National Eating Disorders Awareness Week 2026: Build a care route that recognizes eating disorders in every body and connects people to appropriate help without stigma

Recognition system

Replace stereotype-driven recognition with governed observation and assessment.

Recognition errors do not arise from one misconception. They can be produced by campaign imagery, intake questions, eligibility rules, clinical templates, referral scripts, coverage policies, and informal judgments about who “looks sick.” A weight-neutral route does not ignore medical status. It prevents body size from becoming a proxy for whether symptoms, behaviors, impairment, or medical risk deserve attention.

Experimental vignette research with 180 community participants found that judgments about whether a problem was recognized, how distressing it appeared, and whether treatment should be recommended varied with the depicted person’s weight. The exact direction differed across presentations, which is another reason to avoid replacing one stereotype with another. The study’s central message is that weight cues can change recognition and referral even when the clinical description is held constant.[8] In a separate 2025 study of 42 non-medical mental health professionals, only 28.6% correctly identified the disorder in an assigned vignette; misconceptions about gender, duration, treatment response, and personal blame were also present. The small, single-country sample limits generalization, but the findings show why non-specialist education must include varied diagnoses and presentations rather than a single image of anorexia nervosa.[6]

Healthcare leaders can make recognition safer by defining a short set of observable triggers for further evaluation. These may include changes in eating behavior, compensatory behavior, rapid or concerning health changes, gastrointestinal or endocrine concerns without an adequate explanation, dizziness or fainting, exercise that has become rigid or unsafe, distress about eating or body image, recurrent binge episodes, avoidance based on sensory features or fear of adverse consequences, caregiver concern, and clinician concern. The specific list, wording, and escalation thresholds must be approved by qualified clinical leaders and adapted to the population and setting.

Screening can be part of that design, but it should be treated as a governed decision aid. Leaders should specify where a tool is used, who is eligible, how language and accessibility needs are handled, how results are documented, what positive and negative results mean, and which concerns override the score. They should also monitor whether the tool performs differently across age, language, gender, race and ethnicity, disability, and clinical presentation. A bilingual adolescent study in Italy found similar overall screening prevalence between language versions of the SCOFF while individual item responses differed, reinforcing the need for cultural and language validation rather than assuming that translation alone creates equivalence.[9]

Training should be paired with an action map. Awareness without a next step can leave staff more alert but no more capable of arranging care. A useful curriculum covers respectful inquiry, limits of visual judgment, limits of screening, basic presentation diversity, urgent medical and mental health signals, documentation language, privacy, local referral requirements, and escalation when a receiving service cannot accept the person. Competency should be demonstrated with scenarios and observed practice, not attendance alone.

Equity and access

Design for people who are most likely to be misrecognized, mistrusted, or lost.

Equity cannot be achieved by placing the same referral instructions in front of everyone. People encounter different recognition patterns, financial barriers, prior experiences of discrimination, transportation constraints, caregiver expectations, language needs, digital access, and clinical eligibility rules. The pathway should therefore be universal at the front door and responsive in the support it provides.

Large population data show why an intersectional lens is necessary. An analysis of 251,310 U.S. university students found substantial variation in eating-disorder risk factors, symptoms, and self-reported diagnosis across 35 combinations of gender and racial or ethnic identity. Predicted diagnosis prevalence ranged from 0.3% to 18.3% across strata, and transgender participants of color had higher predicted prevalence than expected from additive gender and race or ethnicity effects alone. Because the study was cross-sectional and college-based, the estimates should not be generalized to all populations. The operational lesson is that an average rate can conceal groups with different needs and different likelihood of recognition.[11]

Trust is also an access variable. In a study of 322 undergraduates, LGBTQ+ participants reported less trust in physicians and more body dissatisfaction and shape or weight overvaluation. Across the full sample, healthcare discrimination was associated with higher body dissatisfaction and shape or weight overvaluation; lower physician trust was also associated with greater overvaluation. The study does not show that discrimination caused eating-disorder symptoms, but it supports a direct leadership question: does the organization measure whether people feel dismissed, stereotyped, or unsafe when they raise concerns?[9]

Access design should include cost and coverage navigation. In a 2026 Australian survey, people with lived experience and carers described delays between symptom onset and treatment, high costs, insufficient treatment sessions, and difficulty with progress reviews. Care received through dedicated eating-disorder benefits was associated with easier referral, better treatment uptake after referral, and fewer access difficulties than other routes. Differences in national financing systems limit direct transfer to the United States, but the study demonstrates the value of asking users where a benefit design solves one barrier and creates another.[2]

Navigation should be practical. The person needs to know whom to contact, what information will be shared, what the receiving program has agreed to do, what to expect while waiting, how to raise a worsening concern, and who can help with authorization, transportation, technology, childcare, interpreter services, or caregiver coordination. When possible, the navigator should verify fit before asking the person to repeat a history. When a referral is declined, the route should capture the reason in a structured way, identify the next owner, and prevent the person from restarting at the beginning.

Virtual care can reduce travel and geography barriers, but it is not automatically equitable or clinically appropriate. A Canadian mixed-methods study of 18 professionals in pediatric programs found virtual multidisciplinary care feasible and generally acceptable, while emphasizing demand, resource limits, uncertainty, and the need for training and hybrid design.[14] A 2021 pilot comparison of 93 people in intensive outpatient treatment found no outcome differences between telehealth and in-person delivery in that program, but the nonrandomized, pandemic-era comparison was small. Leaders should treat virtual care as one governed modality, evaluate who can use it safely, and retain in-person capability where required.[15]

Measurement

Measure the connection, not just the campaign.

A reliable dashboard should answer whether the route is used, whether qualified assessment occurs on time, whether referrals fit and are accepted, whether people reach care, whether urgent needs remain owned, and whether the experience is respectful. Counting page views, training attendance, screening volume, or referrals placed may be useful process information, but none demonstrates completed access.

Definitions matter. “Referral complete” should not mean that an order was entered. A stronger definition could require confirmed acceptance by an appropriate program, a scheduled appointment or documented next action, an interim owner, and communication that the person can understand. Organizations may choose a different definition, but it should be explicit and auditable. Referral declines should be coded by reason, such as clinical mismatch, age, payer, capacity, geography, technology, missing documentation, or person preference. These codes reveal which problems require capacity, contracting, workflow, education, or partnership changes.

Measurement also needs a denominator. Screening rates should use the population for whom the organization has approved the tool, not every encounter. Time-to-assessment should be measured from a defined starting event, such as first documented concern or accepted request. Referral acceptance should use referrals sent to a service that appeared to meet the documented need. Completed connection should distinguish attended appointments, clinically appropriate alternate plans, declined care, inability to contact, and unresolved barriers.

Stratification should be clinically and ethically governed. Leaders may examine results by site, entry setting, age group, payer, language, race and ethnicity, gender identity, rurality, disability access need, or other locally relevant factors. The purpose is to identify system differences, not to label people or create body-size targets. Small-cell suppression, access controls, consent, and privacy requirements should be built into reporting.

Patient-experience questions should be short and specific. Instead of asking only whether a person was satisfied, ask whether the team took the concern seriously, avoided assumptions about appearance or identity, explained the next step, helped resolve a practical barrier, and provided a contact for worsening concerns. Combine survey responses with interviews, complaints, compliments, focus groups, navigation notes, and lived-experience review. Quantitative and qualitative evidence answer different questions.

Leaders should review the dashboard in a recurring forum that can act. A monthly operating review may be appropriate during the pilot, with a quarterly governance review once the route stabilizes. The forum should include clinical, operational, access, equity, patient-experience, and data owners. Each exception should lead to a named action, due date, and verification method. Improvement is not complete when a policy is edited; it is complete when the route performs better for the people using it.