Clinical Trial Methodology

clinical trial methodology

About this technology

Overview Applications Services & Fees Our experts Contact

Applications

Our group has extensive experience in the design and analysis of standard randomized controlled trials, but also a strong focus on innovative trial designs and methods. We are currently most active in the following areas:

 

 

Adaptive trials offer a more flexible and efficient alternative to standard randomized trials, enabling prespecified adaptations such as early stopping for futility or efficacy. Bayesian statistical methods are particularly well suited to these trials, allowing ongoing learning through updating posterior and predictive distributions as data accrues. Bayesian adaptive methods are also central to modern platform studies, supporting efficient concurrent evaluation of multiple interventions in the same population. We have extensive expertise in the design and analysis of Bayesian adaptive trials and are, for instance, involved in the Bayesian-adaptive multi-arm AMICO trial in exercise-oncology and the RECLAIM platform trial for post-Covid.

Early-phase clinical trials aim to identify safe and efficacious doses for further clinical studies. These trials generally use an adaptive design where doses of newly included participants are based on toxicity, and potentially efficacy, outcomes of earlier trial participants. In the ONCODE-ACCELERATOR  (https://www.oncodeaccelerator.nl/) project, we develop and implement innovative statistical methods for adaptive dose-finding in early-phase oncology trials, enabling smarter and more efficient identification of the best dose(s) for the next phases of clinical drug development.

 

Treatment de-escalation aims to reduce toxicity and morbidity, by lowering dose, intensity, or duration, while preserving acceptable efficacy. Trials designed to evaluate the impact of de-escalation are methodologically challenging, as non-inferiority or equivalence designs often require unfeasibly large sample sizes. In the CROSSROADS project, we investigate ways to increase efficiency and feasibility of a de-escalation trial in colorectal cancer by comparing Bayesian and frequentist methods that allow incorporation of historical and external data.

 

Many clinical trials can benefit from incorporating high-quality external data, such as control arm data from previous trials in the same disease, in the analysis. When these data are sufficiently comparable, dynamic borrowing methods allow them to be incorporated into the trial analysis in a prespecified and adaptive way. Such approaches can substantially increase statistical power and improve the precision of treatment effect estimates. Our group is at the forefront of methodological research on dynamic borrowing of external controls, with active projects including a public–private partnership and involvement in the SHARE-CTD consortium.

Trial-within-cohorts (TwiCs) studies embed randomized comparisons within real-world observational cohorts, enabling efficient and pragmatic evaluation of new interventions. Eligible cohort participants are randomized to be offered the new intervention or continue usual care, with only those offered the new intervention being informed about the outcome of randomization. Our group has led publications on the methodological challenges of TwiCs studies and will co-design a TwiCs study within the DIRECT-DCIS project.

Master protocols accelerate discovery of effective interventions by testing multiple targeted therapies in parallel using a shared trial infrastructure. Umbrella trials evaluate several targeted interventions within a single disease, while basket trials assess one targeted therapy across different diseases with patients sharing a common genotype or phenotype. Our group has contributed to methodological studies comparing the performance of different methods of pooling in basket studies. We were further involved in PREFERABLE-II study, which uses an innovative umbrella trial design to evaluate effectiveness of tailored exercise interventions for cancer survivors, and a basket trial study to evaluate effectiveness of a common medical treatment targeting behavioral manifestations in different rare genetic neurodevelopmental disorders.

The stepped-wedge trial design is a pragmatic trial design that is often used in implementation research, enabling sequential roll-out of a new intervention in different hospitals, geographic regions or other clusters, and combining within- and between cluster comparisons to evaluate effectiveness. Our group has extensive experience in the design and analysis of stepped-wedge trials and is currently co-leading the design of a large-scale stepped-wedge study within the PEN-CONNECT project, evaluating an intervention to reduce cardiovascular risk among people with complex multimorbidity in Mozambique.

Conducting adequately powered randomized trials is especially challenging in rare diseases. Specifically for ALS, but broadly applicable to other progressive diseases, we have developed innovative trial methodologies that use disease progression modeling, integrate treatment effects across multiple endpoints (e.g. using joint models), and incorporate patient preferences using the Patient-Ranked Order of Function (PROOF). We are applying joint models in multiple phase 3 ALS trials. The PROOF endpoint is used as a supportive endpoint in various ALS trials. We increase efficiency through interim analyses, such as in the MAGNET phase 3 trial of lithium carbonate which stopped early for futility and reduced sample size and trial duration by 30–50%.

Trials with decentralized elements offer a scientifically robust approach to conducting clinical research while reducing participant burden. By enabling study procedures to take place at participants homes or direct surroundings instead of at traditional clinical sites, often leveraging mobile and wearable technologies, such designs can support efficient recruitment, sustained participant engagement, and improved retention. Decentralized approaches also facilitate the inclusion of more diverse and representative patient populations. We were the academic lead of the TRIALS@HOME project, in which we performed the pan-European the RADIAL proof-of-concept trial and systematically evaluated operational feasibility of clinical trials with varying degrees of decentralization.