Hacker News Reader

Comment by throwaway936482 | Hacker News Reader

The problem with the Giesecke approach is that it relies on 2 assumptions being true. 1) that infection brings long term immunity to the currently circulating strains. 2) that covid-19 will not mutate into a new strain with equivalent pathogenicity to which those with immunity to the current strain are no longer immune. If either of these are false then you will not get a meaningful form of immunity in the population. Currently we don't have any evidence that either assumption is true so pursuing this approach carries an increased risk for very little benefit.

View on HN

To point 1, you’re right we don’t know for sure, but past experience with Coronaviruses (4 are endemic plus SARS and MERS) suggests that 0.5-3 years is not unreasonable. Past that the question becomes how much immunity, I.e. even if not completely immune, do you get a milder illness. Again we don’t know for sure but it’s a good Bayesian prior to assume yes, recovering from a legitimate SARS-COV-2 infection yields useful immunity for a meaningful period.

For 2, it’s unlikely to mutate to the point of not being recognized within a relevant timeframe. We can track how fast it’s mutating reliably, and it’s not dangerously quick. The good news is if we have immunity, it should last for several years if considering only what we can see about the mutations. Follow the NextStrain project and Trevor Bedford on Twitter for a smarter analysis than I can provide here.

Regarding point 2 - it is more likely that mutations would actually cause sars-cov-2 to be less pathogenic. Futher if you judge merits of approach solely by looking at death counts you miss the bigger picture: what is the total cost of the approach when you include economy and other indirect impacts, eg. mental issues caused by more severe isolation.

This is wrong. The concept Giesecke has not considered this is a little bizarre.

We have studied corona viruses before so we have information about mutations and immunity. We have also been watching covid-19 for these for 4 months.

So this is 100% untrue - "Currently we don't have any evidence that either assumption is true"

"for very little benefit." - We are talking millions of millions of lives, so I'm not sure why you'd say this. The lockdowns are killing millions, a lot of them are the very poor.

We need to plan for the fact it might mutate or we are not seeing immunity stick. But that's different to throwing out Giesecke approach because we don't know something with certainty.

On the other hand, if we don't get long term immunity from the virus we wont get it from a vaccine either. In my opinion there is three paths through this. Two of them depends on immunity. 1. Slow down the spread with soft lockdown, let it pass and get immunity. Will take a long time. 2. Try to severly limit spread with hard lockdown. Either to open up and do lockdown again as necessary or stay in lockdown. Untill vaccine. This is a long road. Optimistic figures is a vaccine somewhere second half of next year. 3. Contact tracing and severe quarantine for infected and contacts until the virus is eradicated. Quick, only possible if the spread is limited. You can't open your country to others until they have done the same or a vaccine is here. The unicorn exit is of course every county doing this and a total eradication of the virus. A country could possibly change track from strategy 2 to 3 if the spread is down really low and contact tracing is in place. Testing without tracing wont do it.

So soft lockdown in maybe a year or more, deaths will be in the 0,5% vicinity, more in some countries, less in some. Hard lockdown in the same timespan as above. Less deaths but will you have any society to return to? If you do hard lockdown for a while and then lighten up you're in situation 1 basically or forced to lock down soon again. Number three is very attractive. Had we all been prepared and had plans for this like South Korea and being island nations with easily shut borders like NZ it would have been simpler. But most countries were not and are not any of that.

South Korea is effectively a island nation in this context: ain't nothing much passing over their land border unless something goes severely wrong in North Korea.

Very true

To summarize - it's not an approach based on fact and solid knowledge.

To clarify, the Giesecke approach is merely to use the clearer goal of keeping the case load manageable to the health services.

I don't know who came up with the idea that the goal would be heard immunity. Arguably, a "lockdown until vaccine" strategy is more clearly focused on (artificially) reaching herd immunity.

The lockdowns were sold on “flatten the curve” — which they have done, dramatically so, if we take ex ante predictions as accurate. They were not supposed to be the new normal.

True, my mistake. In my defence, there is a lot of "reasons" for the lockdowns floating around in the global conversation.

This particular error on my part I regarding the reasons for the lockdown I attribute to Tomas Pueyo and his hammer-and-dance amateur epidemiology blog.

There's a third assumption that I find particularly pernicious: that the long-term health effects of getting and surviving COVID-19 don't need to be part of the equation. We're learning more every day about how the disease not only affects the lungs but also the heart, the brain, the kidneys, and so on. Treatments, unlike vaccines, are likely to be mere months away, and don't just include drugs. We're already learning about better ways to deal with the oxygen starvation that is COVID-19's hallmark.

Every person who merely delays getting COVID-19 until better treatment is available is a win. Even if it succeeds by other metrics, the Swedish approach will fail (has already failed) by this one.

Every approach to Covid-19 relies on certain assumptions, so this is hardly unique to Giesecke. For example, most lockdown approaches assume we will have a vaccine in 12-18 months, and/or that we will be able to suppress secondary outbreaks after lockdown ends (via test and trace or other methods). Neither of these assumptions are guaranteed either.

Reply on news.ycombinator.com