Comparison of Patient Identification Methodologies in Chronic Inflammatory Demyelinating Polyneuropathy Using United States Administrative Claims Data

Speaker(s)

Blein C1, Arvin-Berod C2, Gelinas D3, Barrera-Sierra SBS3, Ulla H4, Ward C5, Sato M6, Goyal A5
1argenx, Zwijnaarde, Belgium, 2argenx, Ghent, East Flanders, Belgium, 3Argenx, US, Boston, MA, USA, 4ZS Associates, Haryana, India, 5ZS Associates, New york, NY, USA, 6ZS Associates, New York, NY, USA

Presentation Documents

OBJECTIVES: Complexities of diagnosing chronic inflammatory demyelinating polyneuropathy (CIDP) experienced in clinical practice is also reflected in real-world data. Relying solely on diagnostic codes for patient selection may lead to misrepresentative cohort selection. To explore patient identification methods for CIDP in claims data to better align with diagnostic processes in clinical practice.

METHODS: In step 1, published algorithms were evaluated. In step 2, key CIDP diagnostic evaluations were identified using electronic health records. In step 3, clinical expert consensus was reached to evaluate a nerve conduction test requirement. In step 4, the hypothesis was tested using a US claims database (Komodo Health closed claims, Jan 2016 – Dec 2020). In cohort 1, patients with CIDP were selected using standard methods: ≥2 claims with CIDP diagnosis, ≥30-≤365 days apart (first as index date). In cohort 2, a requirement for ≥1 nerve conduction test was added, which needed to be present either after the index date and before another CIDP diagnosis, or ≤90 days before the index date. All patients had continuous enrollment ±1 year pre- and post-index date. Cohort sizes, baseline characteristics, and CIDP treatments during year 1 post-index were evaluated.

RESULTS: Overall, 6303 patients were identified for cohort 1. A subset of 3409 patients (54%) fulfilled criteria for cohort 2. Mean age, gender, and race/ethnicity proportions were similar across cohorts. While 26% of patients were untreated for CIDP according to cohort 1, a considerably lower proportion (19%) of patients with CIDP were observed to be untreated in year 1 in cohort 2.

CONCLUSIONS:

Our results suggest that in CIDP, identifying patients with a higher certainty of diagnosis may not only impact how prevalence is estimated, but also patient burden assessments. In future studies, inclusion of disease-specific clinical diagnostic procedures may be critical to identify increasingly relevant study cohorts.

Code

EPH8

Topic

Epidemiology & Public Health, Study Approaches

Topic Subcategory

Disease Classification & Coding, Public Health

Disease

Neurological Disorders, Rare & Orphan Diseases