Approaches used to model patient and carer health-related quality of life in economic models of rare disease treatments in NICE appraisals

Providing robust health-related quality of life (HRQoL) evidence for rare disease treatments (RDTs) is difficult due to the challenges involved in generating and interpreting evidence for rare diseases. Our objective was to systematically review the approaches used to model patient and carer HRQoL in economic models in RDT appraisals published by the National Institute for Health and Care Excellence (NICE). We identified appraisals completed between 2011 and 2023 and recorded HRQoL modelling characteristics from appraisal documents, including the approach used to derive the patient health state utility values (HSUVs) used in the final appraisal, and whether and how carer HRQoL was included. We mapped the approaches to estimating patient and carer HRQoL against NICE’s hierarchy of preferred HRQoL methods. We identified a total of 111 indications. Overall, we found heterogeneity in the approaches used to model patient and carer HRQoL. Regarding patient HRQoL, EQ-5D data was not available from a relevant study in 59.5% of indications in the final appraisal, and EQ-5D was not deemed appropriate by the committee in 6.3% of indications. EQ-5D from a relevant study was available and used in 34.2% of indications. Various sources of patient utilities were used, including EQ-5D values from a relevant study, estimates from the published literature (EQ-5D or non-EQ-5D), values derived from statistical mapping, vignette studies, or proxy conditions, combinations of several types, or values based on assumptions. Approaches to modelling patient utilities were also discussed in three-quarters of final appraisal documents, suggesting a considerable influence on the final recommendation. The impact on carer HRQoL was included quantitatively in the final appraisal in 26.1% of indications; for 3.6% of indications, the committee considered it qualitatively; and for one indication (0.9%) both quantitatively and qualitatively. Our findings suggest that, given the NICE hierarchy, patient and carer HRQoL was not very well captured in RDT appraisals. If the NICE hierarchy is viewed as a taxonomy of possible approaches, our findings also suggest that it does not accommodate some of the approaches used to estimate HRQoL in RDT appraisals. These issues raise questions about the consistency with which HRQoL is modelled and reflect the challenges of generating robust HRQoL evidence for RDTs.