Collaborative Research Projects
Challenges can occur throughout the clinical development lifecycle and Cytel consultants help address these. Our teams of analytical, process-oriented software engineers and statisticians work with our customers to answer varied questions such as:
How can I better make Go/ No Go Decisions?
How can I make my patient enrolment forecasts more reliable?
We improve trial design, enhance decision-making and support and logistics management through custom software solutions.
Merck were unable to deploy their innovative methods for conducting dose finding trials when they couldn’t find commercial software to test and validate their models. They approached Cytel for a collaborative solution.
Cytel's cross-functional team of biostatisticians and application engineers translated the Merck’s original concept into a comprehensive functional specification. The Cytel statistical team then conducted research to identify, refine and optimize several methods for simulating dose-adaptive studies. A single, integrated software solution was ultimately created which allowed the client’s trial design and clinical teams to simulate and compare various adaptive design scenarios.
Via Cytel’s custom software solution, the sponsor’s trial planners now have a clear picture of the operational “trade-offs” of each adaptive design option. Simulation results also provide compelling evidence supporting regulatory approval of the proposed trial.
Program and Portfolio Optimization
The lifescience industry faces ongoing R&D productivity and innovation challenges. But how can we reduce costs and accelerate development without compromising the drug or device candidate’s chances of success? We believe a successful portfolio optimization strategy combines three key approaches:
Comparing development options based on the Expected Net Present Value (ENPV)
Use of powerful simulations to determine which strategic options maximize the ENPV
Adaptive Financing achieved through the alignment of investors' incentives and optimistic interim looks
Our consultants apply their statistical knowledge to design trials with a sound risk profile, and use forecasting expertise to identify the trials with the greatest potential.
Our biotech client had limited resources for their lead asset’s pivotal trial. They wanted to avoid high up-front costs as well as attract external investors.
Cytel’s experts proposed an adaptive design with reassessment of sample size at interim analysis. The sample size would only be increased if the results were in the ‘promising zone’. In the absence of overwhelming efficacy, a Promising Zone is defined as conditional power between 30% and 80%. If interim results were not in the Promising Zone the study should continue without sample size increase.
The adaptive design ‘de-risked’ the study and improved transparency- the client obtained external investment. A flexible, novel financing structure was linked with the interim results in a staged approach, conserving resources for launch.
Collaboration between sponsors and regulatory agencies is essential for the success of a clinical drug development program. Cytel has clinical trial statisticians with extensive experience in regulatory settings and can provide knowledge and guidance as you prepare for your regulatory interactions. Each statistician is highly knowledgeable about relevant and emerging methods associated with clinical trials. Importantly, they are also skilled communicators and can articulate complex statistical issues to both non-statisticians and regulatory reviewers.
Our client wanted to avoid a large and expensive cardiology trial. They asked for our help to design an innovative alternative, and to support their interactions with the FDA.
First, we created adaptive trial simulation software to predict probable outcomes of a variety of study approach scenarios. An innovative adaptive group sequential design with early-stopping points for efficacy or futility was selected. We provided the simulation software to the FDA/CDER statistical review committee members in advance of the review meeting, and developed a detailed description of the methodology. The Cytel statistical consultant attended the FDA statistical review meeting with the sponsor representative and presented the design.
The design was successfully defended at the FDA meeting.
Valid statistical interpretation of most clinical trials relies on valid inference from multiple statistical analyses. When more than one analysis yields “statistically significant” results, the overall false positive rate for sets of multiple analyses can inflate above acceptable levels, typically 5%. It is therefore critical that multiplicity issues are properly handled. Our consultant statisticians have extensive knowledge and experience in designing clinical trials and analyzing data to address multiplicity. .
Missing data can bias analysis results in unknown ways, and makes interpretation of trials with missing data difficult. Prevention of missing data can be addressed through careful design and conduct , and handled with analysis that makes full use of information on all randomized participants attends to the assumptions about the nature of the missing data underlying estimates of treatment effects. Our consultant statisticians can advise on the best ways to address these issues and are experienced in approaches including: single imputation techniques; multiple imputation techniques; methods for assessing assumptions regarding missing data; Assessment of patterns of missingness; Sample size calculations accounting for missing data.
DMC and Independent Statistical Committee Support
DMCs are increasingly used in industry-sponsored clinical trials. Cytel are equipped to deliver the range of DMC statistical requirements.
Our consultant statisticians have an extensive track record of participating with and on DMCs including acting as the independent statistician. We are experts in the statistical methods employed in those clinical trials which typically involve DMCs.
The Independent Statistical Center provides timely and accurate statistical analyses and summaries of accumulating clinical trial data to the DMC. Our experienced statisticians, programmers and data managers design, program, assemble, and validate interim analysis packages. Our extensive statistics and programming resource pool ensures adherence to timelines and responsiveness to additional requests.
Maintaining the security of unblinded information to the DMC and the ISC is critical. Cytel’s ACES software facilitates sponsor-committee communications and interim data handling in a secure, regulatory compliant, and audit-trailed environment.
Our client’s innovative Phase 2 dose finding study required multiple adaptation analyses to an Adaptive Design Review committee.
We created a team of Cytel statisticians and programmers to serve as the Independent Statistical Center. Our proprietary ACES system was used to facilitate sponsor-committee communications and interim data handling according to regulatory guidance.
Our experienced, flexible team were able to provide timely responses to requests for exploratory analyses by the Adaptive Design Review Committee.
Clinical drug development is a risky, and expensive process. Adaptive clinical trial designs can mitigate risk and pave the way for flexible interim options. They can also accelerate development, reduce costs and improve the quality of the information yielded by the clinical trial.
Adaptive designs require expert knowledge. Under pressure to obtain orphan drug exclusivity, our client came to us for help when their original design was rejected by the regulators.
Cytel statisticians analyzed the original design to assess the reasons for the original FDA rejection. We identified three viable alternative designs: a single four arm trial; two separate trials; and a two-stage adaptive confirmatory design. Using simulations we were able to compare the operating alternatives of the three alternatives, and ultimately selected the two-stage confirmatory design.
The two stage design was accepted by the FDA and the client easily met the deadline for the FDA submission.
Cytel biostatisticians have experience working in this complex and increasingly important area, which is intrinsically linked to the field of big data. Pattern recognition is an approach to machine intelligence based on statistical modeling of data. The objective is to identify a link between a response (for example disease status) and a set of features of the (biological) process of interest. This link is then used for predicting the response of new observations of the feature set which will tend to be high-dimensional or ‘’Big Data’.
With a revolution taking place in healthcare information, from genomics to use of wearable devices, pattern recognition may be applied in a variety of fields including medical diagnostic device development, genome studies, and personalized medicine.
The sponsor wanted to build and validate a medical device based on a statistical classiﬁer that detects abnormality in a certain physiological function.
The Cytel consultant reviewed and summarized current FDA guidelines to provide understanding of the route to market and regulatory requirements, and conducted a feasibility analysis using wavelet based classification. Then, the consultant helped design a POC trial which collects data for building the classiﬁcation algorithm using supervised machine learning methods. The consultant also advised on building of the statistical classiﬁer using segmentation algorithms, feature extraction, model building, and estimating the classifier’s predictive-accuracy using training-test data splits.
The sponsor has gained preliminary understanding of the population and the operating characteristic of the classiﬁer. With Cytel’s involvement, planning is underway for the forthcoming validation trial.