[08/19/20 Update: Actual Death and Hospitalization Data Made Current in Plots. Our predictions from late May are exceptionally accurate indicating NYC indeed reached the Herd Immunity Threshold in April]
New York City has flattened the curve and is well positioned for a full recovery. Here we align a model with current COVID-19 death and hospitalization data from the NYC department of public health: [NYC COVID 19 Data] The outbreak conditions in this simulation are aligned with a high base reproduction rate, R0, of 4, primarily driven by subway spread as documented by Jeffrey Harris: The Subways Seeded the Massive Coronavirus Epidemic in New York City. We examine an unmitigated outbreak (reference), a “light switch” and two variations of an “age bin” recovery model as described in our post: Age Based Recovery Strategy: An Optimal Balance
New York City is Ready!
The current R(t) for NYC is well below 1. Antibody testing completed in late April indicates that 25% of NYC population was infected: Almost a quarter of NYC residents test positive for COVID-19 antibodies: Cuomo. Our model indicates that the number is now closer to 33% considering infection growth over the month of May. The infection has been suppressed and is rapidly dying out.
A summary of analysis results are presented below for four cases. The “unmitigated” case is a reference case without a lockdown (hypothetical). The “light switch” model is a hard return to normal with protection for the vulnerable. The two “age bin” models show low restrictions on younger folks (reopening schools and businesses) and encouragement for older age folks to practice a higher degree of social distancing (people are naturally doing this). This analysis shows NYC can get back to work now and manage COVID-19 without depending on a vaccine to kill the virus. These totals below for susceptible (uninfected), recovered and deaths and infection and fatality rates are at the conclusion of the simulations when the infection has died out. The simulation population starts with 1 Million discrete agents. A scaling factor of 8.4 is used to represent the total population of NYC.
Final Simulation Results for NYC (May 2021), June 2020 Reopening Cases, Agent Population of 1 Million, 8.4 X Scale.
Cumulative and Daily Deaths versus Data
The analysis is fit to current death statistics and shows excellent correlation with both daily deaths and cumulative deaths. The “light switch” approach does show a minor second comeback. The “age bin” approaches minimize future fatalities.
Reproduction Factor Over Time: R(t)
The spread of the COVID-19 in NYC is well under control. The “light switch” approach kicks up slightly above 1 and dies off quickly. The “age bin” approach maintains negative growth. Eventually the disease dies out in all cases without the need for a vaccine.
Hospitalizations per Day
Actual NYC hospitalizations per day are shown for May 2020 onward are compared to modeled hospitalizations for the “light switch” and “age bin” scenarios. There is excellent correlation to actual data. The “age bin” approaches show a continued decline without any second wave.
Active Infection Tracking
Active infections show regrowth with the “light switch” scenario and a steady decline with the “age bin” scenarios. Infections peak in the range of 1.4 Million in March (16% of the population). By late summer with the “age bin” approach, the peak will have been reduced by over 50X with approximately 25,000 active infections (0.3% of the population). Note that less than a third of those currently infected are actually infectious if we take out those who are in the non-infectious incubation period and those who are quarantined, on average the infectious period is in the range of 6 days, but the average infection from exposure to recovery/death is closer to 20 days.
Time Phased Recovery Parameters
The time phased input parameters for each simulation are shown below. In the 3 recovery scenarios (“light switch”, “age bin 1”, “age bin 3”) a 25% improvement in mortality outcome is assumed due to improvements in treating critically ill people (phased in during the month of May and included in the calibration to actual deaths).
Summary Population Output Parameters
The summary outputs for each June recovery scenario are shown below. In each case a base simulation population of 1 million with a seed of 1125 infected individuals was used.
Age Bin 1 NYC June Recovery Summary: 1 Million Base Population 1125 Seed Count, Through May 2021
General Input Parameters
Core Simulation Parameters (All Cases) Demographic Bins (All Cases)