Statistical modelling in clinical trials
Vladimir V. Anisimov
Abstract: A large clinical trial for testing pharmaceutical drug usually involves hundreds of patients to be recruited by many clinical centres. Patients recruited into the trial are randomized to prescribed medical treatments and at some stages, their responses are statistically analysed for taking a decision on the effectiveness of the drug. This process requires using various statistical techniques at different stages of the trial.
In the first part of the talk, the basic statistical problems that appear in drug development process in pharmaceutical industry are briefly described.
The second part is dealing with the patient recruitment stage. This stage is very costly and affected by various uncertainties in input information, variation in clinical centres and stochastic fluctuations of the recruitment over time. A large proportion of trials fail to complete recruitment before deadline.
A statistical methodology for modelling and predicting patient recruitment and drug supply in multicentre clinical trials is proposed. Patient’s flows in different centres are viewed as delayed Poisson processes with random unknown rates. The problems of modelling, estimating and predicting with credibility bounds the number of patients over time and the recruitment time are considered. The developed technique also allows re-assessing the number of clinical centres required to complete recruitment in time with a given confidence (adaptive adjustment), predicting study/centre performance, and evaluating the amount of drug supply which is needed to cover patient demand with a given risk of undersupply. The technique has been validated for many real trials and is on the way of implementation. Illustrative examples are considered.
SafetyWorks Ontology Construction
Author: G. H. Merrill
Abstract: The SafetyWorks project is oriented towards developing methodologies for the use of large observational data sources in drug safety signal screening and evaluation. The SafetyWorks application, as described here, is a prototype software application (or a set of prototypical components) implementing those methodologies. A feature of the overall SafetyWorks methodologies is the use of large formal biomedical ontologies for the purposes of data normalization and the exploration and inferencing of class effects pertaining to medical conditions and drugs. This paper describes the methods used in the creation and annotation of those ontologies within GSK’s prototype application of the SafetyWorks methodologies.
Note (Added Sept. 15, 2009):
This paper describes the technology employed in the GlaxoSmithKline SafetyWorks project to create the drug and medical conditions ontologies employed in SafetyWorks. It was delivered to ProSanos Corporation as part of a licensing agreement and technology transfer covering their use of this and other intellectual property developed by GlaxoSmithKline, and it is hereby placed in the public domain.
Towards Automating an Inference Model on Unstructured Terminologies: OXMIS Case Study
Author: J. L. Painter
Abstract: Most modern biomedical vocabularies employ some hierarchical representation
that provides a “broader/narrower” meaning relationship among the “codes”
or “concepts” found within them. Often, however, we may find within the
clinical setting the creation and curation of unstructured custom vocabularies
used in the everyday practice of classifying and categorizing clinical data
and findings.
A significant and widely used example of this lies in the General Practice
Research Database which makes use of the Oxford Medical Information Systems
(OXMIS) coding scheme to represent drugs and medical conditions. This scheme
is intrinsically unstructured, is generally regarded as disorganized, and is
not amenable to comparison with other hierarchically structured medical coding
schemes such as ICD-9, MedDRA, or SNOMED CT. In order to improve processes of
data analysis and extraction, we define a semantically meaningful representation
of the OXMIS codes by way of the UMLS Metathesaurus. A structure-imposing
ontology mapping is created, and this process provides a complete illustration
of a general semantic mapping technique applicable to unstructured biomedical
terminologies.
