US-Based Drug Cost Parameter Estimation for Economic Evaluations.

Area of Science:

• Health Economics
• Pharmacoeconomics
• Drug Pricing Analysis

Background:

• Prescription drugs constitute over 10% of national health expenditures in the US.
• Inconsistent assessment of drug costs in economic evaluations stems from unknown true acquisition costs.
• Current practices rely on single drug cost estimates and assumed distributions for uncertainty analysis.

Purpose of the Study:

• To propose a novel set of Rules based on pharmacy practice for calculating heterogeneous drug product costs.
• To develop an empirical distribution of drug costs for more realistic sensitivity analyses and transparent parameter computation.
• To demonstrate the application of these Rules using diverse drug examples and compare with existing cost estimates.

Main Methods:

• Developed Rules based on clinical equivalence, pill burden reduction, appropriate package size, and uniform weighting of substitutable products.
• Generated empirical cost distributions for three distinct drugs using the proposed algorithmic process.
• Compared derived empirical distributions and their ranges with previously reported cost estimates.

Main Results:

• Empirical distributions exhibited significant heterogeneity in shape, including multiple modes and varied ranges.
• Previously published cost estimates and sensitivity analysis ranges did not accurately reflect the derived empirical distributions.
• For lisinopril, the empirical mean cost was $444 (range: $23-$953), differing from a published estimate of $305 (range: $51-$523).

Conclusions:

• The proposed Rules offer a simple, transparent method for estimating drug product costs and their variability.
• This approach facilitates more accurate sensitivity analyses in economic evaluations.
• The empirical distribution methodology is adaptable and easily integrated into standard sensitivity analysis techniques.