Value of information (VoI) methods have been developed with the aim of efficiently informing research decisions, but have remained largely separate from reimbursement, pricing and implementation processes. In this session we bring together recent research that provides links to harmonize VoI methods with these previous separate areas to allow a unified framework to inform optimal decisions in practice.

In the first paper we take Occams’ Razor to VoI methods to consider their usefulness at different level of sophistication in informing decisions of optimal research design and prioritization, namely:

(1) Is further research for a specific health technology assessment (HTA) potentially worthwhile?
(2) Is a given research design for a specific HTA worthwhile?
(3) What is the optimal research design for a specific HTA?
(4) How can research funding be best prioritized across alternative HTAs?

This establishes that despite its simplicity the expected value of perfect information (EVPI) alone does not allow any of questions (1) to (4) to be addressed. Further, EVPI has no necessary relationship to the expected value of sample information (EVSI), expected cost or expected net gain (ENG) of research and can hence be actively misleading if used in isolation to inform (1) to (4). However, parametric methods are shown to enable simple calculation of EVSI and ENG for potential trial and research designs to fully inform questions (1) to (4). Further, these methods allow optimization under real decision making conditions, allowing for time, the option value, and opportunity costs, of delay and optimal global trials across jurisdictions and imperfect implementation as demonstrated previously (Eckermann and Willan 2007, 2008a, 2008b, 2009).

The second paper extends this framework to demonstrate the profound effect of appropriately accounting for imperfect implementation in applying VoI methods to estimate EVSI, ENG and optimal trial design. For the usual case of interest with prior positive while uncertain INB, accounting for imperfect implementation is shown to increase ENG of trial designs, given an expected increase in strength of evidence with further trial information. This paper firmly establishes the relationship between the expected value of sample information and degree of implementation.

The final paper demonstrates advantages to jointly undertaking risk sharing arrangements in conjunction with global trial designs in order to optimally address joint research, pricing and adoption decisions. While adopting and trialing a new therapy with positive while uncertain net benefit is usually infeasible within a jurisdiction, adopting while trialling (AT) is feasible across jusridictions. The abilty to AT in turn enables robust risk sharing where prospective trial evidence can be used to inform prospective payments without selection biases inherent in performance measurement in practice. Global trials with risk sharing to mirror local price-uncertainty tradeoffs are shown as mutually beneficial to payer and manufacturer in avoiding their respective opportunity costs of delay with local trials and the option value of evidence forgone and associated lower degree of implementation with local adoption and no trial.

In combination these papers demonstrate methods for harmonizing VoI methods to optimally inform research, reimbursement and pricing (risk sharing) decisions.