No abstract

In this paper, we consider compartmental disease transmission models and discuss the importance of determining model parameters that provide an insight into disease transmission and prevalence. After a brief review and comparison of the performance of some heuristic approaches, the paper introduces three approaches including an optimization approach, a physics informed deep learning and a statistical inference method to estimate parameters and analyse disease transmission. The deep learning framework utilizes the hidden physics of infectious diseases and infer the latent quantities of interest in the model by approximating them using deep neural networks. The performance of the deep learning method is validated against representative small and big data sets corresponding to a well-known benchmark example and the results indicate that deep learning is a viable candidate to determine model parameters. The paper also presents the need for researchers to understand different types of dependencies exhibited in a typical data set and discovering the most appropriate approaches for statistical inference. Specifically, in this work we apply a time-series inferential method with a variety of statistical models. Our results indicate the efficiency and importance of statistical inference methods for researchers to understand and analyse time-series data to make confident predictions.

In this paper, we consider compartmental disease transmission models and discuss the importance of determining model parameters that provide an insight into disease transmission and prevalence. After a brief review and comparison of the performance of some heuristic approaches, the paper introduces three approaches including an optimization approach, a physics informed deep learning and a statistical inference method to estimate parameters and analyse disease transmission. The deep learning framework utilizes the hidden physics of infectious diseases and infer the latent quantities of interest in the model by approximating them using deep neural networks. The performance of the deep learning method is validated against representative small and big data sets corresponding to a well-known benchmark example and the results indicate that deep learning is a viable candidate to determine model parameters. The paper also presents the need for researchers to understand different types of dependencies exhibited in a typical data set and discovering the most appropriate approaches for statistical inference. Specifically, in this work we apply a time-series inferential method with a variety of statistical models. Our results indicate the efficiency and importance of statistical inference methods for researchers to understand and analyse time-series data to make confident predictions.

Background The relationship between healthcare service accessibility in the community and incarceration is an important, yet not widely understood, phenomenon. Community behavioral health and the criminal legal systems are treated separately, which creates a competing demand to confront mass incarceration and expand available services. As a result, the relationship between behavioral health services, demographics and community factors, and incarceration rate has not been well addressed. Understanding potential drivers of incarceration, including access to community-based services, is necessary to reduce entry into the legal system and decrease recidivism. This study identifies county-level demographic, socioeconomic, healthcare services availability/accessibility, and criminal legal characteristics that predict per capita jail population across the U.S. More than 10 million individuals pass through U.S. jails each year, increasing the urgency of addressing this challenge. Methods The selection of variables for our model proceeded in stages. The study commenced by identifying potential descriptors and then using machine learning techniques to select non-collinear variables to predict county jail population per capita. Beta regression was then applied to nationally available data from all 3,141 U.S. counties to identify factors predicting county jail population size. Data sources include the Vera Institute’s incarceration database, Robert Wood Johnson Foundation’s County Health Rankings and Roadmaps, Uniform Crime Report, and the U.S. Census. Results Fewer per capita psychiatrists (z-score = -2.16; p = .031), lower percent of drug treatment paid by Medicaid (-3.66; p < .001), higher per capita healthcare costs (5.71; p < .001), higher number of physically unhealthy days in a month (8.6; p < .001), lower high school graduation rate (-4.05; p < .001), smaller county size (-2.66, p = .008; -2.71, p = .007; medium and large versus small counties, respectively), and more police officers per capita (8.74; p < .001) were associated with higher per capita jail population. Controlling for other factors, violent crime rate did not predict incarceration rate. Conclusions Counties with smaller populations, larger percentages of individuals that did not graduate high school, that have more health-related issues, and provide fewer community treatment services are more likely to have higher jail population per capita. Increasing access to services, including mental health providers, and improving the affordability of drug treatment and healthcare may help reduce incarceration rates.

Background Young adulthood is a period of increasing independence for the 40% of young adults enrolled in U.S. colleges. Previous research indicates differences in how students’ health behaviors develop and vary by gender, race, ethnicity, and socioeconomic status. George Mason University is a state institution that enrolls a highly diverse student population, making it an ideal setting to launch a longitudinal cohort study using multiple research methods to evaluate the effects of health behaviors on physical and psychological functioning, especially during the COVID-19 pandemic. Results Mason: Health Starts Here was developed as a longitudinal cohort study of successive waves of first year students that aims to improve understanding of the natural history and determinants of young adults’ physical health, mental health, and their role in college completion. The study recruits first year students who are 18 to 24 years old and able to read and understand English. All incoming first year students are recruited through various methods to participate in a longitudinal cohort for 4 years. Data collection occurs in fall and spring semesters, with online surveys conducted in both semesters and in-person clinic visits conducted in the fall. Students receive physical examinations during clinic visits and provide biospecimens (blood and saliva). Conclusions The study will produce new knowledge to help understand the development of health-related behaviors during young adulthood. A long-term goal of the cohort study is to support the design of effective, low-cost interventions to encourage young adults’ consistent performance of healthful behaviors, improve their mental health, and improve academic performance.