Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Launch Supplement

University of Waterloo

$12,500.00

Statistical Methods for Analyzing Binary Bivariate Outcome Data from Multiple Primary Studies

EF 2024-25

Résumé
Whereas aggregate data meta-analysis (ADMA) combines summary statistics such as the standardized sample average and its measure of variability from each primary study to make inferences about an unknown population parameter (e.g., the sensitivity and specificity of a new diagnostic test), individual participant data meta-analysis (IPDMA) uses line-by-line participant-level raw data. The standard models for these approaches to meta-analysis describe the variability in sensitivity and specificity across studies using the bivariate normal distribution. However, this idealistic assumption poses many statistical challenges, including the inability to account for asymmetrical data and outlying or influential studies. Moreover, although there are similar thought processes for methods for IPDMA and approaches for ADMA, it is vital to develop robust statistical methods tailored to each approach in meta-analysis. This is because of the apparent dissimilarity in data types between these two methods, making it clear that a tool designed for one does not readily apply to the other. For example, although the methods for IPDMA account for participant-level covariates to describe variability or estimate effect sizes for subgroups of participants, methods for ADMA do not. Therefore, the long-term goal of this research program is to develop and validate robust and flexible statistical models for both ADMA and IPDMA of diagnostic test accuracy (DTA) data. In the short term, this research program aims to develop novel and flexible statistical methods for both meta-analysis approaches that account for data with skewed distributions or contain outlying or influential studies, use objective statistical approaches to detect unusual studies, and are computationally optimal in calculating accurate and precise model parameters. Thus, in the near term, my research group and I will develop and validate robust Generalized Linear Mixed Model (GLMM)-based bivariate random effects models that accommodate skewness, outliers, and influential studies; objective inference methods based on statistical models and approaches to identify outlying and influential studies; and computational algorithms to estimate parameters of the standard and our proposed models for both meta-analysis approaches described above. As an area of research that has not been actively pursued in Canada, likely due to the absence of programs in most institutions, the proposed research program will benefit Canada by making it one of the hubs for this emerging and active area of research by fostering new and existing collaborations inside and beyond borders; practitioners by making the proposed methods easily accessible via freely available statistical software packages to make informed decisions using flexible methods; and students in my research team by training them to become all-round statisticians with sought-after skills and readying them for work in Canadian research, academic or government institutions.
Programme
Discovery Launch Supplement
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
DGECR-2024-00343
Name nom
Negeri, Zelalem
Department département
Statistics and Actuarial Science
Organization id
33
Institution établissement
University of Waterloo
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
12500
Program id
DGECR
Program name en
Discovery Launch Supplement
Program name fr
Supplément Tremplin vers la découverte
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1508
Committee name en
Mathematics and Statistics
Committee name fr
Mathématiques et statistique
Area of application code
1211
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Medical and health sciences
Area of application fr
Sciences médicales et de la santé
Research subject code
3005
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Biostatistics
Research subject fr
Biostatistiques
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Statistical Methods for Analyzing Binary Bivariate Outcome Data from Multiple Primary Studies
Keywords
robust statistical methods; bivariate modelling; random effects modelling; sensitivity; specificity; meta-analysis; outlier detection; diagnostic test accuracy studies; accommodating outliers,; computational statistics methods
Application summary
Whereas aggregate data meta-analysis (ADMA) combines summary statistics such as the standardized sample average and its measure of variability from each primary study to make inferences about an unknown population parameter (e.g., the sensitivity and specificity of a new diagnostic test), individual participant data meta-analysis (IPDMA) uses line-by-line participant-level raw data. The standard models for these approaches to meta-analysis describe the variability in sensitivity and specificity across studies using the bivariate normal distribution. However, this idealistic assumption poses many statistical challenges, including the inability to account for asymmetrical data and outlying or influential studies. Moreover, although there are similar thought processes for methods for IPDMA and approaches for ADMA, it is vital to develop robust statistical methods tailored to each approach in meta-analysis. This is because of the apparent dissimilarity in data types between these two methods, making it clear that a tool designed for one does not readily apply to the other. For example, although the methods for IPDMA account for participant-level covariates to describe variability or estimate effect sizes for subgroups of participants, methods for ADMA do not. Therefore, the long-term goal of this research program is to develop and validate robust and flexible statistical models for both ADMA and IPDMA of diagnostic test accuracy (DTA) data. In the short term, this research program aims to develop novel and flexible statistical methods for both meta-analysis approaches that account for data with skewed distributions or contain outlying or influential studies, use objective statistical approaches to detect unusual studies, and are computationally optimal in calculating accurate and precise model parameters. Thus, in the near term, my research group and I will develop and validate robust Generalized Linear Mixed Model (GLMM)-based bivariate random effects models that accommodate skewness, outliers, and influential studies; objective inference methods based on statistical models and approaches to identify outlying and influential studies; and computational algorithms to estimate parameters of the standard and our proposed models for both meta-analysis approaches described above. As an area of research that has not been actively pursued in Canada, likely due to the absence of programs in most institutions, the proposed research program will benefit Canada by making it one of the hubs for this emerging and active area of research by fostering new and existing collaborations inside and beyond borders; practitioners by making the proposed methods easily accessible via freely available statistical software packages to make informed decisions using flexible methods; and students in my research team by training them to become all-round statisticians with sought-after skills and readying them for work in Canadian research, academic or government institutions.