Challenges of COVID-19 Case Forecasting in the US, 2020–2021

PLoS Computational Biology, Journal Year: 2024, Volume and Issue: 20(5), P. e1011200 - e1011200

Published: May 6, 2024

During the COVID-19 pandemic, forecasting trends to support planning and response was a priority for scientists decision makers alike. In United States, coordinated by large group of universities, companies, government entities led Centers Disease Control Prevention US Forecast Hub ( https://covid19forecasthub.org ). We evaluated approximately 9.7 million forecasts weekly state-level cases predictions 1–4 weeks into future submitted 24 teams from August 2020 December 2021. assessed coverage central prediction intervals weighted interval scores (WIS), adjusting missing relative baseline forecast, used Gaussian generalized estimating equation (GEE) model evaluate differences in skill across epidemic phases that were defined effective reproduction number. Overall, we found high variation individual models, with ensemble-based outperforming other approaches. generally higher larger jurisdictions (e.g., states compared counties). Over time, performed worst periods rapid changes reported (either increasing or decreasing phases) 95% dropping below 50% during growth winter 2020, Delta, Omicron waves. Ideally, case could serve as leading indicator transmission dynamics. However, while most outperformed naïve model, even accurate unreliable key phases. Further research improve indicators, like cases, leveraging additional real-time data, addressing performance phases, improving characterization forecast confidence, ensuring coherent spatial scales. meantime, it is critical users appreciate current limitations use broad set indicators inform pandemic-related making.

Language: Английский

---

Impact of close interpersonal contact on COVID-19 incidence: Evidence from 1 year of mobile device data

Science Advances, Journal Year: 2022, Volume and Issue: 8(1)

Published: Jan. 7, 2022

Close contact between people is the primary route for transmission of SARS-CoV-2, virus that causes coronavirus disease 2019 (COVID-19). We quantified interpersonal at population level using mobile device geolocation data. computed frequency (within 6 feet) in Connecticut during February 2020 to January 2021 and aggregated counts events by area residence. When incorporated into a SEIR-type model COVID-19 transmission, rate accurately predicted cases towns. Contact explains initial wave infections March April, drop June August, local outbreaks August September, broad statewide resurgence September December, decline 2021. The fits dynamics better than other mobility metrics. data can help guide social distancing testing resource allocation.

Language: Английский

---

Lessons Learned from the Lessons Learned in Public Health during the First Years of COVID-19 Pandemic

International Journal of Environmental Research and Public Health, Journal Year: 2023, Volume and Issue: 20(3), P. 1785 - 1785

Published: Jan. 18, 2023

(1) Objectives: to investigate the main lessons learned from public health (PH) response COVID-19, using global perspective endorsed by WHO pillars, and understand what countries have their practical actions. (2) Methods: we searched for articles in PubMed CINAHL 1 January 2020 31 2022. 455 were included. Inclusion criteria PH themes COVID-19 pandemic. One hundred forty-four finally included a detailed scoping review. (3) Findings: 78 available, cited 928 times 144 articles. Our review highlighted 5 among regions: need continuous coordination between institutions organisations (1); importance of assessment evaluation risk factors diffusion identifying vulnerable populations (2); establishment systems assess impact planned measures (3); extensive application digital technologies, telecommunications electronic records (4); periodic scientific reviews provide regular updates on most effective management strategies (5). (4) Conclusion: found this could be essential future, providing recommendations an increasingly flexible, fast efficient healthcare emergency such as

Language: Английский

---

Using simulation modelling and systems science to help contain COVID‐19: A systematic review

Systems Research and Behavioral Science, Journal Year: 2022, Volume and Issue: 40(1), P. 207 - 234

Published: Aug. 19, 2022

This study systematically reviews applications of three simulation approaches, that is, system dynamics model (SDM), agent-based (ABM) and discrete event (DES), their hybrids in COVID-19 research identifies theoretical application innovations public health. Among the 372 eligible papers, 72 focused on transmission dynamics, 204 evaluated both pharmaceutical non-pharmaceutical interventions, 29 prediction pandemic 67 investigated impacts COVID-19. ABM was used 275 followed by 54 SDM 32 DES papers 11 hybrid papers. Evaluation design intervention scenarios are most widely addressed area accounting for 55% four main categories, COVID-19, pandemic, evaluation societal impact assessment. The complexities demand models can simultaneously capture micro macro aspects socio-economic systems involved.

Language: Английский

---

Wastewater surveillance provides 10-days forecasting of COVID-19 hospitalizations superior to cases and test positivity: A prediction study

Infectious Disease Modelling, Journal Year: 2023, Volume and Issue: 8(4), P. 1138 - 1150

Published: Oct. 31, 2023

The public health response to COVID-19 has shifted reducing deaths and hospitalizations prevent overwhelming systems. amount of SARS-CoV-2 RNA fragments in wastewater are known correlate with clinical data including cases hospital admissions for COVID-19. We developed tested a predictive model incident New York State using data. Using county-level surveillance covering 13.8 million people across 56 counties, we fit generalized linear mixed predicting new from concentrations April 29, 2020 June 30, 2022. included covariates such as vaccine coverage the county, comorbidities, demographic variables, holiday gatherings. Wastewater correlated per 100,000 up ten days prior admission. Models that had higher power than models only, increasing accuracy by 15%. Predicted highly observed (r = 0.77) an average difference 0.013 (95% CI [0.002, 0.025]) predict future is accurate effective superior results case alone. lead time could alert take precautions improve resource allocation seasonal surges.

Language: Английский

---

Semiparametric modeling of SARS-CoV-2 transmission using tests, cases, deaths, and seroprevalence data

The Annals of Applied Statistics, Journal Year: 2024, Volume and Issue: 18(3)

Published: Aug. 6, 2024

Mechanistic models fit to streaming surveillance data are critical understanding the transmission dynamics of an outbreak as it unfolds in real-time. However, model parameter estimation can be imprecise, and sometimes even impossible, because noisy not informative about all aspects mechanistic model. To partially overcome this obstacle, Bayesian have been proposed integrate multiple streams. We devised a modeling framework for integrating SARS-CoV-2 diagnostics test mortality time series data, well seroprevalence from cross-sectional studies, tested importance individual streams both inference forecasting. Importantly, our incidence accounts changes total number tests performed. rate, infection-to-fatality ratio, controlling functional relationship between true case fraction positive time-varying quantities estimate these parameters nonparametrically. compare base against modified versions which do use counts or demonstrate utility including often unused apply integration method COVID-19 collected Orange County, California March 2020 February 2021 find that 32-72% County residents experienced infection by mid-January, 2021. Despite high infections, results suggest abrupt end winter surge January was due behavioral level accumulated natural immunity.

Language: Английский

---

EINNs: Epidemiologically-Informed Neural Networks

Proceedings of the AAAI Conference on Artificial Intelligence, Journal Year: 2023, Volume and Issue: 37(12), P. 14453 - 14460

Published: June 26, 2023

We introduce EINNs, a framework crafted for epidemic forecasting that builds upon the theoretical grounds provided by mechanistic models as well data-driven expressibility afforded AI models, and their capabilities to ingest heterogeneous information. Although neural have been successful in multiple tasks, predictions well-correlated with trends long-term remain open challenges. Epidemiological ODE contain mechanisms can guide us these two tasks; however, they limited capability of ingesting data sources modeling composite signals. Thus, we propose leverage work physics-informed networks learn latent dynamics transfer relevant knowledge another network which ingests has more appropriate inductive bias. In contrast previous work, do not assume observability complete need numerically solve equations during training. Our thorough experiments on all US states HHS regions COVID-19 influenza showcase clear benefits our approach both short-term learning over other non-trivial alternatives.

Language: Английский

---

COVID-19 Testing and Case Rates and Social Contact Among Residential College Students in Connecticut During the 2020-2021 Academic Year

JAMA Network Open, Journal Year: 2021, Volume and Issue: 4(12), P. e2140602 - e2140602

Published: Dec. 23, 2021

During the 2020-2021 academic year, many institutions of higher education reopened to residential students while pursuing strategies mitigate risk SARS-CoV-2 transmission on campus. Reopening guidance emphasized polymerase chain reaction or antigen testing for and social distancing measures reduce frequency close interpersonal contact, Connecticut colleges universities used a variety approaches reopen campuses students.To characterize institutional reopening COVID-19 outcomes in 18 college university across Connecticut.This retrospective cohort study data cases contact from that had during year.Tests performed per week student.Cases student mean (95% CI) student.Between 235 4603 attended fall semester each Connecticut, with fewer at most spring semester. In census block groups containing residence halls, move-in resulted 475% CI, 373%-606%) increase 561% 441%-713%) compared 7 weeks prior move-in. The association between test case rate was complex; tested infrequently detected few but failed blunt transmission, whereas more frequently prevented further spread. 2020, additional associated decrease 0.0014 -0.0028 -0.00001).The findings this suggest that, era available vaccinations highly transmissible variants, should continue use mitigation control on-campus cases.

Language: Английский

---

A Mathematical and Computational Model for Multiple COVID-19 Waves Applied to Kenya

Journal of Applied Mathematics and Physics, Journal Year: 2025, Volume and Issue: 13(04), P. 1323 - 1351

Published: Jan. 1, 2025

Language: Английский

---

Modeling county level COVID-19 transmission in the greater St. Louis area: Challenges of uncertainty and identifiability when fitting mechanistic models to time-varying processes

Mathematical Biosciences, Journal Year: 2024, Volume and Issue: 371, P. 109181 - 109181

Published: March 25, 2024

We use a compartmental model with time-varying transmission parameter to describe county level COVID-19 in the greater St. Louis area of Missouri and investigate challenges fitting such processes. fit this synthetic real confirmed case hospital discharge data from May December 2020 calculate uncertainties resulting estimates. also explore non-identifiability within estimated set. determine that death rate infectious non-hospitalized individuals, testing initial number exposed individuals are not identifiable based on an investigation correlation coefficients between pairs how ties back into parameters find it inflates uncertainty estimates our parameter. However, we do R0 is highly affected by its constituent components associated quantity smaller than those parameters. Parameter values will always be some work highlights importance conducting these analyses when models data. Exploring identifiability crucial revealing much can trust

Language: Английский

---