A systematic review of the methodology used to study weight change among young adults attending college.

Introduction: Young adulthood is a sensitive developmental period that is high-risk for weight gain. Ample research has focused on weight gain among college students; however meta-analyses report <2 kg pooled estimates of weight gain, which is in the range of normal weight fluctuation, and there is disagreement in the literature regarding common predictors of weight gain. These limitations pose a major barrier to targeted obesity prevention efforts. The present study reviewed the literature assessing college weight gain with a focus on three methodological factors that could contribute to variability in the literature: 1) use of an evidence-supported definition of weight gain (>2 kg or ≥3%); 2) weight measurement protocols; and 3) including weight/BMI in analyses of predictors of weight change.

Methods: Three databases were systematically searched. Studies were included in the review if the primary goal was to determine magnitude of weight change and/or test predictors of weight change during the academic year, and they reported weight at 2+ time points.

Results: A total of 81 studies were included in the review. Most studies (90%; 73/81) did not use an evidence-supported definition of weight gain. Studies that used an evidence-supported definition reported estimates of gain among students who gained weight to be beyond the range of normal weight fluctuation (4.0-7.5 kg), and occurred in a subset (<32%) of participants. Studies that did not use an evidence-supported definition reported weight gain to be 2.0-4.5 kg, and occurred in the majority >50% of students. Most studies that measured height and weight (71%; 42/59) did not use a fasting protocol and the majority (63%; 37/59) did not conduct measurements at the same time of day. A higher percentage of studies that used a standardized measurement protocol reported weight change >2 kg (44% vs 20%). A lower percentage of studies that used a standardized measurement protocol had substantial variability in weight change estimates (50% vs 69%). The majority of studies that tested predictors of weight gain (74%; 42/57) included weight/BMI as a covariate in analyses.

Conclusions: The body of literature examining weight change among college students suffers from limitations that may have contributed to overestimations in the percent of students who gain weight, and simultaneous underestimations of the magnitude of weight gain among those who gain weight. Weight gain may be limited to approximately 30% of students in a sample, and weight gain among this subset of students may be substantial (>4 kg). Going forward, use of both an evidence-supported weight gain definition and fasting measurement protocol will likely enhance accuracy in characterizing weight gain among college students, as well as improve researchers' ability to detect important predictors of weight gain.