*3.3. Weight Contingencies and Weight Goals*

Most experts (88.2%) reported setting standardized weekly contingents for weight gain, ranging between 300 g and 800 g per week. The most frequent weight gain goals are 700 g per week (44.8%), 500 g per week (37.3%) and 800 g per week (10.4%). The determination of the designated weight goal differs between the institutions: 46.1 percent of experts indicated individually negotiating the weight goal with the patients, and 39.5 percent of experts indicated that the weight goal is orientated at normal or close to normal weight with BMIs ranging between 17 and 19 kg/m2. One expert indicated using different BMIs according to the age group of the patient for determining normal weight. For 15.8 percent of experts, weight goals were adapted to the planned duration of treatment. Only 6.6 percent of experts reported on not having a determined weight goal.

#### *3.4. Revisiting, Changing and Terminating Contingency Contracts*

Experts reported on revisiting the contingency contract with the patients at determined time points (47.4%), predominantly during ward rounds. Some other cases, e.g., if weight loss occurred, also made it necessary to revisit contingency contracts. For a smaller proportion of experts, revisiting the contingency contract occurred routinely after weighing the patient (28.9%), in the event of negative consequences (28.9%) or in the event of positive consequences (25.0%). Only 27.6 percent of experts reported on revisiting the contingency contract in each session.

Changing contingency contracts in the course of the inpatient treatment seemed to be handled quite differently: About one third of the experts (32.9%) reported on changing contingency contracts when patients lost weight and/or dropped below a certain BMI or when patients could not catch up to the required amount of weight gain anymore (31.6%). Some experts (7.8%) reported that changes/adaptions of the contingency contract were not intended whereas other experts reported on individually adapting contingency contracts over the course of treatment. Individually adapting might for instance take the form of temporarily changing from weight gain to weight maintenance. Some experts also reported on discharging patients from the ward for motivational reasons if weight goals were repeatedly not achieved. Patients were then offered the possibility of a re-admission after one or two weeks if they achieved some weight gain on their own.

About half of the experts (48.6%) reported that terminating contingency contracts did not happen in their institution, whereas 39.2 percent indicated that contingency contracts were terminated in special cases. These include the achievement of normal weight, somatic reasons (e.g., refeeding syndrome, infections) or if other symptoms gain priority (e.g., impulsive behavior).

#### *3.5. Consequences*

Consequences mostly depended on weight loss (90.8%) and weight gain (86.8%). One quarter of experts also reported that consequences could depend on symptoms like vomiting/purging, exercising/physical activity and eating behavior. Consequences were routinely applied after checking weight, either after every weighing (31.6%) or every second weighing (47.4%). Experts reported on choosing positive consequences themselves (23.7%) or letting the patient choose positive consequences from a list (17.1%) or freely (26.3%). In about a quarter of cases (24.7%), consequences were already determined in the contingency contract or were negotiated with the patient (13.0%). In regards to negative consequences, 36.8% experts reported determining the consequences themselves, as opposed to letting patients choose from a list (21.1%) or freely (11.8%).

Most frequently used positive consequences were the cessation of ward restriction (84.2%), being able to temporarily leave the hospital (82.9%) and the cessation of a liquid diet. Other mentioned positive consequences were: extension of treatment opportunities (e.g., patients could also participate in art or music therapy), cessation of accompanied eating, cessation of nasogastric feeding, and opportunities to temporarily leave the ward. When patients could choose their own positive consequences, chosen consequences included: buying themselves something nice, having their hair done, having a meal outside of the hospital, meeting friends, taking a bath, watching a movie/going to the cinema, bringing one's musical instrument to the ward and using the music room.

The most frequently used negative consequences were restriction to the ward (86.8%) and additional high caloric nutrients (69.7%). Further mentioned negative consequences were movement bans, nasogastric feeding, closely accompanied eating, and restrictions on using the phone or having visitors. The ultimate negative consequence was discharge from the hospital.

#### *3.6. Overall Effectiveness and Factors of Success from the Experts' Points of View*

Overall effectiveness of contingency contracts in the inpatient treatment of patients with AN was rated as 'effective for the most part' (*M* = 5.72, *SD* = 0.74). Greater clinical work experience was associated with a higher appraisal of the relevance of contingency contracts for the inpatient treatment of patients with AN (*rs* = 0.328, *p* = 0.006).

Among the factors experts rated as important for the success of a contingency contract were general factors such as therapeutic alliance (*M* = 6.67, *SD* = 0.53), empathy of the therapist (*M* = 6.58, *SD* = 0.62) and motivation of the patient (*M* = 6.53, *SD* = 0.67). Factors such as having a written record of the contingency contract (*M* = 6.56, *SD* = 0.67) and having a copy of the contingency contract available for the patient (*M* = 6.47, *SD* = 0.71) were also rated as important. For a detailed rating of factors of success, see Figure 1.

**Figure 1.** Factors of success of contingency contracts for anorexia nervosa (AN) by expert ratings. Factors of success were rated on a 7-point Likert scale from 1 'not important at all' to 7 'very important'.

#### *3.7. Emotions Experienced by the Experts during the Contingency Contract Negotiation and Emotional Burden*

On average, experts rated the overall significance of contingency contracts for the inpatient treatment of patients with AN as 'significant for the most part' to 'very significant' (*M* = 6.27, *SD* = 1.00). They reported on not experiencing the contingency contract process (preparation, negotiation and conclusion) as emotionally straining (*M* = 3.95, *SD* = 1.65), however there was a significant correlation between experiencing emotional strain and the amount of clinical work experience in years of *rs* = −0.355, *p* = 0.002. This indicates that when clinical work experience increases, emotional strain during the contingency contract process decreases.

Experts reported mainly experiencing a sense of responsibility (*M* = 5.20, *SD* = 1.09), compassion (*M* = 4.80, *SD* = 1.09) and strain (*M* = 4.54, *SD* = 1.20) during the negotiation of contingency contracts. Other emotions (tension, relaxation, ambivalence, frustration, anger and rejection) were reported as being experienced 'rarely' to 'occasionally'.

#### *3.8. Group Differences*

Potential differences in appraising contingency contracts in the inpatient treatment of patients with AN were tested between occupational groups (medical doctors vs. psychologists) and between therapeutic orientations (behavior therapy vs. psychodynamic therapy). Regarding differences between occupational groups, there were no significant differences in emotions experienced during the negotiation of contingency contracts (all *U*s > 406.50, all *p*s > 0.171). However, medical doctors rated the ethical tenability of contingency contracts, especially regarding the application of negative consequences such as restriction to the ward, higher than psychologists (*U* = 373.50, *p* = 0.008).

Regarding the emotions experienced during the contingency contract process, differences in how ambivalence was experienced were found between therapeutic orientations (*U* = 314.00, *p* = 0.060). Specifically, psychodynamic therapists experienced more ambivalence while negotiating a contingency contract. There were no group differences for the other listed emotions (all *U*s > 362.00, all *p*s > 0.302). For the ratings of factors of success, there was only one significant difference between the therapeutic orientations: Behavioral therapists rated recording the contingency contract in a written form as more important compared to psychodynamic therapists (*U* = 359.00, *p* = 0.019).

### **4. Discussion**

This study analyzed characteristics, utilization and appraisal of contingency contracts for weight gain in AN in German university hospitals specializing in the treatment of eating disorders. Experts were asked about their preparation, negotiation, conclusion and revisions of contingency contracts for patients with AN, their overall rating of effectiveness, as well as experienced emotions and possible emotional strain during this process.

#### *4.1. Similarities of Contingency Contracts in Specialized Eating Disorder Centers*

As expected, the majority of patients with AN receive a contingency contract in the participating institutions. Although not following a published manual, utilization in all centers follows internal guidelines or manuals. The most commonly used weight goals range between 500 and 800 g per week and are therefore in line with current recommendations of treatment guidelines for eating disorders [9,20].

Consequences are usually dependent on weight gain and/or weight loss. Only a few experts reported also putting consequences on other eating disorder related behaviors such as excessive exercising or vomiting. Having weight gain or weight loss as a sole focus of contingency contracts for patients with AN presumably originates from early behavioristic approaches of contingency management [18]. In light of a holistic treatment approach however, it seems advisable to consider other eating disorder related behaviors such as excessive exercising or vomiting as part of the contingency contract as well.

In sum, the present study uncovered basic characteristics of contingency contracts shared by the majority of experts. A typical contingency contract in specialized eating disorder centers in Germany can therefore be described as presented in Figure 2.

**Figure 2.** Characteristics of typical contingency contracts in specialized German university centers; AN = anorexia nervosa.

#### *4.2. Differences of Contingency Contracts in Specialized Eating Disorder Centers*

The three major aspects, in which participating institutions differ, are their definitions of weight goals, when and how contingency contracts are revisited and the choice of consequences. The definition of weight goals and the revisiting of contingency contracts possibly reflect the different self-developed manuals of the eating disorder centers. One half of experts negotiate weight goals with the patient, also taking aspects such as planned duration of stay into consideration. Another 40% orientate themselves toward a BMI value that should be achieved (low normal). One possible explanation for different BMI goals, ranging between 17 and 19 kg/m2, is the continued discussion surrounding which BMI cut-off should be used to indicate non-anorectic weight for patients with AN [1,21].

Herzog and colleagues [22] showed that lower weekly weight goals (500 g) led to a higher achieved weight at the end of treatment, compared to higher weekly weight goals (750 g). In contrast, there are studies showing that higher weekly caloric intake led to higher overall weight gain (e.g., [23]). A recent systematic review [24] demonstrated that higher calorie refeeding is not associated with increased risk of the refeeding syndrome, at least for mildly and moderately affected patients. However, inpatient therapy is mainly indicated for severely ill patients and caution regarding caloric intake in the first days of treatment should be applied. Hence, for severely malnourished patients, there is no evidence to change current approaches. However, the long-term impact of different approaches is unknown [24], therefore no clear recommendation can be made from the literature concerning what (weekly) weight goals or BMI goals should be set [9].
