SUMMARY The Collaborative ICPRS clinically-oriented predoctoral/postdoctoral administered development current proposal titled “Sequential Modeling for Prediction of Periodontal Disea ses: An I ntra- Practice-Based Research Study ( ICPRS)” is developed in response to RFA-DE-23-012 . The will seek to develop various tasks in 4 component sections. In Component 1 , we seek to train 10 faculty from Temple University Kornberg School of Dentistry (TUKSoD) nd ten students in Applied Clinica l Research by offering a graduate certificate program by the College of Public Health (CPH) on facilitating skills in applied patient-oriented research a . The 15-credit certificate will focus . In Component 2, we aim to strengthen collaborations between TUKSoD, CPH and the Medical School for faculty and student training, and overall research support for the execution of Component 4 of this proposal. Our proposed project was collaboratively developed by individuals with backgrounds in Periodontology, General and Oral Epidemiology, Oral Microbiome, Health Informatics, Biostatistics and Public Health, and Physicians who will assist in extracting medical metrics relevant to management of periodontal diseases (PD) from real-time medical records in the Pennsylvania Health Share Exchange (PA HSX). In Component 3, we aim to expand clinically-oriented faculty participation in grant funded research activities, in the annual research day, and in national research meetings by increasing the number of faculty-student mentoring partnerships to conduct small-scale research projects. Finally, in Component 4, we will develop longitudinal predictive models of PD incidence and progression using machine learning approaches and microbiome/metagenomic sequencing strategies. This component has the following aims: Aim1 seeks to develop clinical decision tools that utilize matched medical-dental patient datasets from TUKSoD and the PA HSX EHR records. Aim1a will focus on developing a Medical Continuity of Care Record (CCR) that will provide patients' up-to-date medical health metrics that can be used in the provision of care at the dental school. Using dental patient's matched records, Aim1b will build an AI Empowered Prediction Model for PD. Aim2 will build on findings from Aim1b to assess the potential of the subgingival and salivary microbiomes as additional predictors of PD incidence and progression. Aim 2a will validate a dysbiosis index, applied to both subgingival plaque and saliva samples, as a predictor of PD in a cohort study, while Aim2b will identify potential high resolution, salivary/plaque metagenomic biomarkers of PD progression via a nested case-control study. The proposed application builds on prior research conducted at TUKSoD and is directed towards increasing participation of clinically-oriented faculty and pre/postdoctoral dental students in the school's research activities, coupled with training in Applied Clinical Research including health informatics, opening a new pathway to further research in dental education as well as promote faculty professional development, expand the research academic workforce, and provide opportunities to conduct student-mentored practice-based research.

PROJECT NARRATIVE Prediction models to assess periodontal disease (PD) risk have been developed. However, their clinical use is rare because of their suboptimal performance and absence of oral microbiome data, social determinants and behavioral factors, and accurate medical history. This application seeks to use real-time medical data from a health share exchange to develop a continuity of care report for oral health care, to build a risk prediction models for PD using machine learning approaches including subgingival and salivary microbiome data and to identify potential high resolution biomarkers of PD progression for better disease control and prevention.