Presentient’s Trial_Architect package develops your trial protocol and statistical analysis plan with you, supported by our network of domain experts. Our patent-pending subgroup taxonomisation unlocks data-driven exclusion criteria. Rigorously simulates BRAKES’ performance on your specific design to assess suitability and quantify benefits.
Then hands you an optimised statistical analysis plan.
Finding hidden subgroup signals
Functional Taxonomy algorithm - the image here shows hierarchical clustering and hidden patterns in clinical trial genetics data for Triple Negative Breast Cancer patients (unpublished data).
No drug works well in all patients. Subgroup identification tools are currently repurposed from other industries and are suboptimal for the characteristics of clinical trials (ie “small” samples, noisy metrics, heterogeneous data types & unknown underlying structure).
Presentient's patented subgroup analysis methodology has been developed specifically for these characteristics. It taxonomises patients into functional categories without biases/limitations inherent to conventional techniques.
Its high performing innovations include:
Eigenspectral cleaning stripping noise dominating ~90% of raw correlation signals
Distribution-free (no assumption of normality)
Automatically determines the number and hierarchy of subgroups without a priori input.
From safety resolution to stronger efficacy signal
The drug works well, but only in a subgroup that conventional methods cannot isolate. (synthetic data)
Presentient’s patented methodology identifies important subgroups that conventional analyses miss. This shows where treatment effect is being driven or diluted, and whether safety signals are confined to specific patient populations, using data already collected. The output directly supports go/no-go decisions with evidence that a defined subgroup is responding, even where the intent-to-treat population has not yet separated. This is critical for putting the best case forward for extension and Phase 3 funding conversations.
Stay blinded, decide sooner
Want to speed up your clinical trials with early decision making? Want to run 20+ automated interim analyses while eliminating operational bias?