Clinical Trial Design – How achieve better outcomes for companies and patients
Clinical trials have become increasingly complex programs with overall increasing costs and diminishing successes. Strangely, the Biopharma industry sticks to the same ground-losing game with reliance on traditional drug development processes and human-intensive work. This whitepaper explains how to improve the clinical trial design work leveraging best practices and latest technologies. It helps biopharma companies bring more, better and faster medications to patients, reduce wasting money and achieve better business outcome.
Advanced clinical trial design makes effective use of available data, simulates outcomes, enables increase insights and better decision making. It helps understand what data is needed ahead of trials and how to obtain that data. It quantifies assumptions and provides more objective, quantitative assessment of information. It reduces uncertainty in clinical development plans. It improves on the legacy traditions of just "following gut instincts". It helps work smarter, making better use of resources, reduces valuable time by eliminating unnecessary arms or trials. It helps identify assets for even greater impact. It helps generate opportunities while doing the necessary homework and applying latest technologies. It achieves greater probability of success, leads to more assets getting to market faster.
Importance of simulation
1) Simulate the data that might be seen in the study, the results of the proposed analysis, and decisions that would be taken as a result.
2) Generate the data to be used in the simulations from various sources including disease models, PK/ PD models, in vitro and animal experiments, reported results in literature, prior studies.
3) Simulate not just the next trial, but planned sequence of trials in order to understand the possible trade-offs (in time, allocation of resources between development phases, and impact on overall likelihood of success).
Clinical Trial Design Maturity Model
Companies should improve their capabilities along a maturity model. The following proposed model has four levels. Each build upon the next.
Level 1 – Traditional level of Trial Design: Focus is on "type-1 error control" (to avoid that we do not wrongly exclude a potential good drug), "Power" (minimum sample size to reject the null hypothesis) and "sample size". This maturity level may support a regulatory submission. However, may not be adequate even for confirmatory trials. Additional concerns are increasing the scope: Comparative effectiveness, evidence of long-term safety, identification of the correct treatment population. The traditional level of trial design does not help answer key questions such as maximum tolerable dose, minimum efficacious dose, type of patient group where the drug works.
Level 2 - Simulation: There is no alternative to "in silico" simulations for running more complex questions of trial design. Create various scenarios of how the drug may perform. Draw multiple sets of sampled observations from each scenario. Simulate the planned trial on each set. Study individual simulations to understand how the planned trials behave. Use summaries of the results over many simulations to estimate its ability to answer the questions of interest.
Trial designers can estimate a design's ability to correctly answer the key questions of the development program. They can effectively and efficiently answer more complex questions of current drug development: balancing side-effects and efficacy, identifying the right treatment population, proving comparative effectiveness and testing multiple endpoints. They can understand additional trial consideration that are either ignored or managed on an ad hoc basis (ex: effect of subjects that drop out on the statistical analysis).
Simulations enable richer interactions between trial designers and the rest of the clinical team. Individual simulations serve as case studies for discussing a proposed design's performance. It enables the team to anticipate, consider key questions and offer opinions without necessary understanding the statistics. (Ex: In case of such data results at an interim point, the team may not want to continue to recruit that sub-population). A biostatistician can then translate these opinions into modifications of the statistical design.
Level 3 – Model driven Simulation: "Model-based drug development" and "Translational medicine" leverage the significant advances in understanding of drug and disease behavior. This maturity stage incorporates this work and knowledge into the trial design process. This maturity Level 3 derives scenarios from prior data an & models including parameter values from estimated ranges, population covariates, treatment regimens. It then creates sampled responses, proposed design and operating characteristics such as success/ failure outcomes, Treatment/ population selection for Phase 3, stop/ go.
Benefits: 1) Understand better the performance of proposed trial analysis plan. It is first simulated using data that the models predicted rather than simple statistical approximations. 2) Understand better the risks of basing decisions at the end of trials or at interims, on short-term endpoints such as biomarkers, surrogates, predictors or early measures of final outcome. 3) Empower the analysis by exploiting more the correlation between different endpoints (example safety and efficacy). 4) Estimate the expectation of success of the trial. 5) Assess the extent to which the current models are well enough estimated to use them to guide development.
Level 4 – Simulating the clinical Trial program: Simulate not just the next trial, but successive clinical trials and the decisions made between them. Simulate each program strategy using the same datasets of subject outcomes generated from the early development models. Leverage the results summarized by the probability of successful registration, cost of Phase 2, cost of phase 3, probability of failure in phase 3 and the expected NPV.
Benefits: 1) enable quantitative decision-making based on estimates of key metrics for different program strategies. 2) Weigh and compare various program trade-offs on a common, comprehensive playing field. 3) Guide team discussion and decision making among the many groups involved. Break down silos by making visible the impact of each specialist's role on the program. Help each team member understand how to best contribute to the overall success of the program.
What it takes to do it…
Implement advance Clinical trial design requires notable changes to Culture, Structure, People, Processes and Technology for Biopharma organizations. It is important to first understand what behaviors need to be encouraged to support the program simulation environment. Milestones and performance targets need to change to optimize behavior across the development program. Drug development should be understood in terms of reducing, but not eliminating key uncertainties. It requires to establish a modeling & simulation Center of Excellence. People's roles, motivations, skill sets, attitudes, habits, roles, job descriptions and performance expectations will need to evolve with regards to modeling & simulations and using latest technologies. It requires a more hypothesis and quantified data empowered thinking. It requires the willingness to work with the latest technologies.
Read more on clinical trials here...
Example of how Clinical Trial design work in practice
Review development options for the candidate compound. Pre-clinical toxicity results may show less than expected toxicity, but extrapolation of animal tests to humans may indicate that higher than expected exposure is necessary for efficacy. Team brainstorms the underlying models to explain the data to date and creates scenarios of possible patterns of results in the following clinical trials. The modeling & simulation team selects the exemplar values for the parameters generates databases of virtual patients with predicted clinical endpoint values over the complete range of different treatment options. Run initial simulations of the drug development program through phase 1 to 3 using the results from the different databases of virtual patients. The probability of success, even under optimistic weightings for the scenarios may proof disappointing. The team may find reason in poor choices of treatment population and treatment regime for phase 3. This approach may help reduce uncertainties early in the program and help identify the right focus and questions to ask.
As result, the team may develop several alternative/ complementing strategies 1) Conduct a larger Phase 1 trial in patients rather than healthy volunteers and include early efficacy endpoints. 2) After Phase 1 trial, perform a small study to test toxicity in patients as the highest dose and with longer exposures; along with a second study to refine the Pharmacokinetic/ Pharmacodynamic (PK/PD) model instead of the conventional Phase 2a PoC study 3) After conventional Phase 1 trial go straight to a combined Phase 2a/b trial, testing more doses and using a number of interim looks at the data to check whether the trial should be stopped for futility.
Specify representative trials for each strategy. Define key decisions for each phase. Simulate and optimize each strategy to maximize the expected return of the program within fixed Phase 1 & 2 budgets. Calculate financial returns for each scenarios accounting for the length & cost of development, probability of success, value of selected treatment option and population.
Average the skeptics and optimists weightings for each scenario accounting for likelihood of success, probable time to registration if successful, likely cost incurred. Calculate the expected Net Present Value (eNPV) of the different strategies. Treat the figures only as relative measures of merit given the enormous uncertainty in the revenue estimate. Rank strategies. The team may find best plan for developing the compound with a 40-50% greater chance of success compared to the original drug development program.
Read more on how to successfully launch biopharma drugs
Clinical Trial Management System (CTMS)
CTMS enables better/ timelier information sharing more informed forecasting & planning, and better partner relationships. Many organizations do not have a CTMS. Their business processes evolved around non-shared personal spreadsheets of data and informal communications. Early adopters of CTMS often customized their systems heavily and potentially embedded business processes that may no longer be best practice.
A good approach to implement CTMS: 1) Agree on best practice workflow for the organization 2) Identify and eliminate redundant processes and supporting systems. 3) Get buy-in for implementation from all stakeholder groups 4) Understand data flow, eliminate data duplication, and optimize data sharing. 5) Define scope of operations 6) Develop user requirements for selected processes. 7) Draft RFP to engage with technology vendors.
Read more on how to leverage technologies for Life Science companies
Conclusion
If you are interested in this or other Healthcare / Life Sciences topics, you can reach out to me directly at alexwsteinberg@hotmail.com or WeChat (ID: alexwsteinberg2 ).
About the author: Alex Steinberg comes out of a family of doctors, scientists and other health care professionals who have dedicated their lives to improve the health & well-being of people around the world. Alex drives digital transformation, innovation and intelligent automation efforts for the largest brand companies in China.
Special credits: This article leverages extensively text, content and graphics from ResultWorks, a professional services company offering strategy innovation, integrated business process analysis, information transformation, knowledge management, and change management consulting services for the life sciences industry. Special thanks to their extremely valuable business, medical and scientific contribution!
Legal disclaimer: This article represents my personal opinion and does not reflect that of my current/ previous employer(s) or clients. The article intends to increase awareness, understanding and dialog about Health Care and Life Science issues. It does not present any offer or advice in a legal sense. Markets and technology change quickly and information gets out-of-date. The reader is advised to always seek individual analysis & consultation.