Presentation (CTI)

OBJECTIVES: The utilisation of real-world evidence (RWE) is becoming increasingly important in the decision making of reimbursement authorities. However, the extent and manner of its incorporation into health technology assessment (HTA) submissions remain unclear. The objective of this study was to analyse Pharmaceutical Benefits Advisory Committee (PBAC) decision making to understand the history of, and value placed on RWE in the decision making for reimbursement of pharmaceuticals in Australia.
METHODS: A database was created of all submissions to the PBAC including cost-utility analyses (CUAs) from March 2020 to July 2024 (n=341) based on Public Summary Documents. Relevant data was extracted, such as the ICER, disease severity and evidence base i.e., use of randomized/non-randomized evidence. Summary statistics were used to analyse the use of RWE in PBAC submissions and decision making over time and its impact on favourable PBAC recommendations. The probability of recommending a medicine for funding was estimated using multivariate probit regression models.
RESULTS: 17.6% of the CUAs from March 2020 to July 2024 used non-randomised evidence (60/341). Of the submissions based on non-randomised evidence, the proportion considered by the PBAC to have a “significant” treatment effect appears to be increasing over time (from 17% in 2020 to 88% in 2024). The proportion of recommendations is also higher for submissions with non-randomized evidence compared to randomized evidence, where RWE was more likely to be used for submissions with rare conditions. Despite this, regression results suggest that when holding all other variables constant at their mean, the probability of a recommendation increases by 23.91 percentage points for submissions with randomized evidence compared to non-randomized evidence (p<0.05).
CONCLUSIONS: RWE use in PBAC decision making seems to be constant over the last five years. However, acceptance of RWE appears to be growing in certain circumstances, especially for rare conditions.