Daily New Cases per 100k People by US State from the Johns Hopkins COVID-19 Resource Webpage
The claims of growing concern of statistical projections of increasing resurgence of COVID-19 in Pennsylvania has been made by Governor Wolf based on projections made by "his" expert on COVID-19 statistics, The PolicyLab at CHOP, The Children's Hospital of Philadelphia.
The Wolf Administration has used those projections (POLICYLAB’S COVID-19 RESPONSE) to tighten restrictions on the citizenry and businesses, much to the consternation of a great number of the Members of The General Assembly and potential harm to small businesses and the citizens.
We visited PolicyLab's website and found that the projections were not available online on Friday the 17th of July in the early afternoon.
Where did this new source of "accurate" projections come from?
Why is Dr. Wolf not using the well-verified reports and projections that nearly everyone else in the country uses, the University of Washington's respected Institute for Health Metrics and Evaluation (IHME)?
Alternately, of course, there is the widely described Johns Hopkins Models at their Corona Virus Resource Center; it shows no evident projection of COVID-19 growth in PA as shown in the figure above. This study, not a model’s projection, shows existing cases where PA has a very small uptick in a tiny case history list.
IHME has vastly different projections for Pennsylvania online at: https://covid19.healthdata.org/united-states-of-america/pennsylvania than that claimed by the Wolf Administration. It shows no significant growth projected in Covid cases, hospitalizations or deaths for the entire State whereas the PolicyLab has much more dire projections.
Which Model is More Accurate?
Interestingly, a recent publication of comparisons of the performance of several hundred statistical models used in projecting COVID-19 results is online at: IHME COVID-19 Model Comparison Team. Predictive performance of international COVID-19 mortality forecasting models. MedRxiv. 14 July 2020. doi:10.1101/2020.07.13.20151233.
Some of the finding are explained as follows:
They "reviewed 384 published and unpublished COVID-19 forecasting models, and evaluated seven models for which publicly available, multi-country, and date-versioned mortality estimates could be downloaded.
"These included those modeled by: DELPHI-MIT (Delphi), Youyang Gu (YYG), the Los Alamos National Laboratory (LANL), Imperial College London (Imperial), and three models produced by the Institute for Health Metrics and Evaluation (IHME), a curve fit model (IHME-CF), a hybrid curve fit and epidemiological compartment model (IHME-CF SEIR), and a hybrid mortality spline and epidemiological compartment model (IHME - MS SEIR).
"Collectively models covered 171 countries, as well as the 50 states of the United States, and Washington, D.C., and accounted for >99% of all reported COVID-19 deaths on July 11th, 2020"
(NOTE THAT THE EFFORTS AT POLICYLAB WERE NOT REPORTED BY IHME!)
Spoiler Alert:
"For the most recent models, released in June, at four weeks of forecasting the best performing model was the IHME-MS SEIR model, with a cumulative median absolute percent error of 6.4%, followed by YYG (6.5%) and LANL (8.0%)."
So why does Pennsylvania use a less-well known model in their projections, one with no statement of estimated error? Strange thing, indeed.
No comments:
Post a Comment