Marin’s projected coronavirus peak still hard to pinpoint
Author: Richard Halstead
Publication: Marin IJ
Marin health officials are taking longer than initially expected to forecast how large a surge of coronavirus patients the county will see and when the peak will arrive.
Dr. Lisa Santora, Marin County deputy public health officer, told county supervisors on March 31 that she expected to have the numbers later that week. On Friday, Santora said she couldn’t predict when final estimates might be completed.
Forecasts of how many coronavirus patients the county will have when the number of local infections peaks is critical for determining whether there are a sufficient number of intensive care beds, ventilators and medical personnel to respond.
Santora says she wants to collect additional local data before completing the modeling in an effort to improve the accuracy of the forecast.
One important decision the county must make in preparing the forecast is which model to use. Santora said the four best models have been produced by the Institute for Health Metrics and Evaluation in Seattle, the University of Pennsylvania, Stanford University and Imperial College London in the United Kingdom.
“That is our struggle now,” Santora said. “They have all developed their own analytical models to churn out when localities would expect to see a surge and what would be the height of surge. There is a tremendous variation between all four models.”
Santora said the county has plugged its local data into all three U.S. models and come up with a wide range of predictions. The peak of the surge in Marin varied from late April to late July. The maximum number of hospitalizations ranged from 200 to 1,500. The maximum number of deaths ranged from 80 to 2,000.
Santora said the county is contemplating entering into a contract with Stanford to produce more detailed modeling that is tailored to Marin. She said she is uncertain when that work might be completed.
The Seattle institute, which is funded by the Bill and Melinda Gates Foundation, made headlines this week by revising its projections downward for the number of hospital beds, ventilators and other equipment that will be needed to respond to the pandemic in the United States. The institute also said the number of deaths from the virus might peak sooner than expected in some states.
California is entering its fourth week of a statewide stay-at-home order that has shuttered schools and non-essential businesses. Last week, the state saw its first daily decrease in intensive care hospitalizations, a key indicator of how many health care workers and medical supplies the state needs, according to Gov. Gavin Newsom.
The rate of all virus hospitalizations slowed last week. The number of patients in the ICU statewide dropped 1.9% on Wednesday to 1,132, Newsom said Thursday.
“I caution anybody to read too much into that,” Newsom said of the decline in ICU hospitalizations. “But nonetheless, it is encouraging.”
The institute in Seattle is forecasting that the number of daily coronavirus deaths in California will peak at 67 on April 15.
Some skeptics, however, believe the institute’s model is overly optimistic. For example, they say the model combines mortality data out of China with mortality data from U.S. cities and counties to make its predictions. But U.S. jurisdictions so far have not implemented the same kind of drastic lockdown as China did.
The model also assumes that all states will extend social distancing through the end of May and that states that haven’t yet implemented such measures will do so in the next seven days.
More recently, the Seattle institute’s updated modeling has incorporated data from Italy and Spain.
Santora said most of the models use similar inputs such as current hospitalization rates, the rate at which the number of coronavirus infections are doubling and estimates of compliance with social distancing recommendations in the locale. She said the University of Pennsylvania’s model is predicting the highest number of hospitalizations and deaths. The District of Columbia has decided to use that model.
Santora said another factor complicating Marin’s modeling is that more infections than previously expected are resulting from travelers from New York. She said the prior working hypothesis was that most of the Bay Area’s infections would be caused by travelers from China.
“I think we are seeing some indications that we are flattening the curve,” Santora said, “but we have yet to determine the role of domestic travel from New York, before it was restricted. That confounds some of the modeling.”
Although a final forecast for Marin is still in the works, the county is assuming a surge of patients will materialize over the next two to six weeks.
“We really do believe we’re still on the tail of the curve up,” Santora said.
She said that while the state’s social distancing is flattening the curve, it might also be delaying the arrival of the outbreak’s local peak.
The county’s Department of Health and Human Services is working with MarinHealth Medical Center, Kaiser Permanente Medical Center, Sutter Health’s Novato Community Hospital and Kentfield Rehabilitation Hospital to prepare.
According to a statement issued by the county on Tuesday, even though the MarinHealth, Kaiser and Sutter hospitals are licensed for a total of 398 beds, they operate only a combined total of 239 beds under normal circumstances. For a hospital bed to be usable, it must be matched with necessary equipment and medical personnel.
“Collectively, Marin hospitals have identified opportunities to increase staffed bed capacity by about 67% to approximately 400 staffed beds, which would include nearly doubling intensive care unit beds,” the statement said.
Read the original publication here: