Profiling with CourseMap AI
Discover how we profile trial participants and boost trial power
CourseMap AI: what it delivers for trials
CourseMap AI delivers concrete, trial-ready outputs that improve the efficiency and interpretability of neurodegenerative clinical trials.
For each participant, CourseMap AI generates individualized prognostic profiles estimating expected disease progression over time. By accounting for expected disease progression, these covariates reduce unexplained variability without altering treatment effect size.
Seemless integration within existing workflows
CourseMap AI is designed to integrate seamlessly into standard clinical trial processes. Blinded baseline trial data are securely transferred, models are executed under predefined conditions, and patient-level prognostic covariates are generated.
These outputs are delivered in formats directly usable by sponsor biostatistics teams and can be incorporated into standard analysis frameworks (e.g. MMRM or LME), just like other pre-specified covariates like sex or age.
- No training or setup required for the sponsor's teams
- Data directly usable by biostatistics teams
- Methodological support for result interpretation
Locked, versioned and reproducible execution
All CourseMap AI models are executed in a locked and versioned manner. Once a configuration is defined, it remains fixed throughout execution, ensuring full traceability and auditability.
This prevents data leakage, preserves trial integrity, and guarantees that outputs can be reliably reproduced and reviewed.
We only access blinded baseline data from trial participants. The model has no access to treatment allocation or to any follow-up data.
Stronger signals, better decisions, more efficient trials.
The immediate benefit of incorporating prognostic covariates is a more precise estimate of the treatment effect, with reduced variance.
In Phase 2 trials—not designed to demonstrate efficacy—this additional precision enables early efficacy signals to be interpreted with greater confidence. Complementary analyses using prognostic covariates can therefore meaningfully support go / no-go decisions when assessing whether to invest in a larger Phase 3 program.
When applied prospectively to the planning of a Phase 3 trial designed for 80% statistical power, the impact can be substantial. A reduction in outcome variance of 50% can increase the actual probability of success from 80% to 97%, assuming the treatment is effective. Alternatively, the same gain in power can be used to reduce the required sample size by up to 50% while maintaining the original power level at 80%.
Both European and US regulators expressed their support for correcting continuous outcomes with prognostic covariates. Qairnel’s team is investigating other use cases of digital twins for clinical development. Contact us for details.
- Smarter go/no-go decision after phase 2
- Higher probability of success or lower sample size in pivotal phase 3
Ready to boost your trial with AI?
Let’s discuss how Qairnel’s predictive models can help you optimize patient selection and reduce trial variability.