Stopping Trials for Futility Ranjit Lall (May 2009)

Slides:



Advertisements
Similar presentations
Interim Analysis in Clinical Trials: A Bayesian Approach in the Regulatory Setting Telba Z. Irony, Ph.D. and Gene Pennello, Ph.D. Division of Biostatistics.
Advertisements

Mentor: Dr. Kathryn Chaloner Iowa Summer Institute in Biostatistics
Phase II/III Design: Case Study
Simulation methods for calculating the conditional power in interim analysis: The case of an interim result opposite to the initial hypothesis in a life-threatening.
NIHR Research Design Service London Enabling Better Research Forming a research team Victoria Cornelius, PhD Senior Lecturer in Medical Statistics Deputy.
Statistical Analysis for Two-stage Seamless Design with Different Study Endpoints Shein-Chung Chow, Duke U, Durham, NC, USA Qingshu Lu, U of Science and.
Futility Analysis A Miscellany of Issues Professor Andy Grieve King’s College London © Andy Grieve.
Bayesian posterior predictive probability - what do interim analyses mean for decision making? Oscar Della Pasqua & Gijs Santen Clinical Pharmacology Modelling.
1 1 Slide STATISTICS FOR BUSINESS AND ECONOMICS Seventh Edition AndersonSweeneyWilliams Slides Prepared by John Loucks © 1999 ITP/South-Western College.
Decision Errors and Power
MPS Research UnitCHEBS Workshop - April Anne Whitehead Medical and Pharmaceutical Statistics Research Unit The University of Reading Sample size.
‘Cohort multiple RCT’ design workshop Clare Relton & Jon Nicholl School of Health and Related Research (ScHARR) Faculty of Medicine University of Sheffield.
COURSE: JUST 3900 INTRODUCTORY STATISTICS FOR CRIMINAL JUSTICE Instructor: Dr. John J. Kerbs, Associate Professor Joint Ph.D. in Social Work and Sociology.
1 Statistical and Practical Aspects of a Non-Stop Drug Development Strategy Karen L. Kesler and Ronald W. Helms Rho, Inc. Contact:
Sample size optimization in BA and BE trials using a Bayesian decision theoretic framework Paul Meyvisch – An Vandebosch BAYES London 13 June 2014.
Evaluating Hypotheses Chapter 9. Descriptive vs. Inferential Statistics n Descriptive l quantitative descriptions of characteristics.
Evaluating Hypotheses Chapter 9 Homework: 1-9. Descriptive vs. Inferential Statistics n Descriptive l quantitative descriptions of characteristics ~
Stopping Trials for Futility RSS/NIHR HTA/MRC 1 day workshop 11 Nov 2008.
1 Psych 5500/6500 The t Test for a Single Group Mean (Part 5): Outliers Fall, 2008.
Clinical Trials Hanyan Yang
Statistics for Managers Using Microsoft Excel, 4e © 2004 Prentice-Hall, Inc. Chap 8-1 Chapter 8 Fundamentals of Hypothesis Testing: One-Sample Tests Statistics.
1Carl-Fredrik Burman, 11 Nov 2008 RSS / MRC / NIHR HTA Futility Meeting Futility stopping Carl-Fredrik Burman, PhD Statistical Science Director AstraZeneca.
PY 427 Statistics 1Fall 2006 Kin Ching Kong, Ph.D Lecture 6 Chicago School of Professional Psychology.
EVIDENCE BASED MEDICINE
Adaptive Designs for Clinical Trials
RANDOMIZED CLINICAL TRIALS. What is a randomized clinical trial?  Scientific investigations: examine and evaluate the safety and efficacy of new drugs.
© 2008 McGraw-Hill Higher Education The Statistical Imagination Chapter 9. Hypothesis Testing I: The Six Steps of Statistical Inference.
HYPOTHESIS TESTING Dr. Aidah Abu Elsoud Alkaissi
Tests of significance & hypothesis testing Dr. Omar Al Jadaan Assistant Professor – Computer Science & Mathematics.
Inference in practice BPS chapter 16 © 2006 W.H. Freeman and Company.
Week 8 Fundamentals of Hypothesis Testing: One-Sample Tests
Adaptive designs as enabler for personalized medicine
Background to Adaptive Design Nigel Stallard Professor of Medical Statistics Director of Health Sciences Research Institute Warwick Medical School
CHAPTER 16: Inference in Practice. Chapter 16 Concepts 2  Conditions for Inference in Practice  Cautions About Confidence Intervals  Cautions About.
Lecture 7 Introduction to Hypothesis Testing. Lecture Goals After completing this lecture, you should be able to: Formulate null and alternative hypotheses.
Statistical Fundamentals: Using Microsoft Excel for Univariate and Bivariate Analysis Alfred P. Rovai Hypothesis Testing PowerPoint Prepared by Alfred.
Encouraging adaptive designs in NiHR funded clinical trials Professor Sallie Lamb Chair, CET Board.
Optimal cost-effective Go-No Go decisions Cong Chen*, Ph.D. Robert A. Beckman, M.D. *Director, Merck & Co., Inc. EFSPI, Basel, June 2010.
Introduction to inference Use and abuse of tests; power and decision IPS chapters 6.3 and 6.4 © 2006 W.H. Freeman and Company.
Simon Thornley Meta-analysis: pooling study results.
The changing landscape of interim analyses for efficacy / futility
Biostatistics Class 6 Hypothesis Testing: One-Sample Inference 2/29/2000.
1 An Interim Monitoring Approach for a Small Sample Size Incidence Density Problem By: Shane Rosanbalm Co-author: Dennis Wallace.
Bayesian vs. frequentist inference frequentist: 1) Deductive hypothesis testing of Popper--ruling out alternative explanations Falsification: can prove.
Introduction to sample size and power calculations Afshin Ostovar Bushehr University of Medical Sciences.
Department Author Bayesian Sample Size Determination in the Real World John Stevens AstraZeneca R&D Charnwood Tony O’Hagan University of Sheffield.
Statistical Inference for the Mean Objectives: (Chapter 9, DeCoursey) -To understand the terms: Null Hypothesis, Rejection Region, and Type I and II errors.
1 Interim Analysis in Clinical Trials Professor Bikas K Sinha [ ISI, KolkatA ] RU Workshop : April18,
Biostatistics in Practice Peter D. Christenson Biostatistician Session 4: Study Size for Precision or Power.
Fall 2002Biostat Statistical Inference - Proportions One sample Confidence intervals Hypothesis tests Two Sample Confidence intervals Hypothesis.
Bayesian Approach For Clinical Trials Mark Chang, Ph.D. Executive Director Biostatistics and Data management AMAG Pharmaceuticals Inc.
Hypothesis Testing Introduction to Statistics Chapter 8 Feb 24-26, 2009 Classes #12-13.
Biostatistics Basics: Part I Leroy R. Thacker, PhD Associate Professor Schools of Nursing and Medicine.
Pilot and Feasibility Studies NIHR Research Design Service Sam Norton, Liz Steed, Lauren Bell.
European Patients’ Academy on Therapeutic Innovation The Purpose and Fundamentals of Statistics in Clinical Trials.
Review: Stages in Research Process Formulate Problem Determine Research Design Determine Data Collection Method Design Data Collection Forms Design Sample.
Chapter 8: Introduction to Hypothesis Testing. Hypothesis Testing A hypothesis test is a statistical method that uses sample data to evaluate a hypothesis.
1 PRIORITY MEDICINES FOR EUROPE AND THE WORLD Barriers to Pharmaceutical Innovation Richard Laing EDM/PAR WHO.
Introduction to Biostatistics, Harvard Extension School, Fall, 2005 © Scott Evans, Ph.D.1 Sample Size and Power Considerations.
1 Chapter 6 SAMPLE SIZE ISSUES Ref: Lachin, Controlled Clinical Trials 2:93-113, 1981.
DSCI 346 Yamasaki Lecture 1 Hypothesis Tests for Single Population DSCI 346 Lecture 1 (22 pages)1.
An Introduction to Clinical Trials and Pharmaceutical Statistics Workshop Robbie Peck University of Bath Student-Led Symposia 16 th Feb 2016.
Core Research Competencies:
CLINICAL PROTOCOL DEVELOPMENT
Introduction to inference Use and abuse of tests; power and decision
Statistical Approaches to Support Device Innovation- FDA View
Lecture 4: Meta-analysis
Aiying Chen, Scott Patterson, Fabrice Bailleux and Ehab Bassily
The DMC’s role in monitoring futility
Optimal Basket Designs for Efficacy Screening with Cherry-Picking
Presentation transcript:

Stopping Trials for Futility Ranjit Lall (May 2009)

OVERVIEW Summary of a 1 day workshop on “Stopping Trial for Futility” Funded by the RSS/NIHR HTA/MRC 11 th November 2008 in London Speakers:  Jon Nicholl (ScHARR, University of Sheffield, Chair person)  John Whitehead (Lancaster University)  Andy Grieve (King’s College London)  Karl-Fredrik Burman (Director of Statistical Science, Astra Zeneca)

WHAT IS FUTILITY? The term ‘futility’ is used to refer to the inability of the trial to achieve its objectives On-going monitoring of trials for safety, treatment efficacy or futility Early interims – review will focus more on safety, quality of conduct and trial integrity rather than efficacy. Later interims - may include formal efficacy or futility analysis

WHAT IS FUTILITY? A futility analysis causes a clinical trial to be stopped as soon as it becomes clear that a negative outcome is inevitable and thus it is no longer worthwhile continuing the trial to its completion.

PUBLICLY FUNDED TRIALS Early phase trials are supported by MRC Later phase trials supported by HTA HTA budget = £88 million pa Nearly all spent on trials Typical design of phase III trials – parallel group RCT, multi-centre, 3-5 years, £1million - £2.5 million, N =

FUTILITY DESIGNS Futility analyses are common in pharmaceutical trials NO HTA funded trials have planned a futility analysis Questions: - Why is this? - Should futility analyses be designed into HTA trials? - If so, which trials? - And when and how should they be done?

Should HTA do futility analyses? YES – (i) As pointed out by Ware, Muller and Braunwald (1985), “……………..early termination for futility could reduce enormous expenditure of resources, human and financial, involved in the conduct of trials that ultimately provide a negative answer regarding the value of the medical innovation….”. (ii) It is ethically wrong to continue to recruit patients to trials with little hope of achieving helpful results

Should HTA do futility analyses? NO – (i) All well conducted trials provide valuable evidence (ii) The cost of designing trials with a planned futility analysis would outweigh any savings

Should HTA do futility analyses? IT DEPENDS – (i)In placebo controlled trials, futility suggests no worthwhile benefit (i.e. little chance of finding evidence that the intervention is better than placebo) (ii)In head-to-head trials, treatment will be far from futile (so is futility analyses inappropriate in head-to –head trials)? (iii)Futility analyses should be conducted in large expensive trials in which early stopping could save substantial resources (iv)Futility analysis most useful where there are several treatments being compared, and there is a desire to eliminate one (for better resource use)

Extension grants 54% of MRC/HTA trials seek extensions Usually due to long set up/slow recruitment 55% of trials with extensions still don’t achieve N Cost of extensions for HTA = £50K - £1 million Current HTA ask for a ‘futility analysis’ in terms of conditional power on application. Almost never done as intended or refused

Fixed Interim analyses Focus Type I error (i.e. Rejecting the H 0 when its true) Methods used include - Pocock (1977) - O’Brien and Fleming (1979) - Haybrittle and Peto (1976) - Lans and DeMets alpha spending function (1983, 1989)

Futility Analysis (cont’d) Two fundamental approaches: (i)Using the above group sequential methods to compute the stopping boundaries for futility (in a similar way to computing stopping boundaries for efficacy) The boundaries can be defined to allow for early stopping for futility only, or to allow for early stopping for either futility or efficacy

Futility Analysis (cont’d) (ii) Conditional power The power of a trial tells whether a clinical trial is likely to have high probability to detect a pre-defined treatment effect of interest. Very low power implies that a trial is unlikely to reach statistical significance even if there is a true effect. One should never begin a trial with low power. However, sometimes low power becomes apparent only after a trial is well under way.

Futility Analysis (cont’d) (ii) Conditional power: based on the probability of a getting a statistically significant final result, conditional on the data collected so far (in the interim) Stochastic curtailment – a frequentist method Predictive power – partially Bayesian methods Predictive probability – fully Bayesian methods

Futility Analysis (cont’d) (a) Stochastic curtailment: At planning stages, sample size based on (a) Difference - d; (b) power (80% or 90%); (c) type I error rate At the interim, recalculate the power to detect a difference d. Calculations involve the observed data and an assumption regarding the distribution of the future data This method assumes that the observed difference between the treatments at the interim stage is the true treatment difference, d.

Futility Analysis (cont’d) (a) Stochastic curtailment: Detect e.g.<20% power no point continuing; Detect e.g.>50% power then continue. This cut-off is pre-specified (choice would depend on the risk/benefit considerations) Take other information into account before deciding on stopping or not extending.

Futility Analysis (cont’d) (b) Predictive power: Conditional power calculations use the observed data and an assumed future treatment effect Not sure what assumptions to make Predictive power approach addresses this by averaging the conditional power function over a range of treatment difference, with weights based on the observed treatment effect

Futility Analysis (cont’d) (b) Predictive probability: Prior probability: based on the interim data and the prior distribution of the parameter for the treatment effect; Statistical model: mathematical model which expresses the relationship of the parameters and the data Posterior probability: Probability of a clinically important treatment effect

Futility Analysis (cont’d) If early results show: Intervention is better than expected conditional power will be high Intervention is worse that expected conditional power will be low (unless your sample size increases)

Futility Analysis (cont’d) The two most challenging aspects: the selection of optimal stopping thresholds; the timing of the analysis both of which require the balancing of various risks.

Ask the audience (questions for the workshop)

Should we plan to do futility analysis in publicly- funded trials ? Yes – for all trials (built into design) 2.Yes – but only for extension requests 3.Never 4.Undecided

Should we plan to do futility analysis in publicly- funded trials? 1.Yes – for all trials (built into design) 1.Yes – but only for extension requests 1.Never 2.Undecided

Should we plan to do futility analysis in publicly- funded trials? May not be appropriate for all trials, but justification for doing/not doing should be clearly laid out in the application Should be made as a requirement by funders (and the funders should clearly lay out what is required by them) Clear distinction needs to be made between trial futility (e.g. lack of recruitment-operational aspect) and treatment futility (lack of efficacy - statistical)

In what circumstances are they appropriate? 1. Only in placebo- controlled trials 2.In both placebo- controlled and head- to head trials 3.Undecided 0 5

In what circumstances are they appropriate? 1.Only in placebo-controlled trials 2.In both placebo- controlled and head to-head trials 3.Undecided

In what circumstances are they appropriate? It may be difficult to predict in advance Futility analyses may not be very appropriate for very large pragmatic trials, where other information (other than efficacy) is important

Should funding bodies insist upon futility analyses before granting extensions? 1.Yes 2.No 3.Undecided 0 5

Should funding bodies insist upon futility analyses before granting extensions? 1. Yes 2. No 3. Undecided

Should funding bodies insist upon futility analyses before granting extensions? Should not be insisted upon, but if investigator refuses to do futility then justification should be given

What is futile? How low can the conditional power be before stopping a trial 1.1% 2.5% 3.10% 4.15% 5.20% 6.30% 7.40% 8.50% 9.It depends 0 5

What is futile? How low can the conditional power be before stopping a trial? 1.1% 2.5% 3.10% 4.15% 5.20% 6.30% 7. 40% 8.50% 9.It depends

What is futile? How low can the conditional power be before stopping a trial? Futility analysis should be focused on effect size (rather than conditional power) A trial may be considered as futile in its ability to recruit patients, but would add information to a evidence based data

Thank you !!