Modelling Adverse Events in Patients Receiving Chronic Oral Corticosteroids in the UK.

ABSTRACT

BACKGROUND: Oral corticosteroids (OCS) are effective anti-inflammatory agents used across a range of conditions. However, substantial evidence associates their use with increased risks for adverse events (AEs), causing high burden on healthcare resources. Emerging biologics present as alternative agents, enabling the reduction of OCS use. However, current modelling approaches may underestimate their effects by not capturing OCS-sparing effects. In this study, we present a modelling approach designed to capture the health economic benefits of OCS-sparing regimens and agents. METHODS: We developed a disease-agnostic model using a UK health technology assessment (HTA) perspective, with discounting of 3.5% for costs and outcomes, a lifetime horizon, and 4-week cycle length. The model structure included type 2 diabetes mellitus, established cardiovascular disease, and osteoporosis as key AEs and drivers of morbidity and mortality, as well as capturing transient events. Quality-adjusted life-years (QALYs), life-years, and costs were determined for OCS-only and OCS-sparing treatment arms. Outcomes were determined using baseline 50% OCS-sparing, considering several OCS average daily doses (5, 10, 15 mg). RESULTS: A treatment regimen with 50% OCS dose-sparing led to lifetime incremental cost savings per patient of £1107 (95% confidence interval £1014-£1229) at 5 mg, £2403 (£2203-£2668) at 10 mg, and £19,501 (£748-£51,836) at 15 mg. Patients also gained 0.033 (0.030-0.036) to 0.356 (0.022-2.404) QALYs dependent on dose. The benefits of OCS sparing were long-term, plateauing after 35-40 years of treatment. CONCLUSIONS: We present a modelling approach that captures additional long-term health economic benefits from OCS sparing that would otherwise be missed from current modelling approaches. These results may help inform future decision making for emerging OCS-sparing therapeutics by comparing them against the cost of such treatments.