Despite accumulated evidence that caries susceptibility varies within the oral cavity, the intra-oral life course of the disease remains relatively understudied. It is well known that caries susceptibility differs between the maxilla an mandible and also across individual teeth and their surfaces, but little is known about the timing and disease transition for these intra-oral characteristics. A unique database that provides detailed longitudinal intra-oral information for the study of caries life course is the Detroit Denal Health Project (DDHP) conducted by the University of Michigan School of Dentistry at Ann Arbor, Michigan. This is a longitudinal study designed to understand the oral health of low-income African-American children (0-5 years) and their main caregivers (14+ years) residing in the city of Detroit, Michigan. Among other information, this study comprises longitudinal investigations involving tooth-level and tooth-surface level information on caries life course which can provide answers to important research questions such as: which teeth or tooth surfaces transition faster from sound to noncavitated lesions and ultimately to cavitated lesions? Which teeth or tooth surfaces are more likely to transition reversely from noncavitated lesions to sound? Are there (left-right or upper-lower) symmetries in the mouth with respect to caries transition times? Are molars more likely to transition faster to higher disease stages in boys than girls? Analyzing caries life course data from the DDHP study is inherently complex because of interval-censored observations, a hierarchical intra-oral correlation and multiple states of caries lesions. There have been several works in the statistical literature which examine related issues and complications, but these works often accommodate only one or two of these complications. To our best knowledge, we are not aware of any statistical model that addresses all these complications simultaneously. This project specifically aims at developing semiparametric multi-state models for clustered interval-censored caries history data at both tooth and tooth-surface levels;and applying the developed models to extensively analyze the data from the DDHP study. Technically, the methodology will be built from the likelihood of multi- state models for interval-censored observations coupled with penalized likelihood estimation. Following the completion of this project, we hope to substantially improve our understanding of early childhood caries in low income inner city African-American children. Most importantly, the acquired knowledge may then facilitate future prevention and treatment of dental caries in these underserved populations.
The proposed research will develop multistate semiparametric regression models for interval-censored correlated life course data in early childhood caries. The success of the project will advance our knowledge of early childhood caries in low-income African Americans, which may then help their prevention and treatment in these underserved populations.
|Li, Chenxi (2016) Cause-Specific Hazard Regression for Competing Risks Data Under Interval Censoring and Left Truncation. Comput Stat Data Anal 104:197-208|
|Li, Chenxi (2016) The Fine-Gray Model Under Interval Censored Competing Risks Data. J Multivar Anal 143:327-344|