Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

24
Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat

Transcript of Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Page 1: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat

Page 2: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Introduction to:

• CRO and Clinical Trial: definitions• TFS Company & Organisation• Data Management & Statistics working flow• Regulatory guidelines• Type of clinical trials • 3 Illustrations: statistically significant?• Conclusions

Page 3: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

What is a CRO?

Chief Risk Officer Cathode Ray Oscilloscope Cro-Magnons

Clinical Research Organization: a service organization that provides support to the pharmaceutical and biotechnology industries in the form of outsourced pharmaceutical research services (for both drugs and medical devices)

Page 4: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

What is a Clinical trial?A clinical trial is a research study to answer specific questions about vaccines or new therapies or new ways of using known treatments. Clinical trials (also called medical research and research studies) are used to determine whether new drugs or treatments are both safe and effective

Page 5: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

TFS -Introduction

Founded in 1996 with headquarters in Sweden

Worldwide ranking no 14*

~ 600 employees

Operations inspected by US FDA, EMA, MHRA (UK) and MPA (Sweden)

Geographical coverage in Europe, USA and Japan

Conducting clinical trials in 40 countries worldwide (December 2012)

Projected revenue €75 million in 2013

Page 6: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

TFS global HQ Sweden

TFS regional HQ Sweden Spain The Netherlands Hungary

TFS country offices Norway Denmark Finland Russia UK France Germany Portugal Italy The Baltics (Estonia, Latvia, Lithuania)

Poland Czech Republic

TFS European locations

Page 7: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Clinical research professionals within FSP models;

Specialist training for clinical research professionals; www.tfsacademy.com

Phase 0/I and PoC trials;

Phase II – IV, NIS trials; 26 countries world-wide

TFS Solutions for entire clinical development cycle

Page 8: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

TFS Project delivery functions

Page 9: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Em m a A lbacarAssociate Unit M anager B io m etrics,Spain

Ram ón DosantosSenior S tatistic ia n

Senior C linical Data M anage r

Juani Zam oraSenior S tatistic ia n

Senior C linical Data M anage r

Eva UsónSenior S tatistic ia n

Senior C linical Data M anage r

M arta F iguerasSenior S tatistic ia n

M ercè V iladric hSenior S tatistic ia n

Jordan Bertsc hSenior S tatistic ia n

Xavier NuñezSenior S tatistic ia n

B iostatistic s

Cristina Lópe zSenior C linical Data M anage r

Senior S tatistic ia n

Daniel M osteiroSenior S tatistic ia n

Senior C linical Data M anage r

Judith O ribeSenior S tatistic ia n

Senior C linical Data M anage r

Rosario PeláezStatistic ian

Clinical Data M anage r

Em ilio Sánche zStatistic ian

Clinical Data M anage r

Laia Pu jantellSenior C linical Data M anage r

M ario P irche rSenior C linical Data M anage r

Data M anagem en t

M ireia Cuella rC linical Data A ssociate

E lisabe t RoquéClinical Data A ssociate

M arta G utiérrezClinical Data A ssociate

Verónica O rteg aClinical Data A ssociate

Laura G arcíaClinical Data A ssociate

M aite RuizC linical Data A ssociate

Data entry people (variable )

Data Assistan t

Rosa M ª A lons oUnit M anager B iom etrics, Spain

Ricard Q uingle sRegional M anaging D irec tor, South Europe

Eva Lundqvis tD irector G lobal P roject Delivery

TFS Barcelona – Biometrics

14 Statisticians!!!!

Page 10: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Data Management working flow

Page 11: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Statistics working flow

Page 12: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Medical research - Regulations

Good Clinical Practice (GCP)An international ethical and scientific quality standard for designing, conducting, recording and reporting trials that involve the participation of human subjectsThe most important sources for GCP-compliant guidelines referring to the EU are the following: - Declaration of Helsinki (1964) - ICH –E6: GCP (1996) - EU Directives 2001/20/EC, 2005/28/EC

Page 13: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Medical research - Regulations

Additional guidelines refer to specific statistical or DM regulations or to other recommendations, such as - ICH –E9: Statistical principles for clinical trials - ICH –E3: Structure and contents of clinical study reports - ICH –E10: Choice of Control Group in Clinical Trials - CDISC Clinical Data Interchange Standards Consortium,

Operational Data model (ODM)

Page 14: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Clinical trials vs. non-interventional studies

OBSERVATIONALSOBSERVATIONALS CLINICAL TRIALSCLINICAL TRIALS

Intervention in the study design- Treatment assigned to the subjects by the

investigator

EpidemiologicalEpidemiologicalDiseaseDisease

Post-AuthorisationPost-Authorisation study (EPA)study (EPA)

Study medicationStudy medication

Disease exposition = treatment? Disease exposition = treatment?

NoNo YesYes

No intervention in the study design- - Treatment exposition without participation of the

investigator → ‘observes’ subjects- No randomisation procedures

Quasi-experimentalQuasi-experimentalClinical TrialsClinical Trials

(Non-randomised)(Non-randomised)

RANDOMISED RANDOMISED Clinical TrialsClinical Trials(experimental)(experimental)

Page 15: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Phase I - Healthy volunteers

- Small sample size (6-30 subjects) - Usually FTIH- Objectives: safety (adverse events), dose range, PK/PD

Phase II- Healthy volunteers / Patients- Larger sample size (20-300 subjects) - Objectives: efficacy, safety, dose-response

Phase III- Patients- Multicentre, Larger sample size, (1000-3000 subjects) - Objectives: confirm efficacy –superiority, non-inferiority?, no safety issues

Phase IV (post-authorisation)- Patients- Multicentre, non-interventional studies- Objectives: optimal use of treatment, risk-benefit, marketing, etc.

Type of clinical trials

Page 16: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

By the awareness of treatment administered - Open-label: both investigators and subjects know which treatment is being administered

- Single-blinded: investigator is aware of the treatment administered, but the subject is not- Double-blinded: neither investigators nor subjects know which treatment is being administered

By time of observation- Retrospective: data from past records is collected in a unique visit, with no follow-up- Cross-sectional: all present data from subjects is collected at a defined time-point- Prospective: subjects are followed over a period of time, collecting data in different visits

By sequence of treatments- Parallel : subjects are randomly assigned to a unique treatment throughout the study- Cross-over: subjects are randomly assigned to a sequence of treatments

Type of clinical trials

Page 17: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Type of clinical trials

By nature of comparator treatment- Placebo-controlled: a group of subjects receives a ‘placebo’ treatment, which is specifically

designed to have no real effect → sometimes is not ethical!

- Active-control: the experimental treatment is compared to an existing treatment → that is clearly better than doing nothing for the subject

By type of comparison - Superiority: the clinical objective of efficacy is to show that the response to the experimental treatment is superior to the comparator treatment → usually superiority to placebo

- Equivalence or non-inferiority: the clinical objective of efficacy is to show that the response to the experimental treatment is at least as good, or not clinically inferior, to the comparator treatment → usually non-inferiority to active control

Page 18: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

- If we do not get a significant difference, what can we then conclude? Only that we have not found evidence to support the existence of a treatment effect

- Estimates may often be better than p-values

Statistically significant?Importance of estimation and its superiority to significance testing

Page 19: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.
Page 20: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.
Page 21: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Statistically significantBecome a statistician: open-minded and objective in the assumptions; precise and analytical in the results. Study Design is crucial !!Become a scientific: interact with your clinical colleagues, do not be only a programmer! Communicate – “Statisticians seem to talk double Dutch”: make yourself and the results understandable to any person with no knowledge of statistics at allBe responsible: our work is key in the outcome of a clinical trial ; the client will listen to you and act from the results you present Work closely with your team – you need the study input from the project leader, the clinical expertise from the medical writer, the knowledge of data from the CRA, and the DB specifications from the DM

Conclusions

Page 22: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Not Statistically significantDon’t look for p-values, think statistically! “You don’t know the power of the dark side”: if your study is underpowered or you carry out statistical analysis of secondary endpoints, beware of the conclusions: the results do not ‘conclude that’ but ‘suggest that’

Conclusions

We are very lucky!! We are (or will be) statisticians!!!!

Page 23: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Some remarks to end...

-Many people use statistics as a drunken man uses a lamp post; beware of p-values

-Statisticians are more rigorous in interpreting statistics but physicians are more imaginative

-Statisticians expect the average but on average people do not expect statisticians

-An idiot with a computer is often more powerful than a statistician with a pencil

-Even if you have a significant relationship with a statistician you may not find it relevant

Guernsey McPearson http://www.senns.demon.co.uk/Confuseus.htm

Page 24: Statisticians Statistically Significant Xavier Núñez, CStat Senior Statistician, CStat.

Any Questions?

Thank you for your patience!

WWW.TFSCRO.COMemail: [email protected]