percentage body fat) was measured using bioelectrical impedance analysis (InBody 720 device, Biospace Co., Ltd., Seoul, Korea) with the participant standing without shoes and socks and wearing lightweight indoor clothes. BMI was classified according to the World Health Organization guidelines: underweight (<18.5), normal (18.5–24.9), and overweight (≥25) [30]. The percentage of fat free mass was calculated by dividing fat free mass by weight.
Grip strength (kg) was measured using a dynamometer (Saehan, SAEHAN Corporation, Korea). During the examination, the participant stood with his legs apart and his elbow at a 90° angle and was instructed to grip the instrument with maximum strength. The measurement was repeated, and the higher result was selected. The mean of both hands was used in the analyses.
Aerobic fitness was assessed by the Polar Fitness Test™ (Polar Electro, Finland), which estimates maximal oxygen uptake (mL/min/kg) based on resting heart rate, heart rate variability, gender, age, height, body weight, and self-assessed physical activity [31].
Statistical analysis
The cut-off scores for DEB were based on the 95th percentile in the Drive for Thinness [32]. This cut-off has also been previously used for the EDI-2 subscales in a sample of adolescents [33]. Thus, a score of ≥11 was classified as having DT. In the Bulimia subscale, the 95th percentile was at 7 points, but because 7 points could be earned by answering sometimes to each question, those having ≥8 points were classified as having BB.
The group differences between those with and without DEB (DT and/or BB) were compared with the independent samples t-test or with the nonparametric Mann–Whitney U test. The Chi-square test was used for categorical variables. The level of significance was retained at the p-value <0.05. All variables significantly associated with DEB in the univariate analysis were entered in the binary forward stepwise logistic regression. The results are presented as odds ratios (OR) with 95 % confidence intervals. The Nagelkerke R-square was used for explaining the variance of risk of having a DEB trait. Statistical analyses were performed using PASW Statistics 18 for Windows.