All of this has come as a surprise to our public health authorities and a welcome surprise to those of us watching. The Institute for Health Metrics and Evaluation, the I.H.M.E., is a well-regarded research center run from the University of Washington. The I.H.M.E. has produced detailed predictions, charting the expected course of this epidemic here in the United States as well as in other countries. Their model has done perhaps more than any other piece of academic research to shape our response to the coronavirus crisis.
So how accurate has that model been? Here's some numbers.
Initially, the I.H.M.E. predicted that on April 4th, the state of New York would require 65,000 hospital beds to handle infected patients. The low-end estimate of what the state would need was nearly 48,000 beds.
In fact, on April 4th, New York had fewer than 16,000 hospitalizations for coronavirus, and many other states fell far below the model's projections as well -- in many states. Over the weekend, the I.H.M.E. updated its model. Its projections across the country have been scaled back dramatically. And yet, they are still significantly overstated.
For example, as of April 7th, the updated I.H.M.E. model predicts that New York will need 25,000 hospital beds. As of Tuesday morning, the real number was just under 17,500. The new model also predicted that as of today, almost 6,600 people would be in intensive care, and the actual number is just under 4,600.
In Florida, the new model predicted 4,000 people would be hospitalized. The reality in Florida is that not even 2,000 are.
And it wasn't just Florida and New York. The I.H.M.E. got it wrong in state after state after state. By the way, for America, this is great news and we should celebrate it. It's much better than we thought. Though unfortunately, on the question of total deaths, the model has been more accurate, though it still tends to overshoot. For example, yesterday, the I.H.M.E. predicted 784 deaths for New York, the state finished the day with about 600.