OBJECTIVES: Ankylosing Spondylitis (AS) is a subtype of spondyloarthritides. Its symptoms include chronic pain and stiffness of the spine, which can result in permanent disability. Patients with active AS who fail conventional therapy are generally treated with TNF-α inhibitors. A novel biologic treatment, the IL-17A inhibitor secukinumab, is approved in the European Union for adults with active AS, who have inadequately responded to conventional therapy. The increasing need to inform Health Technology Assessment (HTA) and other decision makers on cost effectiveness (CE) of biologics in AS established the need for an improved CE framework. The aim of this project was to develop an improved modeling approach, by assessing existing HTA submissions and overcoming shortcomings, to assist decision makers in drawing accurate conclusions.
METHODS: A targeted literature review was conducted to identify submissions for AS treatments to major HTA agencies. The submission models were evaluated based on multiple parameters including proximity to clinical practice, treatment sequencing, and incorporation of adverse events. Following identification and assessment of the shortcomings, an improved modeling framework was developed.
RESULTS: All identified submissions used a similar Markov structure, consisting of a short-term response criterion (change in Bath Ankylosing Spondylitis Disease Activity Index) to determine initial treatment continuation. In the long-term, models allow patient withdrawal from treatment due to adverse events or loss of efficacy. Notable shortcomings addressed by the proposed framework include absence of treatment sequencing and lack of detailed evaluation of adverse events including effect of mortality. Key limitation identified is the lack of data for some treatments. This has to be further addressed to enable full model potential.
CONCLUSIONS: Even though the common Markov structure used in existing HTA submissions for AS treatments incorporates most aspects of disease progression, areas for improvement were identified. The proposed model matches clinical practice and improves decision making for biologic treatments.
