The multi-domain responder index: a novel analysis tool to capture a broader assessment of clinical benefit in heterogeneous complex rare diseases

Abstract

In traditional clinical trial design, efficacy is typically assessed using a single primary endpoint in a randomized controlled trial to detect an expected treatment effect of a therapy in a narrowly selected patient population. This accepted paradigm is based on clinical evaluations that may not actually capture the breadth of the impact of a disease, which is especially true in the setting of complex, multisystem, rare diseases with small, extremely heterogeneous patient populations. The multi-domain responder index (MDRI) is a novel approach that accommodates complex and heterogeneous disease manifestations and evaluates a broad array of clinical disease without impairing the power or rigor of a study to fully understand a treatment. The MDRI sums the scores corresponding to clinically significant thresholds of change for each component domain in each individual patient, capturing the mean clinically meaningful change across multiple domains within individuals. This novel approach combines and then sums the results of independent domain endpoint responder analyses into one responder score to provide a broad basis for the assessment of efficacy. The impact of a treatment across multiple, physiologically independent domains, can be assessed clinically, reducing the adverse impact of heterogeneity on trial outcomes and allowing eligibility criteria to enroll a wider range of patients, ultimately resulting in efficacy and safety assessments of a therapy across a broad group of heterogeneous patients in rare disease programs.Trial registration The following studies are referenced within this manuscript (CLINICALTRIALS.GOV registration numbers): NCT00912925; NCT00146770; NCT00067470; NCT00104234; NCT00069641; NCT02230566; NCT02377921; NCT02432144.

Generalized pairwise comparisons of prioritized outcomes are a powerful and patient-centric analysis of multi-domain scores

“…Tandon and Kakkis [19] have proposed a multi-domain responder index to capture clinically meaningful changes across multiple domains, which is close in spirit to, but technically different than, the GPC approaches (prioritized and non-prioritized) presented in this paper. The multi-domain responder index uses the treatment-induced change in the score of each individual patient to assign them a "responder" status if the change exceeds a minimally important difference (MID).…”

“…Multi-domain responder index [19] Has better statistical power than a single responder analysis Requires ability to de ne "responder" status for all outcomes Our simulations have several important limitations. We have ignored the effect of age on the rates of change of the scores for the ve domains considered.…”

“…Such a distribution-based approach is convenient because it makes all domains comparable in the analysis (which can be especially important for a non-prioritized approach in which all domains are considered with equal weights). However, the approach illustrated in our simulations could just as easily use absolute changes considered to be clinically meaningful in each domain, just as in the development of a multi-domain responder index [19]. Considerations driving the choice between distribution-based thresholds vs. anchor-based thresholds fall beyond the scope of the present paper [9].…”

“…[8]. Moreover, a multi-domain responder index can capture clinically meaningful changes across multiple domains within individuals and, as such, have better statistical power and clinical relevance than a single outcome measure [19].…”

Scaling genetic resources: New paradigms for diagnosis and treatment of rare genetic disease

“…Gene therapy vectors can be considered a platform technology where the viral backbone is a near constant that can deliver multiple therapeutic genes to address multiple disorders, albeit typically one at a time. Thus, regulatory approaches for approval that view such technology as a single delivery system to treat multiple disorders will lead to faster implementation responder index with staggered start of transition of patients randomized to treatment or placebo followed by transition to open label (Cadaoas et al, 2020;Fox, Volpe, Bullaro, Kakkis, & Sly, 2015;Tandon & Kakkis, 2021). Regardless of study design, trial design and assessment of efficacy are enormously dependent on a firm understanding of the natural history of the disease.…”