health behaviors that influence college freshman weight change

Before beginning their first year at a northeastern public university in fall, 2011, incoming freshmen were re- cruited to participate in a 1-year study on physical activity, incentives, and height/weight change; this study was called Burn & Earn. Students were re- cruited viamailings to their homes dur- ing summer, 2011; 117 participants enrolled. Over the course of their fresh- man year, data on students’ height and weight were collected 4 times: once at the beginning and once at the end of each semester. More information about this study and its results can be found elsewhere.9,10 The University of Vermont’s Committee on Human Research in the Behavioral and Social Sciences approved the study protocol.

For the current study, participants from the 2011 Burn & Earn study were

1

Table. Mean Weight, Height, and Body Mass Index (BMI) Values for Participants at

Students experience continued weight gain over all 4 years of college.

2 Pope et al Journal of Nutrition Education and Behavior � Volume -, Number -, 2016

contacted again in spring, 2015, when they were seniors, and asked to partici- pate in an extension of the original study involving anthropometric assess- ments and information about their cur- rent health behaviors. Respondents scheduled a study assessment visit, when they completed a food frequencyques- tionnaire (FFQ), an alcohol use disor- ders identification test (AUDIT), and a general questionnaire. The general questionnaire was exploratory and as- sessed behaviors previously shown to be related to weight, including phys- ical activity frequency, living situation, grocery shopping habits, specific diet, relationship status, and work situa- tion.11,12 At the assessment visit, study personnel alsomeasured students’ height and weight. A digitally calibrated scale (model BW800S; Tanita, Arlington Heights, IL) was used to measure parti- cipants’ weight; height (in centimeters) was assessed using awall-mounted sta- diometer (Seca, Hanover, MD). Upon completion of the assessment visit, students were compensated $50 for participation. The FFQs were submit- ted to Nutrition Quest (Berkeley, CA) for analysis and the remainder of the data was compiled in the university laboratory. Changes in height,weight, andBMIwere calculatedby comparing 2011–2012 data with 2015 data.The researchers used SPSS (version 23, IBM Corporation, Armonk, NY, 2015) to analyze the data. A paired- samples t testwasperformedtoexamine anthropometricchangesbetweenfresh- man and senior years. An independent- samples t testwas employed to examine differences inweight trajectorybetween originally overweight or obese students and originally normal weight students. The researchers performed regression analyses to determine associations be- tween health behaviors and changes in BMI/weight.

the Beginning and End of Freshman Year and End of Senior Year (n ¼ 86)

Weight, kg Height, cm BMI, kg/m2

Beginning of freshman year (fall, 2011)

66.94 (14.02) 168.23 (8.88) 23.54 (4.00)

End of freshman year (spring, 2012)

68.26 (13.74) 168.51 (9.00) 23.98 (3.93)

End of senior year (spring, 2015)

71.32 (15.60)a,b 169.13 (8.95)a,b 24.84 (4.46)a,b

aSignificantly different from beginning of freshman year to end of senior year (P < .001, paired-samples t test); bSignificantly different from end of freshman year to end of senior year (P < .001, paired-samples t test). Note: Data are shown as mean (SD).

RESULTS

Of the original 117 participants, 86 re- turned to the laboratory for follow-up measurements; this represented a 73.5% retention rate. Analyses of weight, BMI, and height change were performed using only the 86 partici- pantswhohad complete data. Boxplots of weight and BMI change indicated several outliers; however, there was no scientific justification for treating out-

liers differently. Histograms and q–q plots of weight change, BMI change, and the residuals of weight and BMI change indicated that the assumption of normality needed for valid t tests could be accepted. No significant differ- ence was observed in BMI at the begin- ning of freshman year between those who participated in the senior year study and thosewho did not participate (F1,115 ¼ 0.14, P ¼ .69). Of the 86 stu- dents who completed the study, 58 were women and 28weremen. Average agewas21.5years.Ninety-threepercent of participants were white and 7%were Asian. Four students reported beingHis- panic or Latino. As seen in the Table, there were significant increases in mean BMI, weight, and height from the beginning of freshman year to the end of senior year, as well as from the end of freshman year to the end of se- nior year. Mean weight gain was 4.38 kg over 4 years, 2.94 kg of which occurred after freshman year. Mean height increase from the beginning of freshman year to the end of senior year was 0.9 cm, when men gained 1 cm on average andwomengrewby0.85cm.

The number of overweight (BMI ¼ 25–29.9) and obese (BMI$30) students increased from 20 at the beginning of freshmanyear to35by theendof senior year. At the conclusion of senior year, 41% of the sample was overweight or

obese (Figure). Only 11.6% of students maintainedweightwithin1 kgof fresh- man year weight. A total of 21% of stu- dents maintained weight within 3% of freshman year weight over 4 years. Ste- vens et al13 proposed a 3% weight change as the definition of weight maintenance. Twenty-two students, or 26% of the sample, maintained BMI within �3% of their freshman year BMI. There was no significant differ- ence in total change in BMIover 4 years between those who entered their fresh- man year in the normal BMI range (n ¼ 66) and those who entered with a BMI in the overweight/obese categories (n¼ 20) (P ¼ .93).

Based on the results of the FFQ, stu- dents’ average caloric intake was 1,889 kcal/d. Mean fruit and vegetable consumption was 1.06 and 1.46 cups/d, respectively. Students consumed an average of 203 kcal/d of alcohol. Women averaged 164 kcal of alcohol, whereas men averaged 284 kcal of alcohol. The AUDIT results showed that 62% of the sample consumed alcohol 2–3 times/wk, and the most frequent response for number of drinks when drinking was 3–4 drinks. The AUDIT scores also indicated that 96.5% of participants were at very low risk for problem drinking behavior. Regression analyses indi- cated that there were no significant as- sociations between weight change or BMI and lifestyle factors, including how students commuted to campus, relationship status, or where they ob- tained their food. Only 15% of partici- pants regularly met physical activity guidelines of achieving 30 minutes of moderate physical activity 5 times/wk.

Overweight/Obese Fall 2011

Overweight/Obese Spring 2015

Healthy Weight Fall 2011

Healthy Weight Spring 2015

= Overweight/Obese to Healthy Weight

= Healthy Weight to Overweight/Obese

= Stayed Overweight/Obese

*49 students remained in the healthy weight BMI category throughout the study and are not depicted in the figure.

Figure. Movement between body mass index categories for participants who started freshman year in the overweight/obese category (fall, 2011) and for those who ended senior year in the overweight/obese category (spring, 2015).

Journal of Nutrition Education and Behavior � Volume -, Number -, 2016 Pope et al 3