Predicting increases in COVID-19 incidence to identify locations for targeted testing in West Virginia: A machine learning enhanced approach

During the COVID-19 pandemic, West Virginia developed an aggressive SARS-CoV-2 testing strategy which included utilizing pop-up mobile testing in locations anticipated to have near-term increases in SARS-CoV-2 infections. This study describes and compares two methods for predicting near-term SARS-CoV-2 incidence in West Virginia counties. The first method, Rt Only, is solely based on producing forecasts for each county using the daily instantaneous reproductive numbers, Rt. The second method, ML+Rt, is a machine learning approach that uses a Long Short-Term Memory network to predict the near-term number of cases for each county using epidemiological statistics such as Rt, county population information, and time series trends including information on major holidays, as well as leveraging statewide COVID-19 trends across counties and county population size. Both approaches used daily county-level SARS-CoV-2 incidence data provided by the West Virginia Department Health and Human Resources beginning April 2020. The methods are compared on the accuracy of near-term SARS-CoV-2 increases predictions by county over 17 weeks from January 1, 2021- April 30, 2021. Both methods performed well (correlation between forecasted number of cases and the actual number of cases week over week is 0.872 for the ML+Rt method and 0.867 for the Rt Only method) but differ in performance at various time points. Over the 17-week assessment period, the ML+Rt method outperforms the Rt Only method in identifying larger spikes. Results show that both methods perform adequately in both rural and non-rural predictions. Finally, a detailed discussion on practical issues regarding implementing forecasting models for public health action based on Rt is provided, and the potential for further development of machine learning methods that are enhanced by Rt.

Predicting increases in COVID-19 incidence to identify locations for targeted testing in West Virginia: A machine learning enhanced approach

During the COVID-19 pandemic, West Virginia developed an aggressive SARS-CoV-2 testing strategy which included utilizing pop-up mobile testing in locations anticipated to have near-term increases in SARS-CXoV-2 infections. In this study, we describe and compare two methods for predicting near-term SARS-CoV-2 incidence in West Virginia counties. The first method, Rt Only, is solely based on producing forecasts for each county using the daily instantaneous reproductive numbers, Rt. The second method, ML+ Rt, is a machine learning approach that uses a Long Short-Term Memory network to predict the near-term number of cases for each county using epidemiological statistics such as Rt , county population information, and time series trends including information on major holidays, as well as leveraging statewide COVID-19 trends across counties and county population size. Both approaches used daily county-level SARS-CoV-2 incidence data provided by the West Virginia Department Health and Human Resources beginning April 2020. The methods are compared on the accuracy of near-term SARS-CoV-2 increases predictions by county over 17 weeks from January 1, 2021- April 30, 2021. Both methods performed well (correlation between forecasted number of cases and the actual number of cases week over week is 0.872 for the ML+Rt method and 0.867 for the Rt Only method) but differ in performance at various time points. Over the 17-week assessment period, the ML+Rt method outperforms the Rt Only method in identifying larger spikes. We also find that both methods perform adequately in both rural and non-rural predictions. Finally, we provide a detailed discussion on practical issues regarding implementing forecasting models for public health action based on Rt, and the potential for further development of machine learning methods that are enhanced by Rt.

Relationship Between COVID-19 Cases and Vaccination Rates in New York State Counties

Introduction: Vaccines against SARS-CoV-2 have been developed with unprecedented speed. The phased introduction of vaccines may be serving to offset the impact of new viral variants and policy relaxation. In order to assess the impact of vaccination, we examined a snapshot of vaccination rates across counties in a single state, at a single time point, comparing them with population-adjusted case counts. Methods: We calculated descriptive statistics and bivariate correlations for vaccination rates and cases across counties in New York State (NYS). We conducted a linear regression using cases/100K population per NYS county, frozen at a single snapshot in time, as the outcome variable, predicted by percentage of each county’s population (completed series/two doses), controlling for county population. Results: Percentages with one dose and with two doses were highly correlated (r=.935, P<.001) with one another. Both the one dose and two dose z rates were negatively correlated with cases per 100K population (not significant). Population size was strongly correlated with cases per 100K (r=.715, P<.001). The two-dose vaccination rate was a significant negative predictor of cases per 100K population in NYS counties (β= -.866, P=.031), with each percentage point of completed vaccination nearly equating to one case less in the daily count when controlling for county population size (β =2.732, P<.001). Conclusion: While variants may impact vaccine effectiveness, current vaccination efforts are helping forestall some cases in NYS. Widespread vaccination is still an important goal. Primary care providers, public officials, and public health scientists should continue to urgently promote and support vaccination efforts.

Public Protests and the Risk of Novel Coronavirus Disease Hospitalizations: A County-Level Analysis from California

The objective of this study was to assess the relationship between public protests and county-level, novel coronavirus disease (COVID-19) hospitalization rates across California. Publicly available data were included in the analysis from 55 of 58 California state counties (29 March–14 October 2020). Mixed-effects negative binomial regression models were used to examine the relationship between daily county-level COVID-19 hospitalizations and two main exposure variables: any vs. no protests and 1 or >1 protest vs. no protests on a given county-day. COVID-19 hospitalizations were used as a proxy for viral transmission since such rates are less sensitive to temporal changes in testing access/availability. Models included covariates for daily county mobility, county-level characteristics, and time trends. Models also included a county-population offset and a two-week lag for the association between exposure and outcome. No significant associations were observed between protest exposures and COVID-19 hospitalization rates among the 55 counties. We did not find evidence to suggest that public protests were associated with COVID-19 hospitalization within California counties. These findings support the notion that protesting during a pandemic may be safe, ostensibly, so long as evidence-based precautionary measures are taken.

SARS-CoV-2 antibody seroprevalence among firefighters in Orange County, California

ObjectivesWe conducted serological SARS-CoV-2 antibody testing from October to November 2020 to estimate the SARS-CoV-2 seroprevalence among firefighters/paramedics in Orange County (OC), California.MethodsOC firefighters employed at the time of the surveillance activity were invited to participate in a voluntary survey that collected demographic, occupational and previous COVID-19 testing data, and a SARS-CoV-2 immunoglobulin (Ig)G antibody blood test. We collected venous blood samples using mobile phlebotomy teams that travelled to individual fire stations, in coordination with an annual tuberculosis testing campaign for firefighters employed by OC Fire Authority (OCFA), and independently for firefighters employed by cities. We estimated seroprevalence and assessed several potential predictors of seropositivity.ResultsThe seroprevalence was 5.3% among 923 OCFA personnel tested, with 92.2% participating. Among firefighters self-reporting a previous positive COVID-19 antibody or PCR test result, twenty-one (37%) did not have positive IgG tests in the current serosurvey. There were no statistically significant differences in demographic characteristics between cases and non-cases. Work city was a significant predictor of case status (p=0.015). Seroprevalence (4.8%) was similar when aggregated across seven city fire departments (42%–65% participation). In total, 1486°C fire personnel were tested.ConclusionUsing a strong serosurvey design and large firefighter cohort, we observed a SARS-CoV-2 IgG seroprevalence of 5.3%. The seroprevalence among OC firefighters in October 2020 was lower than the general county population estimated seroprevalence (11.5%) in August. The difference may be due in part to safety measures taken by OC fire departments at the start of the pandemic, as well as differences in antibody test methods and/or duration of antibody response.

Research on Transfer Optimization Model of County Transit Network

County transit is an important mode that connects the county center with the surrounding countryside. This paper addresses the problem of unreasonable transit network planning, inconvenient operational optimizations, and protections in the country transit network system to build the transfer optimization model of the county transit network. The model that maximizes the synchronization reach operates in the “end-point connection”, which is the most suitable layout mode by analyzing the characteristics of county transit passenger flow and for comparing different layout modes. Typical county-level cities in three urban agglomerations in China were chosen as cases to validate the effectiveness and practicability of the proposed model. The case results are compared and analyzed in terms of the network density, departure interval, county population, and economic development level, which give theoretical support for decision-making in the planning, construction, and operation management of public transportation in China’s counties.

Using Synthetic Adjustments and Controlling to Improve County Population Forecasts from the Hamilton–Perry Method

Tayman and Swanson (J Popul Res 34(3):209–231, 2017) found in Washington State counties that a forecast based on the Hamilton–Perry method using a synthetic adjustment (SYN) of cohort change ratios and child-woman ratios had greater accuracy and less bias compared to forecasts holding these ratios constant (CONST). In this paper, we assess the robustness of SYN’s efficacy by evaluating forecast accuracy, bias, and distributional error across age groups in counties nationwide. We also investigate whether forecast errors and their patterns change for SYN and CONST if forecasts by age and gender are adjusted to an independent total population forecast for each county. Our main findings are as follows: (1) SYN lowers forecast error compared to CONST whether the forecasts are controlled or not; (2) controlling also leads to the improvements in forecast error, often exceeding those in SYN; and (3) using SYN and controlling together has the greatest effect in reducing forecast error. These findings remain after controlling for population size and growth rate, but the positive impacts on forecast error of SYN and controlling are most evident in counties with less than 30,000 population and that grow by 15% or more.