OBJECTIVES:

The European Medicine Agency (EMA) has made more than 2,000 decisions regarding orphan designation. For this designation, sponsors must establish that prevalence criteria are met: not more than 5 in 10,000. Lacking methodological guidance, manufacturers use various methods for estimating prevalence, leading to discrepancies across applications. This study aimed to develop a comprehensive, valid epidemiological model that synthesizes all available evidence to estimate prevalence over time.

METHODS:

The epidemiological model was designed and implemented using a DICE (Discretely Integrated Condition Event) simulation model with 5 events and 26 conditions. It simulates a population over time starting with an initial prevalence, adding incident cases and removing deaths, until prevalence stabilizes. Multiple scenarios can be tested: altered initial prevalence, different incidence, mortality, etc. As an example, multiple data sources were obtained via comprehensive literature search regarding prevalence, incidence, and overall survival of ovarian cancer (OC) in the EU. The inputs were fed into the DICE model and prevalence over time was evaluated and compared with EMA submissions.

RESULTS:

A stable OC prevalence of 4.1/10,000 was estimated for the EU and compared with the varying rates submitted to EMA (2.3-4.9/10,000). Multiple scenarios were tested. For example, if incidence was increasing by 2% per year post-2002, prevalence would reach the 5/10,000 cut-off in 2012.

CONCLUSIONS:

The DICE epidemiological model can produce a robust prevalence estimate by synthesizing available data from different sources and allowing for multiple scenarios to be tested. This model can assist sponsors and the EMA in their review by introducing a transparent and flexible tool which can be used as a standard for orphan designation submissions across multiple diseases.