Proposal for a General Outcome-based Value Attribution Framework for Combination Therapies

An increasingly common strategy for treating many cancers is to utilise several medicines with distinct but complementary mechanisms of action in combination or in close sequence (IQVIA, 2019; Bashraheel et al, 2020). Valuing and pricing the components of a combination therapy gives rise to several challenges for companies, payers and HTA bodies particularly when two or more of the products in combination are owned by different companies, each of which seeking a value-based price for their product (Davis, 2014; Greber et al. 2014; Person et al. 2018; Danko et al. 2019). These challenges have been discussed in a number of places (Latimer et al. 2019, Latimer et al. 2020; Latimer et al 2021a; Latimer et al 2021b; Briggs et al. 2020, Briggs et al., 2021) and include (1) competition law issues, which prevent companies from discussing prices with each other, (2) the challenges of implementing different prices for the same product in different uses (for example as a monotherapy and in a particular combination use), and (3) how to attribute value between products in a way that incentivises appropriate innovation, i.e. the development of treatments that can address patient need at prices that represent value for money for health care systems. A consequence of those challenges remaining unaddressed is that effective combination treatments may not be reimbursed, with patients not receiving the care most appropriate for their disease. We explored in an earlier paper (Towse et al. 2022) the challenges of value attribution to conventional HTA approaches. Current payer and HTA-body approaches either assume the price of the backbone therapy is unchanged or require some arbitrary reduction in the price of products being used in combination. In the situation of an unchanged price, we considered the typical situation where the combination leads to an increase in treatment duration for the backbone therapy used as part of the combination. In this case it is only possible for any add-on therapy to be cost-effective, even at zero price, in very restricted assumptions. In such circumstances it is very unlikely that the add-on therapy will be cost-effective at a price that reflects its contribution to the value of the combination. This paper seeks to contribute to the emerging debate on this policy challenge by articulating an approach to solving the value attribution problem. The solution we propose assumes the price of the backbone therapy will differ in combination use, as do the two alternative proposals from the literature that we compare our approach with. We also set out in our earlier paper (Towse et al. 2022) criteria that a value attribution solution should meet. The solution should be: Universal, with the solution allowing for value attribution for most possible combination therapy configurations; Logical and symmetrical with the solution being neutral to each combination constituent, regardless of which is the backbone or the add-on; and Complete, by which we mean that the solution will always produce an attribution of the full value of the combination between the component parts. We reiterate our view that the order of backbone and add-on sequence should not, in principle, impact the value attribution between the combination constituents. No product is “first” in the combination. The combination is only created at the point where all the component parts are present. It therefore follows that the value of A in (A+B) must be the same as the value of A in (B+A) when the clinical regimen (A+B) is identical to (B+A) in terms of the value it delivers. However, the proposal we put forward uses weights which could be adjusted to give some preference to a “first” product if that was the preference of the payer or of companies negotiating a combination price attribution. The paper is structured as follows. We set out our assumptions and then our proposed solution. We then compare this approach with two others proposed in the literature. Finally, we discuss issues in applying our approach, notably how the challenge of partial information could be overcome, other implementation issues, and whether some of our assumptions could be relaxed.