Development of a face-validated conceptual model structure for economic evaluation in phenylketonuria, through expert...

Clinical complexity and limitations of available evidence make cost-effectiveness analyses (CEAs) in phenylketonuria (PKU) challenging, with no consensus model structure emerging from historical CEAs. This research aimed to elicit perspectives from PKU medical experts regarding approaches for CEA modeling, to inform development of a face-validated model structure.

In a modified e-Delphi study, clinical experts rated their agreement/disagreement with 14 statements on clinical outcomes, diet, target population, blood phenylalanine (Phe)-level classification, and treatment discontinuation. Separately, a survey of a broader sample of experts elicited perspectives on the prevalence of PKU comorbidities, and biological plausibility of the association of surrogate outcomes with comorbidity risk. Insights informed development of the conceptual structure and parameters for a CEA model.

In the modified e-Delphi study, participants (N = 11) achieved consensus (≥ 70% agreement/disagreement) on statements on the associations of blood Phe levels with outcomes, unmet need, mechanisms for the association of PKU with chronic comorbidities, and medical management approaches. In the survey (N = 30), ≥ 70% of participants agreed that prevalence of comorbidities is higher in PKU vs. the general population, and that increased natural-protein consumption may reduce comorbidity risk, including for depression, osteoporosis, overweight, anemia, and type two diabetes mellitus. Findings informed development of a CEA model structure comprising five blood Phe level-defined health states (range: ≤120 µmol/L to > 1200 µmol/L), associated with four classes of patient-relevant clinical outcomes in PKU: contemporaneous symptoms/manifestations, intellectual disability, comorbidities, and diet liberalization.

In this research, perspectives were elicited from medical experts in PKU to develop a clinically informed, face-validated conceptual model structure for CEA, and to identify associated parameters. The resulting de novo economic model allows for valuation of numerous important patient-relevant outcomes, enabling modeling of the heterogeneous drivers of unmet need in PKU. The structured expert-elicitation approach used in this research, and documentation of the influence of perspectives elicited on model structure and parameters, support the validity of the model for application in economic evaluation of interventions for PKU.