The NICE Decision Support Unit has just published a report with that very title. And the answer to the question is….BIG!
“Nine cost-effectiveness studies conducted alongside clinical trials were used as case studies. Each had existing analyses based on patient completion of the EQ-5D-3L instrument. In each case, we used the copula models to generate a revised analysis based on estimated 5L scores. We compared directly-observed 3L and estimated 5L (EQG and NDB) results.
The 5L instrument and associated tariff has the effect of shifting mean utility scores further up the utility scale towards full health, and compresses them into a smaller space. Thus, improvements in quality of life tend to be valued less using 5L than equivalent changes measured with 3L. In almost all cases, this means that a switch from 3L to 5L causes a decrease in the incremental QALY gain from effective health technologies and therefore technologies appear less cost-effective. This is true whether the estimation of 5L is based on EQG or NDB data. However, an important exception is where life extension is a substantial element of health gain, the ICER can reduce rather than increase.
Estimated incremental QALY gains reduced by up to 75% when moving from 3L to 5L (EQG dataset) or 87% (NDB dataset).”
You can access it here.