The research team set out to address a persistent gap in the eating disorders literature. Family members and partners who support individuals with anorexia nervosa, bulimia nervosa, and related conditions frequently experience chronic stress, social isolation, and emotional exhaustion, which can escalate into clinically significant anxiety and depression. Previous studies have shown that carers of people with eating disorders report some of the highest levels of psychological distress among informal carers of any patient group, driven in part by the demands of meal supervision, the emotional volatility that often accompanies these illnesses, and the prolonged course of recovery. Yet although a range of interventions has been developed to support these carers—workshops, online programmes, guided self-help, and skills training—the field has lacked a rigorous synthesis of the randomized evidence specifically measuring effects on carers’ mood and anxiety.

To fill that gap, the researchers conducted a systematic review following the PRISMA 2020 reporting guidelines, searching major electronic databases for randomized controlled trials published up to 31 October 2025. Their inclusion criteria were strict: studies had to be randomized, had to evaluate a mental health intervention delivered to carers of individuals with eating disorders, and had to report anxiety or depressive symptoms as outcomes, typically measured with validated instruments such as the Hospital Anxiety and Depression Scale (HADS), the Depression Anxiety Stress Scale (DASS-21), and the General Health Questionnaire. Twelve trials passed the filter. Because the studies were too heterogeneous in design, population, and intervention type for a meaningful pooled statistical analysis, the team performed a structured narrative synthesis, and they assessed methodological quality using the Randomized Controlled Trial of Psychotherapy Quality Rating Scale. Reassuringly, all included trials met at least an average standard of quality, meaning the modest findings cannot easily be dismissed as the product of poorly conducted research.

A central thread running through the review is the Cognitive-Interpersonal Maintenance Model, or CIMM, the theoretical framework underlying most of the evaluated interventions. CIMM proposes that eating disorders are perpetuated not only by factors within the patient but by a self-reinforcing loop of interpersonal interaction: the illness generates carer distress and unhelpful coping responses—criticism, accommodation of symptoms, over-involvement—which in turn feed back into the patient’s symptoms. Interventions built on this model, such as the widely studied Collaborative Care Skills Training Workshops, aim to break this loop by teaching carers communication skills, emotional regulation, and strategies to reduce accommodation. The rationale is elegant: reduce carer burden and the entire caregiving system improves, benefiting both carer and patient. CIMM-based programmes were indeed the most frequently evaluated approaches across the twelve trials.

But the results tell a more complicated story. Among the CIMM-derived interventions, only one trial demonstrated a statistically significant improvement in depressive symptoms, with a p-value of 0.010. The same trial found no significant effect on total DASS-21 scores (p = 0.06) and no significant effect on the anxiety subscale (p = 0.50). In other words, even where a benefit appeared, it was confined to a single outcome domain and left broader measures of psychological distress essentially unchanged. This pattern—benefits that flicker into statistical significance on one scale while remaining invisible on others—recurred throughout the review and helps explain the authors’ cautious conclusions.

Cognitive Behavioural Therapy-based interventions offered a similarly mixed picture. CBT approaches, which target maladaptive thoughts and behaviours related to the caregiving role, showed some promising effects on the HADS total score, with one trial reaching statistical significance at p = 0.033. Yet this benefit was not consistently replicated when the same CBT content was delivered in guided versus unguided formats, raising questions about what actually drives the therapeutic effect: the content of the intervention, the presence of a human guide, or simply the act of engaging with structured material. The distinction matters practically, because self-directed digital programmes are far cheaper and more scalable than professionally guided ones, and health services increasingly favour them precisely for those reasons.

The review also highlighted a video-based skills-training intervention that produced a measurable reduction in carer distress—but only when the video material was combined with professional support, achieving significance at p = 0.030. Taken together, the findings point to a consistent signal: interventions incorporating professional or peer support appeared more promising than fully self-directed approaches. Programmes such as Experienced Carers Helping Others (ECHO), which draws on trained peer supporters who have themselves cared for someone with an eating disorder, exemplify this model. The authors note, however, that direct head-to-head comparisons between guided and unguided formats did not consistently show statistically significant differences, so the evidence for guidance as an active ingredient remains suggestive rather than definitive.

What, then, explains the gap between the intuitive appeal of these interventions and their weak measurable effects on carers’ mood? Several factors deserve consideration. First, carers seeking help may present with distress that is reactive to an ongoing, unresolved crisis—a child mid-restriction, a partner relapsing—such that informational and skills-based input, however valuable, cannot fully lift mood while the underlying illness continues. Second, many trials were likely powered primarily to detect changes in carer skills, accommodation, or patient outcomes, with anxiety and depression as secondary measures, increasing the risk of Type II error. Third, instruments like the DASS-21 and HADS capture symptom dimensions that may be less responsive to brief psychoeducational programmes than constructs like self-efficacy or expressed emotion, which these interventions were explicitly designed to improve. The review’s findings thus do not invalidate the CIMM framework; rather, they suggest that reducing carer burden and treating carer mood may require distinct therapeutic levers.

The authors are candid about these limitations and about the direction the field should now take. They call for future research to prioritize interventions that combine more structured mental health strategies—formal techniques for managing anxiety and low mood—with guided self-help programmes and workshops that provide tailored support responsive to the varying needs of individual carers. The implicit argument is that helping carers cope with the caregiving role is necessary but insufficient; what may be needed is direct, personalised attention to the carer’s own psychiatric symptoms, delivered with the same clinical seriousness afforded to patients. This reframing would represent a meaningful shift: from viewing carers primarily as an untrained extension of the treatment team to recognizing them as individuals whose mental health warrants intervention in its own right.

For the millions of families worldwide navigating the exhausting terrain of an eating disorder, the review offers a realistic rather than despairing message. Support programmes do deliver genuine value—better understanding of the illness, practical strategies for mealtimes and communication, and a sense of not being alone—and those that include a human guide or peer connection appear to work best. But carers and clinicians alike should temper expectations: on the specific question of whether these programmes reduce anxiety and depression, the randomized evidence remains thin, with benefits that are small, selective, and most visible when professional support is woven in. As eating disorder services worldwide increasingly adopt family-inclusive models of care, this review serves as both a validation of the impulse to support carers and a rigorous reminder that good intentions must be matched by interventions engineered to measurably protect the minds of those doing the caring.