In a recent comment published in PNAS*, the researchers retrospectively analyzed the collaborative modeling attempt developed by the University of Texas (UT) at Austin and described by Fox et alto guide public health decisions during the coronavirus disease 2019 (COVID-19) pandemic.
Modeling and prediction attempts have helped local, state and national authorities make public health decisions during the COVID-19 pandemic by raising awareness of the current situation, providing information on severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) characteristics and optimization strategies to mitigate COVID-19.
Attempts at forecasting are the most obvious modeling results to the public since forecasts are frequently highlighted by the media. However, several other modeling attempts have also been instrumental in mitigating COVID-19.
The UT model predicted the health care burden of COVID-19 focusing on hospital metrics (hospital admissions, intensive care unit utilization [ICU] beds and hospital beds) which are important indicators of the health care burden of COVID-19.
For external validation, the team compared its model’s predictions for the cumulative fraction of people infected with an independent data source, Centers for Disease Control and Prevention (CDC) infection rate estimates.
Additionally, the authors carefully integrated mobility data to improve the accuracy of the model. However, the extent to which mobility data accounts for uncertainty depends on the situation, for example, with high vaccination coverage and high mask-wearing sites, mobility data would have a small role.
Model results and interpretation
Hospital admissions data could accurately and timely indicate recent viral transmission and upcoming use of hospital beds and ICU beds in the short term (one to two weeks). On the other hand, case data were poor indicators of the possible burden of COVID-19 on health care, as they showed a significantly weak correlation, likely due to changing trends in care-seeking. and case notification.
The extent of the UT model was commendable. The model not only predicted healthcare burden, but also provided a real-time estimate of the reproduction number, which is an estimate of the transmission rate of pathogens. Thus, it could help provide decision-makers with instant and urgent information that would facilitate resource planning by local hospitals, provide federal and state authorities with requests for additional resources in the event of future pandemic outbreaks, and increase sites of care to improve care capabilities.
Additionally, the availability of experts in the field of modeling and data processing to monitor data from the COVID-19 pandemic has enabled more confident government operations. Experts could make adjustments to the data and appropriately interpret the information taking into account disruptions and unforeseen situations such as the unusual Texas winter freeze during the initial period of 2021.
The authors believed that the mobility-focused mechanistic UT model was an honest, conservative, accurate, externally valid, and sensible attempt at real-time and extended collaborative modeling during the COVID-19 pandemic that was tailored to the needs of special health care highlighted. by city officials.
Moreover, according to the authors, the model had a wide scope because the model could not only predict the short-term prospective burden of COVID-19 on health care, but also provide instantaneous policy feedback. Data adjustments by model experts and incorporation of strong evidence could also inspire more confidence in government for public decision-making in Austin.
The authors of this study believe that the attempt at collaborative modeling, described by Fox et al, is notable for its high accuracy and extensive, trust-based collaboration with Austin city officials. They believe that while this attempt would serve as an exemplary model for future collaborations of the same type, the scalability of the model is questionable.
Identifying the most valuable, real-time data streams for model estimates, incorporating methods other than mobility data to account for uncertainty, would improve epidemic modeling. In addition, the use of “modeling centers” such as the United States (US) COVID-19 Prediction Center and the Scenario Modeling Center which would consolidate model outputs from multiple academic modeling teams and industries and instantly generate results for multiple locations would improve availability. of the model.
Continued investments in data modernization, modeling technology and workforce development would improve model scalability and modeling capacity to enable local jurisdictions to benefit from the models.