OBJECTIVES: A budget impact analysis (BIA) is an important requirement for health technology assessment and reimbursement submissions. However, the methodological literature for BIA is not as developed as it is for other economic evaluation topics. This study objective is to compare a complex patient population budget impact model (BIM) with a typical simplified approach for a new oncology product.

METHODS: Two models were developed, a typical simplified BIA with an annual update of the patient cohort, and a complex BIM with a monthly update. The simplified BIM uses mean estimate of treatment duration. The complex BIM introduces new patients into the model at the start of each month, with the proportion of the cohort remaining on treatment after each time interval determined by the time to treatment discontinuation or death survival function. The total number of patients on treatment in any month is the sum of all patients who remain on treatment from all previous cohorts up to the current month. In both models, total costs are calculated as the product of treatment duration and treatment cost. Common model inputs are used in both models to provide a fair comparison.

RESULTS: The difference in budget impact between the two models was significant, with the largest impact observed in the initial years (Year ≤3: 63-173% and Year >3: 23–42%). This resulted in a difference in cumulative budget impact of 64% over 5 years, driven by the equilibrium in patient numbers initiating and discontinuing treatment in the complex model.

CONCLUSIONS: As with any quantitative analysis, when conducting a BIA all methodological assumptions should be considered. A simple approach, not utilising the survival function directly, may not estimate the budget impact accurately, and a complex approach provides important additional accuracy by ensuring a more accurate timing of transition at the point of discontinuation.