OBJECTIVES

Budget impact analyses (BIA) help to quantify the impact of changes in a given treatment mix on health care payers' budgets due to the introduction of a new intervention or targeted increase of the market share of a specific product. In BIA analyses comparing only two interventions, uncertainty arises from the future market share of a new intervention. However, in multiple technology appraisals, there is a further source of uncertainty: which of the existing interventions will be replaced and in what order? Following international recommendations, this type of uncertainty is usually approached deterministically, giving only a limited impression of the degree of uncertainty. We developed a framework that quantifies both the uncertainty related to the future market share and substitution patterns.

METHODS

: We propose a parameter space whose boundaries are defined by the gradual substitution from most to least expensive, and least to most expensive comparative therapy. However, within those boundaries not all treatment options are equally likely to be replaced. We conducted a hypothetical budget impact analysis, including ten therapies likely to be replaced by a new intervention. Upper and lower boundaries of the possible budget impact over different market shares were calculated as described above. Within a Monte-Carlo simulation, estimates of the expected market share were drawn from a beta distribution. Substitution patterns were informed by a Dirichlet distribution which was adjusted after each iteration.

RESULTS

: Within the defined parameter space of possible budget impact results, the combination of the most likely market share and resulting budget impact can be mapped, taking into account decision-makers’ preferences for substitution. The Monte Carlo analysis allows us to quantify the range of most likely outcomes.

CONCLUSIONS

: Possible applications of our model should be discussed with industry and whether it might be a useful supplement for national and international methodological guidelines for conducting BIA.