Modeling Primary Immunodeficiency Disease Epidemiology and Its Treatment to Estimate Latent Therapeutic Demand for Immunoglobulin

Abstract

Estimating the underlying demand for immunoglobulin (Ig) is important to ensure that adequate provision is made for patients with primary immune deficiency (PID) in the context of the competing demands for Ig and to ensure optimal therapeutic regimens. The concept of latent therapeutic demand (LTD) was used to estimate evidence-based requirements and compared to the actual Ig consumption in different countries. The estimates were performed for common variable immunodeficiency (CVID) and X-linked Agammaglobulinaemia (XLA), the two most commonly studied PIDs using Ig. The LTD model for CVID and XLA was derived using decision analysis methodology. Data for the epidemiology and treatment variables were obtained from peer-reviewed publications, clinical registries and publicly-available patient surveys. Incomplete data records from registries were excluded from analysis. The variables impacting LTD were ranked in order of sensitivity through a tornado diagram. The uncertainty surrounding the variables was modeled using probabilistic distributions and evaluated using Monte Carlo simulation. Treatment dosage and prevalence were determined to be the most sensitive variables driving demand. The average potential usage of Ig for the treatment of CVID and XLA was estimated at 72 g per 1,000 population, which is higher than the estimated Ig usage in CVID and XLA of 27-41 g per 1,000 population in the US. The potential demand for treating CVID and XLA exceeds the currently observed usage of Ig in these disorders. Variable usage in different countries is due to varying prevalence and dosage practices. Under-reporting in patient registries represents a major obstacle to calculating the true prevalence of CVID and XLA. Modeling demand relies heavily upon accurate prevalence and practice estimates which reemphasize the importance of accurate registries and improved registry methods. As better data becomes available, revision of model variables provides opportunities to anticipate and prepare for evolving patient needs.