Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Nov 1.
Published in final edited form as: Eat Disord. 2022 Feb 17;30(6):686–699. doi: 10.1080/10640266.2021.2011648

Understanding non-routine discharge: Factors that are associated with premature termination from higher levels of care in adults with anorexia nervosa

Sasha Gorrell a, Renee D Rienecke b,c, Alan Duffy b, Emma Huston a,d, Philip S Mehler b,e,f, Craig Johnson b, Jamie Manwaring b, Susan McClanahan b, Dan V Blalock g,h, Daniel Le Grange a,i
PMCID: PMC9869711  NIHMSID: NIHMS1862890  PMID: 35175902

Abstract

This study sought to replicate and extend associations between clinical and demographic features at admission and types of premature treatment termination for adults diagnosed with anorexia nervosa (AN) in higher-level-of-care settings. Secondary data analyses examined a study population comprised of adults with AN (N = 565) who were admitted to one of two United States eating disorder treatment centers (April 2015–April 2020) for intensive outpatient, partial hospitalization, residential, or inpatient services. There were no significant differences in the type of non-routine discharge according to level of care. At admission, those with lower BMI were more likely to discharge against medical advice, and those with lower cognitive restraint and elevated binge eating were more likely to discharge against medical advice or by staff-initiated request, respectively. Discharge by parent/patient request was more likely among those who were older or who reported lower baseline desire for muscularity. Overall older age, elevated binge eating, and lower weight, desire for muscularity, and cognitive restraint may be associated with less tolerance/acceptability for AN treatment. Increased understanding of how to better support patients who admit to higher levels of care with these clinical features will contribute to better odds of completion of a full course of treatment.

Introduction

Anorexia nervosa (AN) is a profoundly dangerous and treatment-resistant psychiatric disorder with high mortality rates (Smink et al., 2012), significant costs (van Hoeken & Hoek, 2020), and global disease burden (Erskine et al., 2016). While evidence-based approaches to treatment for adults with AN are generally recommended for their utility (McIntosh, 2020), illness duration extends to over 20 years for over half of those afflicted (Fichter et al., 2017), and more people die from this disorder than nearly any other psychiatric condition (Chesney et al., 2014). Further, those who do not complete treatment do not achieve the same clinical progress as treatment-completing counterparts (Björk et al., 2009). To better increase the chances of a positive prognosis for this high-priority clinical population, it is critical that as a field, we achieve a better understanding of non-routine discharge from care.

Across the eating disorder (ED) literature, non-routine discharge has commonly been referred to as ‘dropout,’ or ‘premature termination from treatment (PTT),’ the latter considered by some to be less pejorative (Björk et al., 2009; Seidinger-Leibovitz et al., 2020). PTT is quite common, with prevalence rates that vary across different levels of care (e.g., 20–73%; Fassino et al., 2009). Recent comprehensive reviews of PTT from across various intensities of adult AN treatment (i.e., outpatient to inpatient) have identified specific factors that contribute to the likelihood of this occurrence. While findings are mixed, some patient-level factors that appear to increase the likelihood of PTT include older age, lower BMI/higher weight suppression at admission, and the presence of binge eating and purging in the context of AN (Gregertsen et al., 2019; Seidinger-Leibovitz et al., 2020; Vall & Wade, 2015). Specifically in adult AN inpatient settings, additional predictors of PTT include lower depressive symptoms, education level, and weight concerns (The EVHAN Group et al., 2016).

Despite efforts to operationalize PTT (Hatchett & Park, 2003), disparate definitions used across studies present a challenge to drawing conclusions or making clinical recommendations (Seidinger-Leibovitz et al., 2020; Vall & Wade, 2015). Although a majority of studies have conceptualized dropout as a binary variable (yes/no), this yields a lack of specificity in the qualitative determination of who instigates discharge, i.e., whether termination is initiated by the patient or by staff, or the degree to which discharge was planned vs. unanticipated (e.g., against medical advice [AMA]). Recent review across various medical diagnoses found associations between AMA discharge and numerous negative selequae including increased readmission, system costs, physician distress, patient stigma, and increased morbidity and mortality (Holmes et al., 2021). To date, no work has included a more nuanced examination of whether particular factors may more likely associate with specific types of PTT (e.g., as a result of resource constraint v. AMA). PTT is a multi-faceted and complex decision that involves an interplay between individual, interpersonal, and system-level factors. From a patient perspective, requesting to leave treatment for any number of factors would qualitatively feel different than if staff requested that one leaves due to inappropriate behavior. Similarly, deciding to leave AMA would feel different than if one was compelled to leave due to resource constraint. To date, no work has specifically examined factors that predict these varied scenarios.

The current study sought to fill this gap in the literature by examining factors suggested by prior work (Gregertsen et al., 2019; Roux et al., 2016; Seidinger-Leibovitz et al., 2020; Vall & Wade, 2015), with different types of patient- or staff-initiated PTT among adults with a diagnosis of AN. A considerable literature has examined premature termination from psychotherapy treatment across broader patient populations (Swift et al., 2017). Given the alarming lethality of restrictive EDs such as AN and bulimia nervosa (Lengvenyte et al., 2021), it is essential that we better understand PTT specifically in patients with EDs with more nuance (i.e., the specific type of PTT) than has been previously explored. No studies to date have specifically examined factors that are associated with PTT in EDs in a manner that fully delineates the type of discharge, relative to who initiates it and why. Therefore, our specific interest was in determining which patient-level factors might be associated with different categories of PTT, and how these findings might inform future efforts to prevent these outcomes. Some reasons for PTT are weighted toward being addressed more on a health system or policy-related level (e.g., lack of resources needed to remain in treatment). However, understanding factors that would lead a patient toward PTT that were in their individual control (e.g., baseline factors that might lead one to be more likely to personally request to leave) will inform considerations for how we might adjust clinical service delivery. Specifically, the patient perspective, reflected in part by volitional PTT, is particularly important to consider as we work toward improving the experience of treatment and ED care delivery (Vinchenzo et al., 2021). Based on prior work, we hypothesized that older age, lower weight status, a diagnosis of AN with binge eating and purging, and elevated AN symptom severity at admission would be more likely to contribute to non-routine discharge. Given that evidence of associations between these specific factors and nuanced categories of discharge have not been published to date, further analyses were considered exploratory.

Methods

Participants and procedure

Secondary data analyses were conducted using a study population comprised of treatment-seeking adults (N = 6024) who were admitted to one of two urban United States (US) national ED treatment centers from April 2015 to April 2020. We limited the sample to those with a DSM-5 diagnosis of AN (n = 3079) and for analyses; we further constricted the sample to those for whom height, weight, and type of discharge status were indicated in the treatment center data base (i.e., recorded as either routine or for another reason) (n = 565). The two centers are part of a larger program that consists of 27 sites across 13 US states and offers four levels of care: intensive outpatient (IOP), partial hospitalization (PHP), residential, and inpatient treatment. These levels of care are considered elevated (i.e., indicated for those with greater disorder severity or less responsive to outpatient care) compared to standard outpatient care, and in the US, these levels of care typically require insurance coverage to pay for treatment. According to standard intake assessment procedures, patients provided clinical and demographic information to center staff at treatment entry. Reasons for discharge were determined by the care teams at each site and entered into the patient record by staff. This study was granted exemption from review by the Sterling Institutional Review Board.

Measures

Diagnosis

Treatment center staff, comprised of master’s level licensed clinicians trained in intake assessment protocol, conducted semi-structured interviews that queried basic demographics, psychiatric and medical history, and ED symptoms. When a symptom was endorsed, the clinician obtained further details to aid in an ED diagnosis, according to DSM-5 criteria (American Psychiatric Association, 2013).

BMI

Height and weight were directly measured by the treatment team in the morning after admission (second day), from which BMI was calculated (kg/m2).

Eating Pathology Symptoms Inventory (EPSI; (Forbush et al., 2013))

The EPSI is a 45-item self-report measure designed to assess ED psychopathology. It contains eight subscales: Body Dissatisfaction, Binge Eating, Cognitive Restraint, Purging, Excessive Exercise, Restricting, Muscle Building (i.e., desire for high muscularity), and Negative Attitudes Toward Obesity. Each item is rated on a 5-point Likert-type scale that ranges from 0 (never) to 4 (very often). The EPSI demonstrates strong convergent and discriminant validity, internal consistency and test–retest reliability across a range of samples (Forbush et al., 2013).

Discharge status

Treatment center staff were trained to determine the reason for discharge according to the following categories: (i) Routine; i.e., when deemed appropriate by the care team, including to a more appropriate care setting, if indicated, (ii) Administrative; i.e., requesting the patient leave as a consequence for unacceptable behavior [e.g., inappropriate sexual behavior] or insufficient program fit, (iii) Resource constraint; i.e., when a patient was unable to pay for treatment, including when insurance or disability insurance claims have been exhausted or denied, (iv) Patient/parent request; i.e., with advance notice to staff of intention to leave, including no medical risk and participation in close-out sessions but prior to the time when staff would determine that discharge was indicated, and, (v) Against Medical Advice (AMA); leaving treatment when not advised due to medical risk factors stated by the medical team, including elopement.

Analytic plan

Means, standard deviations, and frequencies were calculated for baseline demographic variables. Pearson’s correlations and χ2-tests were used to examine associations between admission characteristics of interest (i.e., age, BMI, diagnostic subtype [0 = AN with restriction only, 1 = AN with binge eating and purging]), and non-routine discharge (0 = Routine, 1 = Other). In preliminary analyses, we also evaluated associations between admission level of care (1 = intensive outpatient [IOP], 2 = partial hospitalization [PHP], 3 = residential, 4 = inpatient) and discharge status (calculated both as a binary construct, i.e., Routine v. Other; and delineated, i.e., Routine, Administrative, Resource constraint, Patient/parent request, or AMA). We started with a binary model (Routine vs. Other) as an initial query of PTT defined as Routine vs. Non-routine; our follow-up analyses with delineated categories of ‘Other’ as our outcomes of interest were intended to add to the known evidence in the current literature of any potential differences within the category of ‘Non-routine.’

Age, BMI, and diagnosis were added to a multinomial logistic regression model testing their main effects on discharge status (Routine, Administrative, Resource constraint, Patient/parent request, or AMA), with Routine coded as the reference group. For those for whom baseline EPSI scores were indicated in the data base (n = 349), we then conducted correlation analyses between EPSI subscales and non-routine discharge (0 = Routine, 1 = Other). A second multinomial logistic regression model then tested associations between significant EPSI subscales (limited to include only significant subscales to reduce multicollinearity in the model) and discharge status (Routine, Administrative, Resource constraint, Patient/parent request, or AMA). SPSS v.27 was used for all analyses.

Results

Sample characteristics and preliminary analyses

Participants (N = 565, 91% female) were aged 18–63 (M [SD] = 26.53 [9.09]), reported race/ethnicity as: White (86.7%), Hispanic/Latino (4.4%), Asian (3.0%), Bi/multiracial or mixed race (2.0%), Black/African-American (0.7%), and none (3.1%). The specific reasons for discharge were Routine (n = 382, 67.6%) and Other (n = 183, 32.4%), with Other further defined as: Patient/Parent request (18.8%); Resource constraint (6.7%); Administrative (3.7%); or AMA (3.2%). Roughly half of the sample (n = 306, 54%) were diagnosed with AN-restriction only vs. AN with binge eating and purging (n = 259). A majority of participants were admitted to inpatient care (n = 264, 46.7%), followed by residential (n = 162, 28.7%), PHP (n = 100, 17.7%), or IOP (n = 39, 6.9%).

There were significant differences in the proportion of individuals with Routine vs. Other discharge, according to level of care (IOP, PHP, residential, inpatient), Pearson χ2 = 9.99, df = 3, p = .02. In this model, approximately two-thirds of each of the IOP, PHP and residential groups discharged routinely (74.4%, 74.0%, and 72.8%, respectively) whereas in the inpatient group, 61% discharged routinely, which was a significantly lower rate of routine discharges than the average of the other three groups (p < .05). However, when examining this association with more detailed categories, there were no significant differences in the proportions of individuals within each category of discharge (Routine, Administrative, Resource constraint, Patient/parent request, or AMA), relative to admission level of care (IOP, PHP, residential, inpatient), Pearson χ2 = 19.26, df = 12, p = .08.

Pearson correlations indicated that older age (r = 0.13, p = .003), and lower BMI (r = −0.09, p = .04) were associated with non-routine discharge; there was no significant association between diagnosis and non-routine discharge (r = 0.03, p = .46). In examining EPSI subscales, correlations further indicated that higher scores on Binge Eating (r = 0.16, p = .002), and lower scores on Cognitive Restraint (r = −0.11, p = .04), and Muscle Building (r = −0.13, p = .02) subscales were associated with non-routine discharge. No other EPSI subscales demonstrated significant correlations with non-routine discharge.

Multinomial logistic regression models with outcomes defined by five categories of discharge

Model 1

The full model including age, BMI, and AN diagnosis (restricting or binge/purge) was significant, χ2 = 34.36, df = 12, p = .001 (Table 1). Compared to those with Routine discharge, patients who were older were more likely to discharge by Patient/parent request (B = 0.03, p = .003), and those with lower BMI at admission were more likely to discharge AMA (B = −0.49, p < .001). There were no significant associations among age, BMI, or diagnosis and the likelihood of Administrative or Resource Constraint discharge, compared to Routine.

Table 1.

Multinomial regression testing relations between age, BMI, diagnosis and discharge status (n = 565).

Discharge Status Comparator group: Routine (n = 382) B SE Wald p OR 95% CI
Administrative − 2.34 1.80 1.70 .19
n = 21
Age − 0.01 0.03 0.03 .86 1.00 0.94, 1.05
BMI − 0.02 0.09 0.03 .86 0.99 0.83, 1.17
Diagnosis − 0.32 0.45 0.50 .48 0.73 0.30, 1.77
Resource constraint − 2.52 1.34 3.52 .06
n = 38
Age 0.03 0.02 2.87 .09 1.03 1.00, 1.07
BMI − 0.02 0.07 0.06 .81 0.98 0.87, 1.12
Diagnosis − 0.64 0.35 3.30 .07 0.53 0.27, 1.05
Patient/parent request − 1.21 0.89 1.86 0.17
n = 106
Age 0.03 0.01 8.84 .003 1.03 1.01, 1.06
BMI − 0.05 0.04 1.44 .23 0.95 0.87, 1.03
Diagnosis − 0.09 0.23 0.14 .70 0.92 0.59, 1.43
AMA 4.54 2.41 3.54 .06
n = 18
Age 0.01 0.03 0.10 .76 1.01 0.96, 1.06
BMI − 0.49 0.14 12.51 <.001 0.61 0.47, 0.80
Diagnosis 0.48 0.55 0.77 0.38 1.62 0.55, 4.73
Open in a new tab

Diagnosis coded: 0 = anorexia nervosa with restriction only, 1 = anorexia nervosa with binge eating and purging. Reference group = Routine; i.e., discharge when it is deemed appropriate by the care team, including to a more appropriate care setting, if indicated. Categories of discharge are defined as: Administrative; i.e., discharge based on requesting the patient leave as a consequence for unacceptable behavior or insufficient program fit, Resource constraint; i.e., discharge when a patient was unable to pay for treatment, including when insurance or disability insurance claims are denied, Patient/parent request; i.e., discharge with advance notice to staff of intention to leave, including no medical risk and participation in close-out sessions and, Against Medical Advice (AMA); leaving treatment when not advised due to medical risk factors stated by the medical team, including elopement.

Model 2

The full model including EPSI subscales for Binge Eating, Cognitive Restraint, and Muscle Building was significant, χ2 = 40.05, df = 12, p < .001 (Table 2). Compared to those with Routine discharge, patients with elevated Binge Eating subscale scores were more likely to discharge by Administrative request (B = 0.10, p = .01), or by Resource Constraint (B = 0.13, p < .001). Individuals with lower Muscle Building scores were more likely to discharge by Patient/parent Request (B = −0.08, p = .03), and those with lower Cognitive Restraint were more likely to discharge AMA (B = −0.19, p = .03).

Table 2.

Multinomial regression testing relations between eating pathology and discharge status (n = 349).

Discharge Status Comparator group: Routine (n = 238) B SE Wald p OR 95% CI
Administrative − 4.23 1.11 14.48 <.001
n = 10
Binge Eating 0.10 0.04 6.05 .01 1.10 1.02, 1.19
Cognitive Restraint 0.03 0.10 0.07 .79 1.03 0.85, 1.24
Muscle Building − 0.01 0.07 0.03 .87 0.99 0.86, 1.13
Resource constraint 0 3.89 0.82 22.46 <.001
n = 22
Binge Eating 0.13 0.03 19.63 <.001 1.13 1.07, 1.20
Cognitive Restraint 0.03 0.07 0.21 .64 1.03 0.90, 1.19
Muscle Building − 0.01 0.05 0.01 .91 0.99 0.90, 1.10
Patient/parent request − 0.63 0.40 2.52 .11
n = 69
Binge Eating 0.01 0.02 0.18 .67 1.01 0.97, 1.05
Cognitive Restraint − 0.05 0.04 1.54 .21 0.95 0.89, 1.03
Muscle Building − 0.08 0.04 4.62 .03 0.93 0.86, 0.99
AMA − 1.54 0.78 3.86 .05
n = 10
Binge Eating 0.05 0.04 1.14 .28 1.05 0.96, 1.14
Cognitive Restraint − 0.19 0.09 5.00 .03 0.83 0.67, 0.98
Muscle Building − 0.21 0.15 2.04 .15 0.81 0.61, 1.08
Open in a new tab

Italicized variables refer to subscales of the Eating Pathology Symptoms Inventory. Reference group = Routine; i.e., discharge when it is deemed appropriate by the care team, including to a more appropriate care setting, if indicated. Categories of discharge are defined as: Administrative; i.e., discharge based on requesting the patient leave as a consequence for unacceptable behavior or insufficient program fit, Resource constraint; i.e., discharge when a patient was unable to pay for treatment, including when insurance or disability insurance claims are denied, Patient/parent request; i.e., discharge with advance notice to staff of intention to leave, including no medical risk and participation in close-out sessions and, Against Medical Advice (AMA); leaving treatment when not advised due to medical risk factors stated by the medical team, including elopement.

Discussion

In the current study, we conducted a secondary data analysis to examine admission factors and their association with different types of discharge status from higher levels of care among adults with AN, with the broader aim of determining for whom a non-routine discharge may be more likely. Specifically, we were interested in a nuanced examination of non-routine discharge, which has not been examined in prior literature. AMA and Administrative discharge are patient- or staff-initiated categories of treatment termination, respectively, and as they each subjectively suggest a poorer prognosis (Björk et al., 2009), findings centered on factors associated with these two outcomes were of particular interest.

We examined PTT across levels of care that are generally considered most appropriate for individuals with greater disorder severity, or for those who were less responsive to outpatient services. When evaluated in a model with the outcome of ‘Routine’ vs. ‘Other,’ inpatient level of care had the highest proportion of individuals who discharged non-routinely. However, when evaluated in a model where outcomes were delineated within ‘Other,’ the likelihood of a certain type of PTT (i.e., Routine, Administrative, Patient/parent request, Resource constraint, or AMA) did not differ according to level of care. Together, these findings suggest that although inpatient services may be less likely to be completed as intended, we cannot assume that the highest available level of care (e.g., inpatient vs. IOP) would necessarily be associated with the likelihood of a particular type of discharge (e.g., AMA). We note that within the category of ‘Other,’ which represents any non-routine discharge, there is a factor that partially represents a health-system variable, i.e., resource constraint. It is possible that some individuals were not able to remain in inpatient care as long as might be indicated by clinical staff, due to specificities mandated by their insurance or other financial coverage. It could also be the case that those with more severe presentations are less likely to engage in treatment. While these ideas are speculative, they may partially explain the concerning finding that inpatients were more likely to demonstrate non-routine PTT.

Within this study sample, and in alignment with prior research (Gregertsen et al., 2019; Seidinger-Leibovitz et al., 2020), those with lower baseline BMI were more likely to discharge AMA. Moreover, while the current study did not support prior findings of significant associations between a diagnosis of AN binge/purge subtype and non-routine discharge (Gregertsen et al., 2019; Seidinger-Leibovitz et al., 2020; Vall & Wade, 2015), those with lower EPSI Cognitive Restraint and elevated Binge Eating at baseline were also more likely to discharge by AMA and Administrative request, respectively. These findings may be explained by the possibility that those with lower cognitive restraint may believe that they are ‘well-enough’ to discharge prematurely and that those with binge eating may find the confinement of a higher level of care that requires early behavior change particularly challenging.

Discharge by parent or patient request was more likely among those who were older or who reported lower baseline proclivity for muscle-building. Older age is often an indicator of longer duration of illness, and a severe and enduring course of AN may be a factor that supports increased skepticism about the utility of remaining in treatment from patients and parents alike (Wonderlich et al., 2020). The finding that individuals less invested in high muscularity were more likely to request PTT is surprising, considering that individuals who report compulsive exercise, and perhaps more invested in muscle-building, may prefer to engage in exercise while in recovery and therefore be less tolerant of the confines of a standard treatment program (Dittmer et al., 2018). From the patient perspective, a mismatch between expectations for treatment and the actual experience of ED services may impact decisions to leave prematurely. Recovery from AN can be a slow and difficult process. Ultimately, requesting to leave treatment early is a personal decision, but clinical staff may be able to impact the outcome of this decision by considering that individuals who are older in age may require different levels of connection and communication, and/or adjustment of expectations for the speed or nature of treatment progress. We offer these ideas as speculations about how improved communication about the length and difficulty of AN treatment might help to guide prevention of PTT. While qualitative research might best answer these possibilities, it is also important to keep in mind as we reconsider current clinical practice and ED service delivery that we must prioritize an emphasis on the use of patient perspectives to guide decision making (Stockford et al., 2019; Vinchenzo et al., 2021).

A final finding was that elevated binge eating at admission was more likely to associate with discharge by resource constraint. In our sample, it is possible that some individuals may historically have been at a higher body weight secondary to history of binge eating. Certainly, binge eating can exist in the context of objectively low weight, but some individuals with elevated binge eating at admission may have had a higher goal weight that they may be expected to achieve during treatment (based on prior weight history). Therefore, it is quite possible that some individuals in the current study who had higher baseline binge eating may have less successfully been at a weight that was ‘low enough’ to determine a need for continued care by insurance coverage. While our interpretation of this finding is speculative, we believe it supports a need for increased awareness of the fallacy that low weight is commensurate with ED severity (Garber et al., 2019; Gaudiani, 2018).

This study delineates a specific reason for discharge, with hints that lower weight, older age, less desire for high muscularity and lower restraint/higher binge eating may be associated with less tolerance/acceptability for treatment on the whole. These findings beg the question of how adult AN treatment programs might better accommodate the specific needs of each individual to increase the likelihood of treatment completion. For example, increasing awareness about the severity of eating pathology that spans weight status may help to support shame surrounding binge eating (Bottera et al., 2020; Brockdorf et al., 2020) and the failure of insurance companies to support treatment for those who are not ‘sick enough’ if they are not underweight relative to societal norms (Gaudiani, 2018). The current study cannot specifically answer questions of how we might prevent drop-out from higher levels of care in the future. However, a study strength is that this work begins to identify certain patient-level factors that might be directly associated with a more nuanced definition of PTT. And in so doing, we begin to address possible groups of individuals (e.g., those who are older) who may benefit from improved communication regarding expectations for the length of time and nature of treatment in a higher level of care. We also pave the way for future qualitative studies that might directly query those who are older or with lower baseline BMI, as to how their experience of ED services for AN might be improved so as to decrease rates of PTT.

Our study is not without limitations that bear mention. First, we opted for confining our sample to a population where our variables of interest were known at baseline. While we constricted our sample for reasons that centered around our research interests, this approach may have unintentionally precluded our ability to determine meaningful differences between those individuals with complete data for certain demographic variables at baseline and those who we did not include. Further, our sample is comprised principally of White females. As a consequence, our data cannot address the critically important intersectionality of race, gender, and other demographic indices that may impact ED outcomes (Burke et al., 2020). We also note potential confounds in conducting restrospective secondary analysis of clinic data related to selection bias or misclassification in coding. Relatedly, the small number of individuals who were coded via Administrative or AMA categories was comparatively modest (3.7% and 3.2% of the sample, respectively), and may impact the interpretability of our results. We did not have the ability to meaningfully examine % of expected BMI at admission, the impact of medication on discharge status, time in treatment prior to PTT, number of lifetime admissions, or the possibility of change in level of care during admission, all of which may be important features to examine in future work.

Despite the noted limitations in our approach, this study moves us beyond a definition of PTT that is limited to whether discharge is a binary conceptualization of dropout, and begins to suggest areas in which our clinical services might better support treatment efforts for adults who are in higher levels of care for AN. Specifically, clinical recommendations include considerations of programming that might better support those who begin treatment at a lower weight, who are older (and likely with longer duration of illness), and who engage in binge eating or experience less cognitive restraint. Although a majority of higher-level-of-care programs center around the care of individuals with restrictive EDs, our data suggest that preventing early imprudent exit from treatment may include greater support for patients who present with less inhibitory control over eating behavior. Improved understanding of factors that may increase the likelihood of PTT facilitates better treatment planning, with the broader aim of contributing to better odds of continued clinical progress following discharge (Björk et al., 2009), relative to completion of a full course of indicated treatment.

CLINICAL IMPLICATIONS.

  • Inpatient services showed less routine discharges compared to other levels of care

  • Level of care does not appear to associate with a specific type of discharge

  • Factors associated with premature discharge include older age and elevated binge eating

  • Lower weight, desire for muscularity, and restraint associate with premature discharge

  • Patients who admit with lower restraint and BMI are more likely to discharge against medical advice

Funding

This research received no specific grant from any funding agency, commercial or not-for-profit sectors. Dr. Gorrell is supported by the National Institute of Mental Health (K23MH126201).

Footnotes

Disclosure statement

Dr. Le Grange receives royalties from Guilford Press and Routledge and is co-director of the Training Institute for Child and Adolescent Eating Disorders, LLC. Dr. Blalock receives consulting fees from Eating Recovery Center. Dr. Rienecke receives consulting fees from the Training Institute for Child and Adolescent Eating Disorders, LLC, and royalties from Routledge. All authors report no other biomedical financial interests or potential conflicts of interest.

References

  1. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). 10.1176/appi.books.9780890425596 [DOI] [Google Scholar]
  2. Björk T, Björck C, Clinton D, Sohlberg S, & Norring C (2009). What happened to the ones who dropped out? Outcome in eating disorder patients who complete or prematurely terminate treatment. European Eating Disorders Review, 17(2), 109–119. 10.1002/erv.911 [DOI] [PubMed] [Google Scholar]
  3. Bottera AR, Kambanis PE, & De Young KP (2020). The differential associations of shame and guilt with eating disorder behaviors. Eating Behaviors, 39, 101427. 10.1016/j.eatbeh.2020.101427 [DOI] [PubMed] [Google Scholar]
  4. Brockdorf AN, Tomaso CC, & Keel PK (2020). Sensitivity to punishment and eating pathology among undergraduate women: The mediating role of shame. Eating Behaviors, 36, 101362. 10.1016/j.eatbeh.2019.101362 [DOI] [PubMed] [Google Scholar]
  5. Burke NL, Schaefer LM, Hazzard VM, & Rodgers RF (2020). Where identities converge: The importance of intersectionality in eating disorders research. International Journal of Eating Disorders, 53(10), 1605–1609. 10.1002/eat.23371 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chesney E, Goodwin GM, & Fazel S (2014). Risks of all-cause and suicide mortality in mental disorders: A meta-review. World Psychiatry, 13(2), 153–160. 10.1002/wps.20128 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dittmer N, Voderholzer U, von der Mühlen M, Marwitz M, Fumi M, Mönch C, Alexandridis K, Cuntz U, Jacobi C, & Schlegl S (2018). Specialized group intervention for compulsive exercise in inpatients with eating disorders: Feasibility and preliminary outcomes. Journal of Eating Disorders, 6(1), 27. 10.1186/s40337-018-0200-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Erskine HE, Whiteford HA, & Pike KM (2016). The global burden of eating disorders. Current Opinion in Psychiatry, 29(6), 346–353. 10.1097/YCO.0000000000000276 [DOI] [PubMed] [Google Scholar]
  9. Fassino S, Pierò A, Tomba E, & Abbate-Daga G (2009). Factors associated with dropout from treatment for eating disorders: A comprehensive literature review. BMC Psychiatry, 9 (1), 67. 10.1186/1471-244X-9-67 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Fichter MM, Quadflieg N, Crosby RD, & Koch S (2017). Long-term outcome of anorexia nervosa: Results from a large clinical longitudinal study: FICHTER et al. International Journal of Eating Disorders, 50(9), 1018–1030. 10.1002/eat.22736 [DOI] [PubMed] [Google Scholar]
  11. Forbush KT, Wildes JE, Pollack LO, Dunbar D, Luo J, Patterson K, Petruzzi L, Pollpeter M, Miller H, Stone A, Bright A, & Watson D (2013). Development and validation of the Eating Pathology Symptoms Inventory (EPSI). Psychological Assessment, 25 (3), 859–878. 10.1037/a0032639 [DOI] [PubMed] [Google Scholar]
  12. Garber AK, Cheng J, Accurso EC, Adams SH, Buckelew SM, Kapphahn CJ, Kreiter A, Le Grange D, Machen VI, Moscicki A-B, Saffran K, Sy AF, Wilson L, & Golden NH (2019). Weight loss and illness severity in adolescents with atypical anorexia nervosa. Pediatrics, 144(6), e20192339. 10.1542/peds.2019-2339 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gaudiani JL (2018). Sick enough: A guide to the medical complications of eating disorders (1st ed.). Routledge. 10.4324/9781351184731 [DOI] [Google Scholar]
  14. Gregertsen EC, Mandy W, Kanakam N, Armstrong S, & Serpell L (2019). Pre-treatment patient characteristics as predictors of drop-out and treatment outcome in individual and family therapy for adolescents and adults with anorexia nervosa: A systematic review and meta-analysis. Psychiatry Research, 271, 484–501. 10.1016/j.psychres.2018.11.068 [DOI] [PubMed] [Google Scholar]
  15. Hatchett GT, & Park HL (2003). Comparison of four operational definitions of premature termination. Psychotherapy: Theory, Research, Practice, Training, 40(3), 226–231. 10.1037/0033-3204.40.3.226 [DOI] [Google Scholar]
  16. Holmes EG, Cooley BS, Fleisch SB, & Rosenstein DL (2021). Against medical advice discharge: A narrative review and recommendations for a systematic approach. The American Journal of Medicine, 134(6), 721–726. 10.1016/j.amjmed.2020.12.027 [DOI] [PubMed] [Google Scholar]
  17. Lengvenyte A, Strumila R, Maimoun L, Seneque M, Olié E, Lefebvre P, Renard E, Courtet P, & Guillaume S (2021). A specific association between laxative misuse and suicidal behaviours in patients with anorexia nervosa and bulimia nervosa. Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity. 10.1007/s40519-021-01180-x [DOI] [PubMed] [Google Scholar]
  18. McIntosh VVW (2020). Evidence-based treatments remain the best intervention for good long-term outcome of severe and enduring anorexia nervosa. International Journal of Eating Disorders, 53(8), 1322–1323. 10.1002/eat.23325 [DOI] [PubMed] [Google Scholar]
  19. The EVHAN GroupRoux H, Ali A, Lambert S, Radon L, Huas C, Curt F, Berthoz S, & Godart N (2016). Predictive factors of dropout from inpatient treatment for anorexia nervosa. BMC Psychiatry, 16(1), 339. 10.1186/s12888-016-1010-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Seidinger-Leibovitz FM, Malta AR, Rodrigues L, Bastos RA, Leme PAT, Silva DFA, Turato ER, & Vieira CM (2020). Premature termination of treatment among adults attending treatment for eating disorders: A critical review. Open Journal of Medical Psychology, 09(3), 123–140. 10.4236/ojmp.2020.93010 [DOI] [Google Scholar]
  21. Smink FRE, van Hoeken D, & Hoek HW (2012). Epidemiology of eating disorders: Incidence, prevalence and mortality rates. Current Psychiatry Reports, 14(4), 406–414. 10.1007/s11920-012-0282-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Stockford C, Stenfert Kroese B, Beesley A, & Leung N (2019). Women’s recovery from anorexia nervosa: A systematic review and meta-synthesis of qualitative research. Eating Disorders, 27(4), 343–368. 10.1080/10640266.2018.1512301 [DOI] [PubMed] [Google Scholar]
  23. Swift JK, Greenberg RP, Tompkins KA, & Parkin SR (2017). Treatment refusal and premature termination in psychotherapy, pharmacotherapy, and their combination: A meta-analysis of head-to-head comparisons. Psychotherapy, 54(1), 47–57. 10.1037/pst0000104 [DOI] [PubMed] [Google Scholar]
  24. Vall E, & Wade TD (2015). Predictors of treatment outcome in individuals with eating disorders: A systematic review and meta-analysis: Predictors of treatment outcome in individuals with eating disorders. International Journal of Eating Disorders, 48(7), 946–971. 10.1002/eat.22411 [DOI] [PubMed] [Google Scholar]
  25. van Hoeken D, & Hoek HW (2020). Review of the burden of eating disorders: Mortality, disability, costs, quality of life, and family burden. Current Opinion in Psychiatry, 33(6), 521–527. 10.1097/YCO.0000000000000641 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Vinchenzo C, McCombie C, & Lawrence V (2021). The experience of patient dropout from eating disorders treatment: A systematic review and qualitative synthesis. BJPsych Open, 7 (S1), S299–S299. 10.1192/bjo.2021.792 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Wonderlich SA, Bulik CM, Schmidt U, Steiger H, & Hoek HW (2020). Severe and enduring anorexia nervosa: Update and observations about the current clinical reality. International Journal of Eating Disorders, 53(8), 1303–1312. 10.1002/eat.23283 [DOI] [PubMed] [Google Scholar]

RESOURCES