OBJECTIVES: In Ireland, as part of the Health Service Executive’s (HSE) assessments on medicines for reimbursement, medicines undergo price negotiations with the Corporate Pharmaceutical Unit followed by assessment by the HSE Drugs Group. Data mining techniques can be utilised to extract relevant insights from HTA bodies. This research aims to identify factors which influence the Drugs Group decision to reimburse or not reimburse a medicine using a text mining approach.

METHODS: Text data containing the HSE Drugs Group minutes was obtained for a sample of assessments from 2020 – 2022 (n=65). Medicines under consideration were split into two subsamples, reimbursed or not reimbursed. Natural Language Processing (NLP) was utilised in RStudio^® and Python^® to prepare the data for text analysis. Descriptive text analysis was implemented using Python^® to obtain output of the most common words that appear in each subsample to identify factors that influence the Drugs Group’s decision to recommend a medicine for reimbursement.

RESULTS: From the 65 medicines, 43 (66.15%) were recommended for reimbursement from the Drugs Group’s deliberations. There are similar trends observed across the most frequent words utilised in both positive and negative outcomes at Drugs Group. As expected, “clinical evidence” and “cost effectiveness” appear frequently in both subsamples. “Unmet need” appears frequently in only the reimbursed subsample (0.40 times per medicine), and “price” appears frequently in the not reimbursed subsample (0.54 times per medicine).

CONCLUSIONS: The ability to partake in price negotiations can often result in a medicine’s positive recommendation from the HSE. As expected, unmet need and price are important considerations for reimbursement. Cost effectiveness and clinical evidence are likely to influence the Drugs Group’s decisions, either way. Further regression analysis is required to understand the extent of which these factors influence decisions at Drugs Group.