Dialectical Behavior Therapy and Eating Disorders: The Use of Contingency Management Procedures to Manage Dialectical Dilemmas

Introduction

Several randomized controlled trials have indicated that DBT is an efficacious treatment for suicidal patients diagnosed with borderline personality disorder (Koons et al., 2001; Linehan, Armstrong, Suarez, Allmon, & Heard, 1991; Linehan, Comtois, Murray, Brown, Gallop, & Heard, 2006; Linehan, Heard, & Armstrong, 1993). Indeed, Division 12 (Clinical Psychology) of the American Psychological Association listed DBT as one of four empirically supported treatments (ESTs) for borderline personality disorder (BPD) and the only EST that has “strong” research support for treating BPD (Society of Clinical Psychology, 2013).

Since the inception of DBT, several researchers have adapted and applied it to various populations that stand to benefit from this treatment. Because medical complications associated with eating disorders are common, and can become life-threatening, the treatment hierarchy in DBT provides a useful frame to address the myriad complex therapy issues. Additionally, some theorists have argued that eating disorder symptoms represent a maladaptive method to regulate negative affect (Heatherton & Baumeister, 1991; Safer, Telch, & Agras, 2001; Telch, Agras, & Linehan, 2001). Therefore, because of its efficacy in treating emotion dysregulation and the corresponding maladaptive behaviors, DBT has been suggested as a promising intervention for those with eating disorders to regulate affect, e.g., binge/purge behaviors (Federici, Wisniewski, & Ben-Porath, 2012; Wisniewski, Safer, & Chen, 2007).

DBT Applied to Clients Diagnosed with Eating Disorders: A Review

To date several studies have examined the effectiveness of DBT for the treatment of individuals with eating disorders, including those diagnosed with binge eating disorder (BED), bulimia nervosa (BN) and anorexia nervosa (AN). In the first randomized study of DBT and binge eating disorder, Telch, Agras, and Linehan (2001) randomly assigned women to DBT skills training and a wait-list control condition. Results indicated that 89% of participants who received DBT skills were abstinent from binge eating as compared with only 12.5% in the wait-list control condition. Similarly, Masson, von Ranson, Wallace, and Safer (2013) randomly assigned participants to a DBT or a wait-list control condition. Dialectical behavior therapy was self-directed and consisted of an orientation, a copy of the DBT skills manual, and six 20-minute supportive phone calls over the course of 13 weeks. At the end of treatment 40% of DBT participants abstained from binge eating as compared to 3.3% in the wait-list control condition.

In order to control for the possible nonspecific effects of therapy, Safer, Robinson and Jo (2010) compared DBT with an active comparison group therapy (ACGT) modeled after Markowitz and Sacks’ (2002) manual of supportive therapy for chronic depression. Participants were randomly assigned to either 20 group sessions of DBT or ACGT. Results indicated that reductions in binge frequency were greater and achieved more quickly. Abstinence rates for bingeing were higher for the DBT group than for ACGT group (e.g., 64% vs. 36%, respectively). Despite these earlier gains, reported differences between groups were not maintained upon the three-, six-, and 12-month follow up suggesting that DBT may be responsible for the initial rapid treatment gains but not long-term therapy gains in those with BED.

Given that symptoms of bulimia have been theorized to play a role in regulating affect, several researchers have used DBT to treat individuals with bulimia nervosa. For example, Safer, Telch, and Agras (2001), in a randomized treatment study, assigned individuals diagnosed with binge/purge behaviors to once-weekly individual DBT treatment or a wait-list control group. At the end of 20 weeks, 28.6% of participants in the DBT-treatment group were abstinent from binge eating/purging behaviors as compared with no participants in the wait-list control condition. Hill, Craighead, and Safer (2011) randomly assigned participants to weekly sessions of DBT skills plus appetite-awareness training or to a six-week delay-treatment control. The appetite awareness training done in conjunction with DBT skills assisted clients in identifying and responding to internal hunger and satiety cues. At six weeks, the participants who were receiving DBT plus appetite-awareness training reported significantly fewer bulimic symptoms, had greater abstinence rates from binge/purge behaviors, and were more likely to no longer meet full or subthreshold criteria for BN as compared to the delay-treatment control group. At post treatment, after both groups had received DBT treatment for a total of 12 weeks, 26.9% of the entire sample who had received DBT treatment was abstinent from binge/purge episodes within the last month and 61.5% no longer met criteria for bulimia.

Anorexia nervosa (AN), the eating disorder most refractory to treatment, has received considerably less attention in the DBT literature. In an effort to close this gap, two preliminary uncontrolled studies have been conducted (Lynch, Gray, Hempel, Titley, Chen., & O’Mahen, 2013; Salbach-Andrae, Bohnekamp, Pfeiffer, Lehmkuhl, and Miller, 2008) Salbach-Andrae, Bohnekamp, Pfeiffer, Lehmkuhl, and Miller in their 25-week DBT program, found that women diagnosed with anorexia demonstrated an appreciable weight gain post treatment and all individuals diagnosed with AN-restricting type no longer met diagnostic criteria post treatment. However, approximately half of the sample still met criteria for AN-purging subtype, BN, or eating disorder-not otherwise specified (ED-NOS). Lynch et al. (2013) have developed an adaptation of DBT titled, radically open-DBT (RO-DBT) specifically for those individuals who present with the restricting subtype of AN. In it they target emotional overcontrol. (This adaptation is described more in depth in this volume.) In an uncontrolled trial with women diagnosed with anorexia nervosa-restricting subtype, Lynch et al. (2013) found that after an average of 21.7 weeks of RO-DBT treatment, 35% of these patients were in full remission, and an additional 55% were in partial remission. A significant increase in BMI post treatment was also found.

While the aforementioned studies show promise for the use of DBT in those with eating disorders, none of these studies specifically sought out to research individuals with eating disorders who also present with axis II pathology, such as borderline personality disorder. Approximately 56% of patients with ED present with Axis II pathology (Milos, Spindler, Buddeberg, & Crameri, 2003). Indeed, some researchers have speculated that patients with eating disorders who do not respond to treatment are likely to also be diagnosed with borderline personality disorder (Johnson, Tobin, & Dennis, 1990). Several studies suggest that patients with ED who have comorbid personality disorders are likely to be those who do not respond to traditional ED treatment and are perceived negatively by treatment providers (Woollaston & Hixenbaugh, 2008). Research supports the idea that patients with Axis II pathology are likely to respond to difficult interpersonal situations with anger or lying (Mandal & Kocur, 2013). Our clinical experience with this population supports these data and leads us to believe that those with Axis II pathology are more likely to engage in willful behaviors such as lack of transparency, angry outbursts, lying behavior and refusing medical advice when prescriptive and proscriptive approaches around their ED are employed. While these behaviors may be evident in many individuals with borderline personality disorder, the rule-bound nature of traditional eating disorder programs in which proscriptive and prescriptive behaviors are enforced exacerbates these behaviors and tends to increase willfulness.

The Problem: Being Told How to Manage the ED. the Result: Apparent Compliance vs. Active Defiance

Traditional ED treatment programs are rule bound by design. Patients attending ED treatments generally receive a prescription regarding what, when, and how much they can eat, drink, and move. At the same time other behaviors, such as excessive cutting of food or use of condiments are proscribed. The prescriptive and proscriptive model employed in traditional ED programs is effective for many, but not all ED patients. Specifically, the prescriptions typically encountered in ED treatment (e.g., you must …) result in patients with eating disorders and borderline personality disorder refusing or rebelling against treatment providers (I won’t… you can’t make me….). The typical proscriptions (e.g., you cannot…) result in similar responses (I will….And you can’t stop me!). These reactive responses to being told “what to do” may cause negative impact the therapeutic relationship., be seen by providers and loved ones as signs of “not wanting to get better” and be those that lead patients to be discharged from treatment prematurely. The prescriptions and proscriptions typical of traditional ED treatments may unintentionally lead to a dialectical dilemma, or extreme style of coping, for some patients.

Dialectical Dilemmas in EDs: Apparent Compliance vs. Active Defiance

In standard DBT, Linehan identified three dialectical dilemmas, or behavioral extremes, common in BPD patients: Emotional vulnerability vs. self-invalidation, unrelenting crisis vs. inhibited grieving, and apparent competence vs. active passivity (Linehan, 1993). Within DBT theory, emotionally vulnerable individuals [actions] have been reinforced and therefore [they] learn to alternate between these extremes of over- and under-regulation, thereby continuing to engage in ineffective behavior. In previous writings, we described a common dialectical dilemma of eating behavior: Rigid, over-controlled eating vs. absence of an eating plan (Wisniewski & Kelly, 2003). We have recently identified a second dilemma: apparent compliance vs. active defiance.

The authors suggest that the term apparent compliance describes behavior in which the patient reports engaging in a sufficient amount of a behavior to demonstrate effort but does not engage in it enough to make appreciable change. When engaging in apparently compliant behavior, the ED patient’s behavior and words result in the illusion that she is following through (i.e., complying) with treatment recommendations. As in standard DBT’s apparent competence, when the patient engages in apparently compliant behavior, the environment will often attribute lack of change to not trying or to manipulation. A typical example of apparent compliance is represented in the following example. In a traditional ED program, a client who is suffering from dehydration might receive a prescription to drink 32 ounces of a calorie beverage daily and a proscription to refrain from exercise until this medical problem is resolved. This client may report to her therapist “I am drinking Gatorade every day and haven’t gone to the gym!” Taken at face value, the statement “I am drinking Gatorade every day and haven’t gone to the gym” appears as if the patient is compliant with the treatment recommendations. However, upon further questioning by the therapist, the patient eventually describes that she drank only two ounces of Gatorade each day and was jogging in her neighborhood. So while the statement “I am drinking Gatorade and haven’t gone to the gym” may be true, and it is also apparently compliant behavior.

Active defiance, at the other end of the dialectic, connotes behavior that is willful and in opposition to treatment recommendations. An ED patient is thought to be engaging in actively defiant behavior when she directly refuses to follow treatment recommendations or program limits. The patient who refuses to eat her therapeutic meal after having an argument with another patient may be exhibiting actively defiant behavior.

The authors conceptualize apparent compliance and active defiance as problematic since these behaviors necessitate that the therapist act like a detective to obtain the full clinical picture. If apparently compliant or actively defiant statements are taken at face value, they would mislead the therapist about the patient’s progress and may block the therapist from accurate assessment and recommendations regarding the patient’s problems.

The authors further conceptualize the patient’s apparently compliant or actively defiant behavior in view of social learning theory. Specifically, we theorize that in the development or maintenance of ED behavior, the patient may have learned that apparently compliant behavior distracts people (therapist, family, friends, teacher, or coach) from focusing ED behaviors while actively defiant behaviors may prompt individuals to decrease expected/desired change from the patient. Take for example, the patient, who, after returning from a friend’s house, was asked by her mother “Did you and Jackie order pizza?” When the patient answers yes, mom’s anxiety and focus on patient’s eating decreases and the conversation ends. However, if the mom had asked more questions, she may have found that her daughter’s answer was indicative of apparent compliance, as although the pizza was ordered, the daughter hadn’t eaten any of it! The consequence of this apparently compliant behavior is that mom’s focus on the patient’s eating decreases in that moment and the patient is not blocked from or punished for ED behaviors.

An example of actively defiant behavior is noted in the case of Sue. Sue comes to her individual therapy session and though she states that she is following her meal plan 100% and is not exercising or purging, her weight is down three pounds from the previous week When the therapist recommends that Sue may need to increase food intake, she becomes dysregulated and angry. She states that she is “doing everything that is asked” and therefore she “shouldn’t be expected to eat any more” than she is currently. Without the conceptualization of the angry/aggressive behavior as active defiance, a therapist may “blame the victim,” and see this weight loss as intentional, and fail to understand what help the patient actually needs.

In order to address dialectical dilemmas, DBT therapists must focus on secondary targets. Secondary targets in DBT are those issues addressed after the primary targets (i.e., staying alive, behaviors that interfere with therapy, behaviors that interfere with quality of life), yet still must be tackled throughout treatment for an individual to learn to manage their emotions. For each dialectical dilemma in DBT, there are at least two secondary treatment targets (see Miller, Rathus, & Linehan, 2009, for a more complete discussion) the aim of which includes decreasing maladaptive behaviors and increasing adaptive responses. With respect to the dialectical dilemma of apparent compliance, the therapist needs to target increasing actual compliance and decreasing passive, noncompliant behavior. For active defiance the therapist focuses on an increase in willing, open behaviors and communication and a decrease in refusal. The authors also propose that the therapist’s use of contingency management strategies can aid in the effective targeting of these dialectical dilemmas.

Using Contingency Contracting to Address Apparent Compliance & Active Defiance

Contingency management is a general term in behavior therapy that is based on the notion that the consequences of a behavior influence the probability of the behavior’s recurrence. Thus, it is possible to increase or decrease the frequency of a behavior by influencing its associated consequences. Reinforcement, punishment, extinction, shaping, and contingency contracting are all examples of contingency management. Contingency management has been widely used to treat various psychological problems including substance abuse (Hartzler, Lash, & Roll, 2012), autism (Kohler, et al., 1995), obesity (Stalonas, Johnson, & Christ, 1978), and depression (Brannan & Nelson, 1987) by reinforcing adaptive, skillful behaviors and extinguishing maladaptive behaviors. Contingency management strategies may be a highly effective and valuable intervention for patients with complex and multi-diagnostic presentations or patients with recurrent therapy interfering behaviors (e.g., angry outbursts, lack of weight gain, lying, etc.).

In response to our conceptualization of the dialectical dilemma of apparent compliance vs. active defiance being triggered by being told how to manage ED symptoms, our private group practice treatment center in the Midwest (Cleveland Center for Eating Disorders), has adjusted the way we approach setting and evaluating goals with ED patients who attend our DBT Day Treatment Program (see Federici & Wisniewski, 2011; 2013; Federici, Wisniewski & Ben Porath, 2012 for a more through description of the program and for whom this treatment is appropriate). We propose that a collaborative use of contingency contracting can prevent or directly address issues of Apparent Compliance and Active Defiance in eating disorder patients.

Setting Step up and Step Down Criteria Using Contingency Contracting

In our ED DBT program, we ask patients to make a commitment to DBT for one year at any level of care (weekly DBT individual therapy (IT) and skills group, intensive outpatient program, day treatment program). While our goal is to help patients move themselves to the lowest level of care possible, the treatment of ED behaviors generally requires treatment and accountability at various levels of care over the course of the illness. In standard ED programming, changes in level of care and goals of treatment may be based exclusively on the American Psychiatric Association (APA) practice guidelines for eating disorders (American Journal of Psychiatry, 2000), the program itself, or insurance company criteria. Instead, we propose setting these criteria collaboratively between the patient and her DBT therapist. This model allows the patient to decide how to manage their own behavior. A patient sets goals and criteria for moving levels of care, rather than this being set by the program. We attempt to link the patient’s goals with what we have to offer (DBT treatment). We believe that decreasing arbitrary consequences (something that seems to provoke AC/AD behavior) allows the patient to take ownership of the goal as well as if she is meeting the goal.

When a patient begins DBT for ED treatment at our center, she works with her DBT therapist using contingency contracting to determine how they will know that the patient will need or is ready to step up or down a level of care. These criteria are set collaboratively and consider APA and insurance criteria, case conceptualization, learning history, response to previous treatment and most importantly, the patient’s wise mind (a DBT skill that involves a synthesis of logic and emotion). These criteria include observable information such as weight and vital signs, but also data reported by the patient on DBT diary cards, such as self-harm, suicidality, restriction, binge eating, purging, compulsive exercise and drug use (for a discussion around conceptualization of targets in ED behaviors see Wisniewski, Safer & Chen, 2007).

All attempts are made to set contingencies collaboratively while practicing wise mind (Linehan, 1993), and holding with the therapist’s and patient’s needs/beliefs and understanding of the problem at hand as equally relevant. If a disagreement in criteria arises, the therapist and patient continue to discuss the difference until a synthesis is found or one of the parties offers enough wise-minded evidence to convince the other party to alter her opinion.

In order for this model/intervention to be effective, the patient needs to understand behavior management and theory. We, therefore, teach patients the ways that both classical and operant conditioning work. Patients are taught to notice both the intended and potential unintended consequences of their behavior as well as the fact that consequences can affect behavior even without their awareness. Therefore, patients better understand how to set goals that they want to meet and how to hold themselves accountable for meeting or not meeting the goals, and thereby decreasing the situations that are likely to trigger apparently compliant or actively defiant behaviors.

The therapist’s job in contingency management is not to require the patient to set a particular contingency for a target behavior; rather, the therapist’s goal is to notice with the patient how her choice of contingencies does or/does not lead to the patient’s desired outcome. By having the patient set her own goals and contingencies, thereby decreasing the therapist’s role in prescription or proscription, we believe that this will decrease the need for the patient to employ apparently compliant or actively defiant behaviors.

Apparent Compliance: A Case Example

Let’s consider the case of Mary, who is currently purging several times per day and is trying to decrease this behavior. Mary wants to step down to outpatient care as soon as possible because she wants to get back to her job as a barista. Mary and her DBT therapist agree that during past treatments Mary has lied about symptoms (demonstrating apparent compliance) to be allowed to step down, and she wants to do things differently this time. Mary has decided that decreasing purging to once a day or less would be an indicator (among others) that she is ready to step down from day treatment to outpatient care. Mary believes that the natural consequences of feeling better about herself will motivate her to meet this goal. They also discuss the potential for Mary to report apparently compliant behavior and how they will attempt to block this behavior (rating urges to lie on diary card, asking her friends at work not to call her to cover shifts). The DBT therapist suggests to Mary that relying on natural consequences alone may not be sufficient to elicit change, given how hard it has been in the past for Mary to change this behavior. Mary feels strongly that she is “in a different place” and wants to try to set this goal using the natural consequences for motivation for one week. The DBT therapist and Mary agree that since she is currently medically stable, trying this goal for one week is a reasonable plan.

After this one-week period, Mary and her DBT therapist observe that Mary is purging more than twice a day. As part of DBT treatment, they collaboratively conduct a behavior-chain analysis to understand what is getting in the way of Mary meeting her goal. They discover jointly that the thoughts of “I will feel better about myself if I limit my purging” are fleeting and quickly overwhelmed by the anxiety of not purging. They note urges to lie about purging behaviors are somewhat elevated and discuss this. The DBT therapist then reviews learning theory withy Mary and recounts how new behavioral patterns develop. Based on past personal experience, Mary believes that working to avoid a negative consequence will likely be more motivating for her to change behavior than setting up a reward for limiting purging. Mary also believes that if she (rather than others) controls the negative consequence, then she is less likely to use apparently compliant behaviors. Mary, therefore, decides to look at a picture of tooth decay (a natural negative consequence of purging) for 15 minutes on each day that she purges more than once. Based on previous behavior-chain analyses that Mary and her therapist have conducted on purging episodes, she is aware that one of the intended effects of purging are to “get rid of food” that she has eaten to potentially avoid weight gain. Mary decides that if she purges more than once each day, she will plan to eat to replace the food she purged in order to block this goal. Once these goals are collaboratively set, it is the therapist’s job to gently but firmly guide the patient to hold herself to the criteria they have jointly identified.

Conclusion

There are strong data to support the use of modified, skills-only DBT in treating patients with ED who are diagnosed with BED or BN. While the data are still emerging, there does appear to be promising evidence for the use of DBT in individuals who are also diagnosed with any ED as well as with BPD. Future research in the form of randomized controlled trials will be needed to solidify effectiveness of this model. That being said, there is a need in the literature for papers delineating conceptual and practical strategies for use with this difficult population. The current paper detailed a previously undescribed dialectical dilemma in the ED/DBT literature: apparent compliance vs. active defiance. The authors suggest that the term apparent compliance describes behavior in which the patient reports engaging in or appears to display, a sufficient amount of a behavior to demonstrate effort but not enough to make appreciable change; while active defiance connotes behavior that is willful and in opposition to treatment recommendations. The authors propose the development of this dialectical dilemma in the context of learning theory and offer that the use of collaborative contingency contracting to effectively address these behaviors. While there is some preliminary evidence to suggest that a more flexible approach with ED patients also diagnosed with BPD is effective (Federici & Wisniewski, 2013), future studies should attempt to isolate whether this aspect of treatment may be contribute to better outcome.