For the millions of adults living with unrecognized eating disorders, the window between first symptoms and formal diagnosis represents a critical — and often missed — intervention opportunity. A large matched cohort study now quantifies just how frequently these individuals are already in contact with healthcare, raising urgent questions about whether primary care is adequately equipped to identify disordered eating before it becomes entrenched.
Drawing on linked electronic health records from England's Clinical Practice Research Datalink and Hospital Episode Statistics, researchers tracked over 46,000 adults aged 18–65 with a recorded eating disorder diagnosis or specialist referral, matched three-to-one against controls without eating disorders. In the 24 months preceding diagnosis or referral, individuals with eating disorders visited primary care at nearly double the rate of matched controls, with an incidence rate ratio of 1.96. The elevated contact rate held across all diagnostic subgroups, though it varied meaningfully — from an IRR of 1.78 for anorexia nervosa to 2.45 for other specified presentations. Anorexia nervosa accounted for 43% of cases, bulimia nervosa for 13.8%. The cohort was 79.5% female and predominantly White (83.7%), with Asian and Black individuals comprising smaller proportions, and secondary analyses examined whether ethnicity influenced likelihood of specialist referral.
These findings add quantitative weight to a long-suspected clinical pattern: people with eating disorders are not avoiding healthcare — they are attending it repeatedly, likely presenting with downstream physical symptoms rather than the underlying condition itself. This signals a systemic screening gap, not an access gap. Prior research has documented diagnostic delays of several years for both anorexia and bulimia nervosa, and this study reinforces that those delays occur despite frequent healthcare contact. The ethnic composition findings are particularly worth scrutiny, as underrepresentation of non-White groups may reflect both true prevalence differences and well-documented diagnostic bias. The study's observational, retrospective design limits causal inference, and the CPRD population skews toward England, limiting global generalizability. Still, the scale — nearly 50,000 cases over 13 years — makes this one of the more robust epidemiological characterizations of eating disorder healthcare utilization to date. The incremental but clinically actionable implication: structured eating disorder screening in primary care could intercept a high-contact, high-risk population well before specialist referral.