French Insurance Records Cast Doubt on Many Wilson’s Disease Diagnoses

More than 1,500 patients carrying a diagnostic code for Wilson’s disease in France’s national health insurance system turned out to be a strikingly heterogeneous group, and a new analysis suggests that a large share of them may never have had the condition at all. The study, based on the French National Health Data System (SNDS), examined how reliably different administrative markers—hospitalization records, long-term disease status, and dispensing of specific treatment—identify genuine cases of this rare genetic disorder of copper metabolism. Its conclusion is a caution for anyone hoping to use claims databases to study rare diseases: for many patients identified only through a hospitalization claim, no specific Wilson’s disease treatment or copper monitoring was ever documented, leaving substantial doubt about whether they actually had the disease.

The research team set out to test how well different potential markers perform for case identification, comparing four mutually exclusive patient groups defined by combinations of these markers.

The researchers identified patients in the SNDS with hospitalization or long-term disease (LTD) status associated with a diagnostic code for Wilson's disease (E83.0) between 2009 and 2019. Each patient was followed from first documentation until December 31, 2019, or death. Group A comprised patients hospitalized for Wilson’s disease but with neither LTD status nor specific treatment (702 patients). Group B held patients with LTD status but no treatment (153). Group C had treatment but no LTD status (194), and Group D had both LTD status and treatment (471). In total, 1,520 patients were included in the analysis.

The most revealing metric was copper monitoring. Because Wilson's disease is a disorder of copper accumulation, the researchers used blood copper testing as an indicator of monitoring for the disease. Yet more than 75 percent of untreated patients in Groups A and B were never monitored for copper accumulation at all. The mean number of blood copper tests over the follow-up period was 0.30 ± 2.47 in Group A and 0.76 ± 2.34 in Group B, compared with 3.99 ± 7.66 in Group C and 5.33 ± 7.90 in Group D. The pattern is stark: patients flagged only by a hospitalization code show almost no laboratory footprint of a disease whose management depends on tracking copper.

Mortality data added a second layer of doubt. Patients in Group A—those hospitalized with the diagnostic code but lacking both LTD registration and treatment—frequently carried LTD status for conditions unrelated to Wilson’s disease, and they died at a higher rate than the treated group, with a hazard ratio of 2.87 (95% CI: 1.87–4.41) compared to Group D. Notably, they frequently died of causes unrelated to Wilson’s disease. Patients in Group A frequently died of causes unrelated to Wilson's disease, and the study reports that they frequently carried LTD status for conditions unrelated to the disease.

The authors state that validation studies are needed to find the optimal diagnostic algorithm to identify Wilson's disease cases unambiguously, which they describe as a prerequisite for collecting reliable information on the disease. Group D, with both LTD status and treatment, showed the highest mean number of copper tests and represents the most plausible core of true cases. Group C, treated but without LTD registration, also shows substantial monitoring activity. By contrast, Groups A and B—defined by a single administrative marker without treatment—showed little copper monitoring, and the authors conclude that substantial doubt exists whether these patients actually had Wilson's disease.

The study’s design has limitations worth noting. It relies entirely on administrative data, so the researchers could not review clinical charts, genetic test results, or liver biopsies to confirm diagnoses directly. The authors themselves state that validation studies are needed to find the optimal diagnostic algorithm to identify Wilson’s disease cases unambiguously, which they describe as a prerequisite for collecting reliable information on the disease. In other words, the four-group comparison is a diagnostic exercise, not a definitive census of who has Wilson’s disease in France; it is a demonstration that current coding practices produce a mixed population whose composition varies dramatically depending on which administrative marker is used.

Subject of Research: Case identification of Wilson's disease patients in the French national insurance claims database

Article Title: A population-based study of patients with Wilson’s disease identified in the French national insurance claims database, with a proposal for an exploratory algorithm for improved case identification

Article References: Leboucher, C., Daniel, T., Benichou, B., Barbier, S., & Combal, J.-P. (2026). A population-based study of patients with Wilson’s disease identified in the French national insurance claims database, with a proposal for an exploratory algorithm for improved case identification. Frontiers in Gastroenterology, 5, Article 1901046. https://doi.org/10.3389/fgstr.2026.1901046

Image Credits: AI Generated

DOI: 10.3389/fgstr.2026.1901046

Keywords: Wilson's disease, SNDS, insurance claims database, case identification, copper monitoring, long-term disease status, hospitalization records, diagnostic coding, rare disease epidemiology, France, mortality, validation studies