# Accuracy of Budget Impact Projections in Bulgarian Health Technology Assessment: A Five-Year Validation Study (2020–2025)

## TL;DR

Bulgarian health technology assessments underestimated medicine usage, especially for cancer and immunology drugs, leading to budget overruns that require better data and flexible pricing strategies.

## Contribution

This study empirically validates HTA budget projections in Bulgaria over five years, revealing systematic underestimation of patient uptake and expenditure deviations driven by volume, not price.

## Key findings

- Projected patient volumes for 2019-approved medicines were exceeded by a median factor of 1.6 by 2025, especially in oncology and immunology.
- Expenditure deviations were primarily due to higher-than-expected patient uptake, with immunology and oncology contributing €22.4 million in budget overshoots.
- Cost elasticity was near-proportional overall, but varied significantly between hospital and outpatient settings.

## Abstract

What are the main findings?

Health Technology Assessment (HTA) budget projections in Bulgaria systematically underestimated patient uptake for medicines approved in 2019, with realized volumes exceeding forecasts by a median factor of 1.6 by Year 5, driven primarily by oncology and immunology products (Bavencio, Skyrizi, Tremfya).

Expenditure deviations were predominantly volume-driven rather than price-driven, with immunology contributing +€17.4 M and oncology +€5.0 M in budget overshoots, while overall cost elasticity remained near-proportional (β = 1.002), masking substantial product-specific heterogeneity.

What are the implications of the main findings?

Bulgaria’s HTA system requires strengthened epidemiological infrastructure—including national disease registries and real-world data integration—to improve forecasting accuracy and enable dynamic, indication-responsive budget impact models that adapt to evolving clinical use.

Policymakers should implement adaptive contracting mechanisms (e.g., managed-entry agreements, outcome-based pricing) and continuous post-launch surveillance to better align budget planning with actual utilization patterns, ensuring fiscal sustainability without compromising patient access to innovative therapies.

Background: Budget Impact Analysis is an integral part of the Health Technology Assessment in Bulgaria, informing reimbursement decisions of the National Health Insurance Fund. Inaccurate projections risk both fiscal unsustainability and restricted patient access to innovation. Yet projection accuracy methods remains uncertain, particularly given limited epidemiologic data and evolving clinical use. Objectives: This study aimed to assess the empirical validity of Health Technology Assessment budget-impact projections for medicines approved in 2019 by comparing projected patient volumes and expenditures with real-world National Health Insurance Fund reimbursements through 2025, and to identify drivers of divergence across therapeutic areas and reimbursement channels. Methods: We conducted a retrospective cohort analysis linking 2019 Health Technology Assessment submissions with monthly National Health Insurance Fund claims for both hospital and outpatient reimbursement channels. Actual utilization was calculated as the annualized median number of treated patients per month. Projected costs were derived by multiplying HTA-projected patient volumes by the observed unit cost per patient-month. We quantified deviations using observed-to-projected ratios and absolute gaps and assessed the relationship between projected and actual expenditures using a log–log regression model. Results: By September 2025, realized volumes typically exceeded projections (median ratio 1.6; range 0.02–21.3). Large overshoots were observed for Avelumab, Risankizumab, and Guselkumab; Cobimetinib and Abemaciclib remained below forecast. Expenditure deviations were driven predominantly by volume: immunology (+€17.4 million) and oncology (+€5.0 million) accounted for the largest absolute gaps. Elasticity was near proportional overall (β = 1.002; standard error = 0.24; R2 = 0.50), lower in hospitals (β = 0.79; p = 0.055) and higher in outpatient care (β = 1.30; p = 0.003). Conclusions: Health Technology Assessment Budget Impact Analyses captured broad cost scaling but systematically missed product-specific uptake, with deviations largely volume-driven. Strengthening national registries and real-world data pipelines, and adopting dynamic, indication-responsive contracting and forecasting, could materially improve budget predictability while preserving access to innovation.