Current affiliation: Luxembourg Institute of HealthSubject: StatisticsPrimary location: Luxembourg
Published in:Statistics in Biopharmaceutical ResearchPublication date:2013-08-01
I consider Bob's influence on the science of Drug Regulation and make some suggestions as to how we could make clinical trials even more efficient.
Published in:Medical WritingPublication date:2013-11-22
This is a review of Ben Goldacre's book, Bad Pharma
Published in:SignificancePublication date:2013-04-01
The probability of A given B is not the same as the probability of B given A. The fallacy of assuming that they are the same is surprisingly common.
Published in:Encyclopedia of BiostatisticsPublication date:2005-07-15
Entry in the Encyclopedia of Biostatistics covering design and analysis of cross-over designs
Published in:Not availablePublication date:2003-01-01
Chance, Risk and Health
A book that explains to a lay audience what it is that medical statisticians do.
Published in:SignificancePublication date:2007-06-01
Published in:BiometricsPublication date:2007-03-01
Published in:F1000ResearchPublication date:2012-12-04
Published in:Encyclopedia of Statistics in Behavioral SciencePublication date:2005-10-15
Published in:Encyclopedia of Statistical SciencesPublication date:2006-08-15
My entry to the Encyclopedia of Statistics on the subject of Rawlsian Theory of Justice
Published in:Pharmaceutical StatisticsPublication date:2002-01-01
Academic scientists are represented disproportionately in publications of clinical trials of pharmaceuticals.
Published in:BiometricsPublication date:1989-09-01
Published in:Statistical SciencePublication date:2003-02-01
John Nelder discusses his life and his work
Published in:Pharmaceutical StatisticsPublication date:2003-01-01
Published in:Dicing with DeathPublication date:Not available
The early history of statistics starting with Pascal and Fermat.
Published in:Encyclopedia of Biopharmaceutical StatisticsPublication date:2010-05-01
A practical account of the design and analysis of cross-over trials covering deep inferential issues as clearly as possible but firmly rooted in pr...
Encyclopedia article covering properties, problems and pitfalls of P-values. Shows that many criticisms are justified but some are exaggerated.
Published in:Biometrical JournalPublication date:2005-04-01
Published in:Not availablePublication date:2005-03-03
Published in:Statistics in PracticePublication date:2007-12-18
A book that frankly discusses controversies and issue in drug development without hiding the philosophical difficulties.
Published in:Statistics in PracticePublication date:2002-07-15
Explains in simple terms using real examples how to design and analyse cross-oevr trials
Published in:Statistics in MedicinePublication date:1999-07-30
Published in:Statistics in MedicinePublication date:2005-01-01
Criticises the common habit of turning continuous outcomes in binary ones by using an arbitrary cutpoint.
Published in:Pharmaceutical StatisticsPublication date:2005-01-01
Shows that using baselines values appropriately in analysis of covariance can deliver information that is equivalent to having studied more patients.
Published in:Statistical SciencePublication date:2009-05-01
Invited commentary on Christian Robert et al's extensive analysis of Harold Jeffreys's Theory of Probability
Published in:The LancetPublication date:2010-10-01
Protests that an article in the Lancet misrepresents Fisher's views on significance testing.
Various reviews have found that the probability of acceptance of a negative study is the same as a positive one, They have then wrongly concluded t...
Published in:PLoS MedicinePublication date:2005-07-26
Published in:SignificancePublication date:2006-08-23
Published in:Encyclopedia of Biopharmaceutical Statistics, Second editionPublication date:2003-06-04
Published in:Encyclopedia of Biopharmaceutical Statistics, Third EditionPublication date:2012-12-27
Contribution to 3rd edition of Encyclopedia of Biopharmaceutical Statistics
Encyclopedia entry on P-values discussing their history, properties and interpretation
Published in:SignificancePublication date:2008-03-01
Decsribe the use of Student's t-test.
Published in:Statistics in MedicinePublication date:2003-01-01
Published in:Statistics in MedicinePublication date:2012-12-17
Discusses a number of fallacies of randomisation including the claims that randomisation's value rests on its ability to produce balance, the fact ...
Published in:NaturePublication date:1987-07-02
Chapter eight of Dicing with Death. Explains the historical origins of meta-analysis and how it is used.
Published in:Not availablePublication date:2011-12-16
Christie/Simplicity, Complexity and Modelling
An account of the interdisciplinary research on modelling carried out by a group of scientists working on an EPSRC funded project.
Published in:Simplicity, Complexity and ModellingPublication date:2011-10-17
Statistitical Modelling in the Pharmaceutical Industry
Describe various approaches to model selection.
Published in:EpidemiologyPublication date:2008-09-01
Shows that shrinkage of results is a simple consequence of random variation.
Published in:Biometrical JournalPublication date:2005-02-01
Invited commentary on various articles published in The Biometrical Journal
Describes how WS Gosset (Student) came to develop the t-test.
It is shown that a proposal to modify cross-over trial designs while the trial is running is based on misunderstandings as to how such trials are c...
Published in:Statistics in MedicinePublication date:1992-01-01
Published in:Statistics in MedicinePublication date:2008-09-30
Published in:Pharmaceutical StatisticsPublication date:2008-07-01
The pharmaceutical industry does not exploit its own data enough to plan future work.
Published in:The LancetPublication date:2000-02-01
Published in:Controlled Clinical TrialsPublication date:1991-10-01
Published in:Biometrical JournalPublication date:2010-02-01
I discuss the circumstances under which the analysis of summary data can be as efficient as the analysis of original data. I also make some comment...
Published in:Statistics in MedicinePublication date:1993-12-30
Shows that high posterior probabilities of equivalence are highly dependent on assumptions.
Published in:Statistics in MedicinePublication date:1994-02-15
Various matters to do with analysing clinical trials are considered in an adversarial framework (hence "game with the devil").
Published in:Statistics in MedicinePublication date:1994-04-30
An account of the trial at Kalamazoo that provided the data for Student's illustration of his t-test.
Published in:The American StatisticianPublication date:1990-05-01
Shows that, contrary to previous claims, there are some distributions for which regression is to the mean and not the mode.
Published in:Statistics in MedicinePublication date:2006-01-01
It has been claimed incorrectly that simple analysis of change scores, unlike ANCOVA, is unbiased where baselines differ. I show that this is not c...
Published in:Statistics in MedicinePublication date:1994-09-15
Once the data from a clinical trial are available for analysis it is common practice to carry out ‘tests of baseline homogeneity’ on prognostic cov...
Published in:Statistics in MedicinePublication date:2007-01-01
Shows that various so-called Bayesian prior distributions for the random treatment by trial interaction in meta-analysis are not really Bayesian
Published in:Statistics in MedicinePublication date:2001-01-01
Different approaches to bioequivalence are described and their advantages and disadvantages discussed
Published in:Statistics in MedicinePublication date:2004-01-01
Published in:Controlled Clinical TrialsPublication date:2000-12-01
Shows that for analysing trials with many small centres conventional meta-analysis is a bad idea
Published in:Controlled Clinical TrialsPublication date:1993-02-01
Published in:BiometricsPublication date:1999-12-01
Published in:SignificancePublication date:2012-12-01
The story behind RA Fisher's famous Tea Test
Published in:The LancetPublication date:2001-12-01
Published in:F1000ResearchPublication date:2013-01-21
Failure to pay attention to the quality of submitted studies has led reviewers to conclude that editors are not biased against negative studies.
Published in:Statistics in MedicinePublication date:2014-06-04
In general in statistics, and in particular in meta-analysis, the term 'random effect' is used in two different ways. In meta-analysis it is used o...
Published in:BiometricsPublication date:1996-12-01
Published in:European Journal of Human GeneticsPublication date:2006-07-12
Siblings and even cousins cannot be treated as independent measurements on supposed epigenetic effects on grandparents. Analysis either has to take...
Published in:The American StatisticianPublication date:2008-08-01
A heuristic explanation of Philip Dawid's observation that no adjustment in inference is necessary if it is revealed that a particular treatment fo...
Published in:F1000 - Post-publication peer review of the biomedical literaturePublication date:Not available
Published in:Statistics in MedicinePublication date:2009-11-20
The claim is made that medical statisticians should pay more attention to the measures that are defined to judge the effect of treatments rather th...
Published in:Pharmaceutical StatisticsPublication date:2004-04-01
Published in:Statistics in MedicinePublication date:2009-12-04
Makes some practical points about dose escalation designs.
Published in:Statistics in MedicinePublication date:1996-12-30
Reports an error in the associated paper
Published in:Research Synthesis MethodsPublication date:2011-09-01
Review of Fleiss (2003)
Published in:International Journal of EpidemiologyPublication date:1988-01-01
Discusses the dangers of using methods developed for independent observations when clustering applies
Published in:British Journal of Mathematical and Statistical PsychologyPublication date:2012-09-28
Discusses some theoretical objections to the practical Bayesian data analysis approach of Gelman and Shalizi.
Published in:Statistics in MedicinePublication date:2010-03-08
Compares Athony Atkinson's design to minimimation in the case where there is balancing using a number of binary covariates
Published in:TrialsPublication date:2011-01-01
A number of common myths about randomisation are explained. Many of these are traceable to the fact that critics have failed to realise that thr p...
Published in:Statistics in MedicinePublication date:2010-12-16
Suggests strategies for analysing cross-over trials in infertility
Published in:Statistics in MedicinePublication date:2002-01-01
I show that the low replication probability of P-values is shared by other probability statements.
Chapter with Philip Dawid in Simplicity, Complexity and Modelling.
The introduction to the Simplicity and Complexity in Modelling (SCAM) project explaining the background to the team formation and the genesis and p...
Published in:Pharmaceutical StatisticsPublication date:2007-01-01
Considers disjunctive (at least one endpoint significant) and conjunctive (all enpoints significant) power. Shows that for moderate correlation bet...
Published in:Statistics in MedicinePublication date:1998-12-30
The relationship between choice of model for carry-over and choice of efficient cross-over design is studied in particular with reference to two tr...
Published in:SignificancePublication date:2004-12-01
A simple explanation for a lay audience of statistical issues in bioequivalence
Published in:SignificancePublication date:2008-05-29
Published in:Statistics in MedicinePublication date:1998-08-15
Published in:Statistics in MedicinePublication date:1995-12-30
Covers various controversial matters in clinical trials.
Published in:Pharmaceutical StatisticsPublication date:2008-10-01
The article uses two trials, one large and the other small, to make some points about randomisation.
Published in:Journal of Biopharmaceutical StatisticsPublication date:1998-01-01
Current approaches to modeling crossover trials in two treatments are critically reviewed from the perspective of the practical requirements of the...
Published in:The LancetPublication date:1998-07-01
Proposed FDA requirements for individual bioequivalance do not make sense because the authors of the proposals have lost sight of the purpose of bi...
Published in:Perspectives in Biology and MedicinePublication date:2013-01-01
Explains that the argument that there are indefinitely many confounding factors does not invalidate randomization.
Published in:Statistics in MedicinePublication date:2013-09-17
Extends Anisimov and Fedorov's work on the gamma Poisson model to cover prediction before the trial starts.
Published in:Encyclopedia of Statistical SciencesPublication date:2010-03-31
Published in:BiometricsPublication date:1985-06-01
Maximum likelihood solution for cases of truncated and censored samples.
Published in:Biometrical JournalPublication date:2006-04-01
Published in:Journal of Biopharmaceutical StatisticsPublication date:1993-01-01
Published in:Annals of Allergy Asthma & ImmunologyPublication date:2000-06-01
Published in:Drug Information JournalPublication date:2001-10-01
Shows that differential individual response to treatment is difficult to identify in convential clinical trials
Published in:Journal of Epidemiology and BiostatisticsPublication date:2001-03-01
I consider various (mainly Bayesian) criticisms of P-values.
Published in:Clinical Pharmacology & TherapeuticsPublication date:2010-07-07
I discuss areas in which statisticians would do better if they paid more attention to what pharmacokinetic modellers have done. I also cover some f...
Published in:Statistics in MedicinePublication date:2012-07-24
Examines estimation techniques for pharmacokinetic measurements regarded as unreliable because the concentration is low.
Published in:Journal of the Royal Statistical Society Series A (Statistics in Society)Publication date:2003-10-01
Discusses various statistical issues that are relevant to patents and competition in the pharmaceutical industry.
Published in:Methods and Models in StatisticsPublication date:2004-07-01
A discussion of John Nelder's life and work.
Published in:Statistics in MedicinePublication date:2008-01-01
Published in:Pharmaceutical StatisticsPublication date:2014-06-06
The requirement for pharmaceutical sponsors to develop plans for studying children may seem logical but may in fact harm the interests of those it ...
Published in:Statistics & Probability LettersPublication date:2011-07-01
A suite of macros for performing meta-analysis in SAS is described
Published in:Statistics in MedicinePublication date:1991-07-01
A (probably rather too) critical review is given of the 1st edition of the book by Howson and Urbach, Scientific Reasoning the Bayesian Approach.
Published in:Encyclopedia of Statistical SciencesPublication date:2004-07-15
Published in:Bayesian AnalysisPublication date:2008-01-01
I agree that Gelman's eclecticism makes sense since a pure Bayesian approach is difficult in practice.
Shows that one must apply a penalty for non-orthogonality when using hybrid ANCOVA and rank test approaches.
Published in:Statistics in MedicinePublication date:1997-09-15
I show that, contrary to what had been claimed in Statistics in Medicine previously, it was not reasonable to use the two-stage analysis of AB/BA c...
The choice of covariates used in statistical models should be based on influence on outcome not on association with allocation.
Looks at the problem of a pharmaceutical company deciding which drugs to develop faced with a number of projects that may have different stages of ...
Published in:Statistics in MedicinePublication date:1989-04-01
It is concluded that tests of homogeneity on the covariates should not be performed, that covariate imbalance is just as much a problem for large s...
Published in:Statistics in MedicinePublication date:1991-11-01
Considers the fundamental falsification assymetry in clinical trials especially as this effects claims of equivalence. Shows that the value of blin...
Published in:Journal of Clinical EpidemiologyPublication date:1997-07-01
Explains that a publication that had claimed that the BHAT trial exhibited important variation in response from centre to centre had analysed the d...
Published in:Journal of Clinical EpidemiologyPublication date:1998-12-01
Published in:BMJPublication date:1997-04-19
Argues that the use of placebo run-ins involves a covert rather than agreed deception of patients and may be unethical. It also queries whether suc...
Published in:Statistics in MedicinePublication date:2014-09-04
Lower bounds for the numbers of subjects enrolled in studies implied by the results in terms of confidence intervals for odds ratios (ORs) or relat...
Published in:BMC Medical Research MethodologyPublication date:2009-01-01
Much has been written about the need to identify all studies when performing a meta-analysis. The reverse problem is that of making sure that data ...
Published in:Evidence-Based MedicinePublication date:2011-06-21
Argues that the indefinite unmeasured confounders argument of Worrall is not a valid criticism of probabilistic arguments for randomisation. Furthe...
Published in:PainPublication date:2014-03-01
Discusses various sources of variation in observed response in clinical trials: between patient, within patient and patient-by treatment interactio...
Published in:PainPublication date:2014-09-01
Present general design aspects to consider including patient population, active comparators and placebos, study power, pharmacokinetic–pharmacodyna...
Published in:The StatisticianPublication date:2000-07-01
Published in:Pharmaceutical StatisticsPublication date:2014-10-09
It is shown that ﬁxed-effect meta-analyses of naïve treatment estimates from sequentially run trials with the possibility of stopping for efﬁcacy b...
Published in:Pharmaceutical StatisticsPublication date:2014-10-22
A recent analysis of R&D productivity suggests that there are grounds for ‘cautious optimism’ that the industry ‘turned the corner’ in 2008 and is ...
Published in:Drug Information JournalPublication date:1998-01-01
Various aspects of portfolio management within the pharmaceutical industry are considered. It is shown that some popular notions of risk management...
Published in:Statistics in MedicinePublication date:2000-03-30
Researchers are sometimes surprised to learn that reducing repeated measures on patients to a single summary is often nearly as powerful as the cor...
Published in:Wiley StatsRef: Statistics Reference OnlinePublication date:2014-09-29
Published in:Statistical Methods in Medical ResearchPublication date:2014-02-02
Traces the history of attempts to deal with the fact that variances vary not just systematically but also randomly and speculates about lines of fu...
Published in:CHANCEPublication date:2002-03-01
Published in:Wiley StatsRef: Statistics Reference OnlinePublication date:2014-11-17
Published in:Wiley StatsRef: Statistics Reference OnlinePublication date:2014-12-01
Published in:Statistics in MedicinePublication date:2015-09-28
There are three obvious sources of variation în clinical trials: between patient, within patient and treatment by patient interaction. The third of...
Published in:Nature Reviews Drug DiscoveryPublication date:2015-02-06
Published in:Journal of Applied PhysiologyPublication date:2015-02-05
Published in:SignificancePublication date:2011-08-25
Published in:Controlled Clinical TrialsPublication date:1997-06-01
Published in:Statistics in MedicinePublication date:1995-10-30
Published in:Statistics in MedicinePublication date:1993-06-01
Published in:Statistics in MedicinePublication date:1990-05-01
Published in:PLoS ONEPublication date:2016-12-01
Depending on whether you wish to prove that a treatment has an effect, what its average effect will be in future or what its specific effect will b...
Published in:CHANCEPublication date:2001-01-01
Published in:Pharmaceutical StatisticsPublication date:2017-01-01
Published in:Biometrical JournalPublication date:2017-04-01
A biostatistician's view on the current p
P-values may be over and misused but if there is a replication crisis in science, abandoning them would not solve it. As one of many ways of
Published in:Statistical Methods in Medical ResearchPublication date:2017-09-07
The paper give advice on choosing sample sizes for groups of n of 1 trials. Different possible tasks are considered, since the necessary sample siz...