Monday 09:30 AM‑10:30 AM
A Practical Review of Real Word Data for Effectiveness Decisions – Case Studies of Using Historical Data in Developing Novel Therapies by DIA NEED

Presenter: Chenkun Wang

Abstract: It is well known that the (multiple) randomized controlled trials with concurrent control is the most frequently used design to demonstrate effectiveness and safety of an investigative therapy. However, such trial often runs into feasibility and even ethical issue in life-threatening/rare diseases where no effective therapy is available. Single-arm interventional trial with historical control (HC) as the comparator, highlighted in the recent FDA Real World Evidence (RWE) framework, provides an alternative approach to assess the effectiveness of the investigative therapy in these challenging scenarios. In recent years, quite a few novel therapies in oncology and rare diseases were developed with the support of historical data and several gained approvals based on single-arm pivotal trial with HC as comparator. This tutorial will review selected case studies, learn from their execution and regulatory experience, and provide some practical guidance on incorporating historical data from the stage of trial design to the stage of data analysis.
This tutorial consists of the following 2 parts:
• Roadmap from the decision to include HCs to analyzing data from trials with HCs
• Discussions on NDA/BLA filing cases where HC was used as the comparator in the pivotal trials

Monday 11:00 AM‑12:00 PM
Artificial Intelligence and its Impact on Clinical Trials

Presenter: Stephan Ogenstad

Abstract: Artificial intelligence (AI) is revolutionizing a number of areas in the world. To implement AI into clinical trials will of course not be without difficulties. But, apart from these difficulties what impact would AI have if taken step by step? There are three sets of questions that need to be asked to engineer an AI architecture system. First, what is the utility or reward function to be used for optimization? Second, how does the system learn? What data does it work with, what learning techniques does it use, and what prior knowledge needs to be built into it? Third, how does it maximize its expected reward?

Monday 1:00 PM-2:00 PM
Power Considerations for Clinical Trials in Presence of Multiplicity

Presenter: Kaushik Patra

Abstract: In general, the sample size of a clinical trial is determined based on the power considerations of the primary endpoint. However, important secondary endpoints play a key role in overall success of the study. This tutorial will cover relevant topics pertaining to sample size considerations when multiple endpoints (primary and secondary) are of interest. Methods will be illustrated with real examples.

Monday 2:00 PM-4:00 PM
Data Visualization for Clinical Results

Presenter: Zachar Skrivanek

Abstract: Traditional approaches to drug development rely on generating pages and pages of analysis results to assess the safety and efficacy of drug. Data visualization can succinctly communicate in a single page what might take multiple pages to communicate in tables. Often important insights into the data can only be gained through data visualization; there is no better way.

Similar considerations for creating or interpreting a data visualization have to be made that are also made for a statistical analysis. First, the interpretability of the visualization is highly dependent on the quality of the data and how the data was derived relative to the target population. Second, the visualization needs to account for any covariates and confounding factors just like in a statistical analysis.

Data visualization aims to represent the data in an unbiased way, but at the same time it should be impactful, especially when it is being used to communicate a known result. In this tutorial we will discuss pre-attentive processing which can be leveraged to make a graphic go from good to great. We will describe the hierarchies of perception that should govern the visual idiom being chosen for a particular data visualization. We will make recommendations for color palettes, aspect ratios and other factors that need to be considered when creating a data visualization.

Monday 4:00 PM-5:00 PM
Assessing Treatment Effect Using Propensity Score Matching within the U.K. Population of Crohn's Disease Patients

Presenter: Laura Gunn

Abstract: Randomized controlled trials (RCTs) are a ‘gold standard’ for estimating minimally unbiased treatment effects on health outcomes. However, RCTs are not always feasible and population-based observational studies may be more appropriate, particularly involving health analytics using big data. Assigning individuals at random between groups is not always feasible for ethical/practical concerns. Individuals ‘assign’ themselves to groups – e.g., studies about the impact of smoking cessation programs (SCPs) on blood pressure. Those who decide (not) to undertake such SCP intervention are self-selected, and biases/confounding factors (e.g., income, medical history, ethnicity, gender, age, education) may influence decisions (not) to attend SCPs, leading to potentially biased biostatistical inferences on this type of observational data. However, propensity score matching (PSM) reduces such study design biases.

PSM categorizes quantitatively individuals based on confounding factors associated with their decisions, so that samples of individuals ‘assigned’/matched to intervention(s) and control are similar/balanced across these factors. This transforms observational studies into pseudo-RCTs. For each person self-selected to the intervention, we ‘match’ M individuals with similar confounding characteristics who chose not to undertake the intervention. PSM relies on classification methods to ‘match’ individuals.

The Clinical Practice Research Datalink (CPRD) contains clinical and prescribing data for over 13 million patients in the United Kingdom; participating primary care practices are subject to regular audit to ensure data accuracy and completeness, allowing epidemiological studies of this data to be feasible. Since RCTs evaluating the impact of thiopurine treatment on Crohn’s disease patients is not practical, we used CPRD data to identify 5,640 patients with incident Crohn’s cases diagnosed over a 17-year period with at least an additional 5-year follow-up. Propensity score matching (PSM) is used to reduce bias, obtained in estimates of treatment effects as a result of confounding, between baseline factors and exposure group status. This presentation describes the PSM process, and applies optimal PSM, with a sensitivity analysis implementing additional matching techniques, using data collected from this nationally representative UK population-based study, where impact of duration and timing of thiopurine treatment on the likelihood of surgery is assessed using a Cox proportional hazards model and PSM.

Tuesday
8:30:00 AM-9:30:00 AM
Adaptive Design with Recurrent Event Outcome

Instructor: Pranab Ghosh

Abstract: Composite endpoints combining several recurrent events with clinical interest of similar types often define the primary outcome in cardiology trials. Different events might include time to death, cardiovascular related hospitalizations and others. Commonly used analysis approach counts the number of events within a time period, which may follow a Poisson or a negative binomial distribution. However such approaches might not be the most efficient ones to assess the treatment benefit in certain clinical settings. Recurrent-event analysis using Andersen-Gill model or time-to-first event using Cox model are the two frequently used methodologies for this type of problems. But other statistical methods like negative binomial model and win-ratio test are also valid statistical methods in different clinical trial settings. Our work concentrates on a systemic comparison of these models with application to composite endpoints from different clinical trial settings. We also extends this assessment to adaptive design with possible sample size modification based on interim results.

Tuesday 09:30 AM-10:30 AM
Taking Adherence, PROs and RWD into Account in Clinical Trials

Presenter: Alan Menius

Abstract: With the FDA encouraging Direct-to-Patient clinical trials and using real world data to help augment what is known about a medicine’s efficacy and safety, various technologies are now being piloted to collect real world data. Collecting patient reported outcomes (PROs) in real world settings has become increasingly important to help regulators and payers understand how new medications impact overall quality of life. This information becomes incredibly important to regulators and payers when trying to determine the benefit/risk ratio of a new drug compared to treatments already available to patients.

The ability to collect data 24x7 in every conceivable environment may seem like a panacea to some, but using these new technologies also greatly increases the variety, volume and complexity creating a conundrum of how best to use these data to achieve the primary goals of a particular study. While potentially useful for patients, using too many technologies in a single study to collect data might also introduce patient burden, potentially leading to an increased numbers of dropouts.

Even with the move towards patient-centric studies, adherence remains a very real issue, increasing the number of patients needed to maintain the numbers needed for a successful study as well as potentially impacting the measured efficacy of a new medicine.

This session will explore some of the new technologies and data being implemented in direct to patient trials as well as some strategies for using these data as end-points or supporting information. We will also review the use of clinical trial simulations as a tool to help clinical teams determine the potential impact of adherence and persistence on clinical studies and explore how correlations between adherence, ePRO answers provide powerful information provide powerful inputs for predictive analytics and population health insights.

Tuesday 10:30 AM-11:30 AM
Aggregate Safety Analysis for Causality Assessments of SAEs

Presenter: Lothar Tremmel

Abstract: In order to comply with evolving rules for suspected unexpected serious adverse reactions (“SUSARs”), comprehensive ongoing surveillance of accumulating pre-marketing safety data has become necessary. This presentation gives a brief orientation about reporting regulations, and introduces the ASAP ( aggregate safety analysis plan) as a tool to proactively address the logistical and statistical challenges of this type of data monitoring. Practical experience with implementation for a large, ongoing cardiovascular outcomes trial will also be shared.

Tuesday 01:00 PM-02:00 PM
Overview of Some Key Methods for Observational Settings

Presenters: Shankar Srinivasan

Abstract: We will review issues, concerns and some advantages of using observational data to support conclusions of effect in a group of interest against a control, where there are biases involved in the assignment of subjects to the groups. Propensity score methods for observational data will be presented with emphasis on matching and inverse weighting to remove bias in comparisons to control, and in more general settings involving multiple groups. Other issues governing the use of real world data including missing data will be addressed. Some details on sizing and planning the conduct and analyses of real world data will be presented.

Tuesday 02:00 PM-03:00 PM
Innovative Designs for Drug Development in Rare Diseases

Presenter: Qing Liu

Abstract: The small sample sizes associated with studies of rare diseases can restrict study design options, replication, and the use of inferential statistics, which means that novel and innovative statistical designs may need to be considered to assist in assessing the evidence of the efficacy and safety of a potential treatment. Enrichment, where patients are enrolled on the basis of a prospectively defined characteristic that is believed to improve the probability of detecting of a treatment effect compared with an unselected patient population, is one option. Incorporation of real-world evidence (RWE) of patients receiving a standard of care (SOC) as a control, in the form of internal or external historical or concurrent control, is also critically important for quantifying benefits or risks pertinent to patients as well as increasing the probability of success. A patient centric approach focuses quantifying treatment outcomes of new medical product for each individual patient, either through traditional clinical evaluations or patient reported outcomes. These considerations as well as the practicality of rare disease clinical trials for a particular setting may lead to a unique design choice from a list of options including randomized parallel (blinded or unblinded) group design and single arm trial design with either internal or external natural history controls. In addition to the traditional parallel group design, variations such as a randomized delayed start design (RDS design) or a randomized enrichment design with internal RWE control (RWE-RE design) can also be used. The RDS design, which is a suitable for patients with relatively stable disease condition over the duration of the trial, consists of two stages: for stage 1, patients are randomized to receive a new treatment or a control; for stage 2, patients who received control in stage 1 switch to the new treatment. The RWE RE-design also has two-stages: the first stage is an open label observational study of RWE of a SOC over a suitable duration to quantify disease progression; patients from stage 1 who meet outcome driven enrichment criteria are randomized to receive a new treatment or remain on the SOC. Both the RDS design and RWE-RE design are used for confirmatory rare disease clinical trials.

Tuesday 03:30 PM-04:30 PM
What is Next? Implementing a Multiattribute Model for Evaluating Alternative or Additional Indications in Drug Development

Presenter: Jack Knorr

Abstract: At some point in a drug's development, researchers have to determine which indications they still explore with the drug. The choice often requires assessing a large number of factors, including those pertaining to scientific evidence, commercial viability, and operational feasibility. The factors of importance are often subjective in nature, require a wide variety of subject-matter experts, and vary across different programs. As a result, a statistician is well equipped to quantify and aggregate these factors into usable metrics. In this talk, we will review a published method of prioritizing compounds within discovery chemistry projects. We will then present an extension of this method to prioritize indications for a single compound. Strategies for eliciting information from multiple sources and for assessing uncertainty and variability in the inputs for the model will be explored.

Wednesday 09:30 AM-10:30 AM
New Adaptive Design Guidance

Presenter: Gregory Levin

Abstract: In this presentation, I will describe the new FDA draft guidance on adaptive designs. The 2018 document, which replaced the 2010 draft, provides guidance on the appropriate use of adaptive designs for clinical trials to provide evidence of the effectiveness and safety of a drug or biologic. I will describe the key principles for designing, conducting, analyzing, and reporting the results from a clinical trial with an adaptive design. I will also address a number of special topics, such as the use of simulations in adaptive design planning and the use of Bayesian adaptive design features.