6

I heard about a meta-analysis talking about lockdowns and how their affected border closures have had "little to no effect on COVID-19 mortality":

We use “mortality” and “mortality rates” interchangeably to mean COVID-19 deaths per population. ...

Overall, lockdowns and limiting gatherings seem to increase COVID-19 mortality, although the effect is modest (0.6% and 1.6%, respectively) and border closures has little to no effect on COVID-19 mortality, with a precision-weighted average of - 0.1% (removing the imprecise outlier from Guo et al. (2021) changes the precision-weighted average to -0.2%).

Yet if you look at the data of some countries before and after closing their borders you can see significant COVID mortality changes occurring around the same time of the border closures.

As with Taiwan for example (May to July). I'm not sure of any other countries where closing borders may have been as effective, or if the decline in Taiwan COVID mortality may have been for another reason.

11
  • 2
    This is a bit weird. They get that by assigning a very high weight to one study, ~4,000 for Toya & Skidmore, and the others orders of magnitude lower. They say it could indicate some publication bias. I have not found their methodology for calculating weight, may add an answer later.
    – User65535
    Feb 3 at 14:53
  • 1
    If there's no COVID in your country, then closing the border should reduce the chances of COVID getting in. Vietnam (despite sharing a border with China) had pretty much 0 deaths until July or August 2020? Feb 3 at 22:03
  • 4
    Why would closing the border ever have an effect on mortality I just don’t really understand the premise of this, maybe I’m just being dumb though.
    – Topcode
    Feb 3 at 23:44
  • 2
    @Topcode I'm assuming they are referring to the total number of deaths and not the actual mortality rate of COVID-19 itself once contracted, that wouldn't make much sense imo.
    – Renge
    Feb 4 at 1:23
  • 9
    @user1271772 and conversely if Covid is already widespread in your population then closing the borders will have little effect. It's important to know whether this study controlled for that or not. Study finds "closing stable doors pointless after horses have run off". Study reported as "stable doors found to be useless".
    – Eric Nolan
    Feb 4 at 13:27

1 Answer 1

12

This paper is a "working" (e.g., unfinalized) and non peer-reviewed meta analysis. It is written by economists, not epidemiologists, and is not well-received by epidemiologists.

The conclusions are contrary to most other findings in the field, a result possibly achieved by cherry-picking (and weighting) the data. It ignores everything past mid-2021, so delta and omicron are not represented. It also expands the term "lockdown" to include any non-medical legally required intervention, so mask requirements count as lockdowns. They also didn't consider the delay between initial contagion and fatal outcomes, about 3 weeks.

One author, Hanke, has a long Twitter history (@steve_hanke) comparing lockdowns with fascism, so he (at least) may be biased in his selection and interpretation of data.

See expert reaction to a preprint looking at the impact of lockdowns, as posted on the John Hopkins Krieger School of Arts and Sciences website

4
  • 4
    Data showing epidemiologists did the wrong thing not well-received by epidemiologists! News at 11! Feb 4 at 19:27
  • Thank you for this answer. This paper was discussed in a german newspaper "BILD" and they were talking about it very matter-of-fact like. They also interviewed a person who was anti-lockdown (at least judging by his twitter posts), which made me skeptical to the validity of the paper.
    – Renge
    Feb 4 at 23:29
  • The question is about border closures; I’m not sure how lockdowns are relevant here. Feb 5 at 2:57
  • @RebeccaJ.Stones The question says "talking about lockdowns and how their affected border closures" The paper is mainly about lockdowns. I think the title is the misleading thing here.
    – mmmmmm
    Feb 6 at 14:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .