Machine Learning as a Tool for Early Detection: A Focus on Late-Stage Colorectal Cancer across Socioeconomic Spectrums.

PURPOSE: To assess the efficacy of various machine learning (ML) algorithms in predicting late-stage colorectal cancer (CRC) diagnoses against the backdrop of socio-economic and regional healthcare disparities. METHODS: An innovative theoretical framework was developed to integrate individual- and census tract-level social determinants of health (SDOH) with sociodemographic factors. A comparative analysis of the ML models was conducted using key performance metrics such as AUC-ROC to evaluate their predictive accuracy. Spatio-temporal analysis was used to identify disparities in late-stage CRC diagnosis probabilities. RESULTS: Gradient boosting emerged as the superior model, with the top predictors for late-stage CRC diagnosis being anatomic site, year of diagnosis, age, proximity to superfund sites, and primary payer. Spatio-temporal clusters highlighted geographic areas with a statistically significant high probability of late-stage diagnoses, emphasizing the need for targeted healthcare interventions. CONCLUSIONS: This research underlines the potential of ML in enhancing the prognostic predictions in oncology, particularly in CRC. The gradient boosting model, with its robust performance, holds promise for deployment in healthcare systems to aid early detection and formulate localized cancer prevention strategies. The study's methodology demonstrates a significant step toward utilizing AI in public health to mitigate disparities and improve cancer care outcomes.

COVID-19 Case, Death, and Vaccination Rate Disparities by Race and Ethnicity in Virginia.

This paper tracks trends in COVID-19 case, death, and vaccination rate disparities by race and ethnicity in Virginia during the COVID-19 pandemic. COVID-19 case, death, and vaccination rates were obtained from electronic state health department records from March 2020 to February 2022. Rate ratios were then utilized to quantify racial and ethnic disparities for several time periods during the pandemic. The Hispanic population had the highest COVID-19 case and age-adjusted death rates, and the lowest vaccination rates at the beginning of the pandemic in Virginia. These disparities resolved later in the pandemic. COVID-19 case and death rates among the Black population were also higher than those of the White population and these disparities remained throughout the pandemic. Racial and ethnic disparities changed over time in Virginia as vaccination coverage and public health policies evolved. Year 2 of the analysis saw lower case and death rates, and higher vaccination rates for non-White populations in Virginia. Public health strategies need to be addressed during the pandemic and developed before the next pandemic to ensure that large racial and ethnic disparities are not again present at the outset.

The Health Opportunity Index: Understanding the Input to Disparate Health Outcomes in Vulnerable and High-Risk Census Tracts.

The Health Opportunity Index (HOI) is a multivariate tool that can be more efficiently used to identify and understand the interplay of complex social determinants of health (SDH) at the census tract level that influences the ability to achieve optimal health. The derivation of the HOI utilizes the data-reduction technique of principal component analysis to determine the impact of SDH on optimal health at lower census geographies. In the midst of persistent health disparities and the present COVID-19 pandemic, we demonstrate the potential utility of using 13-input variables to derive a composite metric of health (HOI) score as a means to assist in the identification of the most vulnerable communities during the current pandemic. Using GIS mapping technology, health opportunity indices were layered by counties in Ohio to highlight differences by census tract. Collectively we demonstrate that our HOI framework, principal component analysis and convergence analysis methodology coalesce to provide results supporting the utility of this framework in the three largest counties in Ohio: Franklin (Columbus), Cuyahoga (Cleveland), and Hamilton (Cincinnati). The results in this study identified census tracts that were also synonymous with communities that were at risk for disparate COVID-19 related health outcomes. In this regard, convergence analyses facilitated identification of census tracts where different disparate health outcomes co-exist at the worst levels. Our results suggest that effective use of the HOI composite score and subcomponent scores to identify specific SDH can guide mitigation/intervention practices, thus creating the potential for better targeting of mitigation and intervention strategies for vulnerable communities, such as during the current pandemic.

Assessing Emergency Preparedness and Response Capacity Using Community Assessment for Public Health Emergency Response Methodology: Portsmouth, Virginia, 2013.

OBJECTIVE: For the past decade, emergency preparedness campaigns have encouraged households to meet preparedness metrics, such as having a household evacuation plan and emergency supplies of food, water, and medication. To estimate current household preparedness levels and to enhance disaster response planning, the Virginia Department of Health with remote technical assistance from the Centers for Disease Control and Prevention conducted a community health assessment in 2013 in Portsmouth, Virginia. METHODS: Using the Community Assessment for Public Health Emergency Response (CASPER) methodology with 2-stage cluster sampling, we randomly selected 210 households for in-person interviews. Households were questioned about emergency planning and supplies, information sources during emergencies, and chronic health conditions. RESULTS: Interview teams completed 180 interviews (86%). Interviews revealed that 70% of households had an emergency evacuation plan, 67% had a 3-day supply of water for each member, and 77% had a first aid kit. Most households (65%) reported that the television was the primary source of information during an emergency. Heart disease (54%) and obesity (40%) were the most frequently reported chronic conditions. CONCLUSIONS: The Virginia Department of Health identified important gaps in local household preparedness. Data from the assessment have been used to inform community health partners, enhance disaster response planning, set community health priorities, and influence Portsmouth's Community Health Improvement Plan.

A closer look at the rural-urban health disparities: Insights from four major diseases in the Commonwealth of Virginia.

Health disparities are increasingly recorded in literature, but are much less understood in a rural-urban context. This study help bridges this gap through investigation of four major diseases in the Commonwealth of Virginia: cancer, stroke, cardiovascular disease and chronic obstructive pulmonary disease. We utilize a unique inpatient hospital discharge billing dataset, and construct average patient counts at ZIP-code level over 2006-2008 where covariates from alternative sources are merged (806 ZIP-code areas, 190 urban, 616 rural). Count data regressions are first fitted to identify possible regional-level factors that affect disease incidences. A system of equations with rural-urban specification are then estimated via seemingly unrelated regression techniques to account for possible associations among these diseases and correlations of errors, which is followed by disease-specific nonlinear Blinder-Oaxaca decompositions that compare the respective explanatory powers of observed characteristics and unobserved mechanisms. Results suggest that regional-level factors are significantly correlated with health outcomes in both rural and urban areas. The unknown mechanisms behind these linkages are different between rural and urban areas, and explain even larger proportions of the observed disparities. These findings confirm the role of regional-level factors in generating rural-urban health disparities, and call for further investigations of the causal mechanisms of such disparities that remain largely unknown.