A Travesty of Ages and Games with Models, 5/5/20 Data


One hesitates to speak too strongly during unsettled times, but Monique Chartier’s posts in this space yesterday (here and here) led me to ask a question.  And that question leads to a conclusion that our government officials are choosing public policy very poorly in the face of COVID-19.

Maybe at the beginning of the outbreak, when there were many more unknowns, their harsh reactions seemed justified.  As information has come in, however, failing to adjust course is indefensible.

Specifically, Monique pointed out that the reported case fatality rate of COVID-19 is very low, and the infection fatality rate (attempting to account for the people who have caught the disease and were never tested, probably because they had no significant symptoms) is much, much lower.

Reasonable objections to that point do exist, however.  First, COVID-19 appears to spread more easily than the flu.  Second, it is a new bug for which we have less immunity and fewer treatments.  And third, the danger to vulnerable populations, especially the elderly, appears to be more acute than the flu.  As the equation went early in this pandemic, in a nation with 330 million people, if 60% of the population catches the disease and 2% die, that’s still almost 4 million people.  And if they are mostly elderly, the disease could wipe out a big percentage of our older generations.

Given what we now know, however, the idea that shutting down our entire economy and reopening it at a tip-toe pace is less and less reasonable.  Consider this chart of the case fatality rate by age group in Rhode Island.



In short, there is essentially no risk of death for anybody under 30, and survival rates for those in their thirties or forties are higher than 99.5%.  To give that finding a bit of context, consider the percentage of the United States labor force accounted for by each age group.



If we assume that Rhode Island’s labor force is close to the national average, we can say that around 67% of all workers are at almost no risk of death from COVID-19.  It follows that many of the people who actually need to lock down and hide from COVID-19 are retired and wouldn’t lose any income at all if they had to stay home for several months.  Thus, the population most at risk of the disease is least affected, economically, when we shut everything down, and those most hurt by shutting everything down are the least at risk from the disease.  This is a massive and painful transfer of risk from one age bracket to another.

Carry this over to hospitalizations.  While between 2.5% and 8.6% of Rhode Islanders in their twenties through their forties might end up in the hospital if they catch COVID-19, those with preexisting conditions (often being aware of those conditions) would likely account for most of those.  Moreover, we are now prepared to address an increase in hospitalizations of people who are at very low risk of death.

We should also think about different age groups’ types of work.  Retail and leisure/hospitality are two of the industries most affected by shutdowns, and they’ll probably be among the last to get back to normal as the governor eases restrictions little by little.  Yet, employees in these industries skew even more toward the safer end of the COVID-19 spectrum.



We could keep going.  As American workers move toward retirement, they are more likely to have savings and vacation days.  They are also more likely to be working from home already, or to have the ability to work from home for an extended period

In summary, the Rhode Islanders most at risk of harm from COVID-19 are the least likely to be working for the bulk of their income.  Those in the next age range, older workers, are more likely to be able to avoid catching the disease.  Meanwhile, the workers who are most likely to come into contact with COVID-19 at work are at almost no risk of death from the disease, yet they are the most harmed by our reaction to it.  (Anybody want to guess how the near-certain increase in suicides and overdoses will look from one age group to the next?)

It’s time to rethink our response to this disease.  We are causing massive harm to the generations that are most vulnerable to economic restrictions in order to protect the generations who are at risk of the disease, but could be more easily protected.

As my daily RI hospitalization chart shows, we continue to move out down from the peak (even the newly-defined peak that includes people who are hospitalized for some other reason, but who test positive for the virus).  We can therefore afford to move more quickly.



  • Cases:
    • Projection for 5/5: 9,805
    • Actual for 5/5: 9,933
    • Projection for 5/6: 10,098
  • Hospitalizations:
    • Projection for 5/5: 323
    • Actual for 5/5: 327
    • Projection for 5/6: 312
  • Deaths:
    • Projection for 5/5: 346
    • Actual for 5/5: 355
    • Projection for 5/6: 361

  • Mario

    I’m going to say 329 hospitalizations for tomorrow (+21, -12.5, -6.7), it would have been down but more discharges than expected one day is just borrowing from the next. I have ten deaths for tomorrow, but the last three days have seen an average of 15, not counting the old revisions nor revisions to these days yet to come. This is basically as bad as it was back when we had fifty over three days back in mid-April. I didn’t think that would happen. Just a week ago it looked the deaths were trending down fast, even in retrospect. Now we have a second peak? There’s no hospitalization second peak, and the cases are messy and test-dependent, but behaving themselves nonetheless.

    I never had a good idea for the case growth, I was just always assuming a constant downward path once we hit the peak. But now it’s necessary to build one if I want to project how long this will last, and because everything depends on how much testing there is, and my spreadsheet is just a wacky stew of circular references at this point, I’m going to say that there will be 1803 new negative tests. It’s a weird way to look at it, but if I can get close there I can project an endgame, at least until the social distancing completely collapses.

    As for the risk of opening, I am assuming that only 4% of the state has contracted the virus so far, which would mean we have about 21,000 infected people still walking around with it (I know I had a lower number a few days ago, and it’s still going down, but as people die faster than I anticipated, the higher it turns out to have been). The funny thing is that if you assume that I am overestimating how deadly it is, that means I am underestimating how many carriers we have, and if I’m overestimating the infected population, the deadlier it is. There’s really no way to win.

    So while I don’t think you are wrong when it comes to the relative risk of different groups, I think you are underestimating how easily this can spread and how difficult it would be to keep the one group safe while the virus runs rampant through the other. People still have to work in nursing homes, for instance, and hospitals. If every low risk person is basically being allowed to catch it just because it’s unlikely to kill them, no one outside that group will be safe either.