BACKGROUND: Multiple data sources may be required to obtain necessary sample sizes for studies of rare diseases and treatments. While Medicaid, Medicare, and commercial insurance claims data sources are all available in the US, Medicaid programs for patients with low incomes have differences by state in underlying patient case mix, eligibility criteria, and healthcare coverage.
OBJECTIVE: To describe the differences in patient characteristics and medication-usage patterns between patients with chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), a rare neurological disorder affecting nerve signal transmission, as identified in Medicaid versus Medicare or commercial insurance.
METHODS: Patients with diagnoses of CIDP were identified in Medicaid, Medicare supplementary, and commercial insurance data (2008-2018). Those subsequently initiating treatment with intravenous immunoglobulin (IVIg) were identified through procedure and pharmacy dispensing coding. Demographic and clinical characteristics of patients and patterns of use of IVIg were compared between patients identified in Medicaid and those in other insurance.
RESULTS: Out of 32,090 patients with CIDP diagnoses, 12% were identified in Medicaid, and 88% in Medicare or commercial claims. Medicaid patients were younger (mean age 51.7 vs. 57.8) and more likely to be female (59% vs. 47%). Medicaid patients were more likely to have any of several comorbidities, including asthma, arrythmias, chronic obstructive pulmonary disease, heart failure, diabetes mellitus, hepatitis, hypertension, and renal disorders, but were less likely to have electrodiagnostic/nerve conduction testing. Of those with CIDP, 10% of the Medicaid sample and 13% of the others initiated IVIg treatment. 23% of IVIg initiation was identified in pharmacy claims in Medicaid, while only 10% IVIg initiation in other insurance was. The duration of IVIg use was shorter among those with Medicaid (188 vs. 216 days).
CONCLUSIONS: The characteristics and treatment patterns of patients with CIDP identified in Medicaid differed from those identified in other data sources. Medicaid may be a valuable source of information on an important population, but the suitability of its use and availability of critical study variables should be carefully considered in context to the research question before inclusion in multi-database studies of rare diseases. Assessment of differences in demographics, comorbidities, and treatment patterns could aid understanding of whether Medicaid patients would be best suited for studies including long-term outcomes.
