By Exploristics CDSO, Kim Hacquoil
Simulation is the act or process of imitating the functioning or behaviour of a real-world system or process by means of another system or process, usually a computer program. While it has been successfully incorporated into product development for many industries, the use of simulation remains limited in the development of new medicines. For teams operating in the pressurised reality of a clinical setting using simulations to explore development options before embarking on clinical trial may feel time-consuming, fake and second-best to getting on with the real thing.
However, simulation can provide valuable insights for clinical development as it has in other highly innovative sectors such as car manufacturing and aviation. It offers a safe virtual space in which to explore alternative pathways and methods as well as an opportunity to identify and rectify mistakes. It can be used for study, training, testing, or demonstrating purposes. Building more simulation into the development process has the potential to make an immediate impact in key areas.
Simulation can inform internal development decisions
The use of simulation can improve the internal decision-making on multiple levels:
- Study level: it can help make decisions around the design. For example, how best to utilise adaptive designs within a study.
- Project or program level: it can inform decision-making around how much risk you should discharge at different stages of a drugs development. For example, how much you want to de-risk in early phases of development versus waiting until later.
- Portfolio level: it can support decision-making around how best to ensure a balanced portfolio of compounds in different stages or which program to progress, stop or put on hold.
In all these cases, simulation will let you answer ‘what if’ questions and assess potential future states which can help you make decisions.
Simulation can inform interactions with Regulators
Simulation also offers opportunities to support greater dialogue with Regulators. This may be by utilising real world data or historical external clinical trial data to inform the discussions. Or maybe by using simulations when looking at extrapolation between different populations like adults to paediatrics. Regulators are obviously concerned with type 1 error and simulations along with sensitivity analyses, can help to address the robustness of the results. Recent strategy papers by the FDA and EMA have recognised the value of simulation and model informed drug development in improving development success rates. These have highlighted its benefits in supporting more complex innovative trial design as well as the use of biomarkers and real-world evidence to support decision-making.
When simplification is too removed from reality
You may have designed many studies without simulation and so wonder why you need it. Many intricacies found in real studies are ignored for simplicity at the design stage since real trials are complicated. We often need to make some assumptions when it comes to designing trials. But how do we know which ones to ignore and which ones to explore? Simulation in trial design allows us to explore study assumptions quickly at an early stage before too much time and money has been invested in a particular approach. When studies are potentially complicated, simulation can bring more clarity and support decision makers.
Addressing uncertainty and interrogating choices
Simulation can be enormously helpful in addressing the many uncertainties associated with clinical studies. For example, it can help make decisions about multiple correlated endpoints or complex decision criteria. It can also help address the estimands framework regarding intercurrent events – a simulation is the only way to really explore the intercurrent events that don’t happen at random, for example if there is imbalance between events across different treatment arms. Both the FDA and EMA now recommend considering estimands strategies when drafting a clinical study protocol. Simulation can also easily incorporate data quirks and ensure you are as realistic as possible when assessing the sample size. It can also help explore different adaptive rules and see how they play out, especially when considering recruitment or constrained settings like rare diseases or other subpopulations.
Engaging with statisticians earlier to maximise the benefits of simulation
Many statisticians already use simulation in trial design when it’s not possible to quantify a situation or if there are numerous factors to explore in one go within the same framework. It provides an effective and rapid approach for answering many ‘what if’ questions and exploring multiple potential possible truths. It can also help to identify risks which might not be easily pinpointed otherwise and show how best to mitigate for these. It can be used to aid interpretability of complex situations, check rough and ready sample size calculations, help choose indications or compare development paths. It’s an extremely useful tool in the statistician’s toolbox and has broad application for drug developers.
Unfortunately, often statisticians are not brought into the design process early enough to maximise the benefits that simulation offers in informing the key decisions. Unlocking valuable simulation insights requires dialogue with statisticians at the earliest stages. Statisticians hold the key to a powerful virtual world where development decisions can be robustly tested and de-risked. Other industries have understood this and are reaping the rewards. It’s time our sector benefited more from it too.
Read more:
From cars to clinical trials, how simulation can improve design
Revolutionising clinical trial design with in silico studies using synthetic data
Why simulation is needed to optimise both adaptive and fixed clinical trial designs
Shifting the development paradigm with innovative trial design
Simulation can address recruitment challenges
Statistical Consulting Services
Hear more:
The power of simulations for designing clinical studies and beyond
People always, patients sometimes
Simulations your most powerful study design tool
Watch more:
Can pharma learn from development in other industries?
I just need a simple sample size