OBJECTIVES : Payers require a thorough understanding of patient population size to make evidence-based decisions regarding treatment coverage. For rare diseases, this information is not always readily available; often epidemiologic estimates are reported as a wide range of values or may be derived from populations of limited relevance to a payer’s target population. Epidemiological models can combine available data with necessary assumptions to customize estimates of population size. The objective is to describe an interactive model to estimate the number of Duchenne muscular dystrophy (DMD) patients in the United States (US) relevant to different stakeholders.

METHODS : A model was developed to estimate DMD prevalence by age, diagnosis, and ambulatory status. Two approaches for birth incidence input were explored: (1) a systematic review to identify published estimates and (2) numeric calibration using sum-of-squared errors to align model-predicted estimates with published age-specific prevalence data. Published data describing DMD diagnosis, mortality, and the US male population by age were identified. Diagnosis and mortality data were combined via state-transition modeling to estimate age-specific prevalence based on birth incidence from both approaches. A customized interface was developed to generate results for a given plan size and demographic distribution, and sensitivity analyses were conducted across ranges of uncertain values.

RESULTS : Patient population sizes estimated using the incidence back-calculation approach were well-aligned with published registry data and birth incidence values. The model matched reported trends in prevalence by age, reflecting typical patterns of diagnosis and mortality, and extended these results to plan-specific projections.

CONCLUSIONS : Customized interactive epidemiological models are valuable tools to synthesize available information for rare diseases and present results needed by payers. Model outputs can be used in budget impact analyses and other economic assessments. While this example is specific to DMD, this methodology could be applied for developing evidence-based prevalence tools for other rare diseases.