Skip Internal Navigation
The BBB conducts research in statistical theory and methodology relevant to problems under investigation in the areas of maternal and child health. The Branch also provides statistical support in the design and conduct of clinical trials and consults and collaborates with intramural and extramural scientists on statistical and mathematical problems. The Branch also develops quantitative procedures appropriate for application in biomedical and life sciences and supports the statistical and applied mathematics activities of DIPHR.
The Branch's mission is to:
- Develop original biostatistical and bioinformatics research relevant for the research mission of the Division and Institute.
- Engage in collaborative research with other Division, Institute, and extramural investigators working in research areas relevant for the Division and Institute.
- Provide service to the Division, Institute, NIH, DHHS, and other government agencies via consultation, collaboration, and assistance to advance the scientific discipline of biostatistics and the goals of the Institute.
- Recruit and mentor highly qualified students and trainees at various stages of their careers to position them for professional careers in biostatistical and bioinformatics research.
To explore DIPHR's data sharing opportunities, please visit our
Biospecimen Repository and Data Sharing (BRADS) site.
Chief and Senior Investigator of the Biostatistics and Bioinformatics Branch position (PDF - 105 KB)
- BBB investigators developed a new pattern mixture model framework to predict a binary outcome from a longitudinal sequence of biomarkers. An individualized risk predictor is derived from the model and is shown to be robust to model assumptions. This novel methodology is applied to a fetal growth study, where poor pregnancy outcomes can be predicted using longitudinal ultrasound measurements. (Liu D and Albert, Biostatistics, 2014). PMID: 24831103
- BBB investigators developed new methods for analyzing zero-inflated clustered binary response data, from a survey study of adolescents’ dating violence. Both likelihood-based and estimating equation approaches were proposed and compared in terms of their robustness and efficiency. These models provide an important tool for properly accounting for excessive zero responses due to unknown susceptible populations in public health research. (Fulton, Liu D, Haynie, and Albert, Annals of Applied Statistics, 2015). PMID: 26937263
- BBB investigators developed a semi-parametric integrated approach to estimate ROC curves without a gold standard when a priori information exists on the distributions of test scores and applied it to diagnosing endometriosis. (Hwang and Chen, Journal of the American Statistical Association, 2015). PMID: 26839441
- BBB investigators and collaborators carefully evaluated and compared approaches for analyzing correlated data with informative cluster sizes and proposed remedies when these approaches might fail. (Zhang, Liu, Zhang, Chen, and Albert, Statistical Methods in Medical Research, 2016). PMID: 26113386
- BBB investigators and their collaborators proposed a joint modeling framework for assessing menstrual cycle length and fecundity under a Bayesian framework. An interesting feature of this work is that the model for probability of pregnancy accounts for all the days in a menstrual cycle when intercourse occurred, rather than a fixed window around the time of ovulation. (Lum, Sundaram, Louis and Louis, Biometrics 2016). PMID: 26295923
- BBB investigators proposed an estimation of gap time distribution when the underlying data structure is that of recurrent events, and the observation times are informative with an unknown start time. A semiparametric inference for assessing the gap time distribution was proposed. This method was motivated by the progression of the first stage of labor in nulliparous women and allowed the study team to estimate the duration of each centimeter of dilation among women in labor, given covariates like age and BMI. (Ma and Sundaram Journal of the American Statistical Association, Theory & Methods, 2016). DOI: 10.1080/01621459.2016.1246369
- BBB investigators and their collaborators proposed a marginal rank-based inverse normal transformation approach to normalizing the marginal distribution of heavy-tailed multivariate data before employing a multivariate test procedure. They demonstrated the ability of the procedure to adequately control the type I error rate, as well as increase the power of the test, particularly with data from non-symmetric or heavy-tailed distributions. (Cai, Li and Liu A, Statistics in Medicine, 2016). PMID: 26990442
- BBB investigators and their collaborators proposed an approach based on covariates and random effects for evaluating predictive biomarkers under the potential outcome framework. Under the proposed approach, the parameters of interest are identified by assuming conditional independence of potential outcomes given observed covariates, and a sensitivity analysis is performed by incorporating an unobserved random effect that accounts for any residual dependence. (Annals of Applied Statistics, Zhang, Nie, Song and Liu A, 2014). PMID: 26779295