21Jun16 – Statistical Integrity as an essential part of adaptive clinical trials
Although there are many obstacles that need to be overcome to take an investigational therapeutic successfully though to regulatory approval, the vast majority of issues really come down to two factors – the time and expense associated with clinical trials. Recent estimates suggest it costs almost $2.6bil USD and takes well in excess of 10 years to get a drug approved. Clearly, the need to improve the flexibility and efficiency of clinical trials is paramount. Considering this then, it is not surprising that adaptive clinical trial design has been gaining increasing interest and momentum across the research sector. At it’s simplest, the concept behind adaptive clinical trial design is that through the analysis of accrued data, the study design can be updated whilst the trial is still operational – at a patient, protocol or even clinical development program level. The aim of these changes is to improve the overall chance of successfully completing the clinical trial and expediting the regulatory approval process.
Adaptive clinical trial design is used to create clinical research efficiencies. At a patient level, changes to criteria associated with eligibility, investigational product administration and/or the duration, extent and review of study-based assessments can result in more efficient and efficacious treatment plans. Amendments at a protocol level can include changes to study objectives and endpoints, recruitment targets and procedures pertaining to statistical analysis, resulting in smaller sample sizes needed for data collection. There may even be more widespread applications to the overall therapeutic development program that not only help to answer the scientific hypothesis in question, but minimise the overall resources required. An example of this includes two-stage, two-phase seamless adaptive designs. Interim analysis could reveal such clear and compelling information that a Phase II study is adapted into a Phase III confirmatory study by virtue of a protocol amendment. The benefits to study completion would hopefully then be realised in terms of a faster timeframe and a reduced cost.
By utilising interim study data to support any changes in study design, it could be easily rationalised that these study changes are based on scientifically valid reasoning. Therefore, could adaptive clinical trial design be justified in any context? Unfortunately, it is not quite so simple. Whilst protocol amendments are seen as a norm in clinical trials, there is definitely a need for caution. Without careful consideration of proposed changes from both a scientific and statistical standpoint, there is a very real threat that bias and other negative influences may inadvertently impact on the study. Having a clear methodology behind adaptive trial design activities in conjunction with experienced human capital and quality data management infrastructure is essential to mitigating these influences.
From our experience having the active input of a Statistician across the life of a clinical trial has proven to be invaluable to Clients working in both early and late phase work. From ensuring the most accurate assumptions are included in initial protocol development, to the review of hypothetical adaptations that could be implemented dependent on study progress. Statistical review has helped to ensure maximum efficiencies at all trial time-points. Furthermore, having a Statistician involved in Safety Review Committee (SRC) meetings or more the formal Data Safety Monitoring and Review Board (DSMB) meetings can help to highlight the statistical impact of potential design changes, both positive and negative. This is crucial where other members may not be well versed in the theoretical or practical applications of statistics in clinical research.
When considering adaptive trial design, it is critical to consider the statistical integrity of changes to study design, in addition to the efficiency and flexibility the changes may bring.
Attention to statistical computations and simulations for any proposed changes therefore, should be an essential component of this process. The end result of this it to allow appropriate and efficient decisions to be made about the study design across the trial lifecycle based on the most current information.
25May16 – Genome variations in the design of informed clinical trials
Although pharmacogenetics is a relatively new field of study, it has already made a significant contribution to the standard treatment options for several common cancers. However, it is the positive impact that this science can have on the clinical development processes of potential new cancer therapies that is of greatest interest to companies working in this space.
Prior to the mapping of the human genome just over a decade ago, participants were recruited into oncologic clinical trials predominantly on indication. This would typically involve a clinical diagnosis followed by confirmation of the tumour type via histopathology. The use of pharmacogenetic technology takes this one giant step further. By detecting small variations in genome DNA, the most common of which are single polymorphisms, subsets of patients within a primary indication can be identified. These molecular variations can influence how a drug is metabolised and the extent of adverse events may be experienced. As they can even impact on common endpoints including progression free and overall survival rates, the benefits of pharmacogenetics to drug development can’t be overstated.
In the experience of Southern Star Research, an increasing number of oncology trials conducted on behalf of clients include pharmacogenetic testing. It is typically used to confirm patient eligibility in the first instance before allowing the stratification of trial subjects during data analysis. Ultimately this helps to determine the differences in response to the investigational product across subsets of subjects whilst providing insight into the preferred dosing range to ensure an appropriate risk/benefit profile. Certainly, without consideration of the pharmacogenetic impacts, clinical trial data analysis runs the risk of generating generic conclusions on the safety and efficacy of an investigational therapy across a diagnosis. The positive impacts achieved in one particular subset may be diluted, if not concealed, by data from non-responding subsets. The financial, social and therapeutic implications of this can be detrimental to the development of new therapies. Most importantly, it could prevent the commercialisation of potentially life-saving treatment options that could be invaluable to a percentage of the patient population.
By having an understanding of how individuals may be affected by therapies based on the presence of particular molecular variations, pharmacogenetics offers the ability to design more informed trials. The follow-on affect of this is the identification and assessment of targeted cancer therapies in a faster, safer and more cost-effective way.