Wednesday, June 10, 2020

Latest Pro-Lockdown Studies Are Far Less Than Meets The Eye

The latest two studies that claim that lockdowns somehow prevented millions of COVID-19 cases from occurring are getting a lot of attention right now.  But upon closer examination, we see that there is less here than meets the eye:
  • Only six countries are included in the first study:  China, South Korea, Iran, Italy, France, and the USA.  Conspicuously absent are the UK, Belgium, Spain, Germany, Sweden, Iceland, Taiwan, Japan, and Belarus, which clearly skewed the results.
  • The other study, done by the same Professor Neil Ferguson known for his buggy and discredited model back in March among others included 11 European countries (including many left out of the other study) and did a "simplistic counterfactual" compared to no intervention.
  • Teasing out the precise effects of each specific type of intervention is not easy.
  • The researchers in the second study assume an R value of 3.8 and a doubling time of two days at the beginning of the epidemic, before any intervention, and that it would have remained that high in the absence of such interventions.  In contrast, most other researchers put the basic R value at between 2 and 3, with a doubling time of 3 days--and that makes a very big difference.
  • Epidemics do not keep growing exponentially forever, rather, they famously follow a sigmoidal Gompertz curve even without intervention.
  • There is evidence that the R values in many countries plummeted well before any lockdowns went into effect (Germany's dropped below 1 just days before their lockdown, and Sweden's also did without ever doing a lockdown).  This was due to voluntary behavior changes as well as more modest policy measures.
  • Failing that, the R value will plummet and drop below 1 when the herd immunity threshold is reached regardless.
  • There is evidence that the herd immunity threshold is lower than the naive assumption would put it, and that it would be reached sooner than believed even in an unmitigated scenario.  Also, a LOT more people were infected than the number of confirmed cases, by at least a factor of 10 if not a factor of 80 or 100.
  • The first study only looked at confirmed cases, not hospitalization or death rates.  And confirmed cases are probably the most biased measure of true infection rates that there is.
  • And now the real kicker:  flattening the curve does not actually change the number of cases or deaths, only delays and staggers them.  Except to the extent that it prevents hospitals from being overwhelmed and collapsing, but that generally did not materialize anywhere outside of Lombardy, Italy and some localities of Spain, not even in Sweden.  Moderate social distancing appears to be sufficient to prevent such a collapse, while belated lockdowns utterly failed to prevent it in the very few rare and exceptional cases where it happened.  Thus, the case for lockdowns, well, collapses.
And the best empirical evidence does not pan out in favor of lockdowns:  non-lockdown countries are generally outperforming lockdown countries on average, and within the USA, non-lockdown states have also been outperforming lockdown states in terms of coronavirus case and death rates per capita.

The supposed effectiveness of lockdowns (compared to far less extreme restrictions) in terms of slowing or stopping the spread of coronavirus has been called into serious question lately by other recent studies.
Such studies have found there is at best no correlation, and perhaps a perverse effect between the two defining features of hard lockdowns (stay-home orders and closures of all non-essential businesses) and COVID-19 cases and deaths per capita after other factors such as less-extreme policies are accounted for.  The benefits are thus nothing more than a statistical mirage that does not stand up to scrutiny--much like the supposed benefits of the 21 drinking age vis-a-vis DUI deaths in the long run.

Meanwhile, the collateral damage (economic depression, inequality, poverty, alcohol and other drug abuse, domestic violence, child abuse, loneliness, poor mental health, delayed medical treatment, etc.), which also kills people too by the way, continues to mount with each passing week of lockdown, making an utter mockery of practically all progressive and even basic humanitarian priorities.  And that's to say nothing of the civil rights and liberties, as well as community cohesion, that progressives generally support.  It seems that the "cure" is quickly becoming far worse than the disease as time goes on.

(And that's just for the affluent countries.  For poorer countries, the collateral damage will most likely be at least an order of magnitude worse and deadlier.)

Furthermore, all may not be what it seems in terms of death rates.  In many countries, deaths with COVID and from COVID are conflated, and even non-COVID deaths saw increases during at least the first few weeks of lockdown.  Excess all-cause mortality in so many countries shows a rather suspicious pattern that is strongly suggestive that the government response of lockdowns (along with closely related policies such as artificial restrictions of healthcare and monumentally screwing up with nursing homes) likely caused far more deaths than the virus itself.  In fact, despite the fact that the virus was already circulating much earlier, excess all-cause mortality did not exceed statistical norms in any country until after such lockdowns were imposed.

If that is the case, then that is nothing short of mass murder!  And those are just the short-term effects, with longer-term effects coming down the pike.

Thus, the latest Nature studies should be taken with at least a grain of salt, if not a whole pound.

No comments:

Post a Comment