Grant: Discovery Grants Program - Individual
University of British Columbia
$18,000.00
Statistical Modeling and Monitoring of Longitudinal Responses on Multiple Units
FY 2015-16
Summary
My program of research over the current funding period has concentrated on the development, evaluation and application of statistical methodology for data that is observed repeatedly over an extended period of time on each of multiple units. This focus will continue in the next funding period.
The essential feature of the data structure described is that repeated signals are received from each of multiple sources. Such data arise in many different contexts within the natural sciences (e.g., astronomy, climatology, ecology, geology, geophysics, meteorology, seismology, etc) so, as is almost invariably the case, new statistical methodology developed will have potential broad applicability beyond the specific context in which it is developed.
The data structure described can be thought of as multiple time series, although my interest is in methodology that can also be applied when the data may not be observed according to the same schedule over time for each unit. Statistical scientists usually refer to this as longitudinal data. This data structure is particularly relevant to clinical trials where the different units are human subjects; this specific context motivates most of my proposed research for the next funding period.
This research is motivated by my long-standing collaboration with the UBC MS/MRI Research Group and my involvement with organizations such as the U.S. National Multiple Sclerosis Society that are concerned with improving how clinical trials in multiple sclerosis (MS) are designed, executed and analyzed. The focus of the research is to develop suitable statistical models for the types of longitudinal data collected in clinical trials, to determine more effective designs for clinical trials and to determine more efficient methods of analyzing the types of data collected in clinical trials.
Specific projects proposed include the investigation of different designs for experiments where the primary outcome is a longitudinal count signal (e.g., the count of lesions seen on MRI images of the patient's brain) based on a new model for such data that was developed during the current funding period, the enhancement of a novel conditional probability index-based scheme for detecting unusual increases in a longitudinal count response on the individual units involved in an experiment that was developed during the current funding period, the extension of this scheme to the case of multiple longitudinal responses (potentially of different types), the development, for longitudinal voxel-based lesion responses, of appropriate statistical models that appropriately account for the spatial structure of the voxels within the lesion and of the lesions within the brain, and the investigation of adaptive experimental designs involving blinded sample size re-estimation.
The overall objective of my program of research is to enhance the amount of information derived from clinical trials, through the development of better statistical models, more effective experimental designs and more comprehensive approaches to the analysis of the data collected. These statistical models, experimental designs and methods of statistical analysis are also relevant to other research areas, including many areas of the natural sciences as noted above, but my recent involvement has been primarily in clinical trials, particularly in MS, and I expect this to continue during the next funding period.
Program
Discovery Grants Program - Individual
Recipient
Location
British Columbia, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
RGPIN-2014-04706
Name nom
Petkau, Albert
Department département
Statistics
Organization id
2
Institution établissement
University of British Columbia
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2015
Competition year année de concours
2014
Award amount
18000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
1508
Committee name en
Mathematics and Statistics
Committee name fr
Mathématiques et statistique
Area of application code
1205
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Mathematical sciences
Area of application fr
Sciences mathématiques
Research subject code
3004
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Applied statistics
Research subject fr
Statistique appliquée
Partie
2
Nb partie
5
Application title
Statistical Modeling and Monitoring of Longitudinal Responses on Multiple Units
Keyword
NSERC
Application summary
My program of research over the current funding period has concentrated on the development, evaluation and application of statistical methodology for data that is observed repeatedly over an extended period of time on each of multiple units. This focus will continue in the next funding period.
The essential feature of the data structure described is that repeated signals are received from each of multiple sources. Such data arise in many different contexts within the natural sciences (e.g., astronomy, climatology, ecology, geology, geophysics, meteorology, seismology, etc) so, as is almost invariably the case, new statistical methodology developed will have potential broad applicability beyond the specific context in which it is developed.
The data structure described can be thought of as multiple time series, although my interest is in methodology that can also be applied when the data may not be observed according to the same schedule over time for each unit. Statistical scientists usually refer to this as longitudinal data. This data structure is particularly relevant to clinical trials where the different units are human subjects; this specific context motivates most of my proposed research for the next funding period.
This research is motivated by my long-standing collaboration with the UBC MS/MRI Research Group and my involvement with organizations such as the U.S. National Multiple Sclerosis Society that are concerned with improving how clinical trials in multiple sclerosis (MS) are designed, executed and analyzed. The focus of the research is to develop suitable statistical models for the types of longitudinal data collected in clinical trials, to determine more effective designs for clinical trials and to determine more efficient methods of analyzing the types of data collected in clinical trials.
Specific projects proposed include the investigation of different designs for experiments where the primary outcome is a longitudinal count signal (e.g., the count of lesions seen on MRI images of the patient's brain) based on a new model for such data that was developed during the current funding period, the enhancement of a novel conditional probability index-based scheme for detecting unusual increases in a longitudinal count response on the individual units involved in an experiment that was developed during the current funding period, the extension of this scheme to the case of multiple longitudinal responses (potentially of different types), the development, for longitudinal voxel-based lesion responses, of appropriate statistical models that appropriately account for the spatial structure of the voxels within the lesion and of the lesions within the brain, and the investigation of adaptive experimental designs involving blinded sample size re-estimation.
The overall objective of my program of research is to enhance the amount of information derived from clinical trials, through the development of better statistical models, more effective experimental designs and more comprehensive approaches to the analysis of the data collected. These statistical models, experimental designs and methods of statistical analysis are also relevant to other research areas, including many areas of the natural sciences as noted above, but my recent involvement has been primarily in clinical trials, particularly in MS, and I expect this to continue during the next funding period.