DPHS biostatistics faculty focus in a number of areas of methods development. Highlights include:
- Clinical trial design and analysis - adaptive designs, randomization schemes, non-inferiority trials
- Multivariate models involving large number of predictors using random forests for variable selection
- Threshold identification in multivariate prediction models
- Analysis of zero-inflated count data from complex surveys
- Joint modeling of multiple longitudinal outcomes
- Missing data analysis
- Mixed models under mixture distributions
- Latent growth models
- Zero-inflated and semi-continuous models
- Analysis of geo-referenced survey data
- Bayesian multivariate spatio-temporal small area health modeling
- Spatial censoring in accelerated failure time models
- Variable selection using leverage and 2 stage estimation for geo-referenced heath data
- Sparse network Bayesian multivariate exposure models and health outcomes
- Statistical genetics and Bioinformatics, especially analysis of genome-wide association studies (GWAS) and high throughput sequencing data such as RNA-seq and ChIP-seq
- Integrative analysis of high throughput genetic and genomic data with biomedical big data
- High dimensional data analysis
- Unbalanced cluster designs
- Multivariate truncated models in data missing due to limits of quantification
- Other methodological challenges that arise in a variety of collaborative projects across biomedical research disciplines
