OBJECTIVES: PKU is an inherited, genetic condition leading to elevated blood phenylalanine (Phe) causing neurological damage and cognitive disability. Disease management commonly involves a strict, lifelong, low-protein diet supplemented by Phe-free medical foods; adherence can be challenging, and potential neurological damage associated with uncontrolled Phe may remain. Varied cost-effectiveness analysis (CEA) approaches for health technology assessment (HTA) have been used for two regulatory-approved treatments, sapropterin dihydrochloride and pegvaliase. Reviewing PKU health-economic modeling challenges may inform CEA for future interventions.

METHODS: A targeted review of sapropterin and pegvaliase HTA evaluations was conducted. Seven sapropterin evaluations/CEAs were identified (Australia 2018/2022, Canada 2016, England 2021 company/review group, Ireland 2017, Scotland 2018) and three for pegvaliase (Australia 2022, Canada 2023, Ireland 2022). Publicly available HTA materials were reviewed to identify key modeling challenges, which were presented to a multidisciplinary panel of 8 international medical experts (pediatrics, metabolic, dietetics, genetics) to confirm relevance.

RESULTS: CEA approaches included: lifetime-horizon Markov models (6 evaluations), microsimulations (2), 1yr-horizon decision tree (2). Markov models included 4-5 alive health states (varying definitions); 4 evaluations shared 3 health states (“controlled”, “partially-controlled”, “uncontrolled”), and further included “controlled+Phe tolerance” or “asymptomatic” as a 4^th state. One included 5 blood Phe-range health states. HTA committees opined none of the models fully addressed the clinical dimensions of the decision-problem. Major critiques, confirmed by medical experts, included: 1) uncontrolled Phe effects on long-term outcomes not captured; 2) health-state definitions unaligned with utility/cost impacts; 3) diet-liberalization modeling unsupported/likely overstated; 4) discontinuation modeling lacking/implausible. Additionally, committees noted absence of modeling of PKU-associated comorbidities, impacts on mothers and/or unborn children (maternal PKU), and impacts on caregivers.

CONCLUSIONS: Across the 10 PKU HTA CEAs reviewed, the clinical condition was not appropriately captured; future models will need to address clinical accuracy. The challenges identified are common fundamental challenges in health-economic modeling of rare diseases.