Meta Analytic Approaches to Using Historical Data in

Meta Analytic Approaches To Using Historical Data In-PDF Download

  • Date:14 Sep 2020
  • Views:1
  • Downloads:0
  • Pages:68
  • Size:1.01 MB

Share Pdf : Meta Analytic Approaches To Using Historical Data In

Download and Preview : Meta Analytic Approaches To Using Historical Data In


Report CopyRight/DMCA Form For : Meta Analytic Approaches To Using Historical Data In


Transcription:

Introduction with Examples,Overview of Approaches,Meta Analytic Approaches. Meta Combined and Meta Analytic Predictive Approach. Prior Effective Sample Size,Robustness,More on Meta Analytic Predictive MAP Priors. Conclusions, 2 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. Introduction with Examples, 3 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. 1 Introduction and Examples,Informed Decision Making.
Informed decisions should be based on all,relevant information. In particular when,information is sparse,new information is difficult to obtain. Contextual or complementary data are often, 4 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. 1 Introduction and Examples,Historical Data,These data often referred too as historical data. But they may be come from a parallel experiment,Or from data in the same experiment.
E g in a clinical trial from a similar subgroup,Considering historical and current data is an. example of evidence synthesis,Various aspects to consider. methodological and practical issues and challenges. pros and cons, 5 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. 1 Introduction and Examples,Use of Historical Data Pros Cons. Design historical data are always used, This information puts the current experiment into perspective.
For example information about variability and expected effect. sizsed drives sample size calculations,Analysis historical data are rarely used. However these data can improve the inference for key. parameters,adjusted estimates safeguard against extremes. better precision, 6 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. 1 Introduction and Examples,Use of Historical Data Pros Cons. What is relevant historical data, Requires judgment about similarity of historical and current setting.
Requires interaction between subject matter experts. How to incorporate historical data,Requires a statistically principled approach. How much is the historical data worth, What if historical data and actual data are in conflict. Requires careful evaluation of the reasons, Problem can be mitigated by using a robust statistical approach. 7 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. 1 Introduction and Examples,Clinical Trials,Use of historical data is attractive. Smaller sample sizes e g smaller placebo group,More ethical less placebo patients or.
more scientific trials learn more about new treatment. Decreased costs and trial duration,Historical data various formats e g. for control group only our focus,for effect parameter mean difference risk ratio. aggregate and or individual data, 8 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. 1 Introduction and Examples,Novartis Experience,Use of historical data. In all phase I oncology trials to inform prior distributions. In a substantial percentage of phase II trials,In special cases e g non inferiority trials.
Experience overall positive,However there are challenges. Practical drug development is highly regulated,company internal and external standards. Practical more time needed for study design,Methodological innovative statistics. 9 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. Phase IV Trial,Phase IV transplantation trial,Binary outcome treatment failure. New treatment T vs standard of care C,Standard design requires 450 patents per arm.
Historical data,930 historical controls from 11 internal trials. Can these data be used to make control arm smaller. See N et al 2010, 10 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. Phase IV Trial Control Data from 11 Historical Trials. 11 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. Phase II Design,Phase II Trial in Ulcerative Colitis. Outcome clinical remission at week 8, Placebo data from 4 external trials 363 historical. controls of similar design, 12 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data.
Design of a Phase I Oncology Trial in Japan,Western on going first in human study. Objective determine the maximum tolerated dose MTD. Endpoint frequency of dose limiting toxicity DLT,Phase I study in Japan to find Japanese MTD. Often no ethnic differences, For Japanese trial can we make use of Western data. Dose 100 200 400 800 1500 3000 TOTAL,Patients 5 6 5 9 8 4 37. DLT 0 0 0 0 1 3 4,Tentative Western MTD, 13 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data.
Overview of Approaches, 14 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. 2 Overview of Approaches,Find Relevant Historical Data. 1st step idenfity relevant historical data,Systematic Reviews methodology. E g Cochrane Handbook Higgins and Green 2011,Pocock s 1976 criteria. Inclusion exclusion criteria for patient population. Type of study design,Exact definition of the outcome.
Quality of study execution and management,Potential biases due to time trends. Requires cross functional expertise,A psychological barrier for many statisticians. May not lead to a unique set of trials sensitivity analyses. 15 Bayes Pharma Neuenschwander 11 June 2014 London Meta Analytic Approaches to Historical Data. Meta Analytic Approaches to Using Historical Data in Clinical Trials Bayes Pharma 11 June 2014 London Beat Neuenschwander Head Statistical Methodology Oncology Biometrics and Data Management Novartis Pharma AG Basel Joint work with Heinz Schmidli Satrajit Roychoudhury Novartis Sandro Gsteiger University of Bern Tony O Hagan Sheffiled University David Spiegelhalter Cambridge

Related Books