An Exploration of Incremental Net Health Benefit Analysis: A Quantitative Assessment of the Benefit-Harm Profile of Natalizumab in Relapsing Remitting Multiple Sclerosis

Background Currently, benefit-harm decisions are based on semi-quantitative analysis (benefits and harms are expressed numerically but are not weighted quantitatively) with underlying assumptions criticised as being opaque and subjective. Incremental net health benefit (INHB) analysis has been proposed as a quantitative framework to help regulators assess the balance of benefits and harms of healthcare interventions at a population level. Natalizumab is a new disease-modifying treatment for relapsing remitting multiple sclerosis (RRMS), for which clinical data suggest a potential for greater treatment effect than other disease-modifying therapies but which is associated with a risk of progressive multifocal leukoencephalopathy (PML), a typically fatal neurological condition. Objective To explore the potential value of INHB methodology in a regulatory framework using a topical case study: an analysis of the benefit-harm profile of natalizumab (NA) compared with beta interferon (BI) and no disease-modifying therapy (NT) for the treatment of RRMS. Methods A Markov chain decision-analytic model was constructed to represent RRMS disease progression and the impact of therapy in a cohort of newly diagnosed 30 year olds. Health states were EDSS (expanded disability status scale)-based MS disease severity, PML and death. Model benefit parameters were reduction in the risk of disease progression and relapse rate. The model 'harm' parameter was incidence of PML. The model outcome was quality-adjusted life years (QALYs), which was used to derive the INHB of natalizumab versus comparators. Natural history (MS progression and relapse rates), treatment benefits (hazard ratios for progression and reduction in relapse rates), harm (PML risk) and utility weights were obtained from published studies. Time horizons analysed were 2 years, 25 years and lifetime. QALYs were discounted at 3%. Probabilistic sensitivity analysis was used to explore the impact of uncertainty within the model and threshold analysis provided the incidence rate of PML at which there was zero INHB. Results Positive INHB was found with NA for all time horizons. Versus BI the INHB (in QALYs) was 0.01 at 2 years, 0.29 at 25 years and 0.77 over a lifetime, and versus NT it was 0.03, 1.28 and 2.95 respectively. The likelihood of a positive INHB depended on time horizon versus IB and ranged from 79%-98% and was 100% for all time horizons versus NT. The annual PML risk that resulted in a zero INHB ranged from 0.31% (NA vs IB 25 year analysis) to 2.01% (NA vs NT 2-year analysis). Discussion In this analysis natalizumab's benefits in improving MS progression and relapse appeared to outweigh the risks of PML at a population level. Limitations to this study include the lack of patient level natural history data and the use of short-term (2-year) clinical data. The analysis also omitted inclusion of less serious risks associated with natalizumab, such as injection site inflammation and adverse events associated with beta interferon, such as flu-like reactions, depression, and decreased peripheral blood counts. INHB analysis appears to provide potentially useful supplementary insight for regulators assessing the benefit-harm profile of healthcare interventions at a population level.