Optimizing the Interdisciplinary Design of Adaptive Platform Trials

Adaptive platform trials (APTs) are often complex clinical trials that, ideally, are well suited to answer the motivating clinical questions effectively and efficiently, with the motivating clinical questions and associated treatment arms expected to evolve over time as evidence accumulates. Recently, APTs have played a pivotal role in informing public health policy by efficiently generating compelling evidence regarding the effectiveness of therapies for COVID-19. For APTs to be maximally effective in informing future public health policy, they must be carefully tailored to address the right clinical questions, with the right balance of size, scope, rigor, and flexibility. The design process requires input from clinical and statistical domain experts and often includes input from trial implementation personnel, ethicists, and patient representatives. The design process is inherently iterative, with proposed designs evaluated through trial simulation, the identification of strengths and weaknesses of the proposed design, and revision by the team to address weaknesses. This iterative design process requires effective communication and collaboration between the statistical and clinical domain experts. This session is intended to present a current best practice in facilitating and enhancing the collaborative design process for APTs, including how best to present simulation-based trial performance to the design team and ensure effective interdisciplinary communication. The speakers have extensive experience in leading the design of APTs across multiple therapeutic areas, in both academic and industry settings. The session will begin with a brief presentation by Dr. Lewis on the basic structure of an APT and the tasks and challenges associated with the multidisciplinary design process. The subsequent discussion will be organized by the following themes: (1) considerations in the selection of the study population and primary outcome metric;(2) selecting treatment domains and factors to be compared;(3) trial simulation and communication of performance metrics to both statistical and non-statistical team members;and (4) defining and calibrating interim decision rules. Each of the 4 panel members will outline a recommended approach to facilitating 1 of the 4 design tasks, with examples drawn from their experience. The remaining time (15 min) will be available for a panel question-and-answer period. At the end of the session, the audience will have an understanding of the general organization of, and a process for facilitating, the design process for an adaptive platform trial. Panel Members Roger J Lewis, MD, PhD, is a Senior Physician in the Los Angeles County Department of Health Services, Professor of Emergency Medicine at the David Geffen School of Medicine at UCLA, and the Senior Medical Scientist at Berry Consultants, LLC, a group that specializes in innovative clinical trial design. He is also the former Chair of the Department of Emergency Medicine at Harbor-UCLA Medical Center. Dr. Lewis' expertise centers on adaptive and Bayesian clinical trials, including platform trials;translational, clinical, health services and outcomes research methodology;data and safety monitoring boards, and the oversight of clinical trials. Dr. Lewis was elected to membership in the National Academy of Medicine in 2009. He has authored or coauthored over 270 original research publications, reviews, editorials, and chapters. Dr. Lewis is a Past President of the Society for Academic Emergency Medicine (SAEM) and served on the Board of Directors for the Society for Clinical Trials. He is a fellow of the American College of Emergency Physicians, the American Statistical Association, and the Society for Clinical Trials. Juliana Tolles, MD, MHS, is an Assistant Professor of Emergency Medicine at the Harbor-UCLA Medical Center and the David Geffen School of Medicine at UCLA, and a Medical and Statistical Scientist at Berry Consultants, LLC. Her academic research interests include emergency medical services, resuscitation medicine, and trau a care. She has authored several reviews for Journal of the American Medical Association (JAMA) on statistical methodology and has lectured nationally on research methodology for the Society for Academic Emergency Medicine Advanced Research Methodology Evaluation and Design (ARMED) course. She is also a co-investigator for the Strategies to Innovate Emergency Clinical Care Trials (SIREN) network Southern California site. Kert Viele, PhD, is a Director and Senior Statistical Scientist with Berry Consultants, where he leads Berry Consultants' research enterprise. He is a leader in clinical trial implementation of Bayesian hierarchical modeling, with expertise in platform and basket trials as well as clinical trials incorporating the use of historical information. Prior to joining Berry Consultants in 2010, he was a faculty member at the University of Kentucky, where he received the Provost's Award for Outstanding Teaching and was an investigator for NSF and NIH-funded research. He has developed over 100 custom Bayesian adaptive clinical trials for clients in industry, government, and academia, and currently serves on several data safety monitoring boards for randomized clinical trials. A former editor of the journal Bayesian Analysis, Dr. Viele is also an author of FACTS (Fixed and Adaptive Clinical Trial Simulator), clinical trial simulation software currently licensed to multiple pharmaceutical, academic, and government organizations. William Meurer, MD, MS, is an Associate Professor of Emergency Medicine and Neurology at the University of Michigan Health System. In addition, he serves as a Medical and Statistical Scientist for Berry Consultants, LLC. He works to improve the care of patients with acute neurological disease both through his work on the acute stroke team and as a researcher. His work in the field focuses on the design of clinical trials with adaptive and flexible components. In addition, he is a principal investigator of the National Institutes of Neurological Disorders and Stroke (NINDS) Clinical Trials Methodology Course (http:// neurotrials.training) and a co-investigator in the clinical coordinating center of the Strategies to Innovate Emergency Care Clinical Trials (SIREN) network- also funded by NIH). He was a co-investigator on the Adaptive Designs Accelerating Promising Treatments into Trials (ADAPT-IT) project, as part of the FDA Advancing Regulatory Science initiative with NIH.

Identifying and Examining Ethical Considerations Facing Data Monitoring Committees

The COVID-19 pandemic has exposed numerous unresolved research ethics challenges particularly for Data Monitoring Committees (DMCs). DMCs have worked to ensure the ongoing social value of research as rapid changes occur in health policy and epidemiology and there is substantial pressure to release early findings to the public. Unlike Institutional Review Boards, DMCs are charged with carefully monitoring ongoing research, but with limited ethical guidance and often without representation from all host countries. This article highlights ethical challenges for DMCs and lessons learned from the COVID-19 pandemic. DMCs have long faced high-stakes decisions in clinical trials including whether to continue, modify, or terminate a trial based on emerging trial data. Trial protocols, statistical analysis plans, and data monitoring charters establish principles for DMC decisionmaking;however, there has not been a great deal of systematic examination of the ethical issues faced by DMCs. For example, which ethical considerations should be addressed by DMCs as opposed to Institutional Review Boards or researchers is often unclear. Formal guidance rarely addresses whether DMCs should monitor the representativeness of trial participants as compared with the target population for the intervention. Furthermore, post-trial issues have received limited attention. Should DMCs ensure the accuracy of press releases and manuscripts detailing study findings? How should DMCs determine when to unblind participants after a study is over if it is relevant for their medical decision-making? In this presentation, we will report preliminary results of a qualitative study of DMC members (i.e. statisticians, clinicians, and ethicists). We will highlight persistent controversies, the range of roles DMCs are expected to play in monitoring clinical trials, and variation in formal guidance about the ethical obligations of DMCs. We will also examine the question of whether and when ethicists should serve on DMCs. We will conclude by identifying critical ethical issues facing DMCs that warrant further attention.

Real-World Experience in Conducting Adaptive Platform Trials

Platform trials like RECOVERY, SOLIDARITY, REMAP-CAP, TOGETHER, PRINCIPLE, and many others have dominated the COVID clinical trial effort, changing clinical practice with multiple results in The Lancet, New England Journal of Medicine, and the Journal of the American Medical Association. These large-scale international efforts have enrolled tens of thousands of patients and explored dozens of potential therapies for mild, moderate, and severe COVID-19. These trials are the result of a decade of theory and practice building on the experience of oncology platform trials such as I-SPY2. In addition to COVID-19 and oncology, platform trials are now used or planned in Alzheimer's disease, amyotrophic lateral sclerosis (ALS), antibiotics, psychiatry, and many other clinical areas. Conducting a large-scale platform trial is daunting. While recent platform trial review papers have hundreds of references on the theoretical design issues and research potential of these trials, there is no substitute for the actual real-world experience of implementing an adaptive platform trial, made all the more challenging within a fast-moving global pandemic. This proposed session will investigate the challenges and solutions for successfully executing a platform trial. The proposed speakers bring decades of combined expertise from executing platforms such as ISPY-2, GBM-AGILE, REMAP-CAP, TOGETHER, PRINCIPLE, ANTICOV, HEALEYALS, Precision Promise, and others. They will describe many of the challenges specific to these large global platform trials, and the infrastructure and process needs that underpin these complex trials. The session will consist of 4 speakers with integrated 20-min talks, followed by a question-and-answer period as time allows. Dr. Ed Mills will provide an overview of platform trials and their challenges in relation to simpler clinical trials. Dr. Michelle Detry will sharpen the focus on the complexities specific to trial execution and the interactions between various stakeholders, including publication plans that must account for the perpetual nature of many platform trials. Dr. Anna McGlothlin will then discuss the specific requirements and role of the committee performing the actual interim analyses, and, finally, Dr. Hong Yu will discuss these challenges in the context of the HEALEY-ALS platform trial. All speakers will include examples from their rich realworld experience in implementing complex adaptive platform trials. Speakers (in proposed order) Ed Mills, Cytel COVID-19 has exemplified the utility of platform trials for making clinical and public health decisions in a timely manner. The most useful trials that emerged during the pandemic have been from platform trials. However, with the enthusiasm for platform trials comes the concern that trials should be implemented using methods that many groups are unfamiliar with, such as advanced statistical approaches and implementation of quick changes to the protocols. Challenges exist in interactions with funders, partners managing data sets, and clinical users. This session will use real-world experiences of platform trials in the pandemic to exemplify the utility and challenges of this new approach to clinical evaluation. This talk is for any audience with an interest in clinical trials and will address strategies to promote the use of platform trials while also highlighting the concerns about the quick adoption of this method. Michelle Detry, Berry Consultants Adaptive platform trial designs include interim analyses for pre-specified adaptations, sharing of control arm information, and a perpetual design where investigational agents enter and leave the trial at different time points. This talk will discuss the infrastructure considerations for a perpetual platform trial, ongoing statistical support from both blinded and unblinded statistical teams, setting clear communication channels and firewalls between blinded and unblinded teams, and the role of Data and Safety Monitoring Committee (DSMC) and their interaction with the independent Statistical Analysis Committe that implements the protocol-specified interim analyses. In addition, ongoing platform trials face unique challenges in determining ''who knows what, when.'' Dr. Detry will discuss planning for public releases of results, and data sharing in situations where an investigational arm may complete their trial arm participation, but the control arm data will still be used for ongoing investigational arms. Anna McGlothlin, Berry Consultants A Statistical Analysis Committee (SAC) is a team of unblinded statisticians tasked with performing the interim analyses for an adaptive trial. The unblinded SAC must have expertise in the statistical methodology being utilized and in the complexities of adaptive designs. This talk will describe the role of the SAC during the preparation and conduct of a platform trial. Responsibilities of the SAC include performing analyses in accordance with the pre-specified design, monitoring each analysis to ensure that the adaptive algorithm is performing as intended, building semi-automated reproducible reports to facilitate quick turnaround of interim results, and partnering with the Data and Safety Monitoring Board (DSMB) to aid in the interpretation of interim results. In addition, platform trials may have the unique aspect that arms may complete their final read out while other arms continue in the trial. In some cases, the fully unblinded SAC may be the only group that has complete access to the necessary data to perform the final analysis for an arm while other arms continue. Hong Yu, Massachusetts General Hospital A complex adaptive platform trial requires complex infrastructure, and many of the challenges are not revealed when conceiving or planning the trial, but only in the face of actual implementation. In this talk, we will discuss real-world experiences in developing infrastructure for adaptive platform trials, particularly the HEALEY-ALS platform. The specific challenges to be investigated are the required personnel to implement a platform, the system architecture required to support a perpetual trial design beyond the initial set of therapies, and the monitoring and management plans required to maintain robust data throughout multiple interim analyses and periodic reporting of results. Platform trials must be nimble to fulfill their goal of efficiently exploring multiple therapies as quickly as possible. These infrastructure solutions allow the trial to adapt to changing arms, maintain data quality as well as trial integrity, and support multiple sets of results and publications throughout the trial's duration.

Covid Icu Bed Occupancy Inversely Related to the States' Booster Rates: A Myth or a Reality

INTRODUCTION: The current CDC guidelines recommend COVID-19 vaccine boosters for all eligible individuals to enhance protection. Resources have been allocated to research done regarding the COVID-19 vaccine, and we speculate that there is a correlation between COVID booster rates and number of COVID patients in the ICU. We hypothesize that the states with a higher percentage of the population that received the booster shot will have decreased COVID ICU bed utilization and vice versa. METHOD(S): The percentage of people who received the COVID-19 booster vaccine and the number of ICU beds occupied by patients with COVID-19 per 10,000 population, both stratified by states, were reviewed to determine the pattern of correlation. The data for both the variables was sourced from Becker's Healthcare as it used information from the CDC's data tracker to rank states by their booster rates. The rankings were last updated based on data from July 20th, 2022. The state of Idaho was excluded because the data was not available. Limitations of the study included reporting lags between the states and CDC, the emergence of numerous variants of the virus, and a lack of a standardized timeline across the states. RESULT(S): Pearson Correlation Coefficient was used to determine the pattern of correlation between COVID booster rates and the number of COVID patients in the ICU for all US states. Booster rates was set as x and ICU patients was set as y. The data was analyzed while using the formula r = SIGMA((X - My)(Y - Mx)) / ((SSx)(SSy)). X Values were calculated with SIGMA = 2407.7, Mean = 48.154 and SIGMA(X - Mx)2 = SSx = 2308.544. Y Values were calculated with SIGMA = 5112, Mean = 102.24 and SIGMA(Y - My)2 = SSy = 835103.12. The coefficient of determination, R2, was 0.0611. Our obtained R was -0.25 which means no strong correlation was found. The data was analyzed independently by two statisticians and the same results were obtained. The results failed to confirm our hypothesis and suggested that there was no correlation between COVID booster rates and the number of COVID patients in the ICU. CONCLUSION(S): Based on our results, no correlation was found between the states' COVID booster rates and ICU bed occupancy. Further studies are needed to quantify this association if any as highly virulent COVID strains pose a threat to humanity.

COMPARATIVE ANALYSIS OF EPIDEMIOLOGICAL, FORECASTING AND DEEP LEARNING MODELS FOR COVID19 SPREAD IN INDIA

Covid19 is affecting across many nations and most population of the world. As per WHO there are 270million confirmed with about 5.3 million fatalities as on December 15th, 2021. Many governments, organizations and local bodies have been applying various models in order to estimate the disease spread and appliede varied strategies to curb the spread. There are many models proposed by mathematicians and statisticians for the same. In the current work a comparison is done with mathematical disease spread models SIR, SIRD, classic time series forecasting modelARIMA, and artificial neural network models RNN, LSTM with Covid19 India data. The study investigates the effect of disease containment policies and vaccination drives for Covid19 data in the context of India using SIR Model. All the models are built for multiple time prediction windows starting from 5 days up to 45 days. The models are evaluated with MAE, MAPE and RMSE for multiple states and India level data. It is inferred that the prediction time of 5 days has best results for SIR model. The ARIMA model can predict withacceptable performance up to 30 days. RNN and LSTM models can predict for 5 days within acceptable performance. The best model that can predict longer durations and has good performance is ARIMA model. A detailed report on the model details and performance is the outcome of this study.

Net treatment benefit of OSE2101 in HLA-A2+ non-small cell lung cancer (NSCLC) patients after failure to immune checkpoint inhibitors (IO) in phase III Atalante-1 randomized trial

Background: OSE2101 (Tedopi) is an anticancer vaccine that increased overall survival (OS) (HR 0.59, p=0.017) versus Standard of Care Chemotherapy in the population of interest (PoI N=118) of patients with IO secondary resistance after sequential CT-IO (ESMO 2021 #47LBA). The Net Treatment Benefit (NTB) is an original method combining efficacy and safety endpoints to test the overall improvement in health outcome between 2 treatments (Buyse M. Stat Med 2010). NTB was assessed in the overall population (N=219) from whom OS improvement of OSE2101 (HR 0.86, p=0.35) was lower than in PoI. Methods: NTB was tested by comparing prioritized outcomes using Generalized Paired Wise Comparisons (GPC). The prioritized outcomes were OS, then time to worsening ECOG (threshold=2 months) followed by severe adverse events, progression free survival (shorter vs. longer than 2 months) and Quality of Life (threshold=5 points on Global Health Status of EORTC-QLQC30). Analysis was stratified using the 3 strata of the study (histology, best response to 1rst line, line of prior IO) and enrollment time (before vs during COVID-19). Sensitivity analyses used no stratification, different thresholds of clinical relevance and PoI. Results: In the primary analysis (1088 pairs), NTB was 19% and reached statistical significance in favor of OSE2101 (p=0.035). In unstratified analysis (11120 pairs), NTB was 11% (p=0.188). In the PoI (388 pairs), NTB was 22% (p stratified=0.074) and 28% (p=0.014) in unstratified analysis (3040 pairs). Although the primary analysis was statistically positive, results were not consistent in some sensitivity analyses due to the limited sample size and the impact of stratification factors. Conclusions: An overall improvement in health outcome was observed with OSE2101 in the overall population of advanced NSCLC after IO failure with a NTB of 19% over SoC. In PoI with IO secondary resistance after CT-IO, the NTB was 22%. Post-hoc analyses are ongoing intended to explain the variability of NTB and will be detailed. Clinical trial identification: EudraCT: 2015-003183-36;NCT02654587. Editorial acknowledgement: We thank Pierre Attali (Medical Expert, MD) for his support in the writing of the . Legal entity responsible for the study: Ose Immunotherapeutics. Funding: Ose Immunotherapeutics. Disclosure: M.E. Buyse: Financial Interests, Personal, Officer, Chief Scientific Officer: IDDI;Financial Interests, Personal, Invited Speaker, Board Member: CluePoints;Financial Interests, Personal, Stocks/Shares: IDDI, CluePoints. F. Montestruc: Financial Interests, Personal, Member of the Board of Directors, CEO of the Company: eXYSTAT SAS;Financial Interests, Institutional, Other, Statistician Consultant: AbbVie, Biocodex, Geneuro, Gensight, Guerbet, Imcheck, Ose Immunotherapeutics, Pfizer, Takeda;Non-Financial Interests, Personal, Other, Statistician Consultant and Training: Institut Pasteur. J. Chiem: Financial Interests, Personal, Full or part-time Employment: IDDI. V. Deltuvaite-Thomas: Financial Interests, Personal, Full or part-time Employment: IDDI. S. Salvaggio: Financial Interests, Personal, Full or part-time Employment, Working as a statistician: International Drug Development Institute. M.R. Garcia Campelo: Financial Interests, Personal, Advisory Role: Roche/Genentech, MSD Oncology, AstraZeneca, Bristol-Myers Squibb, Pfizer, Novartis, Takeda, Boehringer Ingelheim, Janssen Oncology;Financial Interests, Personal, Speaker’s Bureau: Roche, AstraZeneca, Bristol-Myers Squibb, Pfizer, Novartis, Takeda, Boehringer Ingelheim, MSD Oncology, Sanofi/Aventis, Janssen Oncology, Amgen;Financial Interests, Personal, Other, Travel, Accommodations, Expenses: Roche/Genentech, MSD Oncology, Pfizer. F. Cappuzzo: Financial Interests, Personal, Invited Speaker: Roche, AstraZeneca, BMS, Pfizer, Takeda, Lilly, Bayer, Amgen, Sanofi, PharmaMar, Mirati, Novocure, OSE, and MSD;Financial Interests, Personal, Advisory Board: Roche, AstraZeneca, BMS, Pfizer, Takeda, Lilly, Bayer, Amgen, Sanofi, Mirati, PharmaMar, Novocure, OSE, Galecto and MS . S. Viteri Ramirez: Financial Interests, Personal, Advisory Board: Merck Healthcare KGAA Germany, Bristol Myers Squibb S.A. U, Puma Biotechnology;Financial Interests, Personal, Invited Speaker: Takeda Farmaceutica España SA, MSD de España SA, AstraZeneca Farmaceutica Spain, Roche Farma SA;Financial Interests, Personal, Expert Testimony: Reddy Pharma Iberia SAU. W. Schuette: Financial Interests, Personal, Other, Honoraria: Roche, MSD, Novartis;Financial Interests, Personal, Advisory Role: Roche, MSD, Novartis. A. Zer: Financial Interests, Personal, Invited Speaker: Roche, BMS, MSD, Takeda, Pfizer, Novartis;Financial Interests, Personal, Advisory Board: AstraZeneca, Steba, Oncohost;Financial Interests, Personal, Stocks/Shares: Nixio;Financial Interests, Institutional, Research Grant: BMS. S. Comis: Financial Interests, Personal, Full or part-time Employment: Ose Immunotherapeutics. B. Vasseur: Financial Interests, Personal, Full or part-time Employment: Ose Immunotherapeutics;Financial Interests, Personal, Other, Actions: Ose Immunotherapeutics. R. Dziadziuszko: Financial Interests, Personal, Advisory Board: Roche, AstraZeneca, Seattle Genetics, Pfizer, Takeda, Regeneron, MSD, Bristol Myers-Squibb, PharmaMar, Bayer;Financial Interests, Personal, Invited Speaker: Boehringer Ingelheim, Foundation Medicine;Financial Interests, Personal, Expert Testimony: Novartis;Financial Interests, Personal and Institutional, Invited Speaker: Roche, AstraZeneca, MSD, Amgen, Celon Pharma, Pfizer, Novartis, Brsitol Myers-Squibb, Eli Lilly, Loxo;Financial Interests, Invited Speaker: BeiGene, Ardigen, Ose Immunotherapeutics;Financial Interests, Personal and Institutional, Other, Subinvestigator and ad hoc Consultant: PDC* line Pharma;Non-Financial Interests, Institutional, Product Samples: Novartis, Pfizer, AstraZeneca, Roche;Other, Travel: Roche, Bristol Myers-Squibb, AstraZeneca. G. Giaccone: Financial Interests, Personal, Advisory Board: Novartis;Financial Interests, Institutional, Research Grant: Karyopharm. B. Besse: Financial Interests, Institutional, Funding: 4D Pharma, AbbVie, Amgen, Aptitude Health, AstraZeneca, BeiGene, Blueprint Medicines, Boehringer Ingelheim, Celgene, Cergentis, Cristal Therapeutics, Daiichi Sankyo, Eli Lilly, GSK, Janssen, Onxeo, Ose Immunotherapeutics, Pfizer, Roche-Genentech, Sanofi, Takeda, Tolero Pharmaceuticals;Financial Interests, Institutional, Research Grant: Chugai Pharmaceutical, EISAI, Genzyme Corporation, Inivata, Ipsen, Turning Point Therapeutics. E. Felip: Financial Interests, Personal, Advisory Board: Amgen, AstraZeneca, Bayer, BeiGene, Boehringer Ingelheim, Bristol Myers Squibb, Eli Lilly, Glaxo Smith Kline, Janssen, Medical Trends, Merck Sharp & Dohme, Pfizer, Puma, Sanofi, Takeda, Merck Serono, Peptomyc, Regeneron, Syneos Health, F. Hoffmann-La Roche;Financial Interests, Personal, Invited Speaker: AstraZeneca, Bristol Myers Squibb, Eli Lilly, Medscape, Merck Sharp & Dome, Peervoice, Pfizer, Springer, Touch Medical, Amgen, F. Hoffmann-La Roche, Janssen, Medical Trends, Merck Serono;Financial Interests, Personal, Invited Speaker, Independent member: Grifols;Financial Interests, Institutional, Invited Speaker, Clinical Trial: F. Hoffmann-La Roche Ltd, Merck Sharp & Dohme Corp, AstraZeneca AB, Daiichi Sankyo Inc, Exelixis Inc, Merck KGAA, Janssen Cilag International NV, GlaxoSmithKline Research & Development Limited, AbbVie Deutschland GmbH & Co KG, Novartis Farmaceutica SA, Bayer Consumer Care AG, Takeda Pharmaceuticals International, Boehringer Ingelheim International GmbH, Pfizer S.L.U., Amgen Inc, Bristol-Myers Squibb International Corporation (BMS), Mirati Therapeutics Inc;Non-Financial Interests, Leadership Role, President Elect (2021-2023): SEOM (Sociedad Espanola de Oncologia Medica);Non-Financial Interests, Member, Member of ESMO Nominating Committee and Compliance Committee: ESMO;Non-Financial Interests, Leadership Role, Member of Board of Directors and the Executive Committee (2017-Sept 2021): IASLC (International Association for the Study of Lung Cancer);Non-Fina cial Interests, Member of Scientific Committee: ETOP (European Thoracic Oncology Platform). All other authors have declared no conflicts of interest.

THE STATISTICS OF HUMAN SEXUALITY

INTRODUCTION AND OBJECTIVE: In 2015, English statistician and academic David Spiegelhalter wrote a book backed by the History of Medicine of the Wellcome Foundation entitled Sex by Numbers: What Statistics Can Tell Us About Sexual Behaviour. This represents an investigation into the statistics of human sexual activity to update the statistics of Kinsey and show the startling influence by the COVID-19 pandemic. METHODS: The investigation of the statistics of sexual activity is as enlightening as it is entertaining and there is a plethora of literature on modern sexual practices. This represents a concerted effort to track down these numbers and this data. The largest pornographic site, Pornhub®, also tracks its own statistics since its founding in 2007. RESULTS: An intriguing statistic is that the average number of sexual partners is about 9.9 (6.6 at Kinsey Center) for males and about 3.4 (4.3 at Kinsey) for females in their lifetimes. The exception to this is with homosexual persons, where the averages are much higher. Nature versus nurture is the classic tale of which is more important, genetics or environment, but most often there is a bit of both behind the scenes. Sexual activity is one of the most difficult topics of historical significance, because it is interleaved with so many socio-religious overtones. The average male loses his virginity at age 16.9, compared to age 17.4 for females. About 1 in 10 married adults sleep alone and not with their married spouse - shades of Dick Van Dyke and Mary Tyler Moore in The Dick Van Dyke Show. Only 29% of females achieve orgasm during sexual encounters, compared to 75% in their male counterparts. Currently, 66% of male college students report having “friends with benefits.” Also, currently 50% of sexually active men and women are infected with HPV at some point in their sexual lives. Sexual activity burns about 100-200 calories in males and about 69 calories in females but the heart rate at orgasm is about 140 and equal in both sexes. CONCLUSIONS: Spiegelhalter dedicated his statistical analysis “to everyone in history who has struggled with sex. And eventually called it a draw.” There is something like 900,000,000 acts of just heterosexual intercourse per year in Great Britain alone or roughly 100,000 per hour. This can all be extrapolated to the 7 billion humans around the world making close to 166,667 copulations per minute (or almost 4,000 per second).

A bayesian exact platform design for multi-arm, phase 2 trials with historical control

The COVID-19 pandemic and associated urgency for rapid testing of new or re-purposed therapeutic agents have exposed many shortcomings of the traditional one-drug at a time approach for clinical trials. Adaptive platform trials refer to trials designed to evaluate multiple interventions (targeting the same endpoints and the same population), with the flexibility to allow for accelerated selection of promising therapeutic agents, early removal of inefficient treatments, and adding new trial arms throughout the duration of the trial. Some form of adaptive randomization can be possibly implemented in adaptive platform trials to allocate more patients to interventions that are performing better than the other interventions (including control). Hence, adaptive platform trials provide an excellent, resource saving alternative compared to more traditional clinical trial designs. Our Bayesian exact design allows for the continuous learning of efficacy and safety profiles from incoming data and with built in safety and efficacy stopping rules, we can accomplish the objectives of safety monitoring, preliminary efficacy assessment, and confirmatory testing within the same trial. Such a design is ideal in situations like the COVID-19 pandemic when many potential antiviral, anti-inflammatory, and anticoagulant therapeutics are available for treatment of similar conditions but have not yet been tested for the population affected by the novel disease. Compared to the existing methods, our proposed design is innovative in the following aspects: (1) our method allows for the joint monitoring of multiple endpoints such as hospitalization, ventilation use, and other co-primary endpoints, thereby increasing the statistical power of the study. (2) In addition to implementing early stopping rules for efficacy and futility, the proposed design also allows the trial to be stopped early for safety based on already accrued data. (3) Traditionally, statisticians have to rely on complex and time-consuming simulations to develop the statistical design and estimate the sample size of platform trials, our method is based on a recursive relationship that calculates the exact probability of stopping the trial for any cause at any stage of the trial, without relying on simulations, ultimately making our design more rigorous. (4) This newly discovered recursive relationship allows for adjusting the adaptive randomization, further increasing statistical power of the trial. (5) Our adjusted adaptive randomization is based on Bayesian Play, the winner strategy allowing for the patients to be allocated into the most promising arm.

Design and implementation of clinical trials in a pandemic-reflecting on the adaptive COVID-19 treatment trial studies

This 1-h session will use examples from the Adaptive COVID-19 Treatment Trial to explore the unique challenges that the pandemic setting raises for clinical trials. The session will begin with three 10-min talks from statisticians on the Adaptive COVID-19 Treatment Trial study team. These talks, described in more detail below, are arranged to outline statistical considerations for trial design, trial implementation, and effective communication of trial results. These will be followed by 7-min talks from two discussants: one member of the Adaptive COVID-19 Treatment Trial Data and Safety Monitoring Board and one member of the National Institute of Allergy and Infectious Diseases COVID-19 Treatment Guidelines Panel. The session is designed to highlight lessons learned by the Adaptive COVID-19 Treatment Trial team and to explore applications of those lessons to other trials in similar settings. The session will conclude with 10 min for questions and comments. Dr Lori Dodd, blinded statistician for the Adaptive COVID-19 Treatment Trial protocol, will be the speaker for talk one-''Planning for the Unknown: Designing trials in a pandemic setting'' (12 min). This talk will provide a brief background of the Adaptive COVID-19 Treatment Trial study platform and will then highlight issues that arise at the start of the clinical trial process during an outbreak. The talk will address considerations for endpoint selection in settings where knowledge of the disease is rapidly evolving. Dr Mat Makowski, unblinded statistician for the Adaptive COVID-19 Treatment Trial protocol, will deliver talk two-''Handling the Curve: Data cleaning and interim monitoring during rapid enrollment'' (12 min). This talk will focus on strategies for handling rapid enrollment during trials in outbreak settings and will discuss how the extraordinarily high number of COVID-19 cases during the conduct of the Adaptive COVID-19 Treatment Trials challenged the usual approach to interim monitoring and what was done to address these challenges. Tyler Bonnett, unblinded statistician for the Adaptive COVID-19 Treatment Trial protocol, will be the speaker for talk three-''Getting the Word Out: Communicating trial results when the world is watching'' (12 min). This talk will use examples from multiple Adaptive COVID-19 Treatment Trial studies to highlight how the study team navigated release of trial results with a focus on the tension between the ethical obligation to quickly disseminate findings and the merits of rigorous data cleaning and peer review. These three speakers will be followed by talks from two discussants. Dr Peter Sasieni, Adaptive COVID-19 Treatment Trial Data and Safety Monitoring Board member, will then provide the first of two discussant talks-''Data Monitoring like Never Before: Reflections from participating in Data and Safety Monitoring Board deliberations during a pandemic.'' This 7-min talk will highlight challenges faced by the Adaptive COVID-19 Treatment Trial Data and Safety Monitoring Board and give guidance on various questions surrounding the role of the Data and Safety Monitoring Board during a pandemic. Dr Birgit Grund, National Institute of Allergy and Infectious Diseases Treatment Guidelines Panel member, will then conclude the speaking portion of the session with the final discussant talk-''Treatments: Weighing emerging evidence for benefit or harm.'' This 7-min talk will address challenges in synthesizing information from multiple clinical trials and observational studies in a pandemic, and challenges in interpreting potentially different signals in subgroups of participants;for example, in ACTT-1, the estimated treatment effect in critically ill patients was different from the effect in less severely ill patients. The remainder of the session (10 min) will be devoted to a question-and-answer session.

How the estimand addendum tothe ICH E9 guideline helps structure clinical objectives, analyses, and conclusions-a series of oncology case studies

Background: The estimand framework aims to increase the dialogue between functional areas working on proper alignment of trial objectives formulation, design, conduct, statistical analyses, and conclusions. The draft addendum on Estimands and Sensitivity Analysis in Clinical Trials to the ICH E9 guideline on Statistical Principles for Clinical Trials was released in August 2017. In December 2019, the final version of the ICH E9 estimand addendum was published. The new framework requires clarity and precision in description of the treatment effect, in particular, explicitly accounting for events which occur after randomization/treatment start and either preclude observation of the variable of interest or affect its interpretation (''intercurrent events''). It highlights the need for a discussion among key stakeholders during the design phase, resulting in more precise clinical trials objective. The estimand framework is anticipated to have a major impact on drug development. The estimand that reflects the trial objectives will determine the trial design, data collection, trial conduct, analysis, and interpretation. This work is a result of a cross-industry working group of statisticians and clinicians working on connecting the ICH E9 addendum concepts to applications in oncology. The working group was formed in February 2018 and currently has 34 members (14 EU + 20 US) from 21 companies to ensure common understanding and consistent definitions for time-to-event estimands in oncology. Methods: Followed by a general introduction to the estimand framework, we illustrate the impact of the addendum by applying it to a series of oncology case studies: Censoring mechanisms: evaluate the use of censoring to handle intercurrent events, related assumptions, and interpretation, discussing the often performed sensitivity analyses and possible alternatives in view of the estimand framework. Treatment switching: describe how the estimand framework allows to explicitly account for different types of treatment switching and offers a systematic and transparent approach for assessment. Solid and hematologic tumors: focus on relevant estimands, intercurrent events, and sensitivity analyses and demonstrate how the estimand framework seeks to increase transparency on the treatment effect of interest and facilitates communications between stakeholder. COVID-19: assess the impact of COVID-19 on the clinical trial objective, propose strategies to handle COVID-19-related intercurrent events, and show how the estimand framework provides a common language to discuss the impact of COVID-19 in a structured and transparent manner. Results: Key findings from this exercise are that the estimand framework: (1) makes implicit assumptions transparent, (2) facilitates the discussions about patients' journeys, (3) seeks to increase transparency on the clinical question of interest and facilitates a precise definition of the treatment effect, (4) prospectively plans the handling of certain intercurrent events, potentially leading to a different data collection strategy, and (5) is useful for structuring discussions about the impact of pandemics and mitigating measures one can take. Conclusion: Recommendations for design, data collection, analysis, and reporting for clinical trials planned post-addendum will be given. Key clinical implications of this work are that the treatment effect reflecting the clinical question posed by a given clinical trial objective will need to be more precisely defined in study protocols, guidelines, and publications. If the estimate is likely to be biased in light of an unforeseen impact like COVID-19, the estimand framework provides various stakeholders a common language to discuss the impact in a structured and transparent manner.

Electronic data capture system modifications to capture COVID-19-related data in the national drug abuse treatment clinical trial network

The COVID-19 public health emergency created significant challenges for the safe conduct of clinical trials. The Clinical Trials Network conducts multi-site substance use treatment studies, of which several are in an active recruitment phase. Investigators within the Clinical Trials Network collaborate with the Clinical Coordinating Center and Data and Statistics Center, both at the Emmes Company, to effectively manage these trials and ensure data quality. In March 2020, the US Food and Drug Administration issued guidance for the conduct of clinical trials during the pandemic;notably, the Food and Drug Administration recommended capturing reasons for missing data, protocol deviations, or modified study procedures as related to COVID-19. The Clinical Coordinating Center and Data and Statistics Center worked with investigators to evaluate and modify study protocols to allow for flexibility in the collection of study assessments (e.g. off-site/home visits and telehealth visits) while ensuring the safety of participants and research staff and maintaining the integrity of trial data. In addition, an interdisciplinary team within the Clinical Coordinating Center and Data and Statistics Center reviewed the Food and Drug Administration guidance, identified changes to case report forms, and proposed these to investigators and sponsor for buy-in and approval. Namely, an existing case report form that previously captured missed visits was expanded to collect relevant information for all visits. The modified case report form is expected to be completed for all visits and was adapted to capture whether a visit occurred outside the prescribed visitwindow, if a visit was missed, and if not, where the visit occurred: in clinic, via telemedicine, and/or offsite. Furthermore, if any portion of the visit occurred at an offsite location, follow-up questions capture the location of urine sample collection, which is often the basis of primary outcome, and the location of medication dispensing/administration. Another change included the addition of COVID-19 response options to case report forms collecting data that have potential to be missing or otherwise impacted by COVID-19. To capture the nuances among (1) active COVID-19 infection, (2) lockdown-related isolation/quarantine, or (3) other COVID-19 factors including fears of exposure, we proposed a set of three response options for impacted forms: ''COVID-19: Illness,'' ''COVID-19: Public health measures,'' and ''COVID-19: Other.'' Likewise, to assess the impact on study procedures and assessments, a question was added to evaluate if a protocol deviation was related to COVID-19. Finally, the Data and Statistics Center included automated data queries for operational concerns related to visit flexibility. For example, if a participant's visit occurred entirely offsite, the system issues an automated query if any incongruous visit data point is indicated as being collected in clinic or via telemedicine (e.g. biospecimen). The interdisciplinary collaboration among the Clinical Coordinating Center, Data and Statistics Center, and investigators allowed for timely updates to the electronic data capture system to capture COVID-19-related data across Clinical Trials Network trials in a harmonized fashion and brought the Clinical Trials Network into compliance with Food and Drug Administration guidance. These data will also allow statisticians to conduct sensitivity analyses to assess the impact of COVID-19 on trial outcomes. This, along with the reasons for missing data, protocol deviations, and assessment changes due to COVID-19 will be included in final study reports of these trials.