OBJECTIVES: The National Institute of Health and Care Excellence (NICE) makes recommendations at a national level for oncology drugs in England. NICE provides a reference case example of how generic utility values are recommended for inclusion within a submission, including a preference for the EuroQol-5-dimension (EQ-5D) quality of life (QoL) measure. This research aims to explore how alternative, disease-specific, QoL measures have been received by external assessment groups (EAGs) in recent oncology submissions to NICE.

METHODS: All publicly available NICE oncology technology appraisals published in the previous year (1^st April 2021–31^st March 2022) were reviewed. The appraisal outcomes, source of QoL data, and the EAG critique of economic evidence were extracted.

RESULTS: 50 NICE oncology technical appraisals were identified, 14 of which were terminated, resulting in 36 appraisals requiring extraction. Generic QoL measures were used in 29/36 submissions; disease-specific QoL measures were used in 6/36 submissions; and one submission used a mixture of both (generic QoL measures for progression-free survival and disease-specific measures for progressed disease health states, respectively). Disease-specific QoL measures were ultimately converted to EQ-5D-3L utility scores using appropriate mapping algorithms, as stipulated in the NICE guidance. 5/7 appraisals using disease-specific QoL measures were ultimately recommended (4/5 recommended as part of the Cancer Drugs Fund) by NICE. The remaining 2/7 appraisals received an optimised recommendation (recommended for a smaller group of patients than the original scope of the appraisal).

CONCLUSIONS: The use of disease-specific QoL measures has been accepted in oncology appraisals when an appropriate mapping algorithm is used to convert to EQ-5D utility scores. The use of disease-specific QoL measures is an option for manufacturers in cases where EQ-5D data have not been collected as part of the clinical trial, subject to the availability of an accepted mapping algorithm.