When COVID-19 Social Distancing Can Relax

New predictions (from the 17th) have some very good news: COVID-19 estimation updates | Institute for Health Metrics and Evaluation


A focus on when US states could consider easing social distancing if containment measures are in place

  • Based on our initial estimates, the earliest that some states may be able to ease presently implemented distancing policies – conditional on strong containment measures – appears to be in early to mid-May (lighter greens in the map below). This means, given the current epidemic trajectories and assuming these states have instituted robust containment strategies (e.g., widely available testing, contact tracing and case-based isolation, restricting mass gatherings), states including Alaska, Hawaii, Idaho, Montana, New Hampshire, North Carolina, Vermont, and West Virginia could potentially fall below 1 prevalent case of COVID-19 per 1,000,000 (1 million) around the first or second week of May.

Are the US actually gearing up for such an extensive program of testing, tracing and isolation? Here in the UK the Government seems oblivious to the need for this. They announced that they are developing a mobile phone App to keep track of people’s contacts and warn if they have been close to someone infected. On its own such an App is going to be insufficient, what is also needed is an army of trained people to go out and talk to people to understand their movements and contacts. Not everybody has a mobile, the elderly, the homeless, illegal immigrants will fall through the cracks of such a scheme and keep providing pools for the virus to persist and spread again. They don’t have plans to test the entire population.

Does anyone know how accurate the IHME projections have been so far?


Predictions of past models are archived and true values are public. They have a change log that shows how they are progressively improving the modeling over time.

There has been a great deal of controversy about this model because decision are being based on it. The primary purpose is to guide planning by health care organizations, and almost every hospital in the country is relying on their forecasts.

1 Like

They have been criticized, one reason being that their models are not based in anything physical. That being said, it seems that the main “criticism” in this article is that the confidence intervals in their predictions are very large, and some people don’t understand that every prediction only makes sense with these uncertainties, instead of a single number.


3 posts were merged into an existing topic: Analyze the models?

Not at the federal level. It seems like the states are being left to fend for themselves while the ‘stable genius in chief’ tweets, undermining any concerted effort. Some states are truly screwed thanks to either poor infrastructure and/or incompetent governors/legislatures.

They are some states and coalitions of states that are gearing up for screening and contact tracing but I’d suspect almost none are within a couple months of hitting their necessary strides. So, probably a very gradual relaxing of restrictions, short term… Again, mostly confined to states with competent leadership.

1 Like

But that assumption simply isn’t true, so it’s not good news.

1 Like