Case studies
From research to practice: AI innovations implemented and validated in real clinical trial settings.
Operational Development of Digital Twin Technology in a Phase 2 Parkinson’s Disease Trial
For a top 10 global pharmaceutical company
Executive Summary
Qairnel collaborated with a top-10 global pharmaceutical company to deploy its digital twin technology during an ongoing Phase 2 clinical trial in Parkinson’s disease. The objective was not exploratory research, but operational integration: delivering prognostic covariates on sponsor timelines and supporting real-time statistical decision-making alongside the sponsor’s biostatistics team.
Within less than one month from contract signature, Qairnel built, validated, deployed, and operationalized a Parkinson’s disease digital twin tailored to the study population, delivering prognostic covariates used directly in the sponsor’s statistical analyses. This collaboration illustrates Qairnel’s ability to operate as an industrial partner embedded in clinical development workflows.
Development Context
The study was a large, multicenter, randomized, placebo-controlled Phase 2 trial evaluating a small molecule in Parkinson’s disease. As is common in PD development, the sponsor faced challenges related to disease heterogeneity and outcome variability, particularly for decision-making at the Phase 2 stage.
The sponsor sought to evaluate whether advanced prognostic modeling could be operationally integrated into their analysis framework—under real development constraints, using blinded data, sponsor infrastructure, and tight timelines.
Operational Challenge
Unlike retrospective or exploratory studies, this collaboration required:
- Rapid deployment under active development timelines
- Full integration within the sponsor’s secure data and analysis environment
- Close coordination with the sponsor’s biostatistics team
- Delivery of validated, locked covariates suitable for inclusion in formal analyses.
Qairnel’s role was to function as an execution partner supporting decision-grade analyses.
End-to-End Execution Timeline
Late October: Contract for collaboration signed.
November 1–8: Qairnel applied its proprietary Course Map AI technology to build a digital twin tailored to the study’s target population. Model training, cross-validation, and internal quality control procedures were completed. The model was locked following validation.
November 8–14: Access to the sponsor’s internal platform was granted. Qairnel deployed its software within the sponsor’s environment and conducted dry-run experiments to verify technical integration, reproducibility, and data flows.
November 14: Qairnel received access to blinded baseline data for all study participants. Data management, harmonization, and quality control procedures were completed.
November 18: Qairnel delivered two prognostic covariates derived from the PD Course Map model, corresponding to two clinical endpoints of interest identified by the sponsor.
November 18–21: The sponsor’s biostatistics team integrated the covariates into their statistical analyses. Qairnel provided methodological support, documentation, and clarification to ensure correct interpretation and use.
November 21–30: At the sponsor’s request, Qairnel performed complementary analyses and verification to support internal review and robustness assessment.
Integration into Decision-Making
A distinctive aspect of this collaboration was the parallel and interactive workflow between Qairnel and the sponsor’s biostatistics team. Analyses were not performed in isolation; instead, Qairnel’s outputs were directly incorporated into the sponsor’s ongoing analytical process. This enabled:
- Rapid feedback loops
- Alignment on assumptions and interpretation
- Immediate assessment of the operational value of prognostic covariates
Qairnel functioned as an embedded analytical partner contributing to the sponsor’s internal discussions.
Outcome and Learning
In mid-December, the sponsor disclosed that the study did not meet its primary or secondary clinical endpoints. While the clinical outcome was negative, the collaboration successfully demonstrated that digital twin–derived covariates can be deployed, validated, and used within real-world Phase 2 decision-making timelines.
Importantly, Qairnel’s work was completed independently of study outcome, reinforcing the objectivity and credibility of the approach.
Why This Matters
This case study demonstrates that Qairnel’s digital twin technology can be:
- Deployed rapidly
- Integrated into sponsor infrastructure
- Operated under development-grade constraints
- Used collaboratively with biostatistics teams
- Aligned with real clinical decision points
For sponsors, this illustrates how digital twins can move from concept to execution within active clinical programs.