Post‐Meal Activity and Eating Disorder Behaviors: An Ecological Momentary Assessment (EMA) Study

ABSTRACT

Objective

Mealtimes are associated with heightened distress among individuals with eating disorders (EDs) and are frequently accompanied by ritualized eating behaviors that maintain psychopathology. In inpatient settings, post‐meal rest is commonly prescribed to prevent compensatory behaviors. Yet, enforced inactivity may exacerbate anxiety and ritual engagement. This study used ecological momentary assessment (EMA) to examine whether structured light post‐meal activity is associated with a reduced likelihood of ritualized eating behaviors in an inpatient ED sample.

Method

Participants were 33 inpatients who completed a 2‐week counterbalanced within‐subject crossover protocol, consisting of 1 week of standard post‐meal rest and 1 week allowing structured light post‐meal activity. EMA prompts were delivered six times daily. Ritualized eating behaviors were assessed following reported eating episodes and coded dichotomously. A binomial generalized linear mixed‐effects model with random intercepts for participants tested whether the likelihood of ritualized eating behaviors differed between conditions.

Results

Participants completed 1503 EMA prompts (81.3% compliance). Across 1192 meal‐related assessments, ritualized eating behaviors were reported in 30.8% of prompts during the rest condition and 24.2% during the post‐meal activity condition, with substantial variability across individuals. Participants were significantly less likely to engage in ritualized eating behaviors during weeks allowing post‐meal activity compared with rest‐only weeks (B = −0.70, SE = 0.19, p = 0.0002), corresponding to an ~50% reduction in odds.

Discussion

The findings suggest that supervised post‐meal movement may help interrupt ritual engagement during a high‐risk period and demonstrate the utility of EMA for examining behavioral mechanisms in inpatient care.

Summary

• The study finds initial support for the use of structured light post‐meal activity as a behavioral intervention in inpatient treatment for eating disorders.

• Introducing post‐meal movement may interrupt habitual eating disorder‐related behaviors and reduce reliance on maladaptive rituals.

• For patients receiving inpatient care, post‐meal movement could improve mealtime outcomes and promote generalization of skills used to cope with post‐meal distress.

1. Introduction

Eating disorders (EDs) are severe psychiatric illnesses characterized by persistent disturbances in eating behavior, body image, and weight regulation that interfere with daily functioning (American Psychiatric Association ²⁰¹³). Among individuals with anorexia nervosa (AN), bulimia nervosa (BN), and binge‐eating disorder (BED), mealtimes often evoke pronounced anxiety, disgust, and loss of control, which can manifest as ritualized eating behaviors (Gianini et al. 2015; Ellis et al. ²⁰²⁴; Giel et al. ²⁰²²). These rituals, such as food manipulation, pacing, restriction, or avoidance, serve to regulate distress but maintain pathological patterns of control and avoidance (Gianini et al. 2015; Long et al. ²⁰¹²). In inpatient ED treatment, mealtimes and the subsequent rest/observation period are highly structured to prevent compensatory behaviors. Many individuals with an ED report that anxiety and distress around food and meals increase during a treatment course (Griffiths et al. 2016; Lawson et al. ²⁰¹³; Long et al. ²⁰¹²), and that weight‐gain leads to a general increase in symptom burden (Kaye et al. 2014). Patients describe that regular weigh‐ins, meal plans, and guidelines around activity can be perceived as scary and anxiety‐provoking (Brunet et al. 2021). This applies across ED diagnoses and at different stages of recovery (Brunet et al. 2021). Thus, enforced inactivity can heighten anxiety, intrusive thoughts, and discomfort, leading patients to request opportunities for activity to alleviate distress.

Although prior research has shown that low‐impact movement and mindfulness‐based activity (e.g., yoga and guided stretching) can reduce anxiety and improve affective regulation in ED populations (Hall et al. 2016), evidence within inpatient settings remains scarce, and mechanisms linking post‐meal activity to reductions in ritual behaviors have not been quantified in real‐time. In a study on how to lower post‐meal anxiety in an inpatient ED setting, Shapiro and colleagues compared the effect of progressive muscle relaxation, guided imagery, self‐directed relaxation to a control (rest) condition in 64 females with AN (Shapiro et al. 2008). The authors found that the three first conditions significantly reduced anxiety and increased relaxation post meals, compared to the rest condition. However, it is not specified whether it had an effect on ED behaviors. Similarly, Griffiths and colleagues (Griffiths et al. 2016) compared three different 15‐min post meal activities; a visuospatial task, a verbal task and a somatic task—to a control task where the participants were required to sit quietly. Twenty‐five hospitalized patients with EDs participated, and the findings revealed a significant improvement in post‐meal experience when engaging in an activity after meals (Griffiths et al. 2016). All three tasks lowered negative affect, intrusive thoughts and images—while increasing positive affect, compared to the control (sitting quietly) condition (Griffiths et al. 2016). However, as in the study by Shapiro and colleagues, it is unknown whether the reduction in negative affect also had an impact on ED behaviors.

One of the challenges with exploring the mechanisms that could perpetuate EDs is that these are largely transient phenomena, such as thoughts and feelings, which can vary from morning to evening—and from one situation to another. Thereby making it difficult to assess an individual's experience of an event using end of day, or end of treatment, questionnaires—which in turn leads to uncertain results and makes it challenging to implement interventions specifically targeted at situations where the need for additional support is highest. A way to circumvent these barriers is to utilize methods where frequent questions and feedback are obtained. One such method is ecological momentary assessment (EMA), where participants report their experiences and behaviors in real time, in the environment they are currently in—with a goal to minimize recall bias and to obtain as ecologically valid data as possible (Schaefer et al. 2020).

EMA offers a methodologically powerful way to study the microprocesses of behavior and affect in naturalistic contexts. It minimizes recall bias and allows repeated assessment of momentary states, enabling the detection of within‐person changes in symptom expression over time (Schaefer et al. 2020). EMA also allows for the estimation of associations using naturalistic data and has been described as an ideal method for investigating behaviors and experiences associated with EDs (Smyth et al. 2001; Engel et al. ²⁰¹⁶).

Prior EMA studies have demonstrated that affective states, cognitions, and urges fluctuate in close temporal proximity to eating episodes and are associated with disordered eating behaviors. For example, increases in negative affect and eating‐related cognitions have been shown to precede ED behaviors, while post‐meal periods are often characterized by heightened distress and cognitive preoccupation (Smyth et al. 2001; Engel et al. ²⁰¹⁶; Wegner et al. ²⁰⁰²; Wonderlich et al. ²⁰¹⁸). Levinson et al. (2018) found that meal‐related cognitions predicted subsequent ED behaviors and vice versa, highlighting dynamic, reciprocal relationships between cognitive and behavioral processes across meals. Similarly, Vanzhula et al. (2020) demonstrated that avoidance‐based coping strategies during meals may reduce anxiety in the short term but are associated with increases in ED behaviors over time.

Together, these findings underscore the importance of examining ED symptoms within their momentary context. However, most EMA research has been conducted in outpatient or post‐discharge samples and has focused primarily on affective and cognitive processes rather than directly assessing ED‐related behavioral outcomes within structured treatment environments. To the auhtors' knowledge, EMA has rarely been used to examine ED rituals and behaviors within structured inpatient treatment settings, despite the potential value of understanding how these processes unfold during treatment.

The aim of the current study was to evaluate whether allowing patients to engage in structured light post‐meal activity would reduce the likelihood of engaging in ED behaviors and rituals during meals. We hypothesized a reduction in ritualized eating behaviors on weeks that allowed for post‐meal activity, compared to standard protocol (post‐meal rest).

2. Method

2.1. Design and Participants

Participants were recruited from two third‐line ED treatment sites in Norway, across three units: The Specialist Eating Disorders Unit (RKSF) at Levanger Hospital and the Regional Department for Eating Disorders (RASP) at Oslo University Hospital—both providing inpatient ED treatment free of charge as per the Norwegian health care system. RKSF is a tertiary treatment facility that is part of the regional adult psychiatric services for patients aged ≥ 16. The treatment approach is grounded in psychodynamic theory but incorporates elements from cognitive behavioral therapy (CBT) and includes both individual and group therapy. Each patient is allocated a primary care team, consisting of two nurses, a psychologist or psychiatrist, and a physiotherapist.

There were two participating inpatient units from RASP. First, the Regional Inpatient Unit for Adults which provides care for individuals aged ≥ 18 with severe EDs who require residential treatment. The treatment approach is based on mentalization‐based therapy. Each patient is supported by a primary clinician (a physician or psychologist), an additional physician, five primary milieu‐based contacts (nurses or milieu therapists), and a dietitian when needed. Patients receive both individual and group therapy.

Second, The Regional Intensive Psychiatric Inpatient Unit at RASP which treats patients aged ≥ 16 at various stages of illness. The unit offers structured 6‐week treatment stays. Treatment is individually tailored and incorporates elements from Enhanced CBT (CBT‐E), mentalization‐based therapy, and emotion‐focused therapy. Each patient is assigned a treatment team consisting of a psychologist or psychiatrist (and occasionally two psychiatrists), along with a dietitian and is also staffed by milieu therapists. Patients receive both individual and group therapy.

Both RKSF and RASP utilize two types of dining arrangements: (1) An independent table, or kitchen, where patients serve themselves and are expected to eat in a relatively typical manner and, (2) a support table where patients are served and receive more assistance during meals. Patients may also have pre‐meal and post‐meal conversations to address recurring challenges and plan strategies for the meal. Patients can gradually transition from the support table to serving themselves as they gain independence. The study included patients from both tables. Across units, the patients were at different stages in their treatment processes and presented with varying ED diagnoses, body mass index (BMI), and challenges related to physical activity. Though physical activity is generally limited due to the intensive treatment setting, individual patient‐level information regarding these differences was not recorded.

Inclusion criteria were age ≥ 16 years, inpatient status, and medical clearance to participate in light post‐meal activity. Diagnoses, height, weight, and BMI were obtained from clinical records. In coordination with the unit contact person, the intervention weeks were initiated as soon as eligible patients expressed interest and provided informed consent to participate. Participants were recruited consecutively.

2.2. Inclusion of Persons With Lived Experience

A user group consisting of three former inpatients with lived experience of EDs contributed to the design and refinement of both the EMA protocol and the post‐meal activity intervention. They were recruited through three different routes: One was a prior study participant who had expressed interest in consulting on future research studies, one was hospital staff with lived experience, and one responded to a social media call for individuals with inpatient ED treatment experience. The group participated in piloting the EMA data collection smartphone application (app) before launch and provided feedback on the EMA protocol and post‐meal activity intervention.

Most of the EMA items were drawn from validated instruments previously used in EMA studies. While the user group did not modify the content of these measures, they did offer additional items to measure the other facets of affect, cognitions, and behaviors that they experienced around mealtimes. This process helped ensure that the EMA assessments were appropriate for an inpatient sample and feasible to complete without increasing distress or participant fatigue. Average completion time for each prompt was ~3–5 min.

Before recording the final videos, a psychomotor physiotherapist filmed a pilot video with samples of the movements which the user group assessed. The movements included trunk rotation, stretching the arms above the head before touching the ground with the hands, wrapping the arms around the trunk before opening them up wide. It also included slow, guided movements to help patients notice posture, muscle tension, and balance, and breathing patterns—and activities such as shifting weight and feeling the support of the floor. The user group provided feedback on the suggested post‐meal activity program, as well as suggesting preferred ways to structure and supervise the sessions. Their input informed refinements to enhance both safety and acceptability within the clinical setting. This collaborative process ensured that the study design was responsive to patient perspectives while maintaining scientific rigor.

2.3. Intervention

The study followed a 2‐week counterbalanced within‐subjects design. Each participant completed 1 week under the standard post‐meal rest protocol (30 min rest) and 1 week under the post‐meal activity condition, during which a brief, structured physiotherapy program was implemented. The order of conditions was counterbalanced across subjects. The intervention, referred to as the Eating Disorder Inpatient Physiotherapy Program (EDIPP), was standardized across units and intervention weeks. The program consisted of 10‐min videos led by a psychomotor physiotherapist with several years of clinical experience working with patients with EDs. In the videos, the psychomotor physiotherapist demonstrated low‐impact activity movements, focusing on gentle trunk rotation, upper‐ and lower‐limb mobility, and standing body‐awareness exercises performed without external load or high‐intensity effort. Each session also incorporated breathing and relaxation components aimed at promoting calmness and interoceptive awareness.

To maintain variation, the program contained three videos that were rotated in a 1–2–3–1–2–3 sequence, so that the same program would not be delivered after two meals in a row. After each of the four daily meals, staff accompanied participating patients to the activity room, where the 10‐min EDIPP session was conducted jointly by staff and patients. Following the session, participants returned to join the remaining part of the unit's standard 30‐min rest period together with other patients and staff. All other aspects of treatment, including therapeutic programming and meal supervision, were identical across the weeks. The EMA component of the study was conducted concurrently but was not part of the intervention itself.

2.4. EMA Procedure

Participants completed EMA prompts via a smartphone app “MittNå,” which sent six prompts at semi‐random times, four of which were anchored around meal times on the unit (breakfast, lunch, dinner, and evening meal). Prompts could arrive before, during, or after meals. Last prompt of the day was sent at 21:30.

2.5. Measures

2.5.1. Baseline Measures

Participants completed baseline measures assessing anxiety, depression, and EDs psychopathology.

The Eating Disorder Examination Questionnaire Short (EDE‐QS, Gideon et al. ²⁰¹⁶) is an 11‐item self‐report instrument assessing ED psychopathology over the preceding 7 days; the Generalized Anxiety Disorder‐7 (GAD‐7, Johnson et al. ²⁰¹⁹; Spitzer et al. ²⁰⁰⁶) consists of 7 questions relating to anxiety during the previous 14 days; the Patient Health Questionnaire‐9 (PHQ‐9, Kroenke et al. ²⁰⁰¹; Langkaas et al. ²⁰²¹) is a 9‐item depression screener covering the previous 14 days; and finally the Difficulties in Emotion Regulation Scale Short Form (DERS‐SF, Kaufman et al. ²⁰¹⁶) includes 18 questions concerning emotion‐regulation difficulties.

2.5.2. EMA Measures

The EMA protocol included items assessing momentary affect, eating‐related cognitions, and behaviors. Affect items were informed by widely used state‐based measures, including the Positive and Negative Affect Schedule (PANAS, Watson et al. ¹⁹⁸⁸; Nes et al. ²⁰¹⁸), which has been shown to be sensitive to momentary fluctuations in emotion and has been extensively used in EMA studies, including in ED populations (e.g., Smyth et al. ²⁰⁰¹; Wegner et al. ²⁰⁰²; Levinson et al. ²⁰¹⁸). These items have also been suggested for Norwegian samples (Nes et al. 2018).

At each post‐meal prompt, participants were first asked, “Have you eaten since the last prompt?” If they responded “yes,” a follow‐up item assessed whether they had engaged in any ED rituals or behaviors during that meal. Items assessing ritualized eating behaviors were administered only following prompts in which participants reported having eaten since the previous assessment. Participants who endorsed ritualized behavior were presented with a checklist of possible behaviors and asked to indicate which applied: “Made an effort to reduce intake,” “Engaged in eating and/or preparation rituals,” or “Other.” Multiple responses were permitted. If “Other” was selected, participants were prompted to provide a brief open‐ended description of the behavior.

The questions were developed collaboratively with the user group to ensure that potentially triggering content was avoided. For analytic purposes, responses were dichotomized such that 1 indicated the presence of any ED ritual or behavior and 0 indicated none reported. Although the EMA protocol included affective and cognitive items informed by validated measures, only the ritualized eating behavior variable was included in the present analyses.

2.6. Ethics

The study was approved by the Regional Committee for Medical and Health Research Ethics (541774/REK sør‐øst) and the local governing body (Data Protection Officer, DPO). Data were collected and stored on the TSD (Tjeneste for Sensitive Data) facilities, a secure data storage service owned by the University of Oslo, operated and developed by the TSD service group at the University of Oslo, IT‐Department (USIT). The study was pre‐registered in ClinicalTrials.gov (IDNCT05730348).

2.6.1. Analyses

All analyses were conducted in R version 4.4.1 (2025‐06‐14) using the lme4 (v1.1‐35.3), performance (v0.12.4), MuMIn (v1.47.5), and emmeans (v1.10.1) packages. A binomial generalized linear mixed‐effects model (GLMM) with a logit link tested whether the likelihood of engaging in ritualized eating behaviors differed between the post‐meal activity week (light, supervised movement allowed) and the rest‐only week (standard inactivity). The model was specified as follows:

$$\begin{array}{c}\mathrm{\text{logit}}\left({\mathsf{p}}_{\mathit{ij}}\right)={\mathsf{\beta }}_{\mathrm{0}}+{\mathsf{\beta }}_{\mathrm{1}}\left({\mathrm{\text{ActivityWeek}}}_{\mathit{ij}}\right)+{\mathsf{u}}_{\mathrm{\text{0}}j},\hfill \end{array}$$

where $p_{\mathit{ij}}$ represents the probability of engaging in a ritualized eating behavior for prompt $i$ within participant $j$, and $u_{0j}$ represents random intercepts for participants to account for repeated observations.

A random‐slopes model, allowing the effect of condition to vary across individuals, was tested but did not converge (singular fit), likely due to the modest sample size and sparse within‐person events. Therefore, a random‐intercept model was retained. To evaluate robustness, a sensitivity model was estimated including meal type (breakfast, lunch, dinner, and evening) and day in treatment as covariates. Odds ratios and 95% confidence intervals were computed using the Wald method, and predicted probabilities were obtained via inverse‐logit transformation.

3. Results

3.1. Patient Characteristics

Participants were 33 adult females receiving specialist inpatient treatment for EDs. The sample had a mean age of 27.5 years (SD = 8.6) and a mean BMI of 18.4 kg/m² (range: 12.4–24.1, SD = 3.8) at baseline. Most participants were diagnosed with AN (n = 23), with smaller numbers diagnosed with atypical AN (n = 4) or BN (n = 6). Fourteen of the included patients had comorbid diagnoses, including moderate depressive episode, generalized anxiety disorder, post‐traumatic stress disorder, and anxiety‐related conditions (including avoidant personality disorder and social phobia). Thus, psychiatric comorbidity was common and consistent with the clinical complexity of third‐line inpatient care.

3.2. Descriptive EMA Results

Across the 14‐day protocol, participants completed a total of 1503 EMA prompts, with an average of 45.5 prompts per participant out of a possible 56 (81.3% overall compliance). Compliance was high across both study weeks, indicating adequate momentary sampling density for within‐person analyses. Ritualized eating behaviors were assessed only on EMA prompts following a reported eating episode. Across the 1192 meal‐related prompts, ritualized eating behaviors were reported in 30.8% of prompts during the rest condition and 24.2% during the post‐meal activity condition, corresponding to an absolute reduction of ~6.6% points in the activity week. This pattern is consistent with the mixed‐effects model results indicating a lower likelihood of ritualized behavior during weeks when post‐meal activity was permitted.

Ritualized eating behaviors varied across participants, with individual endorsement rates ranging from 0% to 100% of meal prompts (M = 31.6%, SD = 32.7%), indicating substantial between‐person variability. The distribution of behaviors suggested that the observed effects were not driven by a small subset of participants but were present across individuals. Descriptive analyses of specific ritual types indicated that eating and preparation rituals were the most frequently reported (22.0% of meal‐related prompts), followed by efforts to reduce intake (14.4%). “Other” ritualized behaviors were infrequently endorsed (2.0%). Inspection of free‐text responses suggested that these “other” behaviors primarily reflected variations of ED–related rituals (e.g., pacing, food manipulation, and delayed eating), rather than non‐maladaptive behaviors.

3.3. Mixed‐Effects Model Results

A mixed‐effects logistic regression model examined whether the likelihood of engaging in a ritualized eating behavior differed between the post‐meal activity week (light, supervised movement allowed) and the rest‐only week (standard inactivity). Results indicated a significant effect of post‐meal condition, such that participants were less likely to engage in ritualized eating behaviors during the post‐meal activity week compared with the rest‐only week, B = −0.703, SE = 0.189, z = −3.71, p = 0.0002. This corresponds to an estimated 50% reduction in the odds of ritualized behavior during the post‐meal activity condition (OR = 0.50, 95% CI [0.34, 0.72]).

Predicted probabilities derived from the model indicated that the likelihood of ritualized behavior decreased from ~26% (95% CI [19%, 32%]) during rest‐only weeks to ~16% (95% CI [10%, 21%]) during post‐meal activity weeks. The intraclass correlation coefficient (ICC = 0.22) indicated that ~22% of the variance in ritualized behaviors was attributable to between‐person differences. Model fit indices suggested adequate explanatory power (marginal R ² = 0.07; conditional R ² = 0.25), and there was no evidence of model instability or convergence problems.

A sensitivity model including meal type (breakfast, lunch, dinner, and evening) and day in treatment as covariates yielded a similar effect of post‐meal condition (B = −0.68, SE = 0.20, p < 0.001), indicating that the findings were robust to adjustment for these factors.

4. Discussion

The aim of the current study was to evaluate whether structured light post‐meal activity reduces the likelihood of engaging in ritualized eating behaviors during meals among inpatients with EDs. Using EMA to capture behaviors as they occurred in real time, we found that allowing post‐meal activity was associated with a significantly lower momentary likelihood of ritual engagement. Specifically, the odds of engaging in ritualized eating behaviors during meals were ~50% lower during weeks in which post‐meal activity was permitted compared with rest‐only weeks. Importantly, ritualized eating behaviors were not rare in this inpatient sample; nearly one‐third of meal‐related EMA assessments included the endorsement of at least one ritualized eating behavior despite the highly structured and supervised treatment environment. This pattern underscores the persistence of ritualized eating during inpatient care and highlights the value of within‐person EMA approaches for capturing clinically meaningful behavioral variability.

This result extends prior research suggesting that movement‐based interventions can support affective stability and reduce anxiety in individuals with AN (Kolar et al. 2020; Toutain et al. ²⁰²²). Furthermore, a meta‐analysis reports positive effects from physical activity for symptomatology as well as physical and mental health (Ng et al. 2013). Although physical activity in ED treatment remains controversial due to risks of compulsive exercise (Minano‐Garrido et al. ²⁰²⁵) the present study suggests that carefully structured, therapist‐guided movement may help mitigate distress. The intervention used in the current study emphasized slow, mindful movement and breathing, aligning with emerging evidence that low‐impact activity can support psychological regulation in AN (Toutain et al. 2022). Furthermore, when asking females with EDs about their experiences with physical activity, they describe it as helpful for recovery and wish for it to be integrated into their ED treatment protocol in order to help them learn how to utilize physical activity in a beneficial and helpful way (Brunet et al. 2021). Studies like the current are an important step in deciphering how movement can be safely included in inpatient treatment for EDs, and the potential effects on ED behaviors. In their recent study, Minano‐Garrido and colleagues (2025) demonstrated that an adapted physiotherapy treatment program for patients hospitalized with severe AN is safe and could positively impact health related quality of life and ED symptomatology—without adverse effects during treatment. The current study expands on this research, and by utilizing EMA we are one step closer to identifying temporal associations between physical movement and ED behaviors.

Importantly, this study is among the first to use EMA to capture ED behaviors in real time within a controlled inpatient environment. EMA allowed for moment‐to‐moment assessment of ritualized behaviors as they occurred across meals, providing greater ecological validity than retrospective reporting. The findings are unlikely to reflect an immediate post‐meal effect of movement, but rather contextual or anticipatory effects of the treatment condition. For example, the availability of structured post‐meal activity may reduce anxiety about the post‐meal period or enhance coping expectations during meals, thereby decreasing reliance on ritualized behaviors. In addition, repeated exposure to post‐meal movement across the activity week may contribute to learning or adaptation processes, such as improved tolerance of post‐meal discomfort or reduced behavioral rigidity.

Prior studies have demonstrated that post‐meal interventions, including progressive muscle relaxation, guided imagery, and visuospatial tasks, can improve affective responses in inpatient populations (Griffiths et al. 2016; Shapiro et al. ²⁰⁰⁸). However, it remains unclear whether such improvements in affect translate into changes in ED behaviors. The present findings extend this literature by showing that a structured, low‐impact physiotherapy program was associated with a reduction in ritualized eating behaviors. Future studies should examine whether these behavioral changes are mediated by improvements in emotion regulation and/or reductions in post‐meal distress.

5. Clinical Implications

Increased understanding of the subjective experience of EDs is essential for developing more effective and precise treatments (Ng et al. 2013). If we can learn more about how EDs symptoms and behaviors are maintained, we can also target treatment interventions specifically aimed at these mechanisms. Previous studies have emphasized the importance of teamwork and creating an inclusive environment for treatment providers who wish to offer physical activity (Brunet et al. 2021). The post‐meal intervention utilized in this study aimed to foster an inclusive environment by engaging both staff and participants to carry out the movements together, and it would be helpful for futures studies to assess the potential impact this had on engagement and experiences of the EDDIP. Since there are currently no guidelines on how to safely include exercise in individuals with AN (Minano‐Garrido et al. ²⁰²⁵; Quiles Marcos et al. ²⁰²¹), and not all units have the staff or experience necessary to guide patients suffering from severe malnourishment through safe exercise routines—it will be important for future research to continue assessing how we can safely reap the potential benefits from including movement during inpatient ED treatments.

6. Limitations and Future Directions

The current study has some important strengths and limitations which should be noted. Regarding strengths, the use of EMA within a highly structured inpatient ED environment represents a novel approach, as this methodology has rarely been implemented in such settings. The overall EMA completion rate (81.3%) supports the reliability of the findings, and the large number of meal‐level observations strengthens the statistical power of the within‐person analyses. The counterbalanced within‐person design is particularly well‐suited to inpatient settings where randomized controlled trials are often difficult to implement, enhancing internal validity by controlling for stable individual differences. Finally, the involvement of individuals with lived experience in the study design and intervention refinement strengthens acceptability, feasibility, and patient‐centeredness, in line with current best practices in clinical research.

With regards to limitations, the study was conducted at two inpatient sites within the Norwegian Health System, which may limit generalizability to outpatient or community settings. Additionally, the sample size was modest, and although within‐person design enhances power, replication in larger samples is warranted. Further, while EMA captured behaviors in real time, the data remain correlational, and thus causal mechanisms cannot be inferred directly. Although the EMA protocol included concurrent assessment of affective states (e.g., distress and emotional responses around meals), these variables were not included in the present analyses. As a result, we were unable to examine potential mediating mechanisms underlying the observed behavioral effects. Future research should incorporate these affective measures to test whether reductions in ritualized behaviors are mediated by changes in post‐meal distress, anxiety, or related emotional processes. The use of a dichotomous measure of ritualized eating behavior represents an additional limitation. While this approach provided a clear and consistent indicator across repeated assessments, it does not capture variability in the frequency, intensity, or duration of behaviors within a meal and may obscure clinically meaningful differences in behavioral expression.

In addition, frequent EMA prompting around meals may introduce reactivity effects. Participants may become more aware of their behaviors or modify their responses due to repeated assessment, particularly in a structured inpatient setting where behaviors are already closely monitored. Although EMA reactivity is generally considered modest, this possibility cannot be ruled out in the present study. EMA data were available only for completed prompts, limiting our ability to distinguish between prompts that were missed versus those not delivered. However, mixed‐effects models are robust to unequal numbers of observations across participants, and compliance was high overall.

Furthermore, the sample was restricted to adult female inpatients, predominantly diagnosed with AN, and the mean BMI was close to the normal‐weight range, suggesting that many participants may have been partially or fully weight‐restored. This limits generalizability to more acutely underweight patients, as well as to males, adolescents, nonbinary individuals, and patients with other ED diagnoses. Additionally, each study condition lasted only 1 week, limiting conclusions regarding the stability and persistence of the observed effects beyond the immediate inpatient context. Finally, the exclusive reliance on self‐reported EMA data, without objective verification of activity or behavior, may introduce reporting bias.

7. Conclusions

The current study finds initial support for the use of structured light post‐meal activity as a behavioral intervention in inpatient treatment for EDs. Post‐meal activity was associated with significantly lower momentary likelihood of ritual engagement. This suggests that introducing post‐meal movement—within medically safe and structured limits—may interrupt habitual ED‐related behaviors and reduce reliance on maladaptive rituals. For patients receiving inpatient care, such interventions could improve mealtime outcomes and promote generalization of skills used to cope with post‐meal distress. Future work should investigate whether these benefits persist post‐discharge and whether they vary by diagnostic subtype or baseline anxiety sensitivity.

Author Contributions

Kristin Stedal: conceptualization, funding acquisition, resources, data curation, methodology, project administration, writing – original draft, writing – review and editing. Line Wisting: conceptualization, resources, methodology, writing – review and editing. Joseph Wonderlich: data curation, formal analysis, methodology, writing – original draft, writing – review and editing. Anne Louise Wennersberg: investigation, resources, writing – review and editing. Annette Skarsbakk: resources, writing – review and editing. Ida Aarmo: resources, writing – review and editing. Selma Øverland Lie: methodology, writing – review and editing. Ingrid Funderud: conceptualization, resources, methodology, writing – review and editing.

Lived Experience Involvement Statement

Persons with lived experiences were included in the design of the study and study materials. A user group consisting of three former inpatients with lived experience of eating disorders contributed to the design and refinement of both the study protocol and the post‐meal activity intervention.

Funding

The study received funding from Stiftelsen Dam, Project ID number 34397.

Disclosure

No artificial intelligence (AI) tools were used in study design, study execution, or manuscript preparation.

Ethics Statement

All participants provided written informed consent before enrollment. The study protocol was approved by the Regional committees for medical and health research ethics (REK) (Approval No: 541774/REK Sør‐Øst).

Conflicts of Interest

The authors declare no conflicts of interest.

Stedal, K. , Wisting L., Wonderlich J., et al. 2026. “Post‐Meal Activity and Eating Disorder Behaviors: An Ecological Momentary Assessment (EMA) Study.” International Journal of Eating Disorders 59, no. 10: 2153–2160. 10.1002/eat.70137.

Data Availability Statement

The datasets used and/or analyzed during the current study are not publicly available due to the sensitivity of the data but can be made available from the corresponding author on reasonable request.