This study analyzed a large sample of YPs with AN to improve knowledge of which patients have a sub-optimal response to FBT. We examined predictors for three different aspects of treatment response: Weight gain trajectory, time to successful treatment completion, and the risk of need for intensified treatment (day program or inpatient stay). A broad range of factors related to AN, to the YP, and to the family were assessed. When tested together, a few factors appeared to overrule others that seemed significant when tested separately.

First, we assessed predictors of weight gain trajectories (analysis 1). The weight trajectories revealed a clear pattern: the largest weight increase occurred within the first 3 months of FBT with slower gains later. This corroborates earlier findings and is aligned with the treatment goal of prioritizing renourishment first (J. [29]). Surprisingly, only a few additional factors influenced weight gain: Comorbid diagnosis of behavioural or emotional disorder and older age at intake was associated with less weight gain, while a history of bullying was associated with more. Unlike Lebow et al. [26], we did not identify distinct weight gain trajectories.

Second, we assessed predictors of time to successful treatment completion at baseline and after 4 weeks (analysis 2a & 2b). YPs with typical AN, higher EDE-global, or lower baseline relative BMI took longer to successful treatment completion. The same applied for YPs with family adversities, pre-existing mental health issues, autism, or behavioural or emotional disorder. In addition, relative BMI after 4 weeks was a stronger predictor of time to successful treatment completion than baseline relative BMI. Additionally, compulsive exercise, higher levels of feeling fat, and several therapist-rated moderating factors at 4 weeks also predicted longer treatment duration.

Third, we examined predictors of intensified treatment (analysis 3). A lower baseline BMI, higher EDE-global, autism, depression, and clinician´s assessment of mother’s capacity to actively support renourishment were associated with an elevated risk of need for intensified treatment.

The role of initial underweight

The degree of initial underweight, and in extension pre-treatment weight loss, was an important predictor: it predicted longer time to successful treatment completion (analysis 2a) and a higher risk of need for intensified treatment (analysis 3). This highlights the importance of early detection and easy access to intervention in YPs showing weight loss. Delays leading to larger weight loss may prolong treatment, potentially worsen prognosis, and increase the need for day- og inpatient care, posing risk and raising health care costs. While unknown factors that caused the pretreatment weight loss may also influence treatment response, lower relative BMI was consistently linked to longer treatment and more frequent intensification even when accounting for other moderators. This reinforces prior calls for early detection and intervention [48]. Efforts like the First Episode and Rapid Early Intervention service for Eating Disorders (FREED) approach [13, 36], family-based intervention for high-risk youth [32], and Family-Based Treatment for Primary Care (FBT-PC) [25], aim to put this knowledge into practice.

The role of weight gain in first month of treatment

The weight gained in the first month of treatment appears crucial, as 4-week relative BMI was a stronger predictor of time to successful treatment completion than initial relative BMI. This aligns with earlier findings that early weight gain predicts remission by end of treatment [10, 24, 35]. These insights have inspired an additional intervention focusing on improving meal support in case of insufficient weight gain in the first month [28]. Our study adds that other mental health challenges in YP in general, and specifically autism and behavioural or emotional disorder affect treatment duration beyond early weight gain. The study is not able to map out the processes by which comorbid symptoms and disorders interferes with the parental endeavors to secure renourishment, and further studies are needed so these processes may be systematically targeted in treatment. One hypothesis is that a higher need for predictability and stronger emotional reactions in some YP may be challenging to contain for families. The finding highlights the need for early screening and individualized interventions that address additional mental health challenges in YP, the additional demands they place on the parental task, and the impact they have on renourishment and regaining eating autonomy.

The role of psychological distress

The finding that a higher baseline EDE-global increased the risk of treatment intensification suggests that, although FBT places responsibility for renourishment on the parents, thus bypassing the YP’s fluctuating motivation, the severity of YPs’ psychological distress still affects treatment outcomes, raising the likelihood of intensification. More severe levels of feeling fat, and YP being less able to take co-responsibility for working against AN as rated by week 4, seem to represent similarly high level of eating-disorder distress, and affected time to successful treatment completion. Introducing targeted interventions for YPs with greater psychological distress, addressing their cognitive and emotional challenges alongside the physical aspects of AN might improve outcome. Similarly, if clinicians by week 4 assessed that parents were less able to help patient through difficult emotions, this predicted longer time to treatment completion. Managing emotional distress appear to be a crucial skill for both YPs and parents to succeed in FBT, alongside the necessary behavioural change. Indeed, evaluating and supporting such skills may easily be overlooked due to the initial focus on somatic danger and the behavioural focus of FBT, but qualitative evidence suggests it could strengthen FBT [54]. For instance, emotion-focused strategies [9, 23, 42] and dialectical behavioural therapy strategies [18, 43] may integrate effectively with FBT.

The role of age

Higher age was linked to less weight gain (analysis 1), but not to time to treatment completion (analysis 2). Other studies also show that older YPs with AN tend to have poorer outcomes from FBT and other treatments [11, 30], thus it is surprising that age only predicted lower weight gain. Given that insufficient weight normalisation is a known risk factor for lower remission rates in adults with AN [5, 38], our findings suggest that older adolescents’ weight trajectories need specific clinical focus. In phase one of FBT, weight restoration relies heavily on parents’ authority, but studies point to it being harder to reassert parental control in older YPs, due to their pre-AN independence. Authors recommend negotiating collaboration and co-influence when treating older YPs with FBT [8, 37]. This pinpoints a clinical dilemma: balancing between securing therapeutic alliance and taking YPs perspective into account, while ensuring sufficient weight gain to protect against long-term risks.

The role of comorbidity

A diagnosis of behavioural or emotional disorder (ICD-10: F90-98) was linked to both less weight gain (analysis 1) and prolonged time to successful treatment termination (analysis 2). In our sample, 7,5% had comorbid F90-98, such as ADHD, supporting prior studies showing higher prevalence of attention-deficit disorders among people with AN [4, 45]. Moreover, ADHD and AN may together worsen mental well-being [2]. Thus, it is unsurprising that behavioural emotional disorders negatively impacted treatment response in this study, calling for specific attention during FBT. However, the small number of YPs with this comorbidity should be noted.

Autism was more common than behavioural or emotional disorders in our sample; 21% had or received an autism diagnosis. Frequency in other AN studies differ between 4 and 31%, perhaps due to different sampling and assessment procedures [20, 51]. A diagnosis of autism in the current study predicted longer time to successful treatment termination and increased risk of intensification, consistent with research showing lower success rates in this population [44]. While necessary, intensified treatment may present risks, for instance as the structured inpatient environment can foster dependence and prolonged hospitalization [22]. On the other hand, if treatment is prolonged because outpatient treatment is not sufficient it may carry a risk of chronification. Balancing when to intensify or continue slower FBT progress is challenging, but we suggest continuing outpatient FBT if gradual progress is evident, while accommodating the YP’s autism-related needs. Autism-friendly adaptations, such as predictability, reduced social-communicative demands, and accommodating sensory challenges, might improve FBT [33] with further guidance offered by the PEACE pathway [19, 46].

Comorbid depression was associated with a higher risk of intensification. While depressive symptoms are a frequent complication of AN, a diagnosis of depression should be assigned only when the depressed mood is beyond what is inherent in AN and/or was present before or persists despite weight normalisation. Our finding suggests that comorbid depression can interfere with treatment and should not be overlooked. An integrated approach targeting both the eating disorder and affective disorder may help prevent intensification.

The role of family factors

No aspects of family adversities or carer status assessed in this study influenced weight gain trajectory. This is noteworthy since carer burden and parental resources naturally have been in focus in FBT literature, as parents are the primary agents of change. While family factors did not impact on weight gain, they impacted on time to successful treatment completion: Hence practical, economic, or health-related adversities, therapist assessment by week 4 that it was difficult for parents to take a leading role in renourishment of YP, or that parents were less able to help YP through difficult emotions, all significantly influenced the time to successful treatment completion. Evidence on family factors in FBT is mixed; an earlier study found that families with separated parents in separate households required longer time in FBT [27], while another study found no significant impact of family structure, income, YP comorbidity, or parents'self-efficacy at baseline on FBT outcome [7]. By week 4 of the study, family adversities lost significance, likely because other variables were stronger. The therapist´s assessment that it was difficult for parents to take a leading role in renourishment of YP, that parents struggled to help YP through difficult emotions and that other treatment challenges existed might better capture how family adversities impact on FBT tasks. Overall, our findings suggest that the strong momentum in early FBT and multidisciplinary team support help families even with additional challenges to take an active stance and initiate weight restoration. However, treatment may be prolonged if other adversities strain family resources, especially if this manifests itself in reduced capacity for taking on a leading role or containing strong emotions. It should be noted, however, that several assessments, including the assessment of parental capacity for taking on a leading role or containing strong emotions, was based on a simple clinician evaluation, and that a comprehensive and validated instrument assessing these factors might give more valid information.

The role of clinician-rated assessments

Several clinician-ratings turned out as significant predictive factors; mother´s capacity to take on an active role in renourishment assessed by intake, YP´s reduced ability to take co-responsibility for working against AN, difficulty for parents to take a leading role in renourishment, and parents´ reduced ability to help YP through difficult emotions assessed by 4 weeks were all significant. While these were generic Likert or dichotomous ratings, not validated scales, they highlight the importance of clinicians´ consideration of factors that may affect therapy. This study suggests that treatment team´s reflections on moderating factors in FBT contain valuable predictive insight and should impact treatment planning in clinical practice, although it would be preferable to supplement by independent assessments to evaluate the effect of bias.

Other findings

Mothers´ but not fathers´ capacity played a role in the risk of intensification. FBT specifically focus on parental alignment and on involvement of both parents. Despite Denmark's focus on gender equality, the lack of a similar association with the father’s ability to take an active role in renourishment likely reflects the fact that it is still primarily mothers who take leave of absence from work and assume the day-to-day responsibility for the refeeding process.

It was surprising that a history of bullying was associated with greater weight gain. This finding should be confirmed in other samples to rule out random findings (type 2 error). A speculative explanation for the role of prior bullying might be that the YPs felt relieved by the break from school and responded well to the time at home with a parent, free from peer-related adversities during renourishment.

Strengths

The hospital unit where the study took place, covers the largest region of Denmark and is the only publicly funded, hospital-based service for youth with eating disorders in the region. The primary strength of this study is its large, geographically representative sample that was monitored prospectively.

Limitations

Main limitations stem from the study being a case series from a single site with no comparison to other treatments or usual care and limited fidelity measures. A further limitation is missing data at varied time points, particularly at the end of treatment. Although unfortunate, this is a common issue in clinical psychiatric cohorts [40]. Moreover, several variables, including the assessment of parental capacity for taking on a leading role or containing strong emotions, and assessment of motivation, was based on a simple and generic clinician evaluation. Comprehensive and validated instruments assessing these factors would give better information and are needed in future studies. Also, assessment of clinical presentation and outcomes were performed by internal clinicians, and not independent researchers. Lastly, this was a naturalistic study with no formal treatment fidelity measures, and with adaptations to manualized FBT, especially regarding no fixed duration, and therefore conclusions may not be fully generalizable to FBT delivery within 20 sessions and with control of adherence to manualized FBT.