De-risk & accelerate

Drug development is risky, expensive and slow because of low trial phase transition, enrollment challenges and timeline delays. Novel AI methodologies applied to real-world data hold the key to faster and safer clinical trials.

Tip arrowTip arrowTip arrow

Challenges

How do we create new drugs in a faster and safer way?

Sub-optimal trial design

Underutilization of real-world data to inform effective trial design.

Patient enrolment for trials

Heterogeneity makes it challenging to recruit and target the right patients to reach sufficient statistical power to meet trial endpoints.

High trial failure rates

Phase transition success rates remain low in clinical trials, especially oncology.

Sub-optimal trial design

Underutilization of real-world data to inform effective trial design.

Patient enrolment for trials

Heterogeneity makes it challenging to recruit and target the right patients to reach sufficient statistical power to meet trial endpoints.

High trial failure rates

Phase transition success rates remain low in clinical trials, especially oncology.

We respond to the rapidly changing regulatory landscape.

New guidelines from regulators encourage use of real-word evidence

FDA (2021) and EMA (2015) published guidelines on the use of RWE in regulatory submissions. Both regulators outlined that covariate adjustment improves the efficiency of analysis and produces stronger and more precise evidence if the covariates are prognostic.

Increased use of real-word evidence in new drug applications

In 2020, 75% of FDA and 40% of EMA new drug applications included a real world evidence study. For example, in 2019, Janssen’s BALVERSATM (erdafitinib) was approved for locally advanced or metastatic bladder cancer using external control arm methodologies.

New guidelines from regulators encourage use of real-word evidence

FDA (2021) and EMA (2015) published guidelines on the use of RWE in regulatory submissions. Both regulators outlined that covariate adjustment improves the efficiency of analysis and produces stronger and more precise evidence if the covariates are prognostic.

Increased use of real-word evidence in new drug applications

In 2020, 75% of FDA and 40% of EMA new drug applications included a real world evidence study. For example, in 2019, Janssen’s BALVERSATM (erdafitinib) was approved for locally advanced or metastatic bladder cancer using external control arm methodologies.

Team members working on laptops

Our AI-powered methodologies

Our AI-powered methodologies

Covariate Adjustment

What is it?

Covariate adjustment is an effective way to increase statistical power in a clinical trial without increasing the sample size. Owkin’s methodology improves power by adjusting the efficacy analysis on prognostic covariates associated with the endpoint of a trial. This approach mitigates part of the variance that would otherwise obscure the treatment effect, reducing the noise, so the signal stands out.

What does it achieve?

Covariate adjustment helps better measure treatment effect. It can be used to reduce the number of patients or events required to reach a targeted statistical power or to increase power when enrolling fewer patients.

Download our whitepaper

Why work with Owkin on Covariate Adjustment?

New guidelines from regulators encourage use of real-word evidence

Leverage our high-quality datasets to discover novel prognostic covariates for adjustment.

Multimodal data

Benefit from our ML prognostic covariates that leverage the power of combining multi-modal data data for smarter trial readouts.

New guidelines from regulators encourage use of real-word evidence

Leverage our high-quality datasets to discover novel prognostic covariates for adjustment.

Multimodal data

Benefit from our ML prognostic covariates that leverage the power of combining multi-modal data data for smarter trial readouts.

Case study

desktop-arrowmobile-arrow
arrow-mobile

Impact

Impact

arrow-desktop

Biopharma

De-risk phase 3 clinical trials

Adopting Owkin’s ML Covariate Adjustment tool at the design stage of a phase 3 trial will maximize the probability of achieving statistical significance.

Accelerate clinical trials

Reducing the number of patients required to reach statistical significance in a phase 3 trial will greatly reduce the cost and duration of the study. It will also allow the drug to reach the market sooner which will benefit patients earlier and increase the ROI.

Our AI-powered methodologies

Our AI-powered methodologies

External Control Arms

What is it?

ECA is a methodology that uses patient-level data from previous clinical trials and real-world datasets as virtual controls in single and dual arm trials. ECAs result in more precise estimates of treatment effect compared to the standard of care, without having to recruit a control group.

What does it achieve?

ECAs help accelerate trials by reducing the overall number of trial participants and providing an objective, quantitative estimate of effect. They inform trial transition decisions and ultimately allow more patients to benefit from novel treatments, sooner. ECAs provide robust evidence to support regulatory submissions following phase 2 single arm trials, resulting in faster approval times.

Our AI-powered methodologies

Our AI-powered methodologies

Counterfactual outcomes

What is it?

A model to predict the RECIST score that each patient would have obtained if treated with the standard of care only.

What does it achieve?

Counterfactual outcomes provide confidence of the relative efficacy of the molecule and informs researchers on which patient profiles are good/bad responders.

Why work with Owkin on synthetic control arms?

Faster data access

Faster-than-standard access to high qualtity multimodal and logitudinal datasets as well as historical clinical trial data.

Cutting edge machine learning

Cutting edge machine learning and epidemiological methods to boost statistical power of clinical trials.

Medical imaging

Machine learning prediction models that leverage medical imaging to avoid traditional matching techniques that can drastically limit sample size and statistical power.

Faster data access

Faster-than-standard access to high qualtity multimodal and logitudinal datasets as well as historical clinical trial data.

Cutting edge machine learning

Cutting edge machine learning and epidemiological methods to boost statistical power of clinical trials.

Medical imaging

Machine learning prediction models that leverage medical imaging to avoid traditional matching techniques that can drastically limit sample size and statistical power.