Enhanced SARS-CoV-2 case prediction using public health data and machine learning models.

Objectives: The goal of this study is to propose and test a scalable framework for machine learning (ML) algorithms to predict near-term severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) cases by incorporating and evaluating the impact of real-time dynamic public health data. Materials and Methods: Data used in this study include patient-level results, procurement, and location information of all SARS-CoV-2 tests reported in West Virginia as part of their mandatory reporting system from January 2021 to March 2022. We propose a method for incorporating and comparing widely available public health metrics inside of a ML framework, specifically a long-short-term memory network, to forecast SARS-CoV-2 cases across various feature sets. Results: Our approach provides better prediction of localized case counts and indicates the impact of the dynamic elements of the pandemic on predictions, such as the influence of the mixture of viral variants in the population and variable testing and vaccination rates during various eras of the pandemic. Discussion: Utilizing real-time public health metrics, including estimated Rt from multiple SARS-CoV-2 variants, vaccination rates, and testing information, provided a significant increase in the accuracy of the model during the Omicron and Delta period, thus providing more precise forecasting of daily case counts at the county level. This work provides insights on the influence of various features on predictive performance in rural and non-rural areas. Conclusion: Our proposed framework incorporates available public health metrics with operational data on the impact of testing, vaccination, and current viral variant mixtures in the population to provide a foundation for combining dynamic public health metrics and ML models to deliver forecasting and insights in healthcare domains. It also shows the importance of developing and deploying ML frameworks in rural settings.

Comparing the effects of decreasing prescription opioid shipments and the release of an abuse deterrent OxyContin formulation on opioid overdose fatalities in WV: an interrupted time series study.

INTRODUCTION: The 2010 release of an abuse deterrent formulation (ADF) of OxyContin, a brand name prescription opioid, has been cited as a major driver for the reduction in prescription drug misuse and the associated increasing illicit opioid use and overdose rates. However, studies of this topic often do not account for changes in supplies of other prescription opioids that were widely prescribed before and after the ADF OxyContin release, including generic oxycodone formulations and hydrocodone. We therefore sought to compare the impact of the ADF OxyContin release to that of decreasing prescription opioid supplies in West Virginia (WV). METHODS: Opioid tablet shipment and overdose data were extracted from The Washington Post ARCOS (2006-2014) and the WV Forensic Drug Database (2005-2020), respectively. Locally estimated scatterplot smoothing (LOESS) was used to estimate the point when shipments of prescription opioids to WV began decreasing, measured via dosage units and morphine milligram equivalents (MMEs). Interrupted time series analysis (ITSA) was used to compare the impact LOESS-identified prescription supply changes and the ADF OxyContin release had on prescription (oxycodone and hydrocodone) and illicit (heroin, fentanyl, and fentanyl analogues) opioid overdose deaths in WV. Model fit was compared using Akaike Information Criteria (AIC). RESULTS: The majority of opioid tablets shipped to WV from 2006 to 2014 were generic oxycodone or hydrocodone, not OxyContin. After accounting for a 6-month lag from ITSA models using the LOESS-identified change in prescription opioid shipments measured via dosage units (2011 Q3) resulted in the lowest AIC for both prescription (AIC = -188.6) and illicit opioid-involved overdoses (AIC = -189.4), indicating this intervention start date resulted in the preferred model. The second lowest AIC was for models using the ADF OxyContin release as an intervention start date. DISCUSSION: We found that illicit opioid overdoses in WV began increasing closer to when prescription opioid shipments to the state began decreasing, not when the ADF OxyContin release occurred. Similarly, the majority of opioid tablets shipped to the state for 2006-2014 were generic oxycodone or hydrocodone. This may indicate that diminishing prescription supplies had a larger impact on opioid overdose patterns than the ADF OxyContin release in WV.

Risk factors for casualty in earthquakes : The application of epidemiologic principles to structural engineering

Past investigations of the public health consequences of natural disasters, such as earthquakes, have generally suffered from a lack of comprehensive data due, in part, to the difficulty faced in mounting a significant data collection effort in the aftermath of such an event. This lack of meaningful data severely hampers the ability of casualty researchers to develop reliable and robust estimates of death and injury in future events; these numbers are critical in planning for mitigation and response activities. The basic data required as imput to these models are risk factors for injury which can only effectively be estimated through a careful epidemiologic study of these injured (and not injured) in past events. This paper outlines the field of earthquake injury epidemiolgy and discusses the application of the collected data. An example of a recent case control study (in progress) is given. Implications for the broader application of these techniques to structural engineering are suggested (AU)

Morbidity and mortality in the Loma Prieta earthquake : A review of recent findings

In 1989, the First International Workshop on Earthquake Injury Epidemiology for Mitigation and Response was held at The Johns Hopkins University under the sponsorship of the National Science Foundation and the National Center for Earthquake Engineering Research. In this worskshop, researches representing a broad range of disciplines gthered to assess the state of the art in earthquake injury epidemiology, and to plot a course for future quantitative research in this important but often neglected area. Since that workshop, a number of significant eartjquakes have occurred in the U.S. and elsewhere and have provided the opportunity for detailed follow-up on casualty. In this paper, a brief review of the findings to date of a detailed case-control study of morbidity and mortality during the Loma Prieta earthquake, partially funded by NCEER, will be given. As the full study is nearing completion at time of writing, only preliminary data are available, but more comprehensive results will be available within the next 12 month period.(AU)

Injuries and building data pertinent to the Loma Prieta earthquake county of Santa Cruz

Past investigation on the public health consequences of natural disaster (e.g., earthquakes) have generally suffered from a lack of comprehensive data due, in part, to the difficulty facedby researchers in mounting a significant data collection effortin the aftermath of such an event. For a meaningful assessment to be made, field data must be gathered not only on the victims of the events but olso on the nature of the damage sustained and the context of the injury.this paper outlines some preliminary data concerning the injury outcomes and the results of a survey conducted on the structures where both hospital and the results of a survey conducted on the structures whereboth hospital and dead cases and injured controls were located when injured (AU)

Chapter 2 : Casualty in earthquakes

Significant resources in research support and effort have been expended on the problem of earthquake hazard mitigation over the past twenty years. Most of this research effort has been directed toward questions of geophysical and structural engineering. While this expenditure of effort has been appropriate in terms of advancing scientific understanding of the underlying phenomena responsible for earthquake losses, the tragic fact is that earthquakes continue to injure and kill human beings. The specific mechanisms of death and injury in earthquakes have not yet been the subject of extensive study. This chapter critically reviews past efforts in the study of earthquake-induced morbidity and mortality, outlines the lessons learned from past events, and offers some suggestions for future research in this area and utilization of current knowledge in the Central and Eastern United States.(AU)

Considerations in the epidemiology of earthquake injuries

In July, 1989, a workshop entitled "Earthquake Injury Epidemiology for Mitigation and Response" was held at The Johns Hopkins University in Baltimore, Maryland, U.S.A. The aim of the workshop was to gather a group of interested professionals, all directly or peripherally interested in the research, planning, mitigation, and response aspects associated with earthquake-induced injuries and deaths, to lay the foundations and begin to develop a research agenda for this emerging field. This was achieved by a combination of presentation of summary papers, discussions in small, multidisciplinary working groups, and plenary wrap-up and discussion sessions. The paper presents a condensation and summary of the workshop, its discussions, and its important conclusions. References are cited where appropriate, much of the content reflects a condensation of the discussions(AU)