Periodontitis, a chronic inflammatory disease of the periodontium, affects an estimated 50% of US adults over the age of 35. To develop appropriate planning for treatment and cure, dental health care professionals must understand the biological, socio- demographic, behavioral and other medical factors that affect tooth loss through periodontal progression. However, the statistical methods employed to understand these come with several interesting challenges. The data are multivariate, non-Gaussian, non- stationary and have missing information which are informative of the oral health status of that oral region. Besides, one can conjecture that periodontal progression can be spatially referenced. Current statistical methods do not address all of these under a unified framework. Goals: Using a Bayesian paradigm, the proposed study will develop robust statistical methods combining all the above challenges, for assessing periodontal disease status and identifying important covariates that are associated. Subjects: The statistical methods will be evaluated on a dataset of 313 dentate subjects who were enrolled in the Gullah African-American (AA) Diabetics (GAAD) Study as part of the SC COBRE for Oral Health. For generalizability, the methods will be investigated on nationally-representative data collected as part of NHANES (1999-2004). Available data and study design: Periodontal status (determined by site-level pocket dept, clinical attachment level, and bleeding on probing), other relevant biological and medical status like smoking habits, brushing and flossing habits, demographics (poverty status) and other parameters have been collected at the Medical University of South Carolina (MUSC) as part of the GAAD study. The Gullah-AA subjects represent an interesting population with minimal genetic admixture whose dental health status remains vastly unknown. NHANES data are publicly available. Significance: The new statistical methods will provide dental researchers enhanced knowledge about how spatial associations might predict periodontal progression in presence of the aforementioned characteristics typical for periodontal data. This will enable researchers to better target risk assessment and prevention strategies, thereby improving health status.

Public Health Relevance

(Public Health Relevance) Data collected during a periodontal examination are complex, with several types of responses taken at various mouth locations, complicated missing data patterns, and might have spatially referenced disease progression. Standard methods currently used do not consider these complications and might provide inefficient and/or biased information. This project proposes robust statistical models addressing these concerns, has the potential to exert tremendous impact from a dental public health perspective, and will bridge the gap between dental practitioners and academic dental basic science researchers for developing novel prevention strategies. The long term goal is to provide dental researchers a better understanding about repeated and longitudinal dental (health) data, so as to prevent and control disease and improve overall dental health.

Agency
National Institute of Health (NIH)
Institute
National Institute of Dental & Craniofacial Research (NIDCR)
Type
Small Research Grants (R03)
Project #
5R03DE021762-03
Application #
8251142
Study Section
Special Emphasis Panel (ZDE1-JR (06))
Program Officer
Harris, Emily L
Project Start
2011-04-04
Project End
2014-03-31
Budget Start
2012-04-01
Budget End
2014-03-31
Support Year
3
Fiscal Year
2012
Total Cost
$145,390
Indirect Cost
$49,105
Name
University of Minnesota Twin Cities
Department
Biostatistics & Other Math Sci
Type
Schools of Public Health
DUNS #
555917996
City
Minneapolis
State
MN
Country
United States
Zip Code
55455
Galvis, Diana M; Bandyopadhyay, Dipankar; Lachos, Victor H (2014) Augmented mixed beta regression models for periodontal proportion data. Stat Med 33:3759-71
Cancho, Vicente G; Bandyopadhyay, Dipankar; Louzada, Francisco et al. (2013) The destructive negative binomial cure rate model with a latent activation scheme. Stat Methodol 13:48-68
Boehm, Laura; Reich, Brian J; Bandyopadhyay, Dipankar (2013) Bridging conditional and marginal inference for spatially referenced binary data. Biometrics 69:545-54
Reich, Brian J; Bandyopadhyay, Dipankar; Bondell, Howard D (2013) A nonparametric spatial model for periodontal data with non-random missingness. J Am Stat Assoc 108:
Van Meter, Emily M; Garrett-Mayer, Elizabeth; Bandyopadhyay, Dipankar (2012) Dose-finding clinical trial design for ordinal toxicity grades using the continuation ratio model: an extension of the continual reassessment method. Clin Trials 9:303-13