Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jul 25.
Published in final edited form as: Int J Eat Disord. 2026 Jan 26;59(5):919–928. doi: 10.1002/eat.70039

Predictive Validity of Purging Disorder: A 10-year follow-up study comparing prevalence and psychosocial outcomes for alternative criteria sets

Holly K Spinner 1, Alissa Haedt-Matt 2, Pamela K Keel 3,*
PMCID: PMC13398342  NIHMSID: NIHMS2197161  PMID: 41588570

Abstract

Objective:

Specific diagnostic criteria for purging disorder (PD) are not currently provided in the DSM-5. The current study aims to compare point prevalence estimates and evidence of predictive validity for four alternative criteria sets for PD for men and women.

Method:

Survey data collected in college (N=3291) and at 10-year follow-up (n=2223) were used in analyses. Point prevalence was assessed at both time points and compared in men versus women as well as in college versus at 10-year follow-up. Predictive validity was assessed using multiple linear regressions to determine prospective associations for each PD definition versus no eating disorder across external validators.

Results:

Point prevalence was greater in women than in men in college across all definitions of PD, and no sex differences were found at follow-up. Point prevalence was greater in college than at follow-up in all but the narrowest definition of PD which required purging, on average, twice per week. Predictive validity was supported for PD defined by self-induced vomiting, laxative or diuretic misuse, with evidence of lower satisfaction with work and career at follow-up.

Discussion:

Together, findings support defining PD through self-induced vomiting, laxative use, or diuretic use to influence body shape or weight, on average, at least once per week. Specifying diagnostic criteria for PD that demonstrate evidence of concurrent and predictive validity will contribute to improved diagnostic reliability, screening, assessment, and support future research efforts to improve treatment.

Keywords: purging disorder, predictive validity, psychosocial outcomes, unhealthy weight control behaviors, other specified feeding, eating disorder

Introduction

Purging disorder (PD) is currently included within the category of “other specified feeding and eating disorder” (OSFED) in the DSM-5 (American Psychiatric Association [APA], 2013). While the DSM-5 defines PD through recurrent purging (self-induced vomiting and misuse of laxatives and diuretics) to influence body shape or weight in the absence of binge eating (APA, 2013), specific diagnostic criteria for PD, similar to other OSFED, are not provided. The DSM-5 workgroup deliberately left OSFED criteria unspecified to offer flexibility to clinicians and encourage empirical approaches in research, and data are needed to address the clinical utility of alternative definitions (Keel, 2024). Establishing the predictive validity of a specific criterion set for PD will reduce heterogeneity in what is considered “PD” and rampant replication problems in the OSFED literature (Attia & Guarda, 2024). Indeed, key recent work on OSFED nosology demonstrated mixed findings on the differences between PD and control groups. Dang et al.’s (2024) systematic review supported significantly elevated general psychopathology in PD compared to controls that was not supported in meta-analyses. Given that all studies individually supported greater pathology in PD included in the meta-analysis, nonsignificant findings reflect within-study sampling error and between-study variability captured by random effects models. As such, more studies are needed to refine how PD is defined.

A minimum frequency criterion for “recurrent” purging has not been specified for PD in the DSM-5. Across studies, minimum thresholds have ranged from an average of once (Allen et al., 2013; Forney et al., 2022; Glazer et al., 2019; Hay et al., 2017; Krug et al., 2022; Krug et al., 2020; Lydecker et al., 2017; Micali et al., 2015; Mitchison et al., 2019; Riesco et al., 2018; Stice et al., 2013; Stice et al., 2025) to twice per week (Hammerle et al., 2016; Keel et al., 2008; Wade et al., 2007), with most recent studies defining recurrent as once per week. Furthermore, it is not uncommon for studies to omit minimum purging frequency used for defining PD (Bodell et al., 2022; Silen et al., 2020). Cross-sectional data show support for the concurrent validity of PD at both a minimum frequency of twice per week (Haedt & Keel, 2010; Keel et al., 2008; Wade, 2007) and once per week (Allen et al., 2013; Davis et al., 2020; Haedt & Keel, 2010; Krug et al., 2022; Krug et al., 2020; Lydecker et al., 2017; Stice et al., 2013; Stice et al., 2025). With both thresholds, PD is associated with greater impairment (Keel et al., 2008; Stice et al., 2013; Stice et al., 2025), eating pathology, and non-eating disorder psychopathology (Allen et al., 2013; Krug et al., 2022; Krug et al., 2020; Lydecker et al., 2017; Wade et al., 2007) compared to non-eating disorder controls. PD defined with a minimum purging frequency of once per week is also associated with greater impulsivity compared to non-eating disorder controls (Davis et al., 2020). Haedt and Keel (2010) found significant concurrent associations between presence of PD, at both frequencies, and lower satisfaction with relationships and school performance and greater perfectionism. In addition, the lower minimum frequency of once per week allowed more individuals to move from the heterogenous Eating Disorder Not Otherwise Specified diagnosis to a diagnosis of PD while retaining homogeneity of the PD category (Haedt & Keel, 2010).

In contrast to the DSM-5’s specification of purging in the form of self-induced vomiting and misuse of laxatives and diuretics, some PD definitions do not include diuretic misuse (Allen et al., 2013; Bodell et al., 2022; Glazer et al., 2019; Krug et al., 2022; Lydecker et al., 2017; Micali et al., 2015; Mitchison et al., 2019; Riesco et al., 2018), and others include non-purging behaviors, such as excessive exercise (Lydecker et al., 2017), detoxes (Mitchison et al., 2019), and fasting (Hammerle et al., 2016). In addition, a broader compensatory eating disorder (CED) has been proposed to capture individuals who use inappropriate purging and non-purging behaviors (e.g., excessive exercise, fasting) to influence body shape or weight in the absence of objective binge eating (Tobin et al., 1997). In support of this approach, Lydecker et al. (2017) found no significant differences in eating pathology between individuals who reported compensatory driven exercise and those who reported vomiting or laxative use at least once a week, and both groups reported more psychopathology than non-eating disorder controls. CED shows similar levels of eating pathology as bulimia nervosa (BN), and greater eating pathology compared to binge eating disorder (BED; Tobin et al., 1997). Moreover, no significant differences were found in non-eating disorder psychopathology between individuals with CED and those with BN (Tobin et al., 1997). Similarly, Watson et al. (2013) found individuals who fit the CED criteria of unhealthy weight control behaviors (UWCBs) at least once per week could not be distinguished from individuals with other eating disorders on eating pathology and other psychiatric comorbidities. Likewise, Haedt and Keel (2010) found a broad definition of PD including UWCBs was associated with greater perfectionism compared to non-eating disorder controls. In contrast, only definitions of PD that included self-induced vomiting, laxative use, or diuretic use demonstrated evidence of impairment in social and educational areas of functioning (Haedt & Keel, 2010). These findings highlight the value of comparing alternative criteria sets when examining the validity of a disorder versus studying one definition in isolation.

The DSM-5 distinguishes three forms of validity in defining a mental disorder: antecedent, concurrent, and predictive (APA, 2013). Predictive validity can be determined through evidence of prognostic value using external indicators that are not diagnostic features of the disorder (Keel & Striegel-Moore, 2009). Prognostic indicators can include impairment in social, occupational, or other important areas of functioning as markers of clinical significance as well as other psychosocial outcomes (APA, 2013). Importantly, studies focused on concurrent validity provide limited support for the clinical utility of alternative definitions (Keel & Striegel-Moore, 2009), and most prospective research on PD has focused on prevalence, incidence, and diagnostic stability rather than outcomes (Allen et al., 2013; Glazer et al., 2019; Silen et al., 2020; Stice et al., 2013; Stice et al., 2025). To our knowledge, only one study has examined prospective associations between PD and impairment (Micali et al., 2015). In adolescents, a diagnosis of PD defined by purging through use of self-induced vomiting or laxatives (but not diuretics) at least weekly and binge eating less than monthly prospectively predicted depression, anxiety, drug use, and non-suicidal self-injury at 2-year follow-up (Micali et al., 2015). To our knowledge, no studies have examined the predictive validity of PD in adults over a 10-year period, and none have examined predictive validity across alternative criterion sets.

Establishing a reliable definition for a proposed diagnostic category is a prerequisite for determining the utility of the diagnosis (Keel & Striegel-Moore, 2009). The current study seeks to clarify the potential diagnostic boundaries for PD by directly comparing differing definitions on their predictive validity. Prior research has established meaningful distinctions between PD and other well-established DSM-5 eating disorders on biological and psychological features (Keel, 2019) and work is needed to indicate a reliable definition for PD. The current study aims to (1) provide updated point prevalence estimates for PD both in college and 10 years later across four PD definitions and (2) examine the predictive validity of each PD definitions through associations with external validators at 10-year follow-up. We hypothesized higher prevalence estimates for PD definitions that include UWCBs compared to definitions restricted to purging behaviors whereas reducing minimum symptom frequency would have a modest impact on point prevalence. Additionally, reflecting greater eating disorder prevalence in women than in men (APA, 2013) and in individuals in their 20’s compared to their 30’s (Hudson et al., 2007; Ward et al., 2019), we predicted higher PD prevalence in women than in men across all definitions and at baseline compared to 10-year follow-up. Finally, we hypothesized that PD defined by purging at least once a week would demonstrate the greatest predictive validity, consistent with cross-sectional findings (Haedt & Keel, 2010).

Methods

Participants

Participants included men (n=1097) and women (n=2194) from a northeastern university randomly selected to complete a survey on health and eating patterns in 1982, 1992, 2002, and 2012 and followed 10 years later (n=2223). The sample had a mean (SD) age of 19.94 (1.73) years in college and 30.46 (1.71) years at follow-up. Participants had a mean (SD) body mass index (BMI) of 22.12 kg/m2 (2.96) in college and 23.14 kg/m2 (3.76) at follow-up. The racial composition of the sample was 6.4% African American or Black, 16.3% Asian, 0.5% American Indian/Alaska Native, 0.4% Native Hawaiian/Pacific Islander, 66.6% White, and 9.9% reported “Other” or did not report their race. Ethnicity was endorsed as Hispanic/Latino by 6.6% of participants. Over two thirds (67.5%) of individuals who completed baseline surveys completed surveys at 10-year follow-up. Participant sex and race were significantly associated with attrition, with greater retention in women (χ2 (1) = 40.93, p < .001) and individuals who identified as White (χ2 (5) = 64.91, p < .001). Ethnicity was not significantly associated with attrition (χ2 (1) = 3.55, p = .06). See Data Analyses for handling of missing data.

Measures

Surveys included questions about the presence and weekly frequency of binge eating, self-induced vomiting, laxative use, diuretic use, diet pill use, fasting, and, beginning with surveys distributed in 1992, compulsive exercise to influence body shape or weight. The current measure of compulsive exercise demonstrated unique prognostic value in predicting both eating pathology levels and eating disorder diagnosis (Holland et al., 2014). The Eating Disorder Inventory (EDI; Garner et al., 1983) Drive for Thinness subscale (α = .92 in college; α = .90 at follow-up) provided items used to determine presence of body image disturbance.

All PD definitions required participants to endorse body image disturbance, absence of current binge eating (objective and subjective) for differential diagnosis from BN, and a BMI ≥ 18.5 kg/m2 for differential diagnosis from anorexia nervosa (AN). Replicating prior work (Haedt & Keel, 2010), body image disturbance included an intense fear of gaining weight or being fat or overconcern with weight or shape based on items from the Eating Disorders Inventory (Garner et al., 1983). Although subjective binge eating is common in PD (Smith et al., 2018) and many individuals who self-report binge eating do not consume objectively large amounts of food, we adopted a conservative approach to our survey-based data by excluding individuals who endorsed current bingeing from the PD diagnoses to ensure no cases of BN were included. Following methods used by Haedt and Keel (2010), four PD definitions were created through varying behavior type (purging vs. UWCBs) and minimum weekly frequency of behaviors (twice vs. once per week) (see Figure 1). For aims related to predictive validity, non-eating disorder controls included individuals who did not meet diagnostic criteria for any DSM-5 eating disorder, including PD by any definition, at baseline.

Figure 1.

Figure 1.

Open in a new tab

Four definitions of purging disorder

Predictive validity was defined as the extent to which individuals with PD exhibited significantly poorer psychosocial outcomes than individuals with no eating disorder at 10-year follow-up. To extend previous evidence of PD’s concurrent validity (Haedt & Keel, 2010), we chose satisfaction with friendships, relationships, work, and career and levels of perfectionism and alcohol use as our psychosocial validators. Participants reported satisfaction with their friendships and romantic relationship (if relevant) at both timepoints and reported satisfaction with schoolwork in college and satisfaction with work position and career choice at follow-up on 7-point Likert scales. Perfectionism was assessed using the EDI (Garner et al., 1983) Perfectionism subscale. Internal consistency was α = .75 in college and α = .76 at follow-up. Higher scores on this subscale have been shown to predict the onset and maintenance of eating disorders (Holland et al., 2013). Finally, participants self-reported alcohol use in number of drinks per week.

Data Analyses

All variables were assessed for distribution assumptions prior to conducting parametric analyses. Point prevalence was calculated for each definition of PD in the full sample and in women and men separately during college and at 10-year follow-up. Sex differences in prevalence were examined using a χ2 statistic or Fisher’s exact test for expected cell sizes less than 5 (Kim, 2017). Differences between baseline and follow-up were examined using related-samples McNemar tests (Lu et al., 2017; Zar, 2010). Following guidance from Zar (2010), McNemar χ2 statistics were calculated by hand. Effect sizes for McNemar tests (ψ) were calculated as the maximum difference between the discordant pairs in the 2 × 2 contingency tables (Mangiafico, 2016; Zar, 2010).

Missing data on external validators at follow-up were examined using Little’s missing completely at random test and found not to be missing completely at random (p < .001; Little, 1988). Missing values were not imputed for relationship satisfaction because responses were conditional on participants being in a relationship. Missing data for remaining external validators were imputed at follow-up using expectation-maximization (Schafer & Graham, 2002), which demonstrates low bias in large samples with high retention (Little, 2024; McNeish, 2017). Using this method, we imputed 32.8–33.8% of data at follow-up on external validators. Analyses using listwise deletion are included in the on-line supplement and did not reveal meaningful differences from findings using imputation.

Predictive validity for each PD definition was assessed using multiple linear regressions with each external validator at 10-year follow-up as the dependent variable and PD vs. non-eating disorder control status (coded as 0 = non-eating disorder control and 1 = PD) as the independent variable, with baseline levels of the outcome and participant sex included as covariates. Negative beta values for which 95% confidence intervals did not include 0 were interpreted as signifying significantly lower satisfaction in PD compared to control participants. Positive beta values for which 95% confidence intervals did not include 0 were interpreted as supporting greater perfectionism and alcohol use in PD compared to control participants. A priori hypotheses were tested with one-tailed p-values because significant findings in the opposite direction would be uninterpretable. Specifically, our definition of predictive validity for PD is contingent on individuals in this diagnostic group exhibiting poorer outcomes than controls, rendering findings in the opposite direction uninterpretable. Each regression model included non-eating disorder controls and participants with the definition of PD being tested. Thus, models with more restrictive definitions excluded participants who met criteria for PD with a broader definition. Given recommendations from Rothman (1990) to avoid artificially inflating Type II error rates, analyses do not employ multiple testing corrections due to the pre-planned and theory-driven nature of the hypotheses. Statistics were conducted in IBM SPSS 29.0.1, and p<.05 was the threshold for statistical significance.

Results

Point Prevalence

Table 1 displays point prevalence estimates for PD definitions. During college, point prevalence estimates modestly increased with reduced minimum symptom frequency. In contrast, point prevalence demonstrated a 9-fold increase for definitions including both purging and non-purging behaviors at both frequency thresholds. Similar results were found when examining changes in prevalence by sex. In women, the range of symptom behaviors included (purging v. both purging and non-purging) produced a clinically meaningful change in point prevalence, from less than 1 in 100 women to approximately 1 in 15 to 20 women. Similar results were found in men, such that expanding definitions to include all UWCBs shifted prevalence from 1 in 1000 men to approximately 1 in 40 to 56 men. In college, PD prevalence was higher in women than in men across all definitions, with small effect sizes.

Table 1.

Summary of prior impairment data across varied definitions for Purging Disorder

Variable Compared with No ED Compared with other EDs
Purging UWCBs Purging UWCBs
2x/week 1x/week 2x/week 1x/week 2x/week 1x/week 2x/week 1x/week
Functional Impairment Increaseda Increaseda; NSa NSa
Distress Increaseda Increaseda Decreaseda
Eating Pathology Increaseda Increaseda Increaseda Increaseda; NSb Increaseda; Decreaseda; NSa Increaseda; NSa
General Psychopathology Increaseda Increaseda Increaseda Increaseda; NSa NSa NSa
Satisfaction with Relationships Decreaseda Decreaseda NSa No differencea
Satisfaction with School Decreaseda Decreaseda NSa NSa
Perfectionism Increaseda Increaseda Increaseda Increaseda
Impulsivity Increaseda; NSa Increaseda NSa Decreaseda; NSa NSa
Quality of Life Decreaseda NSa
Depression Increaseda Increaseda, b Increaseda Decreaseda Decreaseda NSa Decreaseda
Anxiety Increaseda Increasedb Decreaseda Increaseda
Drug Use Increasedb
NSSI Increasedb
Suicidality Increaseda NSa
Recovery Status NSb
ED Presence at Follow-up NSb
Open in a new tab

Note. ED = eating disorder. UWCB = unhealthy weight control behavior. NSSI = non-suicidal self-injury. Increased indicates that those with purging and/or UWCBs demonstrated higher scores on the variable. Decreased indicates that those with purging and/or UWCBs demonstrated lower scores on the variable. NS = nonsignificant and indicates significant differences were not found. Cells with differing findings reflect multiple studies or multiple comparison groups.

a

Cross-sectional associations

b

Prospective associations

At 10-year follow-up, changing minimum behavior frequency produced modest changes in point prevalence and only for the definition that included both purging and non-purging behaviors. In contrast, expanding definitions from purging to include non-purging behaviors created a 25 to 29-fold increase in prevalence for the once per week threshold and twice per week threshold, respectively. Similar results emerged when examining the sample by sex, such that expanding definitions to include all UWCBs produced clinically meaningful change in point prevalence. In women, including all UWCBs increased prevalence from 1 in every 1000 women to 1 in every 32 to 45 women. In men, including all UWCBs increased prevalence from 1 in every 500 men to 1 in every 25 to 31 men. At follow-up, differences in PD prevalence between men and women were not statistically significant across any definition (ps ≥ .17). Related-samples McNemar tests revealed point prevalence decreased significantly from college to 10-year follow-up, for all but the definition that required purging twice weekly. Effect sizes for the narrow definition of PD were large for both symptom frequency thresholds. Effect sizes ranged from small to medium for the broad definitions of PD.

Predictive Validity

Table 2 presents results from multivariable linear regressions predicting negative psychosocial outcomes based on presence of PD vs. no eating disorder in college for each PD definition. Initial analyses of frequency of alcohol use revealed considerable skewness and kurtosis in residuals and a square root transformation was applied to correct for this. PD defined by purging predicted worse work and career satisfaction at both the once and twice weekly thresholds, with small effect sizes. Despite larger sample sizes and increased statistical power for predicting outcomes with more inclusive PD definitions, only one significant prospective association emerged. PD including purging and non-purging behaviors at least once per week predicted higher perfectionism at follow-up, with a small effect size. Similar effect sizes for this association were observed across all four definitions, suggesting equivalent predictive associations between all definitions and perfectionism as an outcome.

Table 2.

Point prevalence across each definition of purging disorder in college and at 10-year follow-up

Purging Unhealthy Weight Control Behaviors
2x/week 1x/week 2x/week 1x/week
n % (95% CI) n % (95% CI) n % (95% CI) n % (95% CI)
College
 Full Samplea 11 0.3 (0.2, 0.6) 16 0.5 (0.3, 0.7) 109 3.3 (2.9, 4.0) 152 4.6 (4.1, 5.4)
 Women 13 0.6 (0.3, 1.0) 19 0.9 (0.5, 1.4) 105 4.8 (4.0, 5.8) 149 6.8 (5.9, 8.0)
 Men 1 0.1 (<.01, 0.5) 1 0.1 (<.01, 0.5) 20 1.8 (1.1, 2.9) 27 2.5 (1.7, 3.7)
 Sex Difference, χ2(1), φ 4.26* .036 7.14* .047 17.02* .073 26.45* .091
10-year Follow-up
 Full Samplea 2 0.1 (<.01, 0.3) 3 0.1 (<.01, 0.4) 58 2.7 (2.1, 3.4) 75 3.5 (2.9, 4.3)
 Women 1 0.1 (<.01, 0.4) 2 0.1 (<.01, 0.5) 35 2.2 (1.6, 3.1) 48 3.1 (2.3, 4.1)
 Men 1 0.2 (<.01, 0.9) 1 0.2 (<.01, 0.9) 21 3.2 (2.0, 4.9) 26 3.9 (2.7, 5.9)
 Sex Difference, χ2(1), φ .40 −.013 .02 −.003 1.79 −.029 1.21 −.024
 Age Difference, McNemar test χ2 (1), ψ 3.6 4 5.4* 4 7.64* 1.63 17.04* 1.88
Open in a new tab
*

p<.05

Notes. Purging included self-induced vomiting, diuretic use, and laxative use. Unhealthy weight control behaviors included self-induced vomiting, diuretic use, laxative use, compensatory exercise, fasting, and diet pill use.

a

Estimates in full sample were weighted by sex to account for greater representation of females (66.7%) in the sample and achieve a 50/50 distribution.

Discussion

Consistent with hypotheses, point prevalence of PD was significantly greater in definitions that included compulsive exercise, fasting, and diet pill use compared to definitions that only included purging. In contrast, shifts in symptom frequency from twice to once per week had little to no impact on point prevalence. Sex differences in PD prevalence were consistent across all definitions in college, with higher prevalence in women compared to men in college but not at 10-year follow-up. In all but the narrowest PD definition, prevalence was greater in college than at 10-year follow-up, pointing to a decrease in eating pathology as individuals age from their 20s to their 30s. Finally, PD defined by purging behaviors showed predictive validity over 10-year follow-up in the forms of significantly worse work and career satisfaction compared to those without eating disorders. For definitions requiring purging, altered minimum symptom frequency thresholds did not produce different evidence of impairment. Therefore, using the lower symptom frequency threshold of at least once per week (as opposed to at least twice per week) captures a greater number of individuals for whom a PD diagnosis provides prognostic value. Additionally, the symptom threshold of at least once per week is consistent with behavioral thresholds used to diagnose BN and BED (APA, 2013). Collectively, findings support the predictive validity of the DSM-5 definition of purging through use of self-induced vomiting, laxative or diuretic misuse, and further support use of a minimum frequency of once per week, on average, to define “recurrent.”

In accordance with Haedt and Keel (2010), findings across various PD definitions align with epidemiological patterns for AN and BN in terms of associations with sex (APA, 2013) and age (Hudson et al., 2007; Ward et al., 2019). Interestingly, no definitions of PD produced sex differences in prevalence at 10-year follow-up. This may indicate sex differences in trajectory of use of UWCBs. This parallels previous evidence that disordered eating tends to decrease or stabilize in women and increase in men with age (Brown et al., 2020; Tiggemann & Lynch, 2001) and eating disorders often take the form of compensatory behaviors, including excessive exercise, in adult men (Kummer & Mangweth-Matzek, 2023). Although statistical power is reduced by the low prevalence of PD at 10-year follow-up, the observed prevalence of PD in men was modestly higher than in women, indicating the non-significant difference is not due to inadequate power.

Moreover, findings suggest definitions including UWCBs create more heterogenous groups that are less distinguishable from non-eating disorder controls on psychosocial outcomes 10-years later. Similar to findings at long-term follow-up of BN (Keel et al., 2000), PD impacted psychosocial adjustment at 10-year follow-up. However, effects were most pronounced for work-related indicators in the current sample of PD whereas prior results for BN indicated greater impairment in interpersonal domains and minimal impact on work. Differences may reflect differences between BN and PD or may reflect that the prior study of BN was restricted to women, and the current study of PD includes women and men. As college is a critical period for career and identity development, the cognitive and physical tolls of purging during emerging adulthood may have lasting impacts on the satisfaction with one’s work and career. This is consistent with literature demonstrating individuals with eating disorders tend to have worse work performance than individuals without eating disorders (Ubhi et al., 2025). Effect sizes ranged from standardized β = −.05 to .04 across all PD definitions, reflecting very small effect sizes according to Cohen (2013). However, just as statistically significant findings can be associated with small effect sizes, small effect sizes can be clinically significant (Carey et al., 2023). On work and career satisfaction, unstandardized beta ranged from b = −.66 to −.57, indicating individuals with PD scored greater than half a scale point below individuals without PD. On Likert-type scales, this gap translates into noticeable differences between whether a person is closer to “sometimes” vs. “usually” experiencing satisfaction. This could indicate individuals who have had PD could benefit from vocational support.

Perfectionism differed between the broad definitions of PD and non-eating disorder controls at follow-up. Although this could be interpreted as greater support for a broader definition of PD, equivalent effect sizes found across all PD definitions for predicting perfectionism at 10-year follow-up suggest differences in statistical significance are related to the effect of reduced prevalence on power of analyses. No significant differences were found between any PD definition and non-eating disorder controls on alcohol use at follow-up. Over two-year follow-up of adolescents with PD, Micali et al. (2015) also found no association between a diagnosis of PD and later binge drinking. Moreover, both findings mirror cross-sectional findings that individuals with PD, by any definition, did not report greater alcohol use than non-eating disorder controls in college (Haedt & Keel, 2010).

This study had several strengths, including the longitudinal design and long-term follow-up in large cohorts of adult men and women. As most research on PD is limited by cross-sectional designs, findings address important gaps in the literature. Additionally, the large sample size powered us to detect the effects for a disorder with a low point prevalence. It is important to acknowledge study limitations and interpret results in the context of these limitations. Of note, compulsive exercise was not assessed in 1982 and therefore data on this are missing for this cohort in college. More individuals may have been included in the broader definitions of PD if data on compulsive exercise had been collected in the first wave of this parent study. We did not assess the predictive validity of defining PD at a frequency of less than once per week. We sought to compare commonly used thresholds, but lower thresholds may reflect definitions used in studies that have not stipulated minimum frequencies (Bodell et al., 2022; Silen et al., 2020) and could have extended evidence of validity to an even lower symptom frequency. Importantly, predictive validity was assessed in the current study through self-reported satisfaction with multiple life domains, offering insight into psychosocial functioning but not directly addressing impairment in physical or psychological functioning, additional components in evaluating clinical significance. Further, diagnoses were based on self-report surveys. Prior work demonstrates the validity of diagnoses based on these surveys against the Structured Clinical Interview for DSM disorders (Keel et al., 2002). Additional evidence supports concurrent validity between survey and interview-based assessments of purging behaviors (Field et al., 2003) and that individuals may even be more likely to disclose purging behaviors through survey-based methods (Mond et al., 2007). However, due to our conservative approach in excluding any binge eating, current results may underestimate PD prevalence and reduce generalizability of findings to individuals with PD with subjective binge eating. Subjective binge episodes are common in PD (Smith et al., 2018) and are linked to elevated distress and impairment in PD (Forney et al., 2014), suggesting that current findings may also underestimate long-term impact of PD on psychosocial outcomes. In addition, systematic differences between the final analytic sample compared to baseline may have impacted generalizability of results to men and those who do not identify as White. Finally, in following recommendations from Rothman (1990), Type I error rates could potentially be inflated due to multiple testing.

Future research should continue to longitudinally examine other important external validators of PD such as medical problems, mental health diagnoses, and other facets of quality of life. While findings from the current study prospectively distinguish PD from individuals without an eating disorder on psychosocial outcomes, future work is needed to prospectively compare PD to other eating disorders. Research examining whether treatment responses differ in PD compared to other eating disorders is needed to guide clinical care. Early work in a case series suggested potentially worse treatment response in PD compared to BN and other OSFED (Waller et al., 2014). Due to insufficient number of PD patients in a recent large transdiagnostic sample (Dixon et al., 2025), results for patients with PD were combined with those presenting with an unspecified feeding or eating disorder. To our knowledge, only one study has compared long-term outcomes between PD and BN (Forney et al., 2022) and supported the clinical significance of PD as defined by purging, on average, at least once per week, but more work is needed comparing PD to all DSM-5 eating disorders on measures of clinical validity. Moreover, effect sizes were small, and many psychosocial validators did not show significant differences at this long duration of follow-up. Future work could examine the predictive validity of PD across a wider variety of external validators and at more intermediate durations of follow-up.

In summary, this study supports the predictive validity defining PD as purging through self-induced vomiting, laxative or diuretic misuse, on average, at least once weekly. A standardized definition of PD with evidence of predictive validity is a valuable tool for both researchers and clinicians. For researchers, standardizing how PD is defined across studies will increase reliability in findings, enhance reproducibility of research, and facilitate better communication about the disorder to a wider audience. For clinicians, standardizing the definition of PD will improve clinical assessment and identification of patient needs. Specifically, a clear threshold for diagnosing PD using a standardized once-per-week symptom threshold will improve assessment reliability, contribute to validity of case identification, and improve communication and treatment coordination across clinical settings.

Supplementary Material

OnlineSupplement

Table 3.

Linear Regressions prospectively predicting external validators from each definition of purging disorder.

Validator Purging 2x/week Purging 1x/week UWCB 2x/week UWCB 1x/week
R2 β SE t df p R2 β SE t df p R2 β SE t df p R2 β SE t df p
Satisfaction
Friendships .16 −.01 .24 −.29 2310 .39 .16 <.01 .20 .21 2316 .42 .16 .01 .08 .61 2421 .27 .16 .01 .07 .49 2472 .31
Relationship .01 .04 1.10 .85 597 .20 .01 .02 .78 .46 598 .32 .01 .02 .21 .45 624 .33 .01 −.03 .17 −.83 643 .20
Worka .03 −.04 .31 −2.12 2285 .02 .03 −.05 .26 −2.18 2291 .02 .03 −.02 .11 −.77 2394 .22 .02 <.01 .09 .01 2443 .50
Careera .02 −.04 .31 −1.92 2285 .03 .02 −.05 .26 −2.31 2291 .01 .02 −.03 .11 −1.62 2394 .05 .02 −.01 .09 −.31 2443 .38
Perfectionism .35 .01 .98 .48 2316 .31 .35 .01 .82 .54 2322 .30 .35 .03 .34 1.51 2427 .07 .35 .03 .29 1.65 2478 <.05
Alcohol Useb .18 <.01 .24 −.004 2303 .49 .18 .01 .21 .41 2309 .34 .18 .01 .09 .60 2413 .27 .18 −.01 .07 −.34 2464 .37
Open in a new tab

Notes. Purging includes self-induced vomiting, laxative abuse, or diuretic abuse to influence weight; UWCB = Unhealthy weight control behaviors and includes purging, diet pill use, fasting, and compulsive exercise; positive regression coefficients indicate that the diagnostic group scored higher on the scale than the non-eating disorder comparison group. Based on a priori hypotheses, we predicted lower scores (negative beta values) for all satisfaction variables and higher values (positive beta values) for perfectionism and alcohol use with a one-tailed test.

a

Models predicting satisfaction with work and career included satisfaction with school work at baseline as a covariate given all participants were full-time students at baseline.

b

Alcohol use was square root transformed to correct for high skewness and kurtosis of residuals

Summary:

  • Individuals with purging disorder experience long-term negative psychosocial outcomes compared to individuals without an eating disorder, supporting the predictive validity of purging disorder.

  • Findings support defining purging disorder as the use of self-induced vomiting, laxatives, or diuretics, on average, once per week to influence body shape or weight in the absence of binge eating.

Acknowledgements:

Research reported in this publication was supported by the National Institute of Mental Health of the National Institutes of Health under Award Number R01MH126990. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was also supported by grants from the William F. Milton Fund and National Institute of Mental Health (R01MH63758, PI: Pamela K. Keel).

Footnotes

Statement on involvement of persons with lived experience: Persons with lived experience may have been involved in the study design or execution, or in the preparation of this manuscript. The execution of this study has involved many individuals over an extended period of time, and no individual has been required disclose personal medical information, which includes mental health history.

Statement on usage of artificial intelligence (AI): AI tools were not used in study design, study execution, or manuscript preparation.

Contributor Information

Holly K. Spinner, Department of Psychology, Florida State University

Alissa Haedt-Matt, College of Health Professions, Rosalind Franklin University of Medicine and Science

Pamela K. Keel, Department of Psychology, Florida State University.

Data availability statement:

The data that support the findings of this study are available through the NIMH Data Archive (NDA) Collection #3990. Access to these data requires completion of a Data Use Agreement and is restricted to users with appropriate participant consent in accordance with privacy regulations.

References

  1. Allen KL, Byrne SM, Oddy WH, & Crosby RD. (2013). DSM-IV-TR and DSM-5 eating disorders in adolescents: Prevalence, stability, and psychosocial correlates in a population-based sample of male and female adolescents. Journal of Abnormal Psychology, 122, 720 –732. 10.1037/aoo34004 [DOI] [PubMed] [Google Scholar]
  2. American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th ed.). [Google Scholar]
  3. Attia E, & Guarda AS. (2024). OSFED Subtypes: The Need for Better Definitions. International Journal of Eating Disorders, 57(10), 2053–2055. 10.1002/eat.24275 [DOI] [PubMed] [Google Scholar]
  4. Bodell LP, Egbert AH, Anaya C, & Wildes JE. (2022). Associations between emotion reactivity and eating disorder symptoms in a transdiagnostic treatment-seeking sample. International Journal of Eating Disorders, 55(10), 1390–1396. 10.1002/eat.23808 [DOI] [PubMed] [Google Scholar]
  5. Brown TA, Forney KJ, Klein KM, Grillot C, & Keel PK. (2020). A 30-Year Longitudinal Study of Body Weight, Dieting, and Eating Pathology across Women and Men from Late Adolescence to Later Mid-Life. Journal of Abnormal Psychology, 129(4), 376–386. 10.1037/abn0000519 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Carey EG, Ridler I, Ford TJ, & Stringaris A. (2023). Editorial Perspective: When is a ‘small effect’ actually large and impactful?. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 64(11), 1643–1647. 10.1111/jcpp.13817 [DOI] [PubMed] [Google Scholar]
  7. Cohen J. (2013). Statistical power analysis for the behavioral sciences (2nd ed.). Routledge. [Google Scholar]
  8. Dang TB, Hughes EK, Dang AB, Lai HY, Lee J, Liu S, Portingale J, Fuller-Tyszkiewicz M, & Krug I. (2024). Taking a deeper dive into OSFED subtypes: A meta-analysis and systematic review. International Journal of Eating Disorders, 57(10), 2006–2040. 10.1002/eat.24280 [DOI] [PubMed] [Google Scholar]
  9. Davis HA, Smith GT, & Keel PK. (2020). An examination of negative urgency and other impulsigenic traits in purging disorder. Eating Behaviors, 36, 101365. 10.1016/j.eatbeh.2020.101365 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dixon L, Price C, Bartel S, Harris A, Schenkels M, Spinella T, Nunes A, Ali SI, Waller G, Wournell J, Gamberg S, & Keshen A. (2025). Delivering brief cognitive behavioral therapy (CBT-T) for eating disorders: Examining real-world outcomes of a large-scale training program. International Journal of Eating Disorders, 58(10), 1946–1956. 10.1002/eat.24498 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Field AE, Barr Taylor C, Celio A, & Colditz GA. (2003). Comparison of self-report to interview assessment of bulimic behaviors among preadolescent and adolescent girls and boys. International Journal of Eating Disorders, 35(1), 86–92. 10.1002/eat.10220 [DOI] [PubMed] [Google Scholar]
  12. Forney KJ, Brown TA, Crosby RD, Klein KM, & Keel PK. (2022). Evaluating the predictive validity of purging disorder by comparison to bulimia nervosa at long-term follow-up. International Journal of Eating Disorders, 55(6), 810–819. 10.1002/eat.23712 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Forney KJ, Haedt-Matt AA, & Keel PK. (2014). The role of loss of control eating in purging disorder. International Journal of Eating Disorders, 47(3), 244–251. 10.1002/eat.22212 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Garner DM, Olmstead MP, & Polivy J. (1983). Development and validation of a multidimensional eating disorder inventory for anorexia nervosa and bulimia. International Journal of Eating Disorders, 2(2), 15–34. 10.1002/1098-108X(198321)2:2<15::AID-EAT2260020203>3.0.CO;2-6 [DOI] [Google Scholar]
  15. Glazer KB, Sonneville KR, Micali N, Swanson SA, Crosby R, Horton NJ, Eddy KT, & Field AE. (2019). The Course of Eating Disorders Involving Bingeing and Purging Among Adolescent Girls: Prevalence, Stability, and Transitions. The Journal of Adolescent Health: Official Publication of the Society for Adolescent Medicine, 64(2), 165–171. 10.1016/j.jadohealth.2018.09.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Haedt AA, & Keel PK. (2010). Comparing definitions of purging disorder on point prevalence and associations with external validators. International Journal of Eating Disorders, 43(5), 433–439. 10.1002/eat.20712 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hammerle F, Huss M, Ernst V, & Burger A. (2016). Thinking dimensional: prevalence of DSM-5 early adolescent full syndrome, partial and subthreshold eating disorders in a cross-sectional survey in German schools. BMJ Open, 6, e010843. 10.1136/bmjopen-2015-010843 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hay P, Mitchison D, Lopez Collado AE, Gonzalez-Chica DA, Stocks N, & Touyz S. Burden and health-related quality of life of eating disorders, including Avoidant/Restrictive Food Intake Disorder (ARFID), in the Australian population. Journal of Eating Disorders, 5. 10.1186/s40337-017-0149-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Holland LA, Bodell LP, & Keel PK. (2013). Psychological factors predict eating disorder onset and maintenance at 10-year follow-up. European Eating Disorders Review: The Journal of the Eating Disorders Association, 21(5), 405–410. 10.1002/erv.2241 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Holland LA, Brown TA, & Keel PK. (2014). Defining features of unhealthy exercise associated with disordered eating and eating disorder diagnoses. Psychology of Sport and Exercise, 15(1). 10.1016/j.psychsport.2013.10.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hudson JI, Hiripi E, Pope HG, & Kessler RC. (2007). The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biological Psychiatry, 61(3), 348–358. 10.1016/j.biopsych.2006.03.040 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Keel PK. (2024). Guidelines for Rigorous and Reproducible Research on Other Specified Feeding or Eating Disorder: Commentary on Dang et al. (2024). International Journal of Eating Disorders, 57(10), 2041–2044. 10.1002/eat.24262 [DOI] [PubMed] [Google Scholar]
  23. Keel PK, (2019). Purging disorder: recent advances and future challenges. Current Opinion in Psychiatry, 32(6), 518–524. 10.1097/YCO.0000000000000541 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Keel PK, Crow S, Davis TL, & Mitchell JE. (2002). Assessment of eating disorders: Comparison of interview and questionnaire data from a long-term follow-up study of bulimia nervosa. Journal of Psychosomatic Research, 53(5), 1043–1047. 10.1016/S0022-3999(02)00491-9 [DOI] [PubMed] [Google Scholar]
  25. Keel PK, Mitchell JE, Miller KB, Davis TL, & Crow SJ. (2000). Social adjustment over 10 years following diagnosis with bulimia nervosa. International Journal of Eating Disorders, 27(1), 21–28. 10.1002/(SICI)1098-108X(200001)27:1<21::AID-EAT2>3.0.CO;2-F [DOI] [PubMed] [Google Scholar]
  26. Keel PK, & Striegel-Moore RH. (2009). The validity and clinical utility of purging disorder. International Journal of Eating Disorders, 42(8), 706–719. 10.1002/eat.20718 [DOI] [PubMed] [Google Scholar]
  27. Keel PK, Wolfe BE, Gravener JA, & Jimerson DC. (2008). Co-morbidity and disorder-related distress and impairment in purging disorder. Psychological Medicine, 38(10), 1435–1442. 10.1017/S0033291707001390 [DOI] [PubMed] [Google Scholar]
  28. Kim H-Y. (2017). Statistical notes for clinical researchers: Chi-squared test and Fisher’s exact test. Restorative Dentistry & Endodontics, 42(2), 152–155. 10.5395/rde.2017.42.2.152 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Krug I, Giles SE, Granero R, Aguera Z, Sanchez I, Sanchez-Gonzalez J, Jimenez-Murcia S, Fernandez-Aranda F. (2022). Where does purging disorder lie on the symptomatologic and personality continuum when compared to other eating disorder subtypes? Implications for the DSM. European Eating Disorders Review, 30(1), 36–49. 10.1002/erv.2872 [DOI] [PubMed] [Google Scholar]
  30. Krug I, Granero R, Giles S, Riesco N, Agüera Z, Sánchez I, Jiménez-Murcia S, Del Pino-Gutierrez A, Codina E, Baenas I, Valenciano-Mendoza E, Menchón JM, & Fernández-Aranda F. (2020). A cluster analysis of purging disorder: Validation analyses with eating disorder symptoms, general psychopathology and personality. European Eating Disorders Review: The Journal of the Eating Disorders Association, 28(6), 643–656. 10.1002/erv.2769 [DOI] [PubMed] [Google Scholar]
  31. Kummer KK, & Mangweth-Matzek B. (2023). Male eating disorders in midlife-possible links between excessive sports and hormones. The Aging Male, 26(1), 2154571. 10.1080/13685538.2022.2154571 [DOI] [PubMed] [Google Scholar]
  32. Little RJ. (2024). Missing Data Analysis. Annual Review of Clinical Psychology, 20, 149–173. 10.1146/annurev-clinpsy-080822-051727 [DOI] [PubMed] [Google Scholar]
  33. Little RJA. (1988). A Test of Missing Completely at Random for Multivariate Data with Missing Values. Journal of the American Statistical Association, 83(404), 1198–1202. 10.1080/01621459.1988.10478722 [DOI] [Google Scholar]
  34. Lu Y, Wang M, & Zhang G. (2017). A new revised version of McNemar’s test for paired binary data. Communications in Statistics – Theory and Methods, 46(20), 10010–10024. 10.1080/03610926.2016.1228962 [DOI] [Google Scholar]
  35. Lydecker JA, Shea M, & Grilo CM. (2017). Driven exercise in the absence of binge eating: Implications for purging disorder. International Journal of Eating Disorders, 51(2), 139–145. 10.1002/eat.22811 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mangiafico SS. (2016). Summary and Analysis of Extension Program Evaluation in R, version 1.23.0, revised 2025. http://rcompanion.org/handbook/
  37. McNeish D. (2017). Missing data methods for arbitrary missingness with small samples. Journal of Applied Statistics, 44(1), 24–39. 10.1080/02664763.2016.1158246 [DOI] [Google Scholar]
  38. Micali N, Solmi F, Horton NJ, Crosby RD, Eddy KT, Calzo JP, Sonneville KR, Swanson SA, & Field AE. (2015). Adolescent Eating Disorders Predict Psychiatric, High-Risk Behaviors and Weight Outcomes in Young Adulthood. Journal of the American Academy of Child and Adolescent Psychiatry, 54(8), 652–659.e1. 10.1016/j.jaac.2015.05.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Mitchison D, Mond J, Bussey K, Griffiths S, Trompeter N, Lonergan A, Pike KM, Murray SB, & Hay P. (2019). DSM-5 full syndrome, other specified, and unspecified eating disorders in Australian adolescents: prevalence and clinical significance. Psychological Medicine, 50(6), 981–990. 10.1017/S0033291719000898 [DOI] [PubMed] [Google Scholar]
  40. Mond JM, Hay PJ, Rodgers B, & Owen C. (2007). Self-report versus interview assessment of purging in a community sample of women. European Eating Disorders Review, 15, 403–409. 10.1002/erv.792 [DOI] [PubMed] [Google Scholar]
  41. Riesco N, Aguera Z, Granero R, Jimenez-Murcia S, Menchon JM, Fernandez-Aranda F. (2018). Other Specified Feeding or Eating Disorders (OSFED): Clinical heterogeneity and cognitive-behavioral therapy outcome. European Psychiatry, 54, 109–116. 10.1016/j.eurpsy.2018.08.001 [DOI] [PubMed] [Google Scholar]
  42. Rothman KJ. (1990). No adjustments are needed for multiple comparisons. Epidemiology, 1(1), 43–46. [PubMed] [Google Scholar]
  43. Schafer JL, & Graham JW. (2002). Missing data: Our view of the state of the art. Psychological Methods, 7(2), 147–177. 10.1037/1082-989X.7.2.147 [DOI] [PubMed] [Google Scholar]
  44. Silen Y, Sipila PN, Raevuori A, Mustelin L, Marttunen M, Kaprio J, & Keski-Rahkonen A. (2020). DSM-5 eating disorders among adolescents and young adults in Finland: A public health concern. International Journal of Eating Disorders, 53(5), 790–801. 10.1002/eat.23236 [DOI] [PubMed] [Google Scholar]
  45. Smith KE, Crowther JH, & Lavender JM. (2018). A review of purging disorder through meta-analysis. Journal of Abnormal Psychology, 126(5), 565–592. 10.1037/abn0000243 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Stice E, Desjardins C, Shaw H, Siegel S, Gee K, & Rohde P. (2025). Prevalence, incidence, impairment, course, and diagnostic progression and transition of eating disorders, overweight, and obesity in a large prospective study of high-risk young women. Journal of Psychopathology and Clinical Science, 134(4), 427–437. 10.1037/abn0000965 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Stice E, Marti CN, & Rohde P. (2013). Prevalence, Incidence, Impairment, and Course of the Proposed DSM-5 Eating Disorder Diagnoses in an 8-Year Prospective Community Study of Young Women. Journal of Abnormal Psychology, 122(2), 445–457. 10.1037/a0030679 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Tiggemann M & Lynch JE. (2001). Body image across the life span in adult women: The role of self-objectification. Developmental Psychology, 37(2), 243–253. 10.1037/0012-1649.37.2.243 [DOI] [PubMed] [Google Scholar]
  49. Tobin DL, Griffing A, Griffing S. (1997). An examination of subtype criteria for bulimia nervosa. International Journal of Eating Disorders, 22(2), 179–186. 10.1002/(sici)1098-108x(199709)22:2<179::aid-eat10>3.0.co;2-x [DOI] [PubMed] [Google Scholar]
  50. Ubhi S, Bani-Fatemi A, Howe A, Safi F, Aniserowicz A, & Nowrouzi-Kia B. (2025). Impact of eating disorders on paid or unpaid work participation and performance: a systematic review. Eating and Weight Disorders – Studies on Anorexia, Bulimia and Obesity, 30(1). 10.1007/s40519-025-01783-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Wade TD. (2007). A retrospective comparison of purging type disorders: Eating disorder not otherwise specified and bulimia nervosa. International Journal of Eating Disorders, 40(1), 1–6. 10.1002/eat.20314 [DOI] [PubMed] [Google Scholar]
  52. Waller G, Gray E, Hinrichsen H, Mountford V, Lawson R, & Patient E. (2014). Cognitive-behavioral therapy for bulimia nervosa and atypical bulimic nervosa: effectiveness in clinical settings. International Journal of Eating Disorders, 47(1), 13–17. 10.1002/eat.22181 [DOI] [PubMed] [Google Scholar]
  53. Ward ZJ, Rodriguez P, Wright DR, Austin SB, & Long MW. (2019). Estimation of Eating Disorders Prevalence by Age and Associations with Mortality in a Simulated Nationally Representative US Cohort. JAMA Network Open, 2(10), e1912925. 10.1001/jamanetworkopen.2019.12925 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Watson HJ, Fursland A, Bulik CM, & Nathan P. (2013). Subjective binge eating with compensatory behaviors: a variant presentation of bulimia nervosa. International Journal of Eating Disorders, 46(2), 119–126. 10.1002/eat.22052 [DOI] [PubMed] [Google Scholar]
  55. Zar JH. (2010). Biostatistical Analysis (5th edition). Pearson. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

OnlineSupplement
NIHMS2197161-supplement-OnlineSupplement.pdf (154.3KB, pdf)

Data Availability Statement

The data that support the findings of this study are available through the NIMH Data Archive (NDA) Collection #3990. Access to these data requires completion of a Data Use Agreement and is restricted to users with appropriate participant consent in accordance with privacy regulations.

RESOURCES