POS0523 FEASIBILITY STUDY OF AN ELECTRONIC HEALTH RECORD-BASED GOUT REGISTRY

Abstract

BackgroundGout is the most common inflammatory rheumatism worldwide. Despite guidelines on acute and chronic management, it remains largely undertreated[1]. To evaluate the current standard of gout care, there is an unmet need for gout registers, especially in the non-rheumatology setting. The vast clinical information available in electronic health record (EHR) allows the implementation of registers to assess clinical indicators and monitor them following quality improvement programs. The Geneva University Hospitals (HUG), a 2000-beds tertiary hospital, provides multi-specialty in- and outpatient services, including care to vulnerable population (inmate, uninsured) with a unified EHR and thus represent an ideal setting for this purpose.ObjectivesTo establish an EHR-based gout registry, to assess the correct identification of gout patients by manual chart review, and to evaluate the objective validity of gout diagnosis based on the ACR-EULAR 2015 gout classification criteria[2].MethodsEHR of patients > 18 years old admitted to or consulting at the HUG between 01.01.2013 and 15.11.2022 were screened based on the presence of at least one of four criteria: ICD-10-GM gout diagnosis (M10), gout-related terms (gout, tophi or podagra) in the list of diagnosis, prescription of urate-lowering therapy (allopurinol, probenecid, febuxostat) among non-leukemia or lymphoma-suffering patient (ICD-10-GM code C81 to C96 were excluded), or uric acid crystal in synovial fluid. For each criterion, a sample of 20 medical charts were fully reviewed. The positive predictive value (PPV) to detect a true gout was established if a gout diagnosis was mentioned by any doctor in any part of the EHR. To further assess the validity of the gout diagnosis, the 2015 ACR-EULAR gout criteria was scored. Gout patients without any acute arthritis during any encounter with the HUG were classified as having an antecedent gout.ResultsThe four criteria identified 7’046 patients suffering from gout, among which 33.4% were deceased. Most patients were identified by the drugs criterion (Figure 1). A large proportion of patients (43.2% of outpatients and 72,8% of inpatients) were identified by the presence of urate-lowering therapy only, without the mention of any gout diagnosis. Furthermore, 9.7% and 2.6% in the out- and inpatient setting respectively had a positive punction that wasn’t associated with a documented diagnosis. PPV were 100% for ICD-10-GM code, list of diagnosis and punction, and 75% for urate-lowering therapy, while PPV based on a combined query (any criterion) was 93.8% to detect a gout patient in the charts. Among the 80 charts reviewed, 55 patients had at least one documented gout attack and 25 had an antecedent gout. Of the 55 patients with an acute gout, which allowed the use of the ACR-EULAR 2015 classification, 39 patients (70.9%) had a positive articular punction for uric acid or the presence of tophus as observed by a trained examiner. The remaining 16 patients (29.1%) had a mean score (SD) of 5.5 (2.39) for the ACR-EULAR 2015 classification (threshold to classify as gout is ≥8). The lack of clinical information available (documented presence of tophus, precise clinical course of symptoms) and lack of diagnostic imaging to assess gout complication might explain the low score, among these patients with a high probability of gout according to the chart review.ConclusionWe demonstrate the feasibility of an EHR-based gout registry, with good positive predictive value and the ability to identify patients without a documented gout diagnosis in their EHR. The next step is to expend the queries to look for gout diagnosis in all of the patient’s documents and imaging reports and to estimate the negative predictive value in a sample of patients without any criteria, yet with risk factors of gout.