OBJECTIVES: Long-term treatment with corticosteroids is an option in many auto-immune conditions. This comes with an increased risk of various short- and long-term adverse events. These events may be avoided with the use of corticosteroid-sparing (CSS) agents. Such benefits should be captured in cost-effectiveness models (CEMs). Hence, for this work, we aimed to find out how the short- and long-term consequences of CSS agents were estimated in previously published or appraised CEMs. METHODS: A systematic literature review was conducted. Two bibliographic databases were included: PubMed and Embase. Additionally, technology appraisals (TAs) of CSS agents published by the National Institute for Care Excellence (NICE) were assessed to capture NICE’s view on the inclusion of a CSS effect in CEMs. Data extraction items included model key characteristics, corticosteroid-specific model inputs and model outcomes. RESULTS: We identified four published CEMs and nine NICE TAs where CSS value was recognized. Two TAs (TA397&TA278) submitted a CEM that incorporated CSS effect. In the other TAs where the CSS effect was not incorporated in the CEM, NICE recognized that the most plausible ICER would be lower if incorporated. This positively impacted the appraisal’s reimbursement decision. Most models used relatively simple assumptions except for TA397, where a cohort-based regression model was used. However, this was criticized due to uncertainty around the link between short-term CS dose reduction and long-term consequences. CONCLUSIONS: The CSS consequences are not frequently captured in CEMs. While NICE recommends and welcomes the inclusion of the CSS consequences, they are critical about the methods that were previously used by manufacturers to do so. Guidance is needed on the incorporation of CSS effect in CEMs.