Modeling longitudinal oral health status and pneumonia risk: secondary data analyses of an integrated dental-medical cohort.

BACKGROUND: Mounting evidence indicates potential associations between poor oral health status (OHS) and increased pneumonia risk. Relative pneumonia risk was assessed in the context of longitudinally documented OHS. METHODS: Electronic medical/dental patient data captured from 2007 through 2019 were retrieved from the integrated health records of Marshfield Clinic Health Systems. Participant eligibility initiated with an assessment of OHS, stratified into the best, moderate, or worst OHS groups, with the additional criterion of 'no pneumonia diagnosis in the past 90 days'. Pneumonia incidence was longitudinally monitored for up to 1 year from each qualifying dental visit. Models were assessed, with and without adjustment for prior pneumonia incidence, adjusted for smoking and subjected to confounding mitigation attributable to known pneumonia risk factors by applying propensity score analysis. Time-to-event analysis and proportional hazard modeling were applied to investigate relative pneumonia risk over time among the OHS groups. RESULTS: Modeling identified associations between any incident pneumonia subtype and 'number of missing teeth' (p < 0.001) and 'clinically assessed periodontal status' (p < 0.01), which remained significant following adjustment for prior pneumonia incidence and smoking. The hazard ratio (HR) for 'any incident pneumonia' in the best OHS group for 'number of missing teeth' was 0.65, 95% confidence interval (CI) [0.54 - 0.79] (unadjusted) and 0.744, 95% CI [0.61 - 0.91] (adjusted). The HR for 'any incident pneumonia' in the best 'clinically assessed periodontal status' group was 0.72, 95% CI [0.58 - 0.90] (unadjusted) and 0.78, 95% CI [0.62 - 0.97] (adjusted). CONCLUSION/CLINICAL RELEVANCE: Poor OHS increased pneumonia risk. Proactive attention of medical providers to patient OHS and health literacy surrounding oral-systemic disease association is vital, especially in high-risk populations.

Variations in temporal trends in non-traumatic dental condition related emergencies.

OBJECTIVE: The objective of the study was to investigate temporal trends in non-traumatic dental condition (NTDC) related emergency visits at Emergency Department (ED), urgent care (UC), and at a Federally Qualified Health Center (FQHC) that providing dental services to a mid-sized rural community. METHODS: Temporal trends over a 9-year period (2008-2016) in NTDC rates at ED, UC, FQHC and in a region around the FQHC were determined. Statistically significant changes (α = 0.05) in the proportion of NTDC visits between FQHC and UC across each of the time points were investigated. RESULTS: Proportion of NTDC ED visits was relatively stable over the study period; whereas those at FQHC exceeded those at UC site beginning 2012 and were significantly (α = 0.05) higher than that of UC subsequently (2015-2016). CONCLUSIONS: NTDCs are preventable dental conditions and the care provided in treating NTDCs in emergency settings is palliative and does not address the underlying conditions resulting in poor outcomes. The results presented elucidate the critical role of FQHCs in significantly reducing NTDC visits. These might be precursors to a potential shift in NTDC care seeking behavior and expected to favorably impact oral health outcomes.

A Methodological Approach to Validate Pneumonia Encounters from Radiology Reports Using Natural Language Processing.

INTRODUCTION: Pneumonia is caused by microbes that establish an infectious process in the lungs. The gold standard for pneumonia diagnosis is radiologist-documented pneumonia-related features in radiology notes that are captured in electronic health records in an unstructured format. OBJECTIVE: The study objective was to develop a methodological approach for assessing validity of a pneumonia diagnosis based on identifying presence or absence of key radiographic features in radiology reports with subsequent rendering of diagnostic decisions into a structured format. METHODS: A pneumonia-specific natural language processing (NLP) pipeline was strategically developed applying Clinical Text Analysis and Knowledge Extraction System (cTAKES) to validate pneumonia diagnoses following development of a pneumonia feature-specific lexicon. Radiographic reports of study-eligible subjects identified by International Classification of Diseases (ICD) codes were parsed through the NLP pipeline. Classification rules were developed to assign each pneumonia episode into one of three categories: "positive," "negative," or "not classified: requires manual review" based on tagged concepts that support or refute diagnostic codes. RESULTS: A total of 91,998 pneumonia episodes diagnosed in 65,904 patients were retrieved retrospectively. Approximately 89% (81,707/91,998) of the total pneumonia episodes were documented by 225,893 chest X-ray reports. NLP classified and validated 33% (26,800/81,707) of pneumonia episodes classified as "Pneumonia-positive," 19% as (15401/81,707) as "Pneumonia-negative," and 48% (39,209/81,707) as "episode classification pending further manual review." NLP pipeline performance metrics included accuracy (76.3%), sensitivity (88%), and specificity (75%). CONCLUSION: The pneumonia-specific NLP pipeline exhibited good performance comparable to other pneumonia-specific NLP systems developed to date.

Identifying Pneumonia Subtypes from Electronic Health Records Using Rule-Based Algorithms.

BACKGROUND: The International Classification of Disease (ICD) coding for pneumonia classification is based on causal organism or use of general pneumonia codes, creating challenges for epidemiological evaluations where pneumonia is standardly subtyped by settings, exposures, and time of emergence. Pneumonia subtype classification requires data available in electronic health records (EHRs), frequently in nonstructured formats including radiological interpretation or clinical notes that complicate electronic classification. OBJECTIVE: The current study undertook development of a rule-based pneumonia subtyping algorithm for stratifying pneumonia by the setting in which it emerged using information documented in the EHR. METHODS: Pneumonia subtype classification was developed by interrogating patient information within the EHR of a large private Health System. ICD coding was mined in the EHR applying requirements for "rule of two" pneumonia-related codes or one ICD code and radiologically confirmed pneumonia validated by natural language processing and/or documented antibiotic prescriptions. A rule-based algorithm flow chart was created to support subclassification based on features including symptomatic patient point of entry into the health care system timing of pneumonia emergence and identification of clinical, laboratory, or medication orders that informed definition of the pneumonia subclassification algorithm. RESULTS: Data from 65,904 study-eligible patients with 91,998 episodes of pneumonia diagnoses documented by 380,509 encounters were analyzed, while 8,611 episodes were excluded following Natural Language Processing classification of pneumonia status as "negative" or "unknown." Subtyping of 83,387 episodes identified: community-acquired (54.5%), hospital-acquired (20%), aspiration-related (10.7%), health care-acquired (5%), and ventilator-associated (0.4%) cases, and 9.4% cases were not classifiable by the algorithm. CONCLUSION: Study outcome indicated capacity to achieve electronic pneumonia subtype classification based on interrogation of big data available in the EHR. Examination of portability of the algorithm to achieve rule-based pneumonia classification in other health systems remains to be explored.

Identifying oral disease variables associated with pneumonia emergence by application of machine learning to integrated medical and dental big data to inform eHealth approaches.

Background: The objective of this study was to build models that define variables contributing to pneumonia risk by applying supervised Machine Learning-(ML) to medical and oral disease data to define key risk variables contributing to pneumonia emergence for any pneumonia/pneumonia subtypes. Methods: Retrospective medical and dental data were retrieved from Marshfield Clinic Health System's data warehouse and integrated electronic medical-dental health records (iEHR). Retrieved data were pre-processed prior to conducting analyses and included matching of cases to controls by (a) race/ethnicity and (b) 1:1 Case: Control ratio. Variables with >30% missing data were excluded from analysis. Datasets were divided into four subsets: (1) All Pneumonia (all cases and controls); (2) community (CAP)/healthcare associated (HCAP) pneumonias; (3) ventilator-associated (VAP)/hospital-acquired (HAP) pneumonias and (4) aspiration pneumonia (AP). Performance of five algorithms were compared across the four subsets: Naïve Bayes, Logistic Regression, Support Vector Machine (SVM), Multi-Layer Perceptron (MLP) and Random Forests. Feature (input variables) selection and ten-fold cross validation was performed on all the datasets. An evaluation set (10%) was extracted from the subsets for further validation. Model performance was evaluated in terms of total accuracy, sensitivity, specificity, F-measure, Mathews-correlation-coefficient and area under receiver operating characteristic curve (AUC). Results: In total, 6,034 records (cases and controls) met eligibility for inclusion in the main dataset. After feature selection, the variables retained in the subsets were: All Pneumonia (n = 29 variables), CAP-HCAP (n = 26 variables); VAP-HAP (n = 40 variables) and AP (n = 37 variables), respectively. Variables retained (n = 22) were common across all four pneumonia subsets. Of these, the number of missing teeth, periodontal status, periodontal pocket depth more than 5 mm and number of restored teeth contributed to all the subsets and were retained in the model. MLP outperformed other predictive models for All Pneumonia, CAP-HCAP and AP subsets, while SVM outperformed other models in VAP-HAP subset. Conclusion: This study validates previously described associations between poor oral health and pneumonia. Benefits of an integrated medical-dental record and care delivery environment for modeling pneumonia risk are highlighted. Based on findings, risk score development could inform referrals and follow-up in integrated healthcare delivery environment and coordinated patient management.

Community Awareness and Oral Cancer Screening in Rural Wisconsin.

Purpose: Oral cancer risks have been shown to be modified by improving public awareness and reducing barriers to preventive care. The purpose for this study was to assess oral cancer knowledge and awareness and provide oral cancer screenings and education to a population of rural farmers in Wisconsin.Methods: Attendees 18 years of age and older at a rural farming exposition in Wisconsin were invited to complete a 12-item oral cancer awareness paper survey and to receive a visual and tactile head and neck examination/ oral cancer screening. Completing both the survey and the screening were optional. Participants also received educational materials on oral cancer. Individuals with abnormal lesions were provided with dental referrals.Results: A total of 236 attendees consented to participate either the survey or oral cancer screening (n=236). Most (72%) reported seeing a dentist in the past six months regardless of insurance status. In spite of having had recent dental encounters, only 28% of women and 46% of men were able to identify at least one risk factor associated with oral cancer. Among participants consenting to the oral cancer screening (n=194), 17% (n=33) presented with oral lesions requiring additional assessment and were recommended for follow-up care.Conclusions: Knowledge and awareness of oral cancer risk factors, signs and symptoms was low among the participants in this rural population despite high rates of dental care access. Oral cancer screenings and education provided in varied settings could improve oral cancer knowledge and awareness and early detection of malignant oral lesions in rural communities.

Longitudinal Observation of Outcomes and Patient Access to Integrated Care Following Point-of-Care Glycemic Screening in Community Health Center Dental Safety Net Clinics.

Introduction: Rates of diabetes/prediabetes continue to increase, with disparity populations disproportionately affected. Previous field trials promoted point-of-care (POC) glycemic screening in dental settings as an additional primary care setting to identify potentially at-risk individuals requiring integrated care intervention. The present study observed outcomes of POC hemoglobin A1c (HbA1c) screening at community health center (CHC) dental clinics (DC) and compliance with longitudinal integrated care management among at-risk patients attending dental appointments. Materials and Methods: POC HbA1c screening utilizing Food and Drug Administration (FDA)-approved instrumentation in DC settings and periodontal evaluation of at-risk dental patients with no prior diagnosis of diabetes/prediabetes and no glycemic testing in the preceding 6 months were undertaken. Screening of patients attending dental appointments from October 24, 2017, through September 24, 2018, was implemented at four Wisconsin CHC-DCs serving populations with a high representation of disparity. Subjects meeting at-risk profiles underwent POC HbA1c screening. Individuals with measures in the diabetic/prediabetic ranges were advised to seek further medical evaluation and were re-contacted after 3 months to document compliance. Longitudinal capture of glycemic measures in electronic health records for up to 2 years was undertaken for a subset (n = 44) of subjects with available clinical, medical, and dental data. Longitudinal glycemic status and frequency of medical and dental access for follow-up care were monitored. Results: Risk assessment identified 224/915 (24.5%) patients who met inclusion criteria following two levels of risk screening, with 127/224 (57%) qualifying for POC HbA1c screening. Among those tested, 62/127 (49%) exhibited hyperglycemic measures: 55 in the prediabetic range and seven in the diabetic range. Moderate-to-severe periodontitis was more prevalent in patients with prediabetes/diabetes than in individuals with measures in the normal range. Participant follow-up compliance at 3 months was 90%. Longitudinal follow-up documented high rates of consistent access (100 and 89%, respectively), to the integrated medical/DC environment over 24 months for individuals with hyperglycemic screening measures. Conclusion: POC glycemic screening revealed elevated HbA1c measures in nearly half of at-risk CHC-DC patients. Strong compliance with integrated medical/dental management over a 24-month interval was observed, documenting good patient receptivity to POC screening in the dental setting and compliance with integrated care follow-up by at-risk patients.

Knowledgeability, attitude, and practice behaviors of primary care providers toward managing patients' oral health care in medical practice: Wisconsin statewide survey.

BACKGROUND: In this study, the authors sought to explore the receptivity, preparedness, and rates of adoption of integrated medical-dental models of care (MOCs) in the practice setting among primary care providers (PCPs) treating patients with diabetes mellitus (DM). METHODS: The authors conducted an anonymous statewide survey targeting PCPs across a range of Wisconsin-based practice settings to evaluate knowledgeability, attitude, practice behaviors, and perceived barriers to oral health screening in a medical setting. Qualitative analytical approaches included thematic analyses applied to evaluate the status of and barriers to integrated medical-dental MOC adoption. RESULTS: The integrated medical-dental MOC adoption rate was 34%. Top perceived barriers to integrated medical-dental MOC adoption included insurance coverage (71%) and care access (70%). A total of 39% indicated competency for educating patients about the association between DM and periodontitis. Although 72% of PCPs indicated optimal periodicity for oral health assessment as frequent, 39% reported frequently conducting such assessments. CONCLUSIONS: Although PCPs indicate receptivity to integrated medical-dental MOCs, PCPs identify suboptimal education, lack of adequate training in oral-systemic disease assessment, and barriers to oral health care access as barriers to integrated medical-dental MOC adoption. PRACTICAL IMPLICATIONS: Integrated medical-dental MOC adoption in care delivery to patients with DM remains below average. Interdisciplinary efforts and education are needed to address identified barriers to care integration.

Development of non-invasive diabetes risk prediction models as decision support tools designed for application in the dental clinical environment.

The objective was to develop a predictive model using medical-dental data from an integrated electronic health record (iEHR) to identify individuals with undiagnosed diabetes mellitus (DM) in dental settings. Retrospective data retrieved from Marshfield Clinic Health System's data-warehouse was pre-processed prior to conducting analysis. A subset was extracted from the preprocessed dataset for external evaluation (Nvalidation) of derived predictive models. Further, subsets of 30%-70%, 40%-60% and 50%-50% case-to-control ratios were created for training/testing. Feature selection was performed on all datasets. Four machine learning (ML) classifiers were evaluated: logistic regression (LR), multilayer perceptron (MLP), support vector machines (SVM) and random forests (RF). Model performance was evaluated on Nvalidation. We retrieved a total of 5319 cases and 36,224 controls. From the initial 116 medical and dental features, 107 were used after performing feature selection. RF applied to the 50%-50% case-control ratio outperformed other predictive models over Nvalidation achieving a total accuracy (94.14%), sensitivity (0.941), specificity (0.943), F-measure (0.941), Mathews-correlation-coefficient (0.885) and area under the receiver operating curve (0.972). Future directions include incorporation of this predictive model into iEHR as a clinical decision support tool to screen and detect patients at risk for DM triggering follow-ups and referrals for integrated care delivery between dentists and physicians.