Show Summary Details
Page of

(p. 33) Why Exposure for Eating Disorders?: A Rationale for Treatment 

(p. 33) Why Exposure for Eating Disorders?: A Rationale for Treatment
(p. 33) Why Exposure for Eating Disorders?: A Rationale for Treatment

Carolyn Black Becker

, Nicholas R. Farrell

, and Glenn Waller

Page of

PRINTED FROM OXFORD CLINICAL PSYCHOLOGY ONLINE ( © Oxford University Press, 2020. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Clinical Psychology Online for personal use (for details see Privacy Policy and Legal Notice).

date: 04 April 2020

Now that we have reviewed how exposure therapy is used to address a variety of anxiety-related problems, we turn our attention to the rationale for applying this treatment to eating disorders (EDs). You will benefit from having a thorough understanding of this rationale in two key ways. First, you will likely experience a greater sense of confidence in your ability to conceptualize and treat a variety of ED cases from an exposure-based framework. Rather than worrying about establishing the “correct diagnosis” for your patient and/or choosing the treatment manual/guide that pertains to that diagnosis, you will understand how exposure-based strategies apply to key transdiagnostic features that underpin many EDs (e.g., Murray et al., 2018). Second, you will be able to convey your understanding to your patients so that they can understand why exposure is a good “fit” with ED treatment. This can increase your patients’ investment in the treatment and optimism about a positive outcome—both of which are associated with good response to treatment. Accordingly, this chapter provides you with an overview of three critical reasons why exposure ought to be a tool you employ consistently in your work with individuals with EDs.

Reason #1: Content Overlap

Put simply, the overlap in content between EDs and anxiety disorders is significant. Indeed, several defensible arguments have been made to conceptualize EDs as a variant of a larger, encompassing anxiety “syndrome” (e.g., Waller, 2008). So extensive is the conceptual and phenomenological overlap between EDs and anxiety disorders that it is beyond the scope of this chapter to review it comprehensively. As such, what follows are descriptions of what we believe to be the four most critical shared features between EDs and anxiety disorders: (a) preoccupation with (p. 34) feared outcomes, (b) avoidance of distressing stimuli, (c) safety behaviors, and (d) the functional relatedness of the preceding symptoms. Having a firm grasp of these features will help you better to “see” EDs through the lens of empirically supported models of how anxiety disorders manifest and persist. This understanding is essential to your effective application of exposure with EDs.

Preoccupation with Feared Outcomes

Individuals suffering from anxiety disorders experience persistent preoccupation with a vast array of feared outcomes. From acquiring a fatal illness to being publicly humiliated, the range of possibilities of feared outcomes is essentially limitless. Many feared outcomes involve consequences that are unquestionably catastrophic (e.g., being responsible for the death of a child). Often, the driving force behind these types of fears is the tendency to overestimate the likelihood of such catastrophic outcomes (e.g., Salkovskis et al., 2000). By comparison, other feared outcomes are less severe (e.g., losing one’s wallet). Nonetheless, the perceived threat of these outcomes causes intense fear for certain individuals because of their tendency to overestimate the cost of such outcomes. That is, individuals who experience pathological anxiety often believe that relatively common mishaps would be unbearably distressing (Wilson & Rapee, 2005). For example, missing a bus (low threat) will be accompanied by thoughts of arriving late at work, failing in a key task as a result, and being fired (high cost). The tendency to overestimate the likelihood and/or severity of negative consequences is a critical driver of fear preoccupation (Clark, 1999). Indeed, many established measurements of anxiety disorder symptom severity assess the extent to which individuals report being preoccupied with feared outcomes (e.g., Steketee, Frost, & Bogart, 1996), indicating the transdiagnostic relevance of this construct across anxiety disorders.

Your ED patients will be just as preoccupied with feared outcomes. Indeed, an increasing body of research has converged on the finding that across ED diagnoses, concern with feared outcomes related to eating is central to ED pathology (Levinson et al., 2017; Murray et al., 2018; Smith et al., 2018). Common examples of feared outcomes in EDs include continuous, exorbitant weight gain, inability to tolerate even low to moderate amounts of “surplus” weight, inability to tolerate anxiety while eating, and losing control over eating. Other less typical feared outcomes in EDs that nonetheless present occasionally (particularly with avoidant/restrictive food intake disorder [ARFID]) include choking or vomiting uncontrollably while eating, experiencing unbearable disgust over the texture of certain foods, and suffering a serious medical consequence as a result of one’s eating habits (e.g., developing heart disease from nominal consumption of butter).

Similar to anxiety disorders, your patient’s preoccupation with feared outcomes in EDs will stem from overestimations of the likelihood and/or severity of potential threats. To illustrate, it is very common for individuals with EDs to overestimate the influence that normative dietary intake will have on weight gain (p. 35) (Waller & Mountford, 2015). More specifically, they believe it is very likely that they will gain a significant amount of weight with normal eating. Similarly, individuals with ARFID who have experienced or witnessed choking on food in the past tend to overestimate the likelihood that this will occur again.

EDs also often involve substantial overestimation of the cost (i.e., severity) of outcomes that are fairly common if not unavoidable. For instance, adolescents with EDs often fear that changes to weight and body shape, as part of pubertal development, will be overwhelming and intolerable. Similarly, for someone with an ED, eating 10 kcal more than “allowed” might be a low threat in an objective sense, but their exaggerated belief in the consequences (e.g., ballooning in weight and losing all friends as a result) makes for a very high potential cost. Not surprisingly, research indicates that the frequency and intensity of preoccupation with feared outcomes is high in both EDs and anxiety disorders, with no significant differences between the two (García-Soriano, Roncero, Perpiñá, & Belloch, 2014).

Avoidance of Distressing Stimuli

In response to their preoccupation with feared outcomes, many individuals with anxiety disorders engage in a predictable pattern of avoiding stimuli that they associate with feared outcomes. Just as the range of fears is limitless, so is the range of things that anxious individuals avoid. These avoidant tendencies serve two chief functions for individuals with anxiety disorders. First, they preclude any encounter with a stimulus assumed to be threatening, thereby minimizing the likelihood of harm, as Janet’s case illustrates.

Janet feared that entering an oncology clinic would cause her to develop cancer. Not surprisingly, Janet avoided going into, or near, oncology clinics as well as related stimuli. For instance, Janet avoided doctors. She was particularly afraid of oncology doctors but also avoided other doctors because they might have come in contact with oncology doctors either via work (“They all go to hospitals at some point”) or by socializing (“Doctors probably hang around other doctors”).

The second major function of avoidant behaviors is to prevent or escape negative emotional experiences that an individual, such as Joshua, believes to be overwhelming or intolerable.

Joshua was afraid of attending temple because he feared being unable to tolerate the anxiety he experienced secondary to doubts over whether or not he was a “true believer” of the Jewish faith. Joshua’s children attended a Jewish school, and he avoided picking them up or attending parent–teacher (p. 36) meetings for the same reason. Joshua also feared that the teachers would detect his “religious flaws” even on the phone, so he made his wife handle all teacher discussions.

Lastly, it is important to point out that not all forms of avoidance are directly observable (i.e., noticeable). Many anxious patients engage in patterns of cognitive avoidance that are more subtle and thus difficult to detect. Perhaps the most notable example of this is the avoidance of trauma-related cognitions (i.e., memories) that is characteristic of individuals with PTSD

If you have worked with individuals with EDs, you have undoubtedly noticed many similar avoidant tendencies. Similar to anxiety disorders, there is a wide array of scenarios and stimuli that individuals with EDs avoid. In many cases, the avoidant behaviors serve the purpose of trying to prevent a feared outcome. Selective food avoidance is a common feature across EDs and is often intended as a means to prevent anticipated weight gain. Indeed, it is most often the highly palatable, calorically dense foods (e.g., pastries) that individuals with EDs such as A’isha “ban” from their diet (Marzola, Nasser, Hashim, Shih, & Kaye, 2013).

A’isha reported eating a very limited diet of egg whites, grilled chicken, pickles, salad greens sprinkled with vinegar, and, occasionally, dry wheat toast. She reported that she had not eaten cheese, ice cream, butter, hamburgers, French fries, cake, cookies, or any oil in three years. She reported believing that consuming even a small amount of these foods would lead her to “blow up like a balloon.” Although she loved lattes, she only drank unsweetened black coffee so she could “see that nothing else had slipped into it.”

For an additional example of trying to avoid a feared outcome that is different from “classic” ED body image disturbances, consider the case of Christopher.

Christopher witnessed his aunt choke while eating and subsequently developed an intense fear of choking. Christopher avoided taking normal-sized bites of food and chewed each bite of food for an extended period of time before swallowing. He also ate at a very slow pace and preferred “wet” or “slippery” foods that were less likely to cause him to choke.

There are also a number of common avoidance behaviors in EDs that together share the core function of escaping from aversive emotional states that are perceived to be unbearable. As noted by Waller and Mountford (2015), many ED patients believe they are not capable of tolerating the emotional distress around learning their weight and subsequently avoid scenarios in which they would become aware of their weight.

(p. 37) Jordan, a transgender person, had lost significant weight over the past two years but feared they might gain it back. Jordan was very anxious about their weight and refused to be weighed. They would not step on a scale at home and refused to be weighed at medical appointments because they were certain that they would “totally melt down” if they found out they had gained any weight. Maintaining an appearance of “having my act together” was very important to Jordan, which made the idea of “melting down” particularly unacceptable.

Related to weight avoidance, many individuals with EDs engage in a broader assortment of body image avoidance behaviors. Together, these behaviors serve the function of suppressing negative emotional experiences related to body image. Common examples of these behaviors include avoiding clothing items that are “revealing” of one’s physique and avoiding certain activities that one associates with body image-related scrutiny (Rosen, Srebnik, Saltzberg, & Wendt, 1991).

Jordan also avoided exercising in a crowded gym, preferring to go at odd hours when it was most likely to be quiet or even empty. In addition, Jordan stayed away from people who evoked perceptions of body image-related inferiority. For instance, they stopped hanging out with a friend who “had a much better body.” Jordan also bowed out of conversations in which weight and/or body shape might become the central topic of discussion.

Finally, although binge eating is often an unintended consequence of restrictive eating, binge eating also can occur secondary to cognitive avoidance of negative emotions. Heatherton and Baumeister (1991) were among the first to propose that when individuals become aware of disliked aspects of themselves and experience negative emotions, they may engage in cognitive avoidance (i.e., escape from their self-awareness). This cognitive avoidance, in turn, leaves them susceptible to impulsive behaviors, such as binge eating. If this occurs repeatedly, individuals learn that binge eating reduces the intensity of their negative emotions, and binge eating becomes negatively reinforced. Additionally, binge eating can be involved in more than just avoidance of negative cognitions and emotions. Many individuals with EDs describe patterns of binge eating that occur in the immediate aftermath of interpersonal conflict (e.g., Luo, Nuttall, Locke, & Hopwood, 2018).

Safety Behaviors

Chapters 1 and 2 made occasional reference to safety behaviors, which we define here as actions taken with the aim of detecting and neutralizing potentially threatening scenarios or stimuli. You will see that A’isha is using a range of such (p. 38) behaviors (checking on her grandmother’s cooking; checking restaurant menus to confirm calorie counts; checking her body).

A’isha weighed herself excessively (e.g., after each time she ate) and also repeatedly checked the appearance of her physique in the mirror to gauge for possible increases in her weight or body size. She reported she was able to “breathe a sigh of relief” if everything seemed the same. In addition, A’isha carefully watched how her grandmother prepared food when she visited to make sure she didn’t add any extra fat or sugar. She noted that her grandmother understood she was on a very limited diet, but she just needed to make sure her grandmother didn’t accidentally “slip up” for her own peace of mind. She also extensively researched nutritional information if she was forced to eat at a restaurant, so that she could make sure she stayed within a “reasonable” calorie limit and relax enough to “pretend to be normal.”

In the field of anxiety disorders, it is well-established that safety behaviors play a crucial role in the development and maintenance of pathological anxiety. There are several important reasons why this is so. First, safety behaviors are temporarily “effective” in relieving distress and, thus, are negatively reinforced. Additionally, safety behaviors function similarly to avoidance (see previous section), in that they prevent individuals from having experiences that would otherwise disconfirm their anticipatory feared outcomes. Related to this, when your patient encounters a feared scenario, engages in safety behaviors, and observes that her feared outcome does not occur, it is likely that she will attribute the nonoccurrence of her feared outcome to her safety behaviors (Salkovskis, 1991). That is, safety behaviors are perceived by the individual as the only thing that prevented what was otherwise an assured negative outcome. For instance, after having to eat out several times in one week, A’isha noted that she would have “definitely gained weight” if not for her research.

Similar to avoidance of distressing stimuli, the use of safety behaviors is pervasive in individuals with EDs. Table 4.1 provides a list of safety/avoidant behaviors that are frequently used in EDs, as well as the common functions that these behaviors serve. Some of these behaviors are broadly aimed at detecting potentially threatening outcomes, such as A’isha’s frequent weight and body checking. Other ED safety behaviors are more directly intended to prevent an anticipated negative outcome from occurring. Purging behaviors and compulsive exercising are two characteristic ways that individuals who are fearful of weight gain attempt to keep themselves from gaining weight.

Table 4.1. Common Safety/Avoidant Behaviors in Eating Disorders and Their Typical Function(s)

Safety/Avoidant Behavior

Typical Function of Behavior

Chapters Referenced

Caloric restriction

Limit caloric intake to prevent weight gain

10, 12, 13

Body/weight checking

Detect potential changes in weight and/or body shape

10, 12, 13

Compulsive exercise

Burn calories to prevent weight gain; modify body physique

10, 12

Self-induced vomiting

Limit caloric intake to prevent weight gain

10, 12, 13

Using laxatives/diuretics

Limit caloric intake to prevent weight gain

10, 12, 13

Wearing baggy clothing

Concealing body size/shape from others


Food-related research

Gain awareness of the caloric content or nutritional value of foods


Eating very slowly

Prevent loss of control over eating; achieve satiation before finishing

10, 14, 15

Eating very quickly

Escape negative emotions that accompany eating

10, 14

Ripping/tearing food

Prevent loss of control over eating; slow the pace of eating


Odd food mixtures

Limit enjoyment of food by making it unpalatable


Taking very small bites

Prevent loss of control over eating; prevent choking on food

10, 14, 15

Manipulating food

Limit caloric intake to prevent weight gain



Overheat to induce sweating and lose weight



Cause body shivering to burn calories


Adjusting body posture

Mitigate distressing physiological arousal; control body appearance

13, 15

Comparing self to others

Assess “acceptability” of one’s physique compared to that of others


Seeking reassurance

Neutralize uncertainty about food content and/or body image

10, 12, 13, 14

It also is common for individuals with EDs to engage in a variety of eating-related safety behaviors to reduce anxiety around eating (Gianini et al., 2015). Common examples of these (often bizarre) eating-related safety behaviors include eating at an exceedingly slow pace, taking only very small bites of food, mixing foods together that are not typically combined, and excessive manipulation of food prior to eating (e.g., using napkins to absorb oil or butter from the surface of (p. 39) (p. 40) food). It is important to note, as Vitousek (2019) has explained, that some of these behaviors can, at times, be a normative response to starvation in very low-weight patients. For instance, slow eating may serve as a mechanism to reduce consumption. However, it also may be a relatively automatic savoring behavior associated with starvation in people who want to eat. Similarly, creation of odd combinations of food has been observed in people who are starved for reasons other than an ED (Tucker, 2006). In contrast, using napkins to absorb oil is quite specific to EDs.

The complexity of behaviors that can serve as safety behaviors highlights the need to engage in careful functional assessment to determine the purpose of any given behavior (Murray et al., 2016) or to develop hypotheses about what that purpose might be. We discuss this further in Chapter 6, but as an immediate example, blotting oil from the surface of food secondary to texture-related disgust/anxiety serves a very different purpose compared to when it is done to reduce calories. As you will see in later chapters, exposure to address the core pathology might look very different based on the different functions potentially associated with any one behavior.

Functional Relatedness of Symptoms

Finally, in addition to EDs and anxiety disorders having a great deal of overlap with regard to symptom content, they also share a commonality regarding the functional relatedness of these symptoms. In other words, the patterns of interaction between individuals’ feared concerns and behavioral responses (i.e., avoidance and safety behaviors) in EDs and anxiety tend to mirror one another. In the case of anxiety disorders, preoccupation with feared outcomes and subsequent avoidance/safety behaviors are typically functionally related to one another, and their consistent pairing leads to them becoming increasingly associated with (and predictive of) one another (Abramowitz, Taylor, & McKay, 2009). For example, an individual who is fearful of contracting serious illness from contact with common objects or surfaces may compulsively wash his hands whenever he touches a feared object or surface. His engagement in the handwashing safety behavior will contribute to more intense preoccupation with his feared outcomes, which, in turn, will instigate further handwashing.

This reciprocal influence between preoccupation with feared outcomes and engagement in avoidance/safety behaviors is also frequently observed in EDs. For instance, among individuals with EDs, distinctive patterns of preoccupation with feared outcomes (e.g., experiencing noticeable changes to body size) are prospectively predictive of specific safety behaviors (e.g., body checking), and vice versa (Levinson et al., 2018). Thus, it appears a similar interactive feedback loop between feared concerns and avoidance/safety behaviors that is a hallmark of OCD and other anxiety problems is also present in EDs. This is especially important in the context of EDs, given recent work showing that reductions in patients’ preoccupation with feared outcomes and safety behaviors during the course of (p. 41) treatment are both prospectively predictive of global ED treatment outcome (Farrell, Brosof, et al., 2019).

Reason #2: Comorbidity

In addition to the overlap in symptom content between EDs and anxiety disorders, there is also substantial comorbidity between the two. A wealth of research demonstrates this comorbidity, with the general pattern of findings indicating that well over half of individuals with EDs experience clinical symptoms of at least one anxiety condition (Kaye, Bulik, Thornton, Barbarich, & Masters, 2004). Some of the most commonly co-occurring anxiety conditions include OCD, social anxiety disorder, and PTSD (Kaye et al., 2004). Unfortunately, there is robust evidence indicating that the presence of comorbid clinical anxiety symptoms confers risk for suboptimal ED treatment outcome (Vall & Wade, 2015). Although the reasons for this are unclear, it has been suggested that engaging in an ED may be one way that individuals regulate the intense distress they experience associated with their anxiety condition. An important finding supporting this theory is that anxiety disorders tend to predate the onset of ED symptoms in most patients (Swinbourne & Touyz, 2007).

At this point, you might be discouraged and asking yourself if there is any hope for your ED patients. After all, if the majority of them are likely experiencing comorbid anxiety disorders, which can act as a barrier to optimal treatment response, how can one muster any hope? Fortunately, there is empirical evidence showing that exposure-based therapy applied simultaneously to EDs and comorbid anxiety problems can be helpful in reducing symptoms of both co-occurring conditions (Simpson et al., 2013). The elevated likelihood of encountering comorbid anxiety problems when working with ED patients is further rationale for ED clinicians to utilize exposure therapy routinely in their practice. From your patient’s point of view, it may be a more feasible proposition to adhere to one set of core treatment guidelines (e.g., confront your feared situations and disengage from safety behaviors and avoidance) rather than trying to juggle multiple disorder-specific treatment protocols at the same time. Lastly, using an exposure-based approach to address your patient’s ED and anxiety disorder may provide opportunity for “layering” different exposure stimuli in the same therapeutic exercise to achieve deepened extinction learning (Culver, Vervliet, & Craske, 2014). We discuss this in detail provide examples throughout the book.

Reason #3: Many Evidence-Based Treatments Are Inherently Exposure-Based

There are several treatment protocols that have demonstrated a solid evidence base in the treatment of EDs. Of note, CBT and family-based treatment (FBT) have amassed the greatest degree of empirical support, with dozens of studies (p. 42) attesting to the efficacy of these treatments. Although both contain a variety of different strategies, it can be argued that the key “active ingredients” of each are inherently exposure-based. The same can also be argued for some other evidence-based therapies—particularly specialist supportive clinical management for AN.

To illustrate, in evidence-based CBT for EDs, patients are encouraged to confront their anxiety around weight and body shape via weekly weighing sessions in which patients are told their weight. This weighing procedure also allows patients the opportunity to disconfirm their expectations of exorbitant weight gain (Waller & Mountford, 2015). Additionally, patients receiving CBT are encouraged to gradually include feared/avoided foods into their diet with increasing regularity over the course of treatment (Fairburn, 2008). Finally, body image-focused exposure (see Chapter 13) can be included as a key component of CBT for EDs. Although CBT contains other nonexposure components, such as cognitive therapy strategies to address the maladaptive cognitions and emotional aspects of EDs, there is evidence to suggest that the exposure-based aspects of CBT are principally responsible for the overall therapeutic benefit (Waller & Raykos, 2019).

The specific mechanisms of action in FBT are unclear. However, it is plausible that the efficacy of FBT is driven, at least in part, by intensive feeding exposure in which food is eaten more frequently, in larger quantities, and with foods that were previously avoided (Hildebrandt, Bacow, Markella, & Loeb, 2012). Moreover, because the treatment is conducted almost entirely in the individual’s natural (i.e., home) environment, as opposed to a clinic or hospital, the learning that takes place in feeding exposure likely generalizes to a greater extent. FBT also explicitly targets safety behaviors and particularly aims to eliminate those that interfere with achievement of improved nutritional stability (e.g., compulsive exercising, purging). Of course, it is important to stress to your patient that all such work needs to be structured to be a learning experience—not just eating and white knuckling the anxiety. Finally, similar to CBT, FBT includes regular exposure to weight on a weekly basis, which likely yields reductions in fear about weight gain.

Aside from CBT and FBT, other ED treatment approaches that are implemented by a wide range of providers (e.g., dietitians, psychologists, physicians, nurses etc.) share the same key objectives of guiding patients in normalization of dietary intake and cessation of unhealthy safety behaviors. Because many ED patients are, at a minimum, reluctant to cooperate with these recommended changes, we contend that an exposure-based approach to these changes represents one of the most practical and achievable courses of action for two primary reasons. First, reluctance to begin confronting feared stimuli and to stop safety behaviors is characteristic of anxiety disorders. In most cases, however, individuals with anxiety disorders can be brought tentatively on board by starting with small changes before graduating to the more wholesale changes that result in the elimination of their pathological fear and avoidance. This mirrors our experience in applying exposure to EDs: most patients are only willing to go along with minor changes at first but become increasingly willing to engage in more significant change as their confidence grows. Second, exposure therapy is a very collaborative treatment, which offers patients a reasonably high degree of agency in decision-making with (p. 43) regard to selecting targets for exposure as well as the pace of treatment. As such, we find exposure to be a useful means of navigating through the typical hesitancy that many ED patients initially bring to the table (pun intended).


In this chapter, we have outlined a theoretically based rationale for using exposure consistently in the treatment of individuals with EDs. Due to the substantial overlap between EDs and anxiety disorders (both in symptom content and in comorbidity between the conditions), exposure therapy is a sound choice for therapeutic intervention. Indeed, the most evidence-based treatments for EDs contain a number of exposure-based strategies that drive much of the therapeutic benefit. Of course, the rationale for using exposure that we have outlined in this chapter means very little if there are no data that provide empirical support for exposure in treatment of EDs. In Chapter 5, we will comprehensively review the extant research that supports exposure as an effective approach for EDs. (p. 44)