Target trial emulation to incorporate real-world data in estimation of the clinical- and cost-effectiveness of biologic treatment

Introduction: In health technology assessment (HTA) of biologic treatments for rheumatoid arthritis (RA), there is limited randomised evidence on treatment effectiveness after first-line treatment failure. We demonstrate how real-world data (RWD) could fill this evidence gap.

Methods: Target trial emulation (TTE) minimises biases in causal analysis of RWD by pre-specifying a protocol for a hypothetical randomised clinical trial (RCT) that would estimate the effect of interest. The application of TTE for HTA was illustrated using RWD from the British Society for Rheumatology Biologics Register for Rheumatoid Arthritis to estimate the effectiveness of rituximab versus non-biologic therapy (NBT) after first-line biologic failure, in terms of European Alliance of Associations for Rheumatology (EULAR) response achievement. The effectiveness estimates from RWD were combined with RCT estimates in a meta-analysis. The pooled estimates were entered into an economic model to estimate the incremental cost-effectiveness ratio (ICER) comparing biologic vs. NBT strategies.

Results: Based on RWD, rituximab was associated with higher probabilities of achieving a moderate or good response (0.215 vs. 0.174), and a good response (0.090 vs. 0.066), compared to NBT. These probabilities were lower than those estimated from RCT data (moderate or good 0.650; good 0.150). The economic model estimated less time on treatment and lower costs associated with biologics when based on RWD compared to RCT data (mean £63,500 vs. £70,000). This resulted in a higher ICER based on RWD compared to RCT data (mean £46,800 vs. £34,700 per quality-adjusted life year gained).

Conclusions: RWD can provide supplemental evidence on treatment effectiveness where randomised evidence is limited. This can make a meaningful difference to cost-effectiveness estimates. Our results are not intended to inform current RA management.