Comparative Analysis of Traditional vs AI-Assisted Health Technology Assessment Dossier Writing

OBJECTIVES: Health Technology Assessment (HTA) dossier writing is currently a time intensive, costly and valuable resource binding process. Artificial intelligence (AI) supported creation of dossiers might be a promising approach to overcome these issues. This study aims to compare the traditional writing of dossiers for drug HTA in Germany (AMNOG) versus an AI-assisted approach in terms of time and cost.

METHODS: For comparisons, DO-BO®, a unique and the first commercially available AI-based dossier writing platform was compared to the traditional process of human dossier writing. Conducted systematic literature reviews were used for both processes. The conventional approach involved solely human dossier writing experts whereas the AI-powered approach employed a pre-trained large language model to generate a draft dossier managed by one dossier writing expert.

RESULTS: The mean time frame for the creation of a traditional dossier is 12 months (range 9 to 18 months), associated with costs of 150000 € (range 120000 to 300000 €) for a standard use case. The study found that the AI-assisted HTA writing reduced the time required for a draft dossier by 53% in comparison to the traditional writing method. This time saving relates to 6,36 months, associated cost reduction of 78000 €. The quality of the AI-assisted dossier based on DO-BO® is comparable to traditional dossier writing.

CONCLUSIONS: AI-assisted HTA writing improves traditional methods, saving time and costs significantly. Future research should refine AI-models and explore their use in different countries' HTA assessments.