It has been emphasized on many occasions that the meme that dominates health technology assessment, the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) is a classic facilitator, is not only tenaciously held but has a high copying fidelity. The holders of the meme accept the ‘for information only’ role of health technology models while rejecting any potential role for hypothesis testing in modeled claims for therapies. Characterizing and indeed central to these modeled claims, are the fabrication of incremental lifetime cost-per-QALY models. These have been a mainstay of technology assessment for 30 years; unfortunately this mainstay, this rock on which the technology assessment meme is built, is fundamentally flawed.
Commentaries which have pointed to the flaw in the technology assessment meme have taken the position that if science is to progress, if new acts are to be discovered, if the processes of conjecture and refutation are to be recognized, then claims made for competing therapy products and devices must be empirically credible, evaluable and replicable. This is rejected in ISPOR modeling.
Demolishing the technology assessment meme: the first step is to point to the absurdity of creating imaginary cost-per-QALY worlds and building on these nonsense claims willingness to pay thresholds to support pricing and access recommendations. Yet, these are believed. The second, and decisive step is to point out that the QALY is a mathematically illogical and impossible construct. This follows from the failure of the meme leadership to recognize the standards of fundamental measurement. Multiplying time spent in a modeled disease state by a generic utility is impossible because the utility scale only has ordinal properties. To create a QALY requires a utility to have ratio properties; a true zero. Utilities such as the EQ-5D-3L fail this requirement as negative utilities are possible.
Further nails in the QALY coffin are in excess to requirements, but include the lack of unidimensionality and interval measurement properties with the observation that from inception those developing the EQ-5D had no idea of the need to meet Rasch measurement standards. In fact, developing a ratio scale is difficult.
The purpose of this section is to report on attempts by journal editors and reviewers to deny the standards of fundamental measurement and, hopefully, rescue the QALY from oblivion. Given the past 30 years and thousands of cost per QALY models this is only to be expected. It is difficult to see the standards for defense: perhaps only by assumption. After all, if the lifetime model is built on assumptions, an assumption of ratio property will not be remarkable. WE can persevere with another generation of modeled claims.