The Use of Health Economic Methods in the Development of New Interventions for Systemic Lupus Erythematosus

The objective of this thesis is to investigate how health economic evaluation methods could be applied to optimise clinical trial designs. The study focuses on trial design during the development of new drugs in the pharmaceutical industry. The methods developed account for many of the motivations, constraints, and uncertainties faced by a pharmaceutical company that have not been factored in to previous Expected Net Benefit of Sampling (ENBS) methods. The study focussed on Systemic Lupus Erythematosus (SLE) because it is a disease area with a history of failed clinical trials, attributed to poor trial design. Almost no health economic modelling studies had been published in this area. The work in this PhD made substantial contributions to developments in the economic modelling for SLE. In chapter 2, I systematically reviewed previous clinical trials in SLE and extracted and summarised their design features. In chapter 3, I identified health economic methods that have been used to design clinical trials and which methods would be adopted. In chapter 4, I reviewed observational cohort studies for SLE to develop (i) a conceptual model of SLE (ii) identify quantitative estimates of the natural history of the disease. I found that there was insufficient data from the published literature to describe the natural history of SLE according to the conceptual model so I undertook statistical analysis of an SLE registry. The analysis produces statistical models to extrapolate disease activity, steroid dose, organ damage incidence and mortality over the short and long term. In chapters 6 and 7, I described a Bayesian Clinical Trial Simulation for a Phase III SLE trial to evaluate the value of alternative research designs and a cost-effectiveness model for SLE. In chapter 8, I compared six trial designs for an SLE Phase III RCT using an ENBS approach adapted to the pharmaceutical perspective. For each simulated trial the expected profits were calculated conditional on expected regulatory and reimbursement approval. The study found that the analysis was extremely time consuming and is currently not feasible with individual patient simulation models. However, the framework for evaluating trial designs from a pharmaceutical perspective use value-based pricing can be applied to trial designs in simple decision problems.