Academic journal article Journal of Eating Disorders

Adolescent Risk Factors for Purging in Young Women: Findings from the National Longitudinal Study of Adolescent Health

Academic journal article Journal of Eating Disorders

Adolescent Risk Factors for Purging in Young Women: Findings from the National Longitudinal Study of Adolescent Health

Article excerpt

Author(s): Eric M Stephen[sup.1], Jennifer Rose[sup.1], Lindsay Kenney[sup.1], Francine Rosselli-Navarra[sup.2] and Ruth Striegel Weissman[sup.1]


The etiologies of eating disorders and their core symptoms are best characterized by multifactorial models involving biological, psychological, interpersonal, and cultural variables [1]. Experts have also noted, however, that the research is limited by the dearth of prospective studies with samples large enough to detect small or medium sized effects, the inclusion of demographically limited samples, and the lack of follow-up periods that are long enough to fully capture the period of development when the risk of the onset of eating disorder symptoms is highest [1, 2].

The present study focused on purging (vomiting or abusing laxatives) to lose weight or keep from gaining weight as a clinically relevant symptom and as a possible proxy for the presence of an eating disorder. Purging behaviors have received increasing scientific attention not only in the context of a full syndrome eating disorder but also as symptoms of clinical relevance in their own right [3-5]. Only a few studies have explored risk factors for purging using longitudinal designs, and most of these studies have focused on just one risk factor. For example, two studies of adolescent girls found that eating dinner [6] or eating meals with their families [7] was associated with a decreased risk for onset of purging or extreme weight control behaviors.

Due to skip patterns used in most epidemiological studies of adults, information about purging behaviors typically was collected only in individuals who scored positively on the core symptom of anorexia nervosa (refusal to maintain a minimum adequate body weight) or bulimia nervosa (recurrent episodes of purging and binge eating) [8]. In contrast, studies of adolescents typically have measured disordered eating symptoms among all participants, thus affording the opportunity to explore correlates of these behaviors. Results show marked gender differences in the incidence and prevalence of purging [9]. For example, in a community sample of over 13,000 boys and girls (ages 9 to 14 years), one-year incidence of purging was 1.3% in girls and 0.3% in boys [6]. In a school-based sample of over 2,500 boys and girls, "extreme weight control behaviors" (which included purging) were also reported to be more common in girls than boys [10]. The present study capitalized upon the availability of data collected as part of the National Longitudinal Study of Adolescent Health (Add Health), which affords the unique opportunity to examine the prevalence and incidence of, as well as risk factors for, purging in a nationally representative sample of United States' (US) youth. Because of the low base rate of males who purged in our sample (only 0.55% of young men reported purging at any wave of study whereas 1.96% of young women reported purging), we excluded male participants from our sample.

Our approach to and selection of potential risk factors was informed by Kraemer and colleagues' conceptual framework and by a corresponding review paper of the empirical studies of risk factors for eating disorders [2, 11]. Specifically, to test whether a variable confers risk (rather than merely representing a correlate), data are needed that were either collected prior to occurrence of the behavior or are of a nature where one may reasonably infer that, even when collected after occurrence, the data are not contaminated by the outcome in question (e.g., gender, race/ethnicity). Therefore, in our study, we defined occurrence, or "incidence cases", as participants who reported purging at a subsequent wave of study, controlling for purging behaviors at Wave I. Given our interest in identifying adolescent risk factors for purging, we limited our sample to adolescent girls who, at study entry (Wave I), were age 18 or younger.

Our study included a considerable range of potential adolescent risk factors including demographic variables, early onset menarche, body image concerns, adverse childhood experiences, coping, and health or mental health problems [2]. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.