Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Dec 1.
Published in final edited form as: Body Image. 2021 Aug 4;39:139–145. doi: 10.1016/j.bodyim.2021.07.002

Multi-State Modeling of Thought-Shape Fusion using Ecological Momentary Assessment

Tyler B Mason 1, Kathryn E Smith 2, Ross D Crosby 3,4, Scott G Engel 3,4, Carol B Peterson 5, Stephen A Wonderlich 3,4, Haomiao Jin 6
PMCID: PMC8654058  NIHMSID: NIHMS1727919  PMID: 34358817

Abstract

Body dissatisfaction (BD) and preoccupation with thoughts of food (PTF) are intertwined and are components of thought-shape fusion. Thought-shape fusion describes the process by which PTF lead to beliefs about weight and shape. To study thought-shape fusion in daily life and explore various transitions between BD and PTF, 30 women with binge eating completed ecological momentary assessment for 14 days. BD and PTF were assessed using continuous rating scales at each prompt. Multi-state modeling, which analyzes micro-temporal transitions between discrete states, was used to examine transitions among four states created with BD and PTF ratings. The four states included low BD/low PTF, low BD/high PTF, high BD/low PTF, and high BD/high PTF. Affect and disordered eating were examined as covariates of state transitions. Results showed high BD states were self-perpetrating, such that when in high BD states, transition to low BD states were less likely. Regarding covariates, positive affect buffered against maladaptive transitions whereas negative affect and disordered eating increased risk. Findings highlighted high BD states as influential, and negative affect and disordered eating as risk factors and positive affect as preventive. This study enhances theory of thought-shape fusion and implicates transitions from BD to PTF as possible underlying transitions.

Keywords: thought-shape fusion, eating disorders, multistate modeling, ecological momentary assessment, body dissatisfaction

Eating disorders are severe psychiatric conditions associated with increased morbidity and mortality (Arcelus et al., 2011; Hudson et al., 2007). Body dissatisfaction (BD) and preoccupation with thoughts of food (PTF) are core components of eating disorder psychopathology and are associated with eating disorder behaviors as well as overall severity (Grilo et al., 2009; Mason & Lewis, 2015). Thought-shape fusion refers to a phenomenon in which thoughts of eating forbidden foods leads to BD or perceptions that one has gained weight (Shafran et al., 1999). Thus, thought-shape fusion suggests that BD and PTF are deeply intertwined, with thought-shape fusion representing times when elevated PTF leads to increases in BD (Shafran et al., 1999). Studies have shown that individuals exhibiting greater thought-shape fusion have higher eating disorder psychopathology and psychological distress (Coelho et al., 2015; Jáuregui-Lobera et al., 2012; Shafran et al., 1999). However, although the definition of thought-shape fusion inherently suggests a momentary, micro-temporal association between PTF and BD (where PTF leads to BD), research has yet to study this phenomenon in the natural environment.

Ecological momentary assessment (EMA) is a methodology that is suitable for studying how processes unfold across the day in individuals’ daily lives (Shiffman et al., 2008). Because many theoretical models and concepts in the eating disorder literature imply momentary processes (e.g., affect regulation, interpersonal; Ansell et al., 2012; Haedt-Matt & Keel, 2011), EMA offers a novel way to test these models and concepts. This is accomplished by having participants complete brief surveys about behaviors, experiences, and contextual variables multiple times across the day for a short period of time (e.g., a week). EMA is advantageous as it limits self-report biases and allows for examining momentary, temporal relations among variables across the day in individuals’ natural environments (Shiffman et al., 2008). While a host of statistical techniques can be applied to EMA data to study micro-temporal dynamics, multistate modeling offers particular advantages to study the sequential processes of thought-shape fusion.

Multistate modeling is a statistical approach that models the dynamic processes in which individuals transition through a series of states in continuous time and can be applied to EMA data (Jackson, 2011; Koslovsky et al., 2018). Multi-state models offer several advantages with respect to more commonly used methods like multi-level modeling in analyzing EMA data. The models are inherently based in a continuous-time Markov chain (Jackson, 2011), which assumes that state transitions may happen at any time between observations and the exact times of transitions are unknown. This underlying assumption matches well with the nature of most EMA designs in which momentary data are assessed a few times a day and no data are available about what happens between assessments. Multi-state models also work well with uneven intervals between EMA assessments and handle missing data naturally by treating EMA nonresponses the same as the missing information between discrete EMA assessments. Another advantage of the models is that estimates can be generated for both the risk of moving from one state to the other and the average time of staying on a state. Those estimates can provide a more comprehensive and dynamic characterization of the process under observation.

With regard to thought-shape fusion, multistate modeling can be used to examine how individuals transition between states of high and low PTF and BD. For each pair of states, multistate modeling can estimate their transition intensities, which describe the instantaneous risk of moving from one state to the other. Multistate modeling has been used to study a wide range of health-related behaviors like smoking, cannabis use, and cancer screening (Jackson et al., 2003; Mayet et al., 2011; Mayet et al., 2012; Uhry et al., 2010). However, to our knowledge, multistate modeling has not been utilized in eating disorder research, and therefore, use of multistate modeling extends statistical approaches used in prior EMA eating disorder studies by allowing for studying dynamical transitions between disorder relevant states. Using multistate modeling to understand thought-shape fusion in daily life has theoretical and clinical implications. This approach will provide insight into the construct of thought-shape fusion, which might lead to new recommendations for strategies for targeting thought-shape fusion. Further understanding of micro-temporal links between PTF and BD will provide new information on the utility of de-coupling PTF and BD in the natural environment as well as the sequencing of targeting PTF or BD. In addition, specifically understanding thought-shape fusion in daily life can inform the development of novel empirically-supported just-in-time adaptive interventions for eating disorders.

In addition to identifying transitions between states, covariates can be included in the multistate model to examine whether and to what extent the variables increase or decrease the risk of moving from one state to the other. Affect, dietary restriction, and occurrence of binge eating are key covariates that may serve as risk and preventive factors for states of BD and PTF. Survey research in a large sample of college students has shown inter-correlations between these factors and BD and PTF (Mason et al., 2015). Further, previous EMA research has shown that upward appearance-related social comparisons were associated with higher negative affect, and women with high BD experienced more thoughts of dieting than women with low BD following such comparisons (Leahey et al., 2007). In addition, EMA research has consistently shown that negative affect and dietary restriction are positively associated with BD and other body image constructs (e.g., body image discrepancy, body checking, appearance-related stress; Heron & Smyth, 2013; Lavender et al., 2013; Mason et al., 2018; Stefano et al., 2016) as well as food- and eating-related thoughts (Seidel et al., 2016; Shingleton et al., 2013).

Given that thought-shape fusion is associated with ED psychopathology and no research has examined the dynamics of thought-shape fusion in daily life, the current study used a multistate modeling approach with EMA data to elucidate micro-temporal transitions among BD and PTF as well as factors that increase or decrease risk of transitioning between these states. Given thought-shape fusion has typically referred to the causal relation from thoughts of eating to BD, thought-shape fusion was conceptualized as PTF leading to BD (i.e., the transition from a state of low BD/high PTF to a state of high BD/high PTF). However, it is also possible that BD leads to thoughts of food—including attempts to suppress these thoughts (Mason & Lewis, 2015). Therefore, we also assessed the extent to which high BD/low PTF leads to high BD/high PTF, which may lend insight regarding the degree to which thought-shape fusion becomes self-perpetuating. In this study, dynamic transitions between BD and PTF were modeled first. Then, covariates were added to models to determine whether positive and negative affect, restriction, and recent binge eating were associated with the likelihood of transitioning between states. In general, this study was exploratory, and therefore had no specific hypotheses. This was a secondary data analysis of a project designed to examine state and trait predictors of women’s binge eating in daily life (Mason et la, 2021; Smith et al., 2019).

Method

Participants and Procedure

Thirty women with binge-eating symptoms were recruited through eating disorder and weight management clinics and registries of participants from prior studies. Inclusion criteria included reporting at least one objective binge eating episode in the past month via clinical interview on the Eating Disorder Examination (Fairburn et al., 2008). Exclusion criteria were: 1) inability to read or speak English; 2) current psychosis determined by the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders - IV (SCID-IV; First et al., 2002); 3) current mania determined by the SCID-IV; 4) acutely suicidal as determined by the Suicide Behavior Questionnaire-Revised (Osman et al., 2001); 5) current medical instability as determined by vital signs and blood pressure; 6) past year severe substance use disorder determined by the SCID-IV; 7) severe cognitive impairment or intellectual disability determined by phone screen; 8) currently pregnant or breastfeeding; 9) current or past four weeks inpatient or partial hospitalization; and 10) changes to eating disorder treatment in the past four weeks.

The study was reviewed and approved by the University of North Dakota Institutional Review Board (#201612–126). Potential participants were screened via phone or in-person at a clinic visit to evaluate initial study criteria. Women meeting initial criteria and who expressed interest in study participation were scheduled for in-person study visit. During this visit, they completed the informed consent process and clinical interviews to assess eligibility criteria; interviews were administered by trained master’s level assessors. Those who met all eligibility criteria completed anthropometrics, self-report questionnaires, cognitive tasks, and were trained on the EMA protocol. Training included a practice EMA survey and instruction about definitions of eating disorder behaviors. Participants were given a document with the eating disorder behavior definitions, which a research assistant described to the participant; they were also able to take this home. Specifically, for binge eating, participants were reminded about the binge-eating episodes that they described during the EDE and informed that they should only record binge eating if it includes overeating and loss of control.

Participants completed 14 days of EMA starting the following day. The Real Time Assessment In the Natural Environment (RETAINE; retaine.org) system was used to administer EMA surveys via text message and web. Each day, participants received five semi-random text messages delivered to their mobile phones, and messages were sent within five pre-determined windows starting in the morning through the evening. These windows varied around five anchor points: 9am, 12pm, 3pm, 6pm, and 9pm. In the text message, women were provided a link to use to complete the EMA survey. Participants had an hour to complete the survey before they could no longer access the survey, which prevented backlogging of EMA reports. A research assistant called participants halfway through the EMA protocol to remind them about compliance and answer questions/concerns. Participants received $110 for completion of in-person clinical interviews and assessments, and $2 per signal that they completed during the EMA protocol.

EMA Measures

Momentary body dissatisfaction.

Four items from the State Self-Esteem Scale (SSES; Heatherton & Polivy, 1991) appearance subscale were used to measure momentary appearance concerns. Participants were instructed to indicate the extent to which they felt satisfied with the way their body looks, dissatisfied with their weight, pleased with their appearance right now, and unattractive at each signal on a scale ranging from 0 (not at all) to 4 (extremely). The first and third items were reverse scored so that higher scores indicated higher levels of body dissatisfaction. An average of the four items was taken, and scores ranged from 0 to 4. This measure has been used in previous EMA research (Leahey et al., 2011).

Preoccupation with thoughts of food (PTF).

PTF was assessed with one item from the Appetite Scale (Kikuchi et al., 2015). At each signal, participants were asked: “RIGHT NOW, how would you describe your preoccupation with thoughts of food?” They rated this item on a visual analog scale anchored with 1 (none) and 100 (intense). This item shows a high degree of within-subjects variability indicating that it varies across the day (Kikuchi et al., 2015).

Affect.

Current positive and negative affect was assessed with the 10-item Positive and Negative Affect Schedule Short Form (PANAS-SF; Thompson, 2007; Watson et al., 1988). Participants indicated their current affective states on a scale from 1 (not at all) to 5 (extremely). The five positive affect and five negative affect items were summed to calculate composite positive affect and negative affect scores at each signal, with higher scores indicating greater positive affect or negative affect intensity.

Eating disorder behaviors.

Binge eating was assessed at each recording by an item asking whether participants engaged in this behavior since their last recording. Restriction was assessed by an item asking whether participants “ate less to control their weight.” Meal-skipping was assessed by an asking whether participants “skipped a meal (or meals) in an effort to control their weight.” Previous eating disorders EMA studies have used behavior checklists where participants indicate whether various eating disorder behaviors occurred (e.g., Engel et al., 2013; Smyth et al., 2007).

Statistical Analysis

The body dissatisfaction and preoccupation with thoughts of food scores were categorized to form states. Based on the distributions in the study sample, cutoffs were chosen to ensure each state has roughly equal number of frequencies. The distribution of body dissatisfaction score in the study sample was negatively skewed, with a median score of 3.75 and no scores less than 1.25 reported. Based on the distribution, we dichotomized the body dissatisfaction score into two levels: ≥3.75 indicating “higher body dissatisfaction” and <3.75 indicating “lower body dissatisfaction”. The preoccupation with thoughts of food score had a U-shaped distribution in the study sample. As a result, the score was dichotomized into two levels with the middle score (i.e. 50) as the cutoff: ≥50 indicating “higher preoccupation with thoughts of food” and <50 indicating “lower preoccupation with thoughts of food”. Using the dichotomized variables, women could experience four different states: a) Higher Body Dissatisfaction and Higher Preoccupation with Thoughts of Food (HBD/HPTF); b) Lower Body Dissatisfaction and Higher Preoccupation with Thoughts of Food (LBD/HPTF); c) Higher Body Dissatisfaction and Lower Preoccupation with Thoughts of Food (HBD/LPTF); and d) Lower Body Dissatisfaction and Lower Preoccupation with Thoughts of Food (LBD/LPTF).

A multistate modeling approach was used to analyze the transitions among the four states: HBD/HPTF, LBD/HPTF, HBD/LPTF, and LBD/LPTF. The approach assumes that data about states are collected at discrete times, but information from the periods between data collections are not available. This is consistent with EMA in which data about states are collected at several random prompts throughout the day. State change can occur anytime in the interval between random prompts, and therefore, it is unknown when exactly state changes occur. State transitions from LBD/HPTF to HBD/HPTF were conceptualized as consistent with the original definition of thought-shape fusion, and state transitions from HBD/LPTF to HBD/HPTF were conceptualized as the exploratory alternative definition of thought-shape fusion. Covariates were included in the model to examine if they increase or decrease the risk of moving from one state to the other. Covariates included negative affect, positive affect, binge eating, restriction, and meal-skipping. Both negative affect and positive affect were coded as dichotomous variables with a cutoff point of 3 in a score range from 1 to 5. Reported binge eating, restriction, and meal-skipping were dichotomous variables. We also examined whether the time of assessment has an effect by adding weekday vs. weekend and daytime (before 6pm] vs. evening (after 6pm) as covariates in a post-hoc examination of the model.

To build the multistate model, we first examined the EMA data to generate the frequency of transitions for each pairs of states. As the observed frequencies of transitions between HBD/HPTF and LBD/LPTF and between HBD/LPTF and LBD/HPTF are small, we set their respective transition intensities, which quantifies the likelihood of transitioning from one state to the other, to zero in the multistate model. Assigning zero transition intensity to transitions with small observed frequency is a practice recommended by Jackson (2011) to improve model fitting and interpretation. Using the model setting described above, EMA data were fitted into the multistate model without any covariates to estimate the transition intensities for each pairs of states. The estimated transition intensity has a value ranging from 0 to positive infinity, with larger value indicating higher likelihood of transition. Covariates are then added to the multistate model to examine their effects on the transition intensity, which are reported as hazard ratios (HRs). The model fitting also estimated the average time of staying in a specific state. The statistical analysis was conducted using R version 3.5.2 with package “msm” for multistate modeling (Jackson, 2011).

Results

Descriptive Statistics

The mean age of the sample was 34.13 years (SD=13.92; Range=19–62), and the mean body mass index was 34.13 (SD=9.47; Range=18.43–57.83). The majority of the sample was White (93%) followed by 7% Black. One person reported Hispanic ethnicity. Most women had some college education (56.7%) or a Bachelor’s degree (33.3%) followed by Master’s degree (3.3%), high school diploma or equivalent (3.3%), and less than high school (3.3%). Eating disorder diagnoses as defined by the Eating Disorder Diagnostic Scale DSM-5 version (Stice. 2014) were 3.3% anorexia nervosa, 46.7% bulimia nervosa, 20.0% binge-eating disorder, 3.3% low frequency binge-eating disorder, 3.3% night-eating syndrome, 23.3% no diagnosis or missing. Assessed via EDE interview, the average number of objective binge-eating episodes in the past 28 days was 12.27 (SD=13.78; Range=1–76). Table 1 shows the frequencies of transitions in this study sample.

Table 1.

Frequencies of transition between states in the study sample

From╲To LBD/LPTF HBD/LPTF LBD/HPTF HBD/HPTF
LBD/LPTF 391 50 84 22
HBD/LPTF 39 220 15 94
LBD/HPTF 80 16 73 28
HBD/HPTF 33 91 23 269
Open in a new tab
*

LBD: Lower Body Dissatisfaction; HBD: Higher Body Dissatisfaction; LPTF: Lower Preoccupation with Thoughts of Food; HPTF: Higher Preoccupation with Thoughts of Food

Micro-Temporal Transitions between PTF and BD

Estimates and 95% confidence intervals (CIs) of the transition intensity and average stay for the multistate model of BD and PTF are shown in Figure 1. The results showed that the average duration of the LBD/HPTF state was 0.12 days (95% CI=0.09–0.15 days). When the transition from the LBD/HPTF state occurred, participants were 4.86 times (calculated by the ratio of transition intensities, i.e., 7.15/1.47=4.86) more likely to transition from the LBD/HPTF state to the LBD/LPTF state rather than moving to the HBD/HPTF state. However, if a participant did transit to the HBD/HPTF state, the average stay on the HBD/HPTF state was 0.31 days (95% CI =0.25–0.38 days), and when in this state, participants were 3.35 times (i.e. 2.51/0.75=3.35) more likely to move to the HBD/LPTF state rather than moving back to the LBD/HPTF state. The average duration of staying on the HBD/LPTF state was 0.27 days (95% CI=0.22–0.35 days) and it was 3.54 times (i.e. 2.80/0.79=3.54) more likely to move from the HBD/LPTF to HBD/HPTF rather than to the LBD/LPTF state. As for the state of LBD/LPTF, the average duration was 0.32 days (95% CI = 0.25–0.41 days) and it was 3.54 times (i.e., 2.44/0.69=3.54) more likely to move to the LBD/HPTF state than to the HBD/LPTF state.

Figure 1.

Figure 1.

Open in a new tab

Transition Intensity and Average Stay for the Multistate Model of Body Dissatisfaction and Preoccupation with Thoughts of Food

The above results suggest that when individuals are in either LBD/LPTF or LBD/HPTF state, it is more likely that they will continue moving between these two states in which they experience relatively lower BD and varying levels of PTF. However, if the state transitions to either the HBD/LPTF or HBD/HPTF (i.e., reflecting potential thought-shape fusion), they will be likely to continue moving between these two states that represent higher BD, rather than transitioning back to a lower BD state.

Predictors of State Transitions

The HR estimate and 95% CI of for covariate effects on the transitions in the multistate model of BD and PTF are shown in Figure 2. The results suggest that positive affect was significantly associated with a lower likelihood of transitioning between the LBD/LPTF and LBD/HPTF states. Positive affect was also associated with a lower likelihood of transitioning from the LBD/HPTF to HBD/HPTF state (i.e., thought-shape fusion). In contrast, negative affect was associated with a higher likelihood of thought-shape fusion (i.e., transitioning between the LBD/HPTF and HBD/HPTF states) and a higher likelihood of transitioning from the LBD/LPTF to HBD/LPTF state. Binge eating was associated with a higher likelihood of transitioning between the LBD/LPTF and HBD/LPTF states. Restricting food intake and skipping meals were associated with a higher likelihood of transitioning from LBD/HPTF to HBD/HPTF state and a higher likelihood of transitioning from the HBD/LPTF to HBD/HPTF state. Post-hoc examination suggested there was no significant effect of the time of assessment on state transitions.

Figure 2.

Figure 2.

Open in a new tab

Hazard Ratios of Significant Covariate Effects on the Transitions in the Multistate Model of Body Dissatisfaction and Preoccupation with Thoughts of Food

Discussion

The present study was the first momentary exploration of the operational definition of thought-shape fusion, which, by definition, suggests a micro-temporal association between PTF and BD (where PTF leads to BD; Shafran et al., 1999). In order to understand the dynamic processes underpinning thought-shape fusion, conceptualized as high PTF leading to high BD, multistate modeling examined micro-longitudinal relationships between states of varying BD and PTF among women with binge eating pathology. Although transitions indicative of the original definition of thought-shape fusion occurred, there was more evidence for BD leading to increased thoughts of food, suggesting that thought-shape fusion may be better characterized as such.

Furthermore, results showed that once participants transitioned into one of the two states that consisted of higher BD, they were more likely to transition between these states as opposed to transitioning out of these states to lower BD. That is, when women felt dissatisfied with their body and were highly preoccupied with thoughts of food, they were more likely to stay in a state of high BD coupled with fluctuating PTF, which could contribute to the maintenance of eating disorder psychopathology. This is consistent with previous research highlighting the centrality of BD in eating disorders (Lantz et al., 2018; Smith et al., 2019) and characterizes thought-shape fusion as a potential self-reinforcing process.

To better understand the context in which these transitions occur, analyses examined the influence of specific covariates that increased the likelihood of transitions among states, particularly those characterized by thought-shape fusion. Results showed that positive affect was a preventive factor, while negative affect, recent occurrence of a binge-eating episode, dietary restriction, and skipping meals were risk factors. Specifically, elevated positive affect reduced risk of moving from states of low BD and PTF to states characterized by high BD and PTF. This finding is consistent with previous literature showing that positive affect improves self-regulation (Tice et al., 2007) and bolsters adaptive coping and self-compassion skills (Fredrickson & Joiner, 2002). Negative affect was associated with a higher likelihood of transitions between states of low BD (with either low or high PTF) to states of high BD (with either low or high PTF). There is considerable research that describes the role of momentary negative affect as a robust precipitant of eating disorder behaviors (Engel et al., 2016; Haedt-Matt & Keel, 2011). Further, negative affect has been associated with poorer coping, increased rumination, as well as greater momentary impulsivity and impulsive behaviors (Engel et al. 2016; Smith et al., 2021; Tomko et al., 2015).

The occurrence of binge eating while in states of low BD and low PTF increased the risk of transitioning into the high and low BD states with low PTF. This finding suggests that harm reduction strategies could be used to dampen the negative impact of binge eating on dynamic BD. After binge eating, many individuals may ruminate about the reasons for the binge eating or may experience guilt over eating (Cowdrey & Park, 2011; Goldschmidt et al., 2018). Restricting food intake and skipping meals increased the probability of transitioning from states of low BD and high PTF to high BD and high PTF. Therefore, restriction and meal skipping may potentiate thought-shape fusion, as reflected by an increased risk that states of high PTF will lead to high BD. In addition, restriction and meal skipping increased the probability of moving from states of high BD and low PTF to high BD and high PTF. It may be that restriction and skipping meals are associated with increases in physiological hunger, which in turn increase thoughts of food and desire to eat, particularly when individuals are already experiencing thoughts about weight gain and BD.

This study provides a number of clinical implications. Broadly, results suggest the important of the development of novel ecological momentary interventions (EMIs; Heron & Smyth, 2014). EMIs are interventions that occur in individuals’ daily life that provide intervention material in real-life when it is most crucial. Psychoeducational strategies could be used from a variety of therapeutic approaches to target risk and maintenance factors. Use of these strategies could “short circuit” a chain of events which could lead to the pernicious cycles underlying thought shape fusion. Momentary coping and emotion regulation strategies and behavioral activation, which have been implemented in previous therapies (e.g., Mazzucchelli et al., 2010; Wonderlich et al., 2014), may be useful in reducing negative affect and bolstering positive affect. Acceptance and commitment strategies, such as cognitive defusion, may be useful for harm reduction after binge eating (Hayes et al., 2006). That is, teaching individuals to accept that the binge eating occurred and committing to trying not to binge eat in the future. Further, strategies focused on promoting regular meals and eating (e.g., Cognitive Behavior Therapy-Enhanced and Integrative Cognitive Affective Therapy; Fairburn, 2008; Wonderlich et al., 2015) may be effective in reducing restriction and meal-skipping. However, it is important to acknowledge that EMIs are relatively new types of interventions and are still not well-understood (Nahum-Shani et al., 2018). For example, research on decision points, or the optimal times to deliver interventions, is needed. It may be too late to intervene when individuals are distressed or in a vulnerable state.

Although this was the first study to our knowledge to apply multistate modeling to EMA data in the context of eating disorder symptoms, there were several limitations worth noting. The between-subjects sample size (i.e., number of participants) was relatively small. Also, the sample was mostly White and all women, which limits generalizability to other racial groups and men. The small observed frequencies of transition that involved changes in both state variables (e.g., transitions between LBD/LPTF and HBD/HPTF) in the current dataset limited our ability to model and estimate the transition intensities of such transitions. Measures were constructed to be brief to keep EMA surveys relatively unobtrusive, which may have not fully captured constructs of interest. Specifically, one item was used to measure PTF, and this item asked about foods in general opposed to foods seen as unhealthy or harmful. While no longer form EMA measures of PTF have been validated, future studies should develop new measures of assessing PTF with EMA and integrate these into EMA research where warranted. It may be useful to understand motivations for thoughts of food, such as due to cravings or physiological hunger, as well as the types of food tied to one’s thoughts. Further, it is unclear if the EMA sampling frequency used in the current study was optimal to measure the process of thought-shape fusion. It is possible that including more EMA prompts per day could lead to different findings and identify more transitions (e.g., Kockler et al., 2018).

The BD and PTF scores were categorized to form states based on their distributions in the study sample. The distribution of BD score was negatively skewed, which makes people with low BD scores less represented in the study sample. This may impact our ability to detect and model the transitions from actual low BD states. The cutoffs may also not generalize to other samples without modification based on distributions in future studies. Further, this study was cross-sectional in that it used one wave of EMA, which limited ability to examine outcomes of various dynamic states. Future studies should use a multi-burst design in the context of a macro-longitudinal study to investigate how dynamic states affect eating disorder outcomes. Finally, while we found associations between state transitions and covariates, there may be other important covariates to study in future EMA studies of thought-shape fusion. Because this was a secondary data analysis, covariates examined were limited to those available in the dataset.

In conclusion, this study elucidated patterns of dynamic states of BD and PTF, and the current findings suggest several key momentary factors that could be targeted in EMIs in the natural environment to reduce the likelihood of states that could potentiate eating disorder psychopathology. In addition, the study supports the salience of thought-shape fusion as a psychotherapeutic target, but questions remain about the ecological validity of the original definition of the thought-shape fusion of the construct. Further larger-scale EMA research with greater sampling frequency could prove useful for better characterizing the operational definition of thought-shape fusion. In addition, more research is needed to determine the most effective strategies to reduce transitions between BD and PTF in daily life.

Highlights.

  • Thought-shape fusion suggests body dissatisfaction (BD) and preoccupation with thoughts of food (PTF) are connected.

  • Multi-state modeling analyzed micro-temporal transitions among low BD/low PTF, low BD/high PTF, high BD/low PTF, and high BD/high PTF.

  • Transitions to states of high BD were self-perpetuating.

  • Positive affect buffered against maladaptive transitions.

  • Negative affect, restriction, and prior binge eating were risk factors for maladaptive transitions.

Acknowledgements

This work was funded partially by the National Institute of Mental Health (T32MH082761).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. Ansell EB, Grilo CM, & White MA (2012). Examining the interpersonal model of binge eating and loss of control over eating in women. International Journal of Eating Disorders, 45, 43–50. 10.1002/eat.20897 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arcelus J, Mitchell AJ, Wales J, & Nielsen S (2011). Mortality rates in patients with anorexia nervosa and other eating disorders: A meta-analysis of 36 studies. Archives of General Psychiatry, 68, 724–731. doi: 10.1001/archgenpsychiatry.2011.74 [DOI] [PubMed] [Google Scholar]
  3. Coelho JS, Ouellet-Courtois C, Purdon C, & Steiger H (2015). Susceptibility to cognitive distortions: The role of eating pathology. Journal of Eating Disorders, 3, 31. 10.1186/s40337-015-0068-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cowdrey FA, & Park RJ (2011). Assessing rumination in eating disorders: Principal component analysis of a minimally modified ruminative response scale. Eating Behaviors, 12, 321–324. doi: 10.1016/j.eatbeh.2011.08.001 [DOI] [PubMed] [Google Scholar]
  5. Engel SG, Crosby RD, Thomas G, Bond D, Lavender JM, Mason T, … & Wonderlich SA (2016). Ecological momentary assessment in eating disorder and obesity research: a review of the recent literature. Current Psychiatry Reports, 18, 37. 10.1007/s11920-016-0672-7 [DOI] [PubMed] [Google Scholar]
  6. Fairburn CG, Cooper Z & O’Connor M (2008). Eating Disorder Examination. In Fairburn CG (ed.), Cognitive Behaviour Therapy and Eating Disorders. Guildford Press, New York. [Google Scholar]
  7. First MB, Spitzer RL, Gibbon M, & Williams JBW (2002). Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Patient Edition. (SCID-I/P). New York: Biometrics Research, New York State Psychiatric Institute. [Google Scholar]
  8. Fredrickson BL, & Joiner T (2002). Positive emotions trigger upward spirals toward emotional well-being. Psychological Science, 13, 172–175. 10.1111/1467-9280.00431 [DOI] [PubMed] [Google Scholar]
  9. Goldschmidt AB, Crosby RD, Cao L, Wonderlich SA, Mitchell JE, Engel SG, & Peterson CB (2018). A preliminary study of momentary, naturalistic indicators of binge‐eating episodes in adults with obesity. International Journal of Eating Disorders, 51, 87–91. 10.1002/eat.22795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Grilo CM, Crosby RD, Masheb RM, White MA, Peterson CB, Wonderlich SA, … & Mitchell JE (2009). Overvaluation of shape and weight in binge eating disorder, bulimia nervosa, and sub-threshold bulimia nervosa. Behaviour Research and Therapy, 47, 692–696. 10.1016/j.brat.2009.05.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Haedt-Matt AA, & Keel PK (2011). Revisiting the affect regulation model of binge eating: a meta-analysis of studies using ecological momentary assessment. Psychological Bulletin, 137, 660–681. doi: 10.1037/a0023660 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hayes SC, Luoma JB, Bond FW, Masuda A, & Lillis J (2006). Acceptance and commitment therapy: Model, processes and outcomes. Behaviour Research and Therapy, 44, 1–25. 10.1016/j.brat.2005.06.006 [DOI] [PubMed] [Google Scholar]
  13. Heatherton TF, & Polivy J (1991). Development and validation of a scale for measuring state self-esteem. Journal of Personality and Social Psychology, 60, 895–910. [Google Scholar]
  14. Heron KE, & Smyth JM (2013). Body image discrepancy and negative affect in women’s everyday lives: An ecological momentary assessment evaluation of self-discrepancy theory. Journal of Social and Clinical Psychology, 32, 276–295. 10.1521/jscp.2013.32.3.276 [DOI] [Google Scholar]
  15. Hudson JI, Hiripi E, Pope HG Jr, & Kessler RC (2007). The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biological Psychiatry, 61, 348–358. 10.1016/j.biopsych.2006.03.040 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Jackson CH (2011). Multi-state models for panel data: The msm package for R. Journal of Statistical Software, 38, 1–29. [Google Scholar]
  17. Jackson CH, Sharples LD, Thompson SG, Duffy SW, & Couto E (2003). Multistate Markov models for disease progression with classification error. Journal of the Royal Statistical Society: Series D, 52, 193–209. 10.1111/1467-9884.00351 [DOI] [Google Scholar]
  18. Jáuregui-Lobera I, Bolaños-Ríos P, & Ruiz-Prieto I (2012). Thought–shape fusion and body image in eating disorders. International Journal of General Medicine, 5, 823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kikuchi H, Yoshiuchi K, Inada S, Ando T, & Yamamoto Y (2015). Development of an ecological momentary assessment scale for appetite. BioPsychoSocial Medicine, 9, 2. doi: 10.1186/s13030-014-0029-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kockler TD, Santangelo PS, & Ebner-Priemer UW (2018). Investigating binge eating using ecological momentary assessment: The importance of an appropriate sampling frequency. Nutrients, 10, 105. 10.3390/nu10010105 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Koslovsky MD, Swartz MD, Chan W, Leon‐Novelo L, Wilkinson AV, Kendzor DE, & Businelle MS (2018). Bayesian variable selection for multistate Markov models with interval‐censored data in an ecological momentary assessment study of smoking cessation. Biometrics, 74, 636–644. 10.1111/biom.12792 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lantz EL, Gaspar ME, DiTore R, Piers AD, & Schaumberg K (2018). Conceptualizing body dissatisfaction in eating disorders within a self-discrepancy framework: a review of evidence. Eating and Weight Disorders-Studies on Anorexia, Bulimia and Obesity, 23, 275–291. 10.1007/s40519-018-0483-4 [DOI] [PubMed] [Google Scholar]
  23. Lavender JM, Wonderlich SA, Crosby RD, Engel SG, Mitchell JE, Crow S, … & Le Grange D (2013). A naturalistic examination of body checking and dietary restriction in women with anorexia nervosa. Behaviour Research and Therapy, 51, 507–511. 10.1016/j.brat.2013.05.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Leahey TM, Crowther JH, & Ciesla JA (2011). An ecological momentary assessment of the effects of weight and shape social comparisons on women with eating pathology, high body dissatisfaction, and low body dissatisfaction. Behavior Therapy, 42, 197–210. doi: 10.1016/j.beth.2010.07.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Mazzucchelli TG, Kane RT, & Rees CS (2010). Behavioral activation interventions for well-being: A meta-analysis. Journal of Positive Psychology, 5, 105–121. 10.1080/17439760903569154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Mason TB, Lavender JM, Wonderlich SA, Crosby RD, Engel SG, Mitchell JE, … & Peterson CB (2018). Examining a momentary mediation model of appearance-related stress, anxiety, and eating disorder behaviors in adult anorexia nervosa. Eating and Weight Disorders-Studies on Anorexia, Bulimia and Obesity, 23, 637–644. 10.1007/s40519-017-0404-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mason TB, & Lewis RJ (2015). Assessing the roles of impulsivity, food-related cognitions, BMI, and demographics in the dual pathway model of binge eating among men and women. Eating Behaviors, 18, 151–155. 10.1016/j.eatbeh.2015.05.015 [DOI] [PubMed] [Google Scholar]
  28. Mason TB, Smith KE, Crosby RD, Engel SG, & Wonderlich SA (2021). Examination of momentary maintenance factors and eating disorder behaviors and cognitions using ecological momentary assessment. Eating Disorders, 29, 42–55. 10.1080/10640266.2019.1613847 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Mayet A, Legleye S, Chau N, & Falissard B (2011). Transitions between tobacco and cannabis use among adolescents: a multi-state modeling of progression from onset to daily use. Addictive Behaviors, 36, 1101–1105. 10.1016/j.addbeh.2011.06.009 [DOI] [PubMed] [Google Scholar]
  30. Mayet A, Legleye S, Falissard B, & Chau N (2012). Cannabis use stages as predictors of subsequent initiation with other illicit drugs among French adolescents: use of a multi-state model. Addictive Behaviors, 37, 160–166. 10.1016/j.addbeh.2011.09.012 [DOI] [PubMed] [Google Scholar]
  31. Nahum-Shani I, Smith SN, Spring BJ, Collins LM, Witkiewitz K, Tewari A, & Murphy SA (2018). Just-in-time adaptive interventions (JITAIs) in mobile health: key components and design principles for ongoing health behavior support. Annals of Behavioral Medicine, 52, 446–462. 10.1007/s12160-016-9830-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Seidel M, Petermann J, Diestel S, Ritschel F, Boehm I, King JA, … & Ehrlich S (2016). A naturalistic examination of negative affect and disorder-related rumination in anorexia nervosa. European child & Adolescent Psychiatry, 25, 1207–1216. 10.1007/s00787-016-0844-3 [DOI] [PubMed] [Google Scholar]
  33. Shafran R, Teachman BA, Kerry S, & Rachman S (1999). A cognitive distortion associated with eating disorders: Thought‐shape fusion. British Journal of Clinical Psychology, 38, 167–179. [DOI] [PubMed] [Google Scholar]
  34. Shingleton RM, Eddy KT, Keshaviah A, Franko DL, Swanson SA, Yu JS, … & Herzog DB (2013). Binge/purge thoughts in nonsuicidal self‐injurious adolescents: An ecological momentary analysis. International Journal of Eating Disorders, 46, 684–689. 10.1002/eat.22142 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Smith KE, Mason TB, Crosby RD, Cao L, Leonard RC, Wetterneck CT, … & Moessner M (2019). A comparative network analysis of eating disorder psychopathology and co-occurring depression and anxiety symptoms before and after treatment. Psychological Medicine, 49, 314–324. doi: 10.1017/S0033291718000867 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Smith KE, Mason TB, Crosby RD, Engel SG, & Wonderlich SA (2019). A multimodal, naturalistic investigation of relationships between behavioral impulsivity, affect, and binge eating. Appetite, 136, 50–57. 10.1016/j.appet.2019.01.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Smith KE, Mason TB, Reilly EE, Hazzard VM, Borg SL, Dvorak R, … & Wonderlich SA (2021). Examining prospective mediational relationships between momentary rumination, negative affect, and binge eating using ecological momentary assessment. Journal of Affective Disorders Reports, 5, 100138. 10.1016/j.jadr.2021.100138 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Stefano EC, Hudson DL, Whisenhunt BL, Buchanan EM, & Latner JD (2016). Examination of body checking, body image dissatisfaction, and negative affect using ecological momentary assessment. Eating Behaviors, 22, 51–54. 10.1016/j.eatbeh.2016.03.026 [DOI] [PubMed] [Google Scholar]
  39. Stice E (2014). Eating Disorder Diagnostic Scale — DSM-5 version. https://www.ori.org/sticemeasures.
  40. Telch CF, Agras WS, & Linehan MM (2001). Dialectical behavior therapy for binge eating disorder. Journal of Consulting and Clinical Psychology, 69, 1061. doi: 10.1037//0022-006x.69.6.1061 [DOI] [PubMed] [Google Scholar]
  41. Thompson ER (2007). Development and validation of an internationally reliable short-form of the positive and negative affect schedule (PANAS). Journal of Cross-Cultural Psychology, 38, 227–242. 10.1177/0022022106297301 [DOI] [Google Scholar]
  42. Tice DM, Baumeister RF, Shmueli D, & Muraven M (2007). Restoring the self: Positive affect helps improve self-regulation following ego depletion. Journal of Experimental Social Psychology, 43, 379–384. 10.1016/j.jesp.2006.05.007 [DOI] [Google Scholar]
  43. Tomko RL, Lane SP, Pronove LM, Treloar HR, Brown WC, Solhan MB, … & Trull TJ (2015). Undifferentiated negative affect and impulsivity in borderline personality and depressive disorders: A momentary perspective. Journal of Abnormal Psychology, 124, 740–753. doi: 10.1037/abn0000064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Trompetter HR, de Kleine E, & Bohlmeijer ET (2017). Why does positive mental health buffer against psychopathology? An exploratory study on self-compassion as a resilience mechanism and adaptive emotion regulation strategy. Cognitive Therapy and Research, 41, 459–468. 10.1007/s10608-016-9774-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Uhry Z, Hédelin G, Colonna M, Asselain B, Arveux P, Rogel A, … & Molinie F (2010). Multi-state Markov models in cancer screening evaluation: a brief review and case study. Statistical Methods in Medical Research, 19, 463–486. 10.1177/0962280209359848 [DOI] [PubMed] [Google Scholar]
  46. Watson D, Clark LA, & Tellegen A (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology, 54, 1063. [DOI] [PubMed] [Google Scholar]
  47. Wonderlich SA, Peterson CB, Crosby RD, Smith TL, Klein MH, Mitchell JE, & Crow SJ (2014). A randomized controlled comparison of integrative cognitive-affective therapy (ICAT) and enhanced cognitive-behavioral therapy (CBT-E) for bulimia nervosa. Psychological Medicine, 44, 543–553. 10.1017/S0033291713001098 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES